Aren/NetBSD
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
2ff97689e1
NetBSD/sys/dev/ic/rtw.c

4270 lines
107 KiB
C
Raw Normal View History

- remove unused variables - move variables inside ifdef sections - ifdef notdef unused code - use __USE for debugging variables
2013-10-18 01:24:24 +04:00
/* $NetBSD: rtw.c,v 1.120 2013/10/17 21:24:24 christos Exp $ */
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
/*-
Accommodate the variety of RF front-ends when radiotapping frames: tap only RSSI when there is a Philips RF front-end. Tap both Barker lock quality and RSSI when there is any other RF front-end. Provide radiotap listeners a more complete picture of channel activity: in promiscuous mode, tap frames who do not pass the CRC32 check. Flag packets that were received with a short preamble. Ask the NIC to pass us 802.11 Control frames. Pass frames to radiotap listeners before stripping the FCS. Re-order operations in rtw_intr_rx() in order to accomplish all of this, taking care not to pass a broken packet to net80211! Do not provide a flags field when tapping xmitted frames. Assert sane Rx packet lengths. Really should check and drop instead of KASSERTing. I will revisit this, soon. Update copyright.
2007-11-17 02:35:19 +03:00
* Copyright (c) 2004, 2005, 2006, 2007 David Young. All rights
* reserved.
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
*
* Programmed for NetBSD by David Young.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
*
* THIS SOFTWARE IS PROVIDED BY David Young ``AS IS'' AND ANY
* EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO,
* THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A
* PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL David
* Young BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
* EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED
* TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
* DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
* ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
* OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY
* OF SUCH DAMAGE.
*/
/*
* Device driver for the Realtek RTL8180 802.11 MAC/BBP.
*/
#include <sys/cdefs.h>
- remove unused variables - move variables inside ifdef sections - ifdef notdef unused code - use __USE for debugging variables
2013-10-18 01:24:24 +04:00
__KERNEL_RCSID(0, "$NetBSD: rtw.c,v 1.120 2013/10/17 21:24:24 christos Exp $");
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
#include <sys/param.h>
At last, I have rtw w/ Philips RF receiving packets. I added some sysctls to aid debugging: * hw.rtw.debug -- enable debugging * hw.rtw.flush_rfio -- Linux voodoo: possibly makes the MAC "flush" bits down the serial bus to the RF * hw.rtw.host_rfio: force the host to bang bits to the RF, instead of the MAC banging bits * hw.rtw.rfio_delay: after telling the MAC to bang bits to the RF front-end, delay rfio_delay microseconds. * hw.rtw.rfprog_fallback: there is this notion of the "RF programming method." I believe the choice influences the polarity/timing of the serial bus used to program the RF front-end. I know the correct choice for Intersil/RFMD/Philips front-ends, only. For all other front-ends, I "fallback" to rfprog_fallback. Make rtw_txdac_enable take an rtw_softc argument. I will probably revert this change. Add some Linux voodoo to rtw_continuous_tx_enable. I will probably revert this change. Important: add rtw_set_rfprog, which sets the correct RF programming method. This change and the following change are probably responsible for making the Philips RF work. Important: RTW_CONFIG1 is an 8-bit register, treat it that way! Important: RTW_BRSR is 16-bit, RTW_CRCOUNT, RTW_PHYDELAY, and RTW_MSR are 8-bit: treat them that way! Vastly simplify rtw_resume_ticks. Note to self: set the LED state to match the power state. Hedge against the possibility that RTW_MSR is protected as RTW_CONFIG[0123] are, meanwhile reworking that section of rtw_init a little. Add sc_anaparm, which isn't used, yet....
2004-12-13 03:48:02 +03:00
#include <sys/sysctl.h>
nuke trailing whitespace
2005-02-27 03:26:58 +03:00
#include <sys/systm.h>
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
#include <sys/callout.h>
nuke trailing whitespace
2005-02-27 03:26:58 +03:00
#include <sys/mbuf.h>
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
#include <sys/malloc.h>
#include <sys/kernel.h>
#include <sys/time.h>
#include <sys/types.h>
Explicitly include <sys/device.h> for device_t.
2008-03-03 15:30:57 +03:00
#include <sys/device.h>
Socket ioctl definitions need sys/sockio.h.
2010-11-15 08:56:29 +03:00
#include <sys/sockio.h>
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
#include <machine/endian.h>
machine/{bus,cpu,intr}.h -> sys/{bus,cpu,intr}.h
2007-10-19 15:59:34 +04:00
#include <sys/bus.h>
#include <sys/intr.h> /* splnet */
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
#include <net/if.h>
#include <net/if_media.h>
#include <net/if_ether.h>
Resolve conflicts in importation of 18-May-2005 ath(4) / net80211(9) from FreeBSD. Introduce compatibility shims (sys/dev/ic/ath_netbsd.[ch], sys/net80211/ieee80211_netbsd.[ch]). Update drivers (an, atu, atw, awi, ipw, iwi, rtw, wi) for the new net80211(9) API.
2005-06-22 10:14:51 +04:00
#include <net80211/ieee80211_netbsd.h>
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
#include <net80211/ieee80211_var.h>
#include <net80211/ieee80211_radiotap.h>
#include <net/bpf.h>
#include <dev/ic/rtwreg.h>
#include <dev/ic/rtwvar.h>
#include <dev/ic/rtwphyio.h>
#include <dev/ic/rtwphy.h>
#include <dev/ic/smc93cx6var.h>
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
static int rtw_rfprog_fallback = 0;
static int rtw_host_rfio = 0;
At last, I have rtw w/ Philips RF receiving packets. I added some sysctls to aid debugging: * hw.rtw.debug -- enable debugging * hw.rtw.flush_rfio -- Linux voodoo: possibly makes the MAC "flush" bits down the serial bus to the RF * hw.rtw.host_rfio: force the host to bang bits to the RF, instead of the MAC banging bits * hw.rtw.rfio_delay: after telling the MAC to bang bits to the RF front-end, delay rfio_delay microseconds. * hw.rtw.rfprog_fallback: there is this notion of the "RF programming method." I believe the choice influences the polarity/timing of the serial bus used to program the RF front-end. I know the correct choice for Intersil/RFMD/Philips front-ends, only. For all other front-ends, I "fallback" to rfprog_fallback. Make rtw_txdac_enable take an rtw_softc argument. I will probably revert this change. Add some Linux voodoo to rtw_continuous_tx_enable. I will probably revert this change. Important: add rtw_set_rfprog, which sets the correct RF programming method. This change and the following change are probably responsible for making the Philips RF work. Important: RTW_CONFIG1 is an 8-bit register, treat it that way! Important: RTW_BRSR is 16-bit, RTW_CRCOUNT, RTW_PHYDELAY, and RTW_MSR are 8-bit: treat them that way! Vastly simplify rtw_resume_ticks. Note to self: set the LED state to match the power state. Hedge against the possibility that RTW_MSR is protected as RTW_CONFIG[0123] are, meanwhile reworking that section of rtw_init a little. Add sc_anaparm, which isn't used, yet....
2004-12-13 03:48:02 +03:00
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
#ifdef RTW_DEBUG
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
int rtw_debug = 0;
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
static int rtw_rxbufs_limit = RTW_RXQLEN;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
#endif /* RTW_DEBUG */
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
#define NEXT_ATTACH_STATE(sc, state) do { \
DPRINTF(sc, RTW_DEBUG_ATTACH, \
("%s: attach state %s\n", __func__, #state)); \
sc->sc_attach_state = state; \
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
} while (0)
Remove useless debug sysctls. For AP scanning, change the channel dwell time from 1 second to 1/5 second.
2004-12-27 09:12:28 +03:00
int rtw_dwelltime = 200; /* milliseconds */
Here is a grotty hack that lets the RTL8180L handle WEP in the receive direction, while software handles WEP in the transmit direction. When net80211 calls rtw's rtw_key_set with a WEP key, I point the key's wk_cipher at our "fake" cipher, rtw_cipher_wep, which is alike to ieee80211_cipher_wep except it provides a different crypto-decapsulation routine, rtw_wep_decap. rtw_wep_decap copies the key passed to it by net80211, clears the key's SWCRYPT flag, and then calls wep_decap. Now wep_decap will decapsulate, but it will *not* re-decrypt. XXX I need to check whether the hardware supports 40-bit WEP, XXX 104-bit WEP, or both, and act accordingly.
2005-06-28 11:19:33 +04:00
static struct ieee80211_cipher rtw_cipher_wep;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
Use device_t and its accessors throughout. Use aprint_*_dev(). Improve PMF-ability. Add a 'flags' argument to suspend/resume handlers and callers such as pmf_system_suspend(). Define a flag, PMF_F_SELF, which indicates to PMF that a device is suspending/resuming itself. Add helper routines, pmf_device_suspend_self(dev) and pmf_device_resume_self(dev), that call pmf_device_suspend(dev, PMF_F_SELF) and pmf_device_resume(dev, PMF_F_SELF), respectively. Use PMF_F_SELF to suspend/resume self in ath(4), audio(4), rtw(4), and sip(4). In ath(4) and in rtw(4), replace the icky sc_enable/sc_disable callbacks, provided by the bus front-end, with self-suspension/resumption. Also, clean up the bus front-ends. Make sure that the interrupt handler is disestablished during suspension. Get rid of driver-private flags (e.g., RTW_F_ENABLED, ath_softc->sc_invalid); use device_is_active()/device_has_power() calls, instead. In the network-class suspend handler, call if_stop(, 0) instead of if_stop(, 1), because the latter is superfluous (bus- and driver-suspension hooks will 'disable' the NIC), and it may cause recursion. In the network-class resume handler, prevent infinite recursion through if_init() by getting out early if we are self-suspending (PMF_F_SELF). rtw(4) improvements: Destroy rtw(4) callouts when we detach it. Make rtw at pci detachable. Print some more information with the "rx frame too long" warning. Remove activate() methods: Get rid of rtw_activate() and ath_activate(). The device activate() methods are not good for much these days. Make ath at cardbus resume with crypto functions intact: Introduce a boolean device property, "pmf-powerdown". If pmf-powerdown is present and false, it indicates that a bus back-end should not remove power from a device. Honor this property in cardbus_child_suspend(). Set this property to 'false' in ath_attach(), since removing power from an ath at cardbus seems to lobotomize the WPA crypto engine. XXX Should the pmf-powerdown property propagate toward the root of the device tree? Miscellaneous ath(4) changes: Warn if ath(4) tries to write crypto keys to suspended hardware. Reduce differences between FreeBSD and NetBSD in ath(4) multicast filter setup. Make ath_printrxbuf() print an rx descriptor's status & key index, to help debug crypto errors. Shorten a staircase in ath_ioctl(). Don't check for ieee80211_ioctl() return code ERESTART, it never happens.
2008-03-12 21:02:21 +03:00
static void rtw_disable_interrupts(struct rtw_regs *);
static void rtw_enable_interrupts(struct rtw_softc *);
static int rtw_init(struct ifnet *);
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
static void rtw_start(struct ifnet *);
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
static void rtw_reset_oactive(struct rtw_softc *);
static struct mbuf *rtw_beacon_alloc(struct rtw_softc *,
struct ieee80211_node *);
static u_int rtw_txring_next(struct rtw_regs *, struct rtw_txdesc_blk *);
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
I have made a small change that suppresses "recovery" from (spurious?) tx timeouts on the beacon queue. The driver's recovery process badly disrupts the MAC's receiver. This needs a closer look, later. Every other fix is by Charles M. Hannum. Charles writes: "With these, the device is quite stable under load, does not experience xmit timeouts, no longer mysteriously freezes, and does not spew a lot of garbage. It does, however, continue to get some "tx fifo underflow"s. global, SYSCTL_SETUP(): a) Removed unused sysctl variables. rtw_io_enable(): b) Change rtw_io_enable to take a softc, and make it call rtw_txring_fixup(). c) Enable the PCI multiple read/write bit. d) Always reset rdb_next to 0 when starting the receive engine. rtw_intr_rx(): e) When processing receive packets, store the next pointer in rdb_next and read it back on the next round, to fix a race condition vs. rtw_tune() during AP discovery and association. f) Remove the special "skip to 0th descriptor" code, because it is no longer needed. rtw_collect_txring(): g) In the "force" case, stop when we catch up to the hardware xmit pointer. rtw_intr_beacon(): h) Resetting tdb_next here looked fishy to me, in light of changes in rtw_collect_txring(). XXX However, this is the one part I haven't tested. rtw_txring_fixup(): i) Add function and line arguments, so we can print out information about the caller. j) In addition to reading the hardware pointer, skip over any additional packets with OWN set -- we may have queued more. rtw_rxring_fixup(): k) Remove this, because it's completely wrong. rtw_intr_ioerror(): l) Don't handle receive overruns here; handle them in the normal receive path. Also changed RTW_INTR_RX and RTW_INTR_IOERROR accordingly. m) Do not fiddle with CR; the transmit engine is still running just fine. n) Do not fiddle with IMR; it was unnecessary. o) Don't touch the receive engine at all; not necessary now that we're leaving CR alone. p) Remove the rtw_do_chip_reset case; it was dangerous and not actually used. rtw_intr(): q) Don't mask off ISR bits when we pass down the ISR value; it's not necessary, and sometimes the other bits useful in debugging output. rtw_seg_too_short(): r) Don't spew here; the caller will do it if necessary. (Stops spewage in an expected case.) rtw_rxdesc_blk: s) Minor change to make the elements have the same order as txdesc_blk. general: t) Added a bunch of debugging output, #ifdef RTW_DIAG, that was useful while investigating the ring management problems. Also moved rtw_txring_fixup() into RTW_DIAG, as I believe the transmit pointer is always kept in sync now, and the function has potential race conditions.
2007-01-06 11:27:53 +03:00
static void rtw_io_enable(struct rtw_softc *, uint8_t, int);
rtw: Try to get hardware WEP to work. It seems to work in the RX direction, but not in the TX direction. The net80211 crypto framework doesn't seem to cope very well with the assymetry (I'm probably missing something), so I will use software WEP for now. net80211: In ieee80211_compute_duration, figure out whether to add the WEP header to the packet overhead by checking the WEP bit in the Frame Control field of the 802.11 header, instead of checking the IEEE80211_F_PRIVACY flag. Also, if the WEP bit is present, assume that the frame described by (wh, len) has already already been WEP encapsulated, and adjust the payload length accordingly. XXX that's a grotty hack that I will have to revisit, later.
2005-06-27 09:49:13 +04:00
static int rtw_key_delete(struct ieee80211com *, const struct ieee80211_key *);
static int rtw_key_set(struct ieee80211com *, const struct ieee80211_key *,
const u_int8_t[IEEE80211_ADDR_LEN]);
static void rtw_key_update_end(struct ieee80211com *);
static void rtw_key_update_begin(struct ieee80211com *);
make rtw.c compile again after the 22-jun-05 net80211 import
2005-09-06 05:59:26 +04:00
static int rtw_wep_decap(struct ieee80211_key *, struct mbuf *, int);
rtw: Try to get hardware WEP to work. It seems to work in the RX direction, but not in the TX direction. The net80211 crypto framework doesn't seem to cope very well with the assymetry (I'm probably missing something), so I will use software WEP for now. net80211: In ieee80211_compute_duration, figure out whether to add the WEP header to the packet overhead by checking the WEP bit in the Frame Control field of the 802.11 header, instead of checking the IEEE80211_F_PRIVACY flag. Also, if the WEP bit is present, assume that the frame described by (wh, len) has already already been WEP encapsulated, and adjust the payload length accordingly. XXX that's a grotty hack that I will have to revisit, later.
2005-06-27 09:49:13 +04:00
static void rtw_wep_setkeys(struct rtw_softc *, struct ieee80211_key *, int);
Misc. changes to improve ad hoc mode and to enable hostap mode: Fix the work-around for the NIC bug where it skips to rx descriptor 0. The driver used to skip to rx descriptor 1. Hopefully this stops the out-of-order packet reception that Charles Hannum saw. When debugging is enabled, print rx-descriptor status flags before printing the rx bit rate. Add a debug message for when a beacon tx buffer reclamation. Reset IFF_OACTIVE when we reset the transmitter. Pass the consolidated LED state, a struct rtw_led_state, to rtw_led_attach. Choose the bit-rate for management frames (1Mb/s) at the same place we choose for all other frames. Do not use the NIC's short preamble or RTS options for management frames. Label beacons for the NIC. Following a Linux driver, take care not to zero arbitrary bits in the TPPOLL register. Use the new idiom for IBSS merges: disable transmitter, kick the state machine. Add a second descriptor to the beacon ring. The NIC seems to like this much better.
2005-03-02 08:20:43 +03:00
static void rtw_led_attach(struct rtw_led_state *, void *);
Use device_t and its accessors throughout. Use aprint_*_dev(). Improve PMF-ability. Add a 'flags' argument to suspend/resume handlers and callers such as pmf_system_suspend(). Define a flag, PMF_F_SELF, which indicates to PMF that a device is suspending/resuming itself. Add helper routines, pmf_device_suspend_self(dev) and pmf_device_resume_self(dev), that call pmf_device_suspend(dev, PMF_F_SELF) and pmf_device_resume(dev, PMF_F_SELF), respectively. Use PMF_F_SELF to suspend/resume self in ath(4), audio(4), rtw(4), and sip(4). In ath(4) and in rtw(4), replace the icky sc_enable/sc_disable callbacks, provided by the bus front-end, with self-suspension/resumption. Also, clean up the bus front-ends. Make sure that the interrupt handler is disestablished during suspension. Get rid of driver-private flags (e.g., RTW_F_ENABLED, ath_softc->sc_invalid); use device_is_active()/device_has_power() calls, instead. In the network-class suspend handler, call if_stop(, 0) instead of if_stop(, 1), because the latter is superfluous (bus- and driver-suspension hooks will 'disable' the NIC), and it may cause recursion. In the network-class resume handler, prevent infinite recursion through if_init() by getting out early if we are self-suspending (PMF_F_SELF). rtw(4) improvements: Destroy rtw(4) callouts when we detach it. Make rtw at pci detachable. Print some more information with the "rx frame too long" warning. Remove activate() methods: Get rid of rtw_activate() and ath_activate(). The device activate() methods are not good for much these days. Make ath at cardbus resume with crypto functions intact: Introduce a boolean device property, "pmf-powerdown". If pmf-powerdown is present and false, it indicates that a bus back-end should not remove power from a device. Honor this property in cardbus_child_suspend(). Set this property to 'false' in ath_attach(), since removing power from an ath at cardbus seems to lobotomize the WPA crypto engine. XXX Should the pmf-powerdown property propagate toward the root of the device tree? Miscellaneous ath(4) changes: Warn if ath(4) tries to write crypto keys to suspended hardware. Reduce differences between FreeBSD and NetBSD in ath(4) multicast filter setup. Make ath_printrxbuf() print an rx descriptor's status & key index, to help debug crypto errors. Shorten a staircase in ath_ioctl(). Don't check for ieee80211_ioctl() return code ERESTART, it never happens.
2008-03-12 21:02:21 +03:00
static void rtw_led_detach(struct rtw_led_state *);
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
static void rtw_led_init(struct rtw_regs *);
static void rtw_led_slowblink(void *);
static void rtw_led_fastblink(void *);
static void rtw_led_set(struct rtw_led_state *, struct rtw_regs *, int);
At last, I have rtw w/ Philips RF receiving packets. I added some sysctls to aid debugging: * hw.rtw.debug -- enable debugging * hw.rtw.flush_rfio -- Linux voodoo: possibly makes the MAC "flush" bits down the serial bus to the RF * hw.rtw.host_rfio: force the host to bang bits to the RF, instead of the MAC banging bits * hw.rtw.rfio_delay: after telling the MAC to bang bits to the RF front-end, delay rfio_delay microseconds. * hw.rtw.rfprog_fallback: there is this notion of the "RF programming method." I believe the choice influences the polarity/timing of the serial bus used to program the RF front-end. I know the correct choice for Intersil/RFMD/Philips front-ends, only. For all other front-ends, I "fallback" to rfprog_fallback. Make rtw_txdac_enable take an rtw_softc argument. I will probably revert this change. Add some Linux voodoo to rtw_continuous_tx_enable. I will probably revert this change. Important: add rtw_set_rfprog, which sets the correct RF programming method. This change and the following change are probably responsible for making the Philips RF work. Important: RTW_CONFIG1 is an 8-bit register, treat it that way! Important: RTW_BRSR is 16-bit, RTW_CRCOUNT, RTW_PHYDELAY, and RTW_MSR are 8-bit: treat them that way! Vastly simplify rtw_resume_ticks. Note to self: set the LED state to match the power state. Hedge against the possibility that RTW_MSR is protected as RTW_CONFIG[0123] are, meanwhile reworking that section of rtw_init a little. Add sc_anaparm, which isn't used, yet....
2004-12-13 03:48:02 +03:00
static int rtw_sysctl_verify_rfio(SYSCTLFN_PROTO);
static int rtw_sysctl_verify_rfprog(SYSCTLFN_PROTO);
#ifdef RTW_DEBUG
I have made a small change that suppresses "recovery" from (spurious?) tx timeouts on the beacon queue. The driver's recovery process badly disrupts the MAC's receiver. This needs a closer look, later. Every other fix is by Charles M. Hannum. Charles writes: "With these, the device is quite stable under load, does not experience xmit timeouts, no longer mysteriously freezes, and does not spew a lot of garbage. It does, however, continue to get some "tx fifo underflow"s. global, SYSCTL_SETUP(): a) Removed unused sysctl variables. rtw_io_enable(): b) Change rtw_io_enable to take a softc, and make it call rtw_txring_fixup(). c) Enable the PCI multiple read/write bit. d) Always reset rdb_next to 0 when starting the receive engine. rtw_intr_rx(): e) When processing receive packets, store the next pointer in rdb_next and read it back on the next round, to fix a race condition vs. rtw_tune() during AP discovery and association. f) Remove the special "skip to 0th descriptor" code, because it is no longer needed. rtw_collect_txring(): g) In the "force" case, stop when we catch up to the hardware xmit pointer. rtw_intr_beacon(): h) Resetting tdb_next here looked fishy to me, in light of changes in rtw_collect_txring(). XXX However, this is the one part I haven't tested. rtw_txring_fixup(): i) Add function and line arguments, so we can print out information about the caller. j) In addition to reading the hardware pointer, skip over any additional packets with OWN set -- we may have queued more. rtw_rxring_fixup(): k) Remove this, because it's completely wrong. rtw_intr_ioerror(): l) Don't handle receive overruns here; handle them in the normal receive path. Also changed RTW_INTR_RX and RTW_INTR_IOERROR accordingly. m) Do not fiddle with CR; the transmit engine is still running just fine. n) Do not fiddle with IMR; it was unnecessary. o) Don't touch the receive engine at all; not necessary now that we're leaving CR alone. p) Remove the rtw_do_chip_reset case; it was dangerous and not actually used. rtw_intr(): q) Don't mask off ISR bits when we pass down the ISR value; it's not necessary, and sometimes the other bits useful in debugging output. rtw_seg_too_short(): r) Don't spew here; the caller will do it if necessary. (Stops spewage in an expected case.) rtw_rxdesc_blk: s) Minor change to make the elements have the same order as txdesc_blk. general: t) Added a bunch of debugging output, #ifdef RTW_DIAG, that was useful while investigating the ring management problems. Also moved rtw_txring_fixup() into RTW_DIAG, as I believe the transmit pointer is always kept in sync now, and the function has potential race conditions.
2007-01-06 11:27:53 +03:00
static void rtw_dump_rings(struct rtw_softc *sc);
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
static void rtw_print_txdesc(struct rtw_softc *, const char *,
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
struct rtw_txsoft *, struct rtw_txdesc_blk *, int);
At last, I have rtw w/ Philips RF receiving packets. I added some sysctls to aid debugging: * hw.rtw.debug -- enable debugging * hw.rtw.flush_rfio -- Linux voodoo: possibly makes the MAC "flush" bits down the serial bus to the RF * hw.rtw.host_rfio: force the host to bang bits to the RF, instead of the MAC banging bits * hw.rtw.rfio_delay: after telling the MAC to bang bits to the RF front-end, delay rfio_delay microseconds. * hw.rtw.rfprog_fallback: there is this notion of the "RF programming method." I believe the choice influences the polarity/timing of the serial bus used to program the RF front-end. I know the correct choice for Intersil/RFMD/Philips front-ends, only. For all other front-ends, I "fallback" to rfprog_fallback. Make rtw_txdac_enable take an rtw_softc argument. I will probably revert this change. Add some Linux voodoo to rtw_continuous_tx_enable. I will probably revert this change. Important: add rtw_set_rfprog, which sets the correct RF programming method. This change and the following change are probably responsible for making the Philips RF work. Important: RTW_CONFIG1 is an 8-bit register, treat it that way! Important: RTW_BRSR is 16-bit, RTW_CRCOUNT, RTW_PHYDELAY, and RTW_MSR are 8-bit: treat them that way! Vastly simplify rtw_resume_ticks. Note to self: set the LED state to match the power state. Hedge against the possibility that RTW_MSR is protected as RTW_CONFIG[0123] are, meanwhile reworking that section of rtw_init a little. Add sc_anaparm, which isn't used, yet....
2004-12-13 03:48:02 +03:00
static int rtw_sysctl_verify_debug(SYSCTLFN_PROTO);
If defined(RTW_DEBUG), provide a sysctl (hw.rtw.rxbufs_limit) for limiting the number of rx buffers an rtw may allocate. Use this sysctl to test the code that copes with buffer exhaustion. Allocate at most RTW_RXQLEN rx buffers, stopping at the sysctl limit. Record in sc_nrxdesc how many were allocated, and put the end-of-ring indication on sc_rxdesc[sc_nrxdesc - 1]. In rtw_init, if no rx buffers could be allocated, log a complaint, clear IFF_RUNNING, and exit with an error. Many changes to accomodate a short rx ring, mainly of the "add a rx-ring length argument" variety. XXX I really should consolidate all of the rx ring variables in one struct and pass that to the rx-ring subroutines. Bug fix: after calling MCLGET, use the (m->m_flags & M_EXT) idiom to check for success, instead of m != NULL. Bug fix: at the top of rtw_start, if IFF_RUNNING is not set, or IFF_OACTIVE is, get out.
2004-12-29 01:21:15 +03:00
static int rtw_sysctl_verify_rxbufs_limit(SYSCTLFN_PROTO);
At last, I have rtw w/ Philips RF receiving packets. I added some sysctls to aid debugging: * hw.rtw.debug -- enable debugging * hw.rtw.flush_rfio -- Linux voodoo: possibly makes the MAC "flush" bits down the serial bus to the RF * hw.rtw.host_rfio: force the host to bang bits to the RF, instead of the MAC banging bits * hw.rtw.rfio_delay: after telling the MAC to bang bits to the RF front-end, delay rfio_delay microseconds. * hw.rtw.rfprog_fallback: there is this notion of the "RF programming method." I believe the choice influences the polarity/timing of the serial bus used to program the RF front-end. I know the correct choice for Intersil/RFMD/Philips front-ends, only. For all other front-ends, I "fallback" to rfprog_fallback. Make rtw_txdac_enable take an rtw_softc argument. I will probably revert this change. Add some Linux voodoo to rtw_continuous_tx_enable. I will probably revert this change. Important: add rtw_set_rfprog, which sets the correct RF programming method. This change and the following change are probably responsible for making the Philips RF work. Important: RTW_CONFIG1 is an 8-bit register, treat it that way! Important: RTW_BRSR is 16-bit, RTW_CRCOUNT, RTW_PHYDELAY, and RTW_MSR are 8-bit: treat them that way! Vastly simplify rtw_resume_ticks. Note to self: set the LED state to match the power state. Hedge against the possibility that RTW_MSR is protected as RTW_CONFIG[0123] are, meanwhile reworking that section of rtw_init a little. Add sc_anaparm, which isn't used, yet....
2004-12-13 03:48:02 +03:00
#endif /* RTW_DEBUG */
I have made a small change that suppresses "recovery" from (spurious?) tx timeouts on the beacon queue. The driver's recovery process badly disrupts the MAC's receiver. This needs a closer look, later. Every other fix is by Charles M. Hannum. Charles writes: "With these, the device is quite stable under load, does not experience xmit timeouts, no longer mysteriously freezes, and does not spew a lot of garbage. It does, however, continue to get some "tx fifo underflow"s. global, SYSCTL_SETUP(): a) Removed unused sysctl variables. rtw_io_enable(): b) Change rtw_io_enable to take a softc, and make it call rtw_txring_fixup(). c) Enable the PCI multiple read/write bit. d) Always reset rdb_next to 0 when starting the receive engine. rtw_intr_rx(): e) When processing receive packets, store the next pointer in rdb_next and read it back on the next round, to fix a race condition vs. rtw_tune() during AP discovery and association. f) Remove the special "skip to 0th descriptor" code, because it is no longer needed. rtw_collect_txring(): g) In the "force" case, stop when we catch up to the hardware xmit pointer. rtw_intr_beacon(): h) Resetting tdb_next here looked fishy to me, in light of changes in rtw_collect_txring(). XXX However, this is the one part I haven't tested. rtw_txring_fixup(): i) Add function and line arguments, so we can print out information about the caller. j) In addition to reading the hardware pointer, skip over any additional packets with OWN set -- we may have queued more. rtw_rxring_fixup(): k) Remove this, because it's completely wrong. rtw_intr_ioerror(): l) Don't handle receive overruns here; handle them in the normal receive path. Also changed RTW_INTR_RX and RTW_INTR_IOERROR accordingly. m) Do not fiddle with CR; the transmit engine is still running just fine. n) Do not fiddle with IMR; it was unnecessary. o) Don't touch the receive engine at all; not necessary now that we're leaving CR alone. p) Remove the rtw_do_chip_reset case; it was dangerous and not actually used. rtw_intr(): q) Don't mask off ISR bits when we pass down the ISR value; it's not necessary, and sometimes the other bits useful in debugging output. rtw_seg_too_short(): r) Don't spew here; the caller will do it if necessary. (Stops spewage in an expected case.) rtw_rxdesc_blk: s) Minor change to make the elements have the same order as txdesc_blk. general: t) Added a bunch of debugging output, #ifdef RTW_DIAG, that was useful while investigating the ring management problems. Also moved rtw_txring_fixup() into RTW_DIAG, as I believe the transmit pointer is always kept in sync now, and the function has potential race conditions.
2007-01-06 11:27:53 +03:00
#ifdef RTW_DIAG
static void rtw_txring_fixup(struct rtw_softc *sc, const char *fn, int ln);
#endif /* RTW_DIAG */
At last, I have rtw w/ Philips RF receiving packets. I added some sysctls to aid debugging: * hw.rtw.debug -- enable debugging * hw.rtw.flush_rfio -- Linux voodoo: possibly makes the MAC "flush" bits down the serial bus to the RF * hw.rtw.host_rfio: force the host to bang bits to the RF, instead of the MAC banging bits * hw.rtw.rfio_delay: after telling the MAC to bang bits to the RF front-end, delay rfio_delay microseconds. * hw.rtw.rfprog_fallback: there is this notion of the "RF programming method." I believe the choice influences the polarity/timing of the serial bus used to program the RF front-end. I know the correct choice for Intersil/RFMD/Philips front-ends, only. For all other front-ends, I "fallback" to rfprog_fallback. Make rtw_txdac_enable take an rtw_softc argument. I will probably revert this change. Add some Linux voodoo to rtw_continuous_tx_enable. I will probably revert this change. Important: add rtw_set_rfprog, which sets the correct RF programming method. This change and the following change are probably responsible for making the Philips RF work. Important: RTW_CONFIG1 is an 8-bit register, treat it that way! Important: RTW_BRSR is 16-bit, RTW_CRCOUNT, RTW_PHYDELAY, and RTW_MSR are 8-bit: treat them that way! Vastly simplify rtw_resume_ticks. Note to self: set the LED state to match the power state. Hedge against the possibility that RTW_MSR is protected as RTW_CONFIG[0123] are, meanwhile reworking that section of rtw_init a little. Add sc_anaparm, which isn't used, yet....
2004-12-13 03:48:02 +03:00
/*
* Setup sysctl(3) MIB, hw.rtw.*
*
Remove LKMs and switch to the module framework, pass 1. Proposed on tech-kern@.
2008-11-12 15:35:50 +03:00
* TBD condition CTLFLAG_PERMANENT on being a module or not
At last, I have rtw w/ Philips RF receiving packets. I added some sysctls to aid debugging: * hw.rtw.debug -- enable debugging * hw.rtw.flush_rfio -- Linux voodoo: possibly makes the MAC "flush" bits down the serial bus to the RF * hw.rtw.host_rfio: force the host to bang bits to the RF, instead of the MAC banging bits * hw.rtw.rfio_delay: after telling the MAC to bang bits to the RF front-end, delay rfio_delay microseconds. * hw.rtw.rfprog_fallback: there is this notion of the "RF programming method." I believe the choice influences the polarity/timing of the serial bus used to program the RF front-end. I know the correct choice for Intersil/RFMD/Philips front-ends, only. For all other front-ends, I "fallback" to rfprog_fallback. Make rtw_txdac_enable take an rtw_softc argument. I will probably revert this change. Add some Linux voodoo to rtw_continuous_tx_enable. I will probably revert this change. Important: add rtw_set_rfprog, which sets the correct RF programming method. This change and the following change are probably responsible for making the Philips RF work. Important: RTW_CONFIG1 is an 8-bit register, treat it that way! Important: RTW_BRSR is 16-bit, RTW_CRCOUNT, RTW_PHYDELAY, and RTW_MSR are 8-bit: treat them that way! Vastly simplify rtw_resume_ticks. Note to self: set the LED state to match the power state. Hedge against the possibility that RTW_MSR is protected as RTW_CONFIG[0123] are, meanwhile reworking that section of rtw_init a little. Add sc_anaparm, which isn't used, yet....
2004-12-13 03:48:02 +03:00
*/
SYSCTL_SETUP(sysctl_rtw, "sysctl rtw(4) subtree setup")
{
int rc;
Change the rest of the sysctl subsystem to use const consistently. The __UNCONST macro is now used only where necessary and the RW macros are gone. Most of the changes here are consumers of the sysctl_createv(9) interface that now takes a pair of const pointers which used not to be.
2005-06-20 06:49:18 +04:00
const struct sysctlnode *cnode, *rnode;
At last, I have rtw w/ Philips RF receiving packets. I added some sysctls to aid debugging: * hw.rtw.debug -- enable debugging * hw.rtw.flush_rfio -- Linux voodoo: possibly makes the MAC "flush" bits down the serial bus to the RF * hw.rtw.host_rfio: force the host to bang bits to the RF, instead of the MAC banging bits * hw.rtw.rfio_delay: after telling the MAC to bang bits to the RF front-end, delay rfio_delay microseconds. * hw.rtw.rfprog_fallback: there is this notion of the "RF programming method." I believe the choice influences the polarity/timing of the serial bus used to program the RF front-end. I know the correct choice for Intersil/RFMD/Philips front-ends, only. For all other front-ends, I "fallback" to rfprog_fallback. Make rtw_txdac_enable take an rtw_softc argument. I will probably revert this change. Add some Linux voodoo to rtw_continuous_tx_enable. I will probably revert this change. Important: add rtw_set_rfprog, which sets the correct RF programming method. This change and the following change are probably responsible for making the Philips RF work. Important: RTW_CONFIG1 is an 8-bit register, treat it that way! Important: RTW_BRSR is 16-bit, RTW_CRCOUNT, RTW_PHYDELAY, and RTW_MSR are 8-bit: treat them that way! Vastly simplify rtw_resume_ticks. Note to self: set the LED state to match the power state. Hedge against the possibility that RTW_MSR is protected as RTW_CONFIG[0123] are, meanwhile reworking that section of rtw_init a little. Add sc_anaparm, which isn't used, yet....
2004-12-13 03:48:02 +03:00
if ((rc = sysctl_createv(clog, 0, NULL, &rnode,
CTLFLAG_PERMANENT, CTLTYPE_NODE, "hw", NULL,
NULL, 0, NULL, 0, CTL_HW, CTL_EOL)) != 0)
goto err;
if ((rc = sysctl_createv(clog, 0, &rnode, &rnode,
CTLFLAG_PERMANENT, CTLTYPE_NODE, "rtw",
"Realtek RTL818x 802.11 controls",
NULL, 0, NULL, 0, CTL_CREATE, CTL_EOL)) != 0)
goto err;
#ifdef RTW_DEBUG
/* control debugging printfs */
if ((rc = sysctl_createv(clog, 0, &rnode, &cnode,
CTLFLAG_PERMANENT|CTLFLAG_READWRITE, CTLTYPE_INT,
"debug", SYSCTL_DESCR("Enable RTL818x debugging output"),
rtw_sysctl_verify_debug, 0, &rtw_debug, 0,
CTL_CREATE, CTL_EOL)) != 0)
goto err;
If defined(RTW_DEBUG), provide a sysctl (hw.rtw.rxbufs_limit) for limiting the number of rx buffers an rtw may allocate. Use this sysctl to test the code that copes with buffer exhaustion. Allocate at most RTW_RXQLEN rx buffers, stopping at the sysctl limit. Record in sc_nrxdesc how many were allocated, and put the end-of-ring indication on sc_rxdesc[sc_nrxdesc - 1]. In rtw_init, if no rx buffers could be allocated, log a complaint, clear IFF_RUNNING, and exit with an error. Many changes to accomodate a short rx ring, mainly of the "add a rx-ring length argument" variety. XXX I really should consolidate all of the rx ring variables in one struct and pass that to the rx-ring subroutines. Bug fix: after calling MCLGET, use the (m->m_flags & M_EXT) idiom to check for success, instead of m != NULL. Bug fix: at the top of rtw_start, if IFF_RUNNING is not set, or IFF_OACTIVE is, get out.
2004-12-29 01:21:15 +03:00
/* Limit rx buffers, for simulating resource exhaustion. */
if ((rc = sysctl_createv(clog, 0, &rnode, &cnode,
CTLFLAG_PERMANENT|CTLFLAG_READWRITE, CTLTYPE_INT,
"rxbufs_limit",
SYSCTL_DESCR("Set rx buffers limit"),
rtw_sysctl_verify_rxbufs_limit, 0, &rtw_rxbufs_limit, 0,
CTL_CREATE, CTL_EOL)) != 0)
goto err;
At last, I have rtw w/ Philips RF receiving packets. I added some sysctls to aid debugging: * hw.rtw.debug -- enable debugging * hw.rtw.flush_rfio -- Linux voodoo: possibly makes the MAC "flush" bits down the serial bus to the RF * hw.rtw.host_rfio: force the host to bang bits to the RF, instead of the MAC banging bits * hw.rtw.rfio_delay: after telling the MAC to bang bits to the RF front-end, delay rfio_delay microseconds. * hw.rtw.rfprog_fallback: there is this notion of the "RF programming method." I believe the choice influences the polarity/timing of the serial bus used to program the RF front-end. I know the correct choice for Intersil/RFMD/Philips front-ends, only. For all other front-ends, I "fallback" to rfprog_fallback. Make rtw_txdac_enable take an rtw_softc argument. I will probably revert this change. Add some Linux voodoo to rtw_continuous_tx_enable. I will probably revert this change. Important: add rtw_set_rfprog, which sets the correct RF programming method. This change and the following change are probably responsible for making the Philips RF work. Important: RTW_CONFIG1 is an 8-bit register, treat it that way! Important: RTW_BRSR is 16-bit, RTW_CRCOUNT, RTW_PHYDELAY, and RTW_MSR are 8-bit: treat them that way! Vastly simplify rtw_resume_ticks. Note to self: set the LED state to match the power state. Hedge against the possibility that RTW_MSR is protected as RTW_CONFIG[0123] are, meanwhile reworking that section of rtw_init a little. Add sc_anaparm, which isn't used, yet....
2004-12-13 03:48:02 +03:00
#endif /* RTW_DEBUG */
/* set fallback RF programming method */
if ((rc = sysctl_createv(clog, 0, &rnode, &cnode,
CTLFLAG_PERMANENT|CTLFLAG_READWRITE, CTLTYPE_INT,
"rfprog_fallback",
SYSCTL_DESCR("Set fallback RF programming method"),
rtw_sysctl_verify_rfprog, 0, &rtw_rfprog_fallback, 0,
CTL_CREATE, CTL_EOL)) != 0)
goto err;
/* force host to control RF I/O bus */
if ((rc = sysctl_createv(clog, 0, &rnode, &cnode,
CTLFLAG_PERMANENT|CTLFLAG_READWRITE, CTLTYPE_INT,
"host_rfio", SYSCTL_DESCR("Enable host control of RF I/O"),
rtw_sysctl_verify_rfio, 0, &rtw_host_rfio, 0,
CTL_CREATE, CTL_EOL)) != 0)
goto err;
return;
err:
printf("%s: sysctl_createv failed (rc = %d)\n", __func__, rc);
}
static int
rtw_sysctl_verify(SYSCTLFN_ARGS, int lower, int upper)
{
int error, t;
struct sysctlnode node;
node = *rnode;
t = *(int*)rnode->sysctl_data;
node.sysctl_data = &t;
error = sysctl_lookup(SYSCTLFN_CALL(&node));
if (error || newp == NULL)
return (error);
if (t < lower || t > upper)
return (EINVAL);
*(int*)rnode->sysctl_data = t;
return (0);
}
static int
rtw_sysctl_verify_rfprog(SYSCTLFN_ARGS)
{
Sprinkle __UNCONST() and rename variables to quiet uiet -Wcast-qual, -Wshadow complaints.
2005-06-10 06:35:34 +04:00
return rtw_sysctl_verify(SYSCTLFN_CALL(__UNCONST(rnode)), 0,
Per discussion on tech-kern and tech-userlevel, move the bit-twiddling macros, __BIT, __BITS, SHIFTIN, SHIFTOUT, and __arraycount() from lib/libkern/libkern.h to sys/cdefs.h. Add a __-prefix to SHIFTIN and SHIFTOUT, and add a manual page for the bit-twiddling macros, bits(3). Make the __BIT and __BITS macros "widthless," as best I can, by changing their type to uintmax_t from uint32_t. XXX The manual page lags this change by a bit. Define __PRIxBIT and __PRIxBITS printf(3) format strings.
2006-08-31 23:24:37 +04:00
__SHIFTOUT(RTW_CONFIG4_RFTYPE_MASK, RTW_CONFIG4_RFTYPE_MASK));
At last, I have rtw w/ Philips RF receiving packets. I added some sysctls to aid debugging: * hw.rtw.debug -- enable debugging * hw.rtw.flush_rfio -- Linux voodoo: possibly makes the MAC "flush" bits down the serial bus to the RF * hw.rtw.host_rfio: force the host to bang bits to the RF, instead of the MAC banging bits * hw.rtw.rfio_delay: after telling the MAC to bang bits to the RF front-end, delay rfio_delay microseconds. * hw.rtw.rfprog_fallback: there is this notion of the "RF programming method." I believe the choice influences the polarity/timing of the serial bus used to program the RF front-end. I know the correct choice for Intersil/RFMD/Philips front-ends, only. For all other front-ends, I "fallback" to rfprog_fallback. Make rtw_txdac_enable take an rtw_softc argument. I will probably revert this change. Add some Linux voodoo to rtw_continuous_tx_enable. I will probably revert this change. Important: add rtw_set_rfprog, which sets the correct RF programming method. This change and the following change are probably responsible for making the Philips RF work. Important: RTW_CONFIG1 is an 8-bit register, treat it that way! Important: RTW_BRSR is 16-bit, RTW_CRCOUNT, RTW_PHYDELAY, and RTW_MSR are 8-bit: treat them that way! Vastly simplify rtw_resume_ticks. Note to self: set the LED state to match the power state. Hedge against the possibility that RTW_MSR is protected as RTW_CONFIG[0123] are, meanwhile reworking that section of rtw_init a little. Add sc_anaparm, which isn't used, yet....
2004-12-13 03:48:02 +03:00
}
static int
rtw_sysctl_verify_rfio(SYSCTLFN_ARGS)
{
Sprinkle __UNCONST() and rename variables to quiet uiet -Wcast-qual, -Wshadow complaints.
2005-06-10 06:35:34 +04:00
return rtw_sysctl_verify(SYSCTLFN_CALL(__UNCONST(rnode)), 0, 1);
At last, I have rtw w/ Philips RF receiving packets. I added some sysctls to aid debugging: * hw.rtw.debug -- enable debugging * hw.rtw.flush_rfio -- Linux voodoo: possibly makes the MAC "flush" bits down the serial bus to the RF * hw.rtw.host_rfio: force the host to bang bits to the RF, instead of the MAC banging bits * hw.rtw.rfio_delay: after telling the MAC to bang bits to the RF front-end, delay rfio_delay microseconds. * hw.rtw.rfprog_fallback: there is this notion of the "RF programming method." I believe the choice influences the polarity/timing of the serial bus used to program the RF front-end. I know the correct choice for Intersil/RFMD/Philips front-ends, only. For all other front-ends, I "fallback" to rfprog_fallback. Make rtw_txdac_enable take an rtw_softc argument. I will probably revert this change. Add some Linux voodoo to rtw_continuous_tx_enable. I will probably revert this change. Important: add rtw_set_rfprog, which sets the correct RF programming method. This change and the following change are probably responsible for making the Philips RF work. Important: RTW_CONFIG1 is an 8-bit register, treat it that way! Important: RTW_BRSR is 16-bit, RTW_CRCOUNT, RTW_PHYDELAY, and RTW_MSR are 8-bit: treat them that way! Vastly simplify rtw_resume_ticks. Note to self: set the LED state to match the power state. Hedge against the possibility that RTW_MSR is protected as RTW_CONFIG[0123] are, meanwhile reworking that section of rtw_init a little. Add sc_anaparm, which isn't used, yet....
2004-12-13 03:48:02 +03:00
}
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
#ifdef RTW_DEBUG
At last, I have rtw w/ Philips RF receiving packets. I added some sysctls to aid debugging: * hw.rtw.debug -- enable debugging * hw.rtw.flush_rfio -- Linux voodoo: possibly makes the MAC "flush" bits down the serial bus to the RF * hw.rtw.host_rfio: force the host to bang bits to the RF, instead of the MAC banging bits * hw.rtw.rfio_delay: after telling the MAC to bang bits to the RF front-end, delay rfio_delay microseconds. * hw.rtw.rfprog_fallback: there is this notion of the "RF programming method." I believe the choice influences the polarity/timing of the serial bus used to program the RF front-end. I know the correct choice for Intersil/RFMD/Philips front-ends, only. For all other front-ends, I "fallback" to rfprog_fallback. Make rtw_txdac_enable take an rtw_softc argument. I will probably revert this change. Add some Linux voodoo to rtw_continuous_tx_enable. I will probably revert this change. Important: add rtw_set_rfprog, which sets the correct RF programming method. This change and the following change are probably responsible for making the Philips RF work. Important: RTW_CONFIG1 is an 8-bit register, treat it that way! Important: RTW_BRSR is 16-bit, RTW_CRCOUNT, RTW_PHYDELAY, and RTW_MSR are 8-bit: treat them that way! Vastly simplify rtw_resume_ticks. Note to self: set the LED state to match the power state. Hedge against the possibility that RTW_MSR is protected as RTW_CONFIG[0123] are, meanwhile reworking that section of rtw_init a little. Add sc_anaparm, which isn't used, yet....
2004-12-13 03:48:02 +03:00
static int
rtw_sysctl_verify_debug(SYSCTLFN_ARGS)
{
Sprinkle __UNCONST() and rename variables to quiet uiet -Wcast-qual, -Wshadow complaints.
2005-06-10 06:35:34 +04:00
return rtw_sysctl_verify(SYSCTLFN_CALL(__UNCONST(rnode)),
0, RTW_DEBUG_MAX);
At last, I have rtw w/ Philips RF receiving packets. I added some sysctls to aid debugging: * hw.rtw.debug -- enable debugging * hw.rtw.flush_rfio -- Linux voodoo: possibly makes the MAC "flush" bits down the serial bus to the RF * hw.rtw.host_rfio: force the host to bang bits to the RF, instead of the MAC banging bits * hw.rtw.rfio_delay: after telling the MAC to bang bits to the RF front-end, delay rfio_delay microseconds. * hw.rtw.rfprog_fallback: there is this notion of the "RF programming method." I believe the choice influences the polarity/timing of the serial bus used to program the RF front-end. I know the correct choice for Intersil/RFMD/Philips front-ends, only. For all other front-ends, I "fallback" to rfprog_fallback. Make rtw_txdac_enable take an rtw_softc argument. I will probably revert this change. Add some Linux voodoo to rtw_continuous_tx_enable. I will probably revert this change. Important: add rtw_set_rfprog, which sets the correct RF programming method. This change and the following change are probably responsible for making the Philips RF work. Important: RTW_CONFIG1 is an 8-bit register, treat it that way! Important: RTW_BRSR is 16-bit, RTW_CRCOUNT, RTW_PHYDELAY, and RTW_MSR are 8-bit: treat them that way! Vastly simplify rtw_resume_ticks. Note to self: set the LED state to match the power state. Hedge against the possibility that RTW_MSR is protected as RTW_CONFIG[0123] are, meanwhile reworking that section of rtw_init a little. Add sc_anaparm, which isn't used, yet....
2004-12-13 03:48:02 +03:00
}
If defined(RTW_DEBUG), provide a sysctl (hw.rtw.rxbufs_limit) for limiting the number of rx buffers an rtw may allocate. Use this sysctl to test the code that copes with buffer exhaustion. Allocate at most RTW_RXQLEN rx buffers, stopping at the sysctl limit. Record in sc_nrxdesc how many were allocated, and put the end-of-ring indication on sc_rxdesc[sc_nrxdesc - 1]. In rtw_init, if no rx buffers could be allocated, log a complaint, clear IFF_RUNNING, and exit with an error. Many changes to accomodate a short rx ring, mainly of the "add a rx-ring length argument" variety. XXX I really should consolidate all of the rx ring variables in one struct and pass that to the rx-ring subroutines. Bug fix: after calling MCLGET, use the (m->m_flags & M_EXT) idiom to check for success, instead of m != NULL. Bug fix: at the top of rtw_start, if IFF_RUNNING is not set, or IFF_OACTIVE is, get out.
2004-12-29 01:21:15 +03:00
static int
rtw_sysctl_verify_rxbufs_limit(SYSCTLFN_ARGS)
{
Sprinkle __UNCONST() and rename variables to quiet uiet -Wcast-qual, -Wshadow complaints.
2005-06-10 06:35:34 +04:00
return rtw_sysctl_verify(SYSCTLFN_CALL(__UNCONST(rnode)),
0, RTW_RXQLEN);
If defined(RTW_DEBUG), provide a sysctl (hw.rtw.rxbufs_limit) for limiting the number of rx buffers an rtw may allocate. Use this sysctl to test the code that copes with buffer exhaustion. Allocate at most RTW_RXQLEN rx buffers, stopping at the sysctl limit. Record in sc_nrxdesc how many were allocated, and put the end-of-ring indication on sc_rxdesc[sc_nrxdesc - 1]. In rtw_init, if no rx buffers could be allocated, log a complaint, clear IFF_RUNNING, and exit with an error. Many changes to accomodate a short rx ring, mainly of the "add a rx-ring length argument" variety. XXX I really should consolidate all of the rx ring variables in one struct and pass that to the rx-ring subroutines. Bug fix: after calling MCLGET, use the (m->m_flags & M_EXT) idiom to check for success, instead of m != NULL. Bug fix: at the top of rtw_start, if IFF_RUNNING is not set, or IFF_OACTIVE is, get out.
2004-12-29 01:21:15 +03:00
}
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
static void
rtw_print_regs(struct rtw_regs *regs, const char *dvname, const char *where)
{
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
#define PRINTREG32(sc, reg) \
RTW_DPRINTF(RTW_DEBUG_REGDUMP, \
("%s: reg[ " #reg " / %03x ] = %08x\n", \
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
dvname, reg, RTW_READ(regs, reg)))
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
#define PRINTREG16(sc, reg) \
RTW_DPRINTF(RTW_DEBUG_REGDUMP, \
("%s: reg[ " #reg " / %03x ] = %04x\n", \
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
dvname, reg, RTW_READ16(regs, reg)))
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
#define PRINTREG8(sc, reg) \
RTW_DPRINTF(RTW_DEBUG_REGDUMP, \
("%s: reg[ " #reg " / %03x ] = %02x\n", \
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
dvname, reg, RTW_READ8(regs, reg)))
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
RTW_DPRINTF(RTW_DEBUG_REGDUMP, ("%s: %s\n", dvname, where));
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
PRINTREG32(regs, RTW_IDR0);
PRINTREG32(regs, RTW_IDR1);
PRINTREG32(regs, RTW_MAR0);
PRINTREG32(regs, RTW_MAR1);
PRINTREG32(regs, RTW_TSFTRL);
PRINTREG32(regs, RTW_TSFTRH);
PRINTREG32(regs, RTW_TLPDA);
PRINTREG32(regs, RTW_TNPDA);
PRINTREG32(regs, RTW_THPDA);
PRINTREG32(regs, RTW_TCR);
PRINTREG32(regs, RTW_RCR);
PRINTREG32(regs, RTW_TINT);
PRINTREG32(regs, RTW_TBDA);
PRINTREG32(regs, RTW_ANAPARM);
PRINTREG32(regs, RTW_BB);
PRINTREG32(regs, RTW_PHYCFG);
PRINTREG32(regs, RTW_WAKEUP0L);
PRINTREG32(regs, RTW_WAKEUP0H);
PRINTREG32(regs, RTW_WAKEUP1L);
PRINTREG32(regs, RTW_WAKEUP1H);
PRINTREG32(regs, RTW_WAKEUP2LL);
PRINTREG32(regs, RTW_WAKEUP2LH);
PRINTREG32(regs, RTW_WAKEUP2HL);
PRINTREG32(regs, RTW_WAKEUP2HH);
PRINTREG32(regs, RTW_WAKEUP3LL);
PRINTREG32(regs, RTW_WAKEUP3LH);
PRINTREG32(regs, RTW_WAKEUP3HL);
PRINTREG32(regs, RTW_WAKEUP3HH);
PRINTREG32(regs, RTW_WAKEUP4LL);
PRINTREG32(regs, RTW_WAKEUP4LH);
PRINTREG32(regs, RTW_WAKEUP4HL);
PRINTREG32(regs, RTW_WAKEUP4HH);
PRINTREG32(regs, RTW_DK0);
PRINTREG32(regs, RTW_DK1);
PRINTREG32(regs, RTW_DK2);
PRINTREG32(regs, RTW_DK3);
PRINTREG32(regs, RTW_RETRYCTR);
PRINTREG32(regs, RTW_RDSAR);
PRINTREG32(regs, RTW_FER);
PRINTREG32(regs, RTW_FEMR);
PRINTREG32(regs, RTW_FPSR);
PRINTREG32(regs, RTW_FFER);
/* 16-bit registers */
PRINTREG16(regs, RTW_BRSR);
PRINTREG16(regs, RTW_IMR);
PRINTREG16(regs, RTW_ISR);
PRINTREG16(regs, RTW_BCNITV);
PRINTREG16(regs, RTW_ATIMWND);
PRINTREG16(regs, RTW_BINTRITV);
PRINTREG16(regs, RTW_ATIMTRITV);
PRINTREG16(regs, RTW_CRC16ERR);
PRINTREG16(regs, RTW_CRC0);
PRINTREG16(regs, RTW_CRC1);
PRINTREG16(regs, RTW_CRC2);
PRINTREG16(regs, RTW_CRC3);
PRINTREG16(regs, RTW_CRC4);
PRINTREG16(regs, RTW_CWR);
/* 8-bit registers */
PRINTREG8(regs, RTW_CR);
PRINTREG8(regs, RTW_9346CR);
PRINTREG8(regs, RTW_CONFIG0);
PRINTREG8(regs, RTW_CONFIG1);
PRINTREG8(regs, RTW_CONFIG2);
PRINTREG8(regs, RTW_MSR);
PRINTREG8(regs, RTW_CONFIG3);
PRINTREG8(regs, RTW_CONFIG4);
PRINTREG8(regs, RTW_TESTR);
PRINTREG8(regs, RTW_PSR);
PRINTREG8(regs, RTW_SCR);
PRINTREG8(regs, RTW_PHYDELAY);
PRINTREG8(regs, RTW_CRCOUNT);
PRINTREG8(regs, RTW_PHYADDR);
PRINTREG8(regs, RTW_PHYDATAW);
PRINTREG8(regs, RTW_PHYDATAR);
PRINTREG8(regs, RTW_CONFIG5);
PRINTREG8(regs, RTW_TPPOLL);
PRINTREG16(regs, RTW_BSSID16);
PRINTREG32(regs, RTW_BSSID32);
#undef PRINTREG32
#undef PRINTREG16
#undef PRINTREG8
}
#endif /* RTW_DEBUG */
void
Miscellaneous changes. Details below. Important changes flagged with []. Using the driver with my Linksys WPC11 ver. 4, it seems to be receiving packets for a change. The WPC11 ver. 4 has a Maxim RF section. My no-name rtw with Philips RF section still does not receive any packets. Keep access-level (analog params > config[0123] registers > none) in sc_access. Add rtw_set_access for changing the access level. Make rtw_continuous_tx_enable and other subroutines use rtw_set_access instead of rtw_config0123_enable and rtw_anaparm_enable. Factor part of the chip-reset code into rtw_chip_reset1. Change the 'struct foo (*bar)[N]'-style arguments to 'struct foo *bar'-style arguments. Consolidate software/hardware Tx/Rx ring setup in rtw_hwring_setup, rtw_swring_setup. Add a new constant, SA2400_OPMODE_DEFAULTS, for the bits that we *always* set in the SA2400 OPMODE register. Factor some code out into rtw_sa2400_calibrate. (Inspired by the Linux driver.) [] When the receiver goes into underrun/overflow state, call a new subroutine, rtw_kick() that stops the Rx/Tx processes, resets the chip, reinitializes the Tx/Rx rings, and restarts Rx/Tx processes. (Inspired by the Linux driver.) [] In rtw_intr_rx, check for too-short packets before calling ieee80211_find_rxnode. I believe this will prevent a repeat of the MCHK exception I saw once on macppc. [] Use seconds-elapased as well as microseconds-elapsed to set the next "due date" for the timeout interrupt. This keeps the driver from programming the timeout to expire too early. [] In rtw_intr, read RTW_ISR at most 10 times, then get out. If the interface is not enabled (RTW_F_ENABLED), then get out. [] In rtw_stop, get out if the interface is not enabled (RTW_F_ENABLED). Block IPL_NET interrupts. Don't read/write any registers if the interface is invalid (RTW_F_INVALID). [] Call rtw_stop in rtw_detach.
2004-12-12 09:37:59 +03:00
rtw_continuous_tx_enable(struct rtw_softc *sc, int enable)
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
{
Miscellaneous changes. Details below. Important changes flagged with []. Using the driver with my Linksys WPC11 ver. 4, it seems to be receiving packets for a change. The WPC11 ver. 4 has a Maxim RF section. My no-name rtw with Philips RF section still does not receive any packets. Keep access-level (analog params > config[0123] registers > none) in sc_access. Add rtw_set_access for changing the access level. Make rtw_continuous_tx_enable and other subroutines use rtw_set_access instead of rtw_config0123_enable and rtw_anaparm_enable. Factor part of the chip-reset code into rtw_chip_reset1. Change the 'struct foo (*bar)[N]'-style arguments to 'struct foo *bar'-style arguments. Consolidate software/hardware Tx/Rx ring setup in rtw_hwring_setup, rtw_swring_setup. Add a new constant, SA2400_OPMODE_DEFAULTS, for the bits that we *always* set in the SA2400 OPMODE register. Factor some code out into rtw_sa2400_calibrate. (Inspired by the Linux driver.) [] When the receiver goes into underrun/overflow state, call a new subroutine, rtw_kick() that stops the Rx/Tx processes, resets the chip, reinitializes the Tx/Rx rings, and restarts Rx/Tx processes. (Inspired by the Linux driver.) [] In rtw_intr_rx, check for too-short packets before calling ieee80211_find_rxnode. I believe this will prevent a repeat of the MCHK exception I saw once on macppc. [] Use seconds-elapased as well as microseconds-elapsed to set the next "due date" for the timeout interrupt. This keeps the driver from programming the timeout to expire too early. [] In rtw_intr, read RTW_ISR at most 10 times, then get out. If the interface is not enabled (RTW_F_ENABLED), then get out. [] In rtw_stop, get out if the interface is not enabled (RTW_F_ENABLED). Block IPL_NET interrupts. Don't read/write any registers if the interface is invalid (RTW_F_INVALID). [] Call rtw_stop in rtw_detach.
2004-12-12 09:37:59 +03:00
struct rtw_regs *regs = &sc->sc_regs;
Cosmetic: change u_int{8,16,32,64}_t to C99-style uint{8,16,32,64}_t, which is preferred.
2005-01-02 07:23:03 +03:00
uint32_t tcr;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
tcr = RTW_READ(regs, RTW_TCR);
tcr &= ~RTW_TCR_LBK_MASK;
if (enable)
tcr |= RTW_TCR_LBK_CONT;
else
tcr |= RTW_TCR_LBK_NORMAL;
RTW_WRITE(regs, RTW_TCR, tcr);
RTW_SYNC(regs, RTW_TCR, RTW_TCR);
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
rtw_set_access(regs, RTW_ACCESS_ANAPARM);
At last, I have rtw w/ Philips RF receiving packets. I added some sysctls to aid debugging: * hw.rtw.debug -- enable debugging * hw.rtw.flush_rfio -- Linux voodoo: possibly makes the MAC "flush" bits down the serial bus to the RF * hw.rtw.host_rfio: force the host to bang bits to the RF, instead of the MAC banging bits * hw.rtw.rfio_delay: after telling the MAC to bang bits to the RF front-end, delay rfio_delay microseconds. * hw.rtw.rfprog_fallback: there is this notion of the "RF programming method." I believe the choice influences the polarity/timing of the serial bus used to program the RF front-end. I know the correct choice for Intersil/RFMD/Philips front-ends, only. For all other front-ends, I "fallback" to rfprog_fallback. Make rtw_txdac_enable take an rtw_softc argument. I will probably revert this change. Add some Linux voodoo to rtw_continuous_tx_enable. I will probably revert this change. Important: add rtw_set_rfprog, which sets the correct RF programming method. This change and the following change are probably responsible for making the Philips RF work. Important: RTW_CONFIG1 is an 8-bit register, treat it that way! Important: RTW_BRSR is 16-bit, RTW_CRCOUNT, RTW_PHYDELAY, and RTW_MSR are 8-bit: treat them that way! Vastly simplify rtw_resume_ticks. Note to self: set the LED state to match the power state. Hedge against the possibility that RTW_MSR is protected as RTW_CONFIG[0123] are, meanwhile reworking that section of rtw_init a little. Add sc_anaparm, which isn't used, yet....
2004-12-13 03:48:02 +03:00
rtw_txdac_enable(sc, !enable);
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
rtw_set_access(regs, RTW_ACCESS_ANAPARM);/* XXX Voodoo from Linux. */
rtw_set_access(regs, RTW_ACCESS_NONE);
Miscellaneous changes. Details below. Important changes flagged with []. Using the driver with my Linksys WPC11 ver. 4, it seems to be receiving packets for a change. The WPC11 ver. 4 has a Maxim RF section. My no-name rtw with Philips RF section still does not receive any packets. Keep access-level (analog params > config[0123] registers > none) in sc_access. Add rtw_set_access for changing the access level. Make rtw_continuous_tx_enable and other subroutines use rtw_set_access instead of rtw_config0123_enable and rtw_anaparm_enable. Factor part of the chip-reset code into rtw_chip_reset1. Change the 'struct foo (*bar)[N]'-style arguments to 'struct foo *bar'-style arguments. Consolidate software/hardware Tx/Rx ring setup in rtw_hwring_setup, rtw_swring_setup. Add a new constant, SA2400_OPMODE_DEFAULTS, for the bits that we *always* set in the SA2400 OPMODE register. Factor some code out into rtw_sa2400_calibrate. (Inspired by the Linux driver.) [] When the receiver goes into underrun/overflow state, call a new subroutine, rtw_kick() that stops the Rx/Tx processes, resets the chip, reinitializes the Tx/Rx rings, and restarts Rx/Tx processes. (Inspired by the Linux driver.) [] In rtw_intr_rx, check for too-short packets before calling ieee80211_find_rxnode. I believe this will prevent a repeat of the MCHK exception I saw once on macppc. [] Use seconds-elapased as well as microseconds-elapsed to set the next "due date" for the timeout interrupt. This keeps the driver from programming the timeout to expire too early. [] In rtw_intr, read RTW_ISR at most 10 times, then get out. If the interface is not enabled (RTW_F_ENABLED), then get out. [] In rtw_stop, get out if the interface is not enabled (RTW_F_ENABLED). Block IPL_NET interrupts. Don't read/write any registers if the interface is invalid (RTW_F_INVALID). [] Call rtw_stop in rtw_detach.
2004-12-12 09:37:59 +03:00
}
Make rtw(4) build w/o RTW_DEBUG. Thanks to Charles Hannum for pointing out my oversight.
2004-12-27 01:59:41 +03:00
#ifdef RTW_DEBUG
Miscellaneous changes. Details below. Important changes flagged with []. Using the driver with my Linksys WPC11 ver. 4, it seems to be receiving packets for a change. The WPC11 ver. 4 has a Maxim RF section. My no-name rtw with Philips RF section still does not receive any packets. Keep access-level (analog params > config[0123] registers > none) in sc_access. Add rtw_set_access for changing the access level. Make rtw_continuous_tx_enable and other subroutines use rtw_set_access instead of rtw_config0123_enable and rtw_anaparm_enable. Factor part of the chip-reset code into rtw_chip_reset1. Change the 'struct foo (*bar)[N]'-style arguments to 'struct foo *bar'-style arguments. Consolidate software/hardware Tx/Rx ring setup in rtw_hwring_setup, rtw_swring_setup. Add a new constant, SA2400_OPMODE_DEFAULTS, for the bits that we *always* set in the SA2400 OPMODE register. Factor some code out into rtw_sa2400_calibrate. (Inspired by the Linux driver.) [] When the receiver goes into underrun/overflow state, call a new subroutine, rtw_kick() that stops the Rx/Tx processes, resets the chip, reinitializes the Tx/Rx rings, and restarts Rx/Tx processes. (Inspired by the Linux driver.) [] In rtw_intr_rx, check for too-short packets before calling ieee80211_find_rxnode. I believe this will prevent a repeat of the MCHK exception I saw once on macppc. [] Use seconds-elapased as well as microseconds-elapsed to set the next "due date" for the timeout interrupt. This keeps the driver from programming the timeout to expire too early. [] In rtw_intr, read RTW_ISR at most 10 times, then get out. If the interface is not enabled (RTW_F_ENABLED), then get out. [] In rtw_stop, get out if the interface is not enabled (RTW_F_ENABLED). Block IPL_NET interrupts. Don't read/write any registers if the interface is invalid (RTW_F_INVALID). [] Call rtw_stop in rtw_detach.
2004-12-12 09:37:59 +03:00
static const char *
rtw_access_string(enum rtw_access access)
{
switch (access) {
case RTW_ACCESS_NONE:
return "none";
case RTW_ACCESS_CONFIG:
return "config";
case RTW_ACCESS_ANAPARM:
return "anaparm";
default:
return "unknown";
}
}
Make rtw(4) build w/o RTW_DEBUG. Thanks to Charles Hannum for pointing out my oversight.
2004-12-27 01:59:41 +03:00
#endif /* RTW_DEBUG */
Miscellaneous changes. Details below. Important changes flagged with []. Using the driver with my Linksys WPC11 ver. 4, it seems to be receiving packets for a change. The WPC11 ver. 4 has a Maxim RF section. My no-name rtw with Philips RF section still does not receive any packets. Keep access-level (analog params > config[0123] registers > none) in sc_access. Add rtw_set_access for changing the access level. Make rtw_continuous_tx_enable and other subroutines use rtw_set_access instead of rtw_config0123_enable and rtw_anaparm_enable. Factor part of the chip-reset code into rtw_chip_reset1. Change the 'struct foo (*bar)[N]'-style arguments to 'struct foo *bar'-style arguments. Consolidate software/hardware Tx/Rx ring setup in rtw_hwring_setup, rtw_swring_setup. Add a new constant, SA2400_OPMODE_DEFAULTS, for the bits that we *always* set in the SA2400 OPMODE register. Factor some code out into rtw_sa2400_calibrate. (Inspired by the Linux driver.) [] When the receiver goes into underrun/overflow state, call a new subroutine, rtw_kick() that stops the Rx/Tx processes, resets the chip, reinitializes the Tx/Rx rings, and restarts Rx/Tx processes. (Inspired by the Linux driver.) [] In rtw_intr_rx, check for too-short packets before calling ieee80211_find_rxnode. I believe this will prevent a repeat of the MCHK exception I saw once on macppc. [] Use seconds-elapased as well as microseconds-elapsed to set the next "due date" for the timeout interrupt. This keeps the driver from programming the timeout to expire too early. [] In rtw_intr, read RTW_ISR at most 10 times, then get out. If the interface is not enabled (RTW_F_ENABLED), then get out. [] In rtw_stop, get out if the interface is not enabled (RTW_F_ENABLED). Block IPL_NET interrupts. Don't read/write any registers if the interface is invalid (RTW_F_INVALID). [] Call rtw_stop in rtw_detach.
2004-12-12 09:37:59 +03:00
static void
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
rtw_set_access1(struct rtw_regs *regs, enum rtw_access naccess)
Miscellaneous changes. Details below. Important changes flagged with []. Using the driver with my Linksys WPC11 ver. 4, it seems to be receiving packets for a change. The WPC11 ver. 4 has a Maxim RF section. My no-name rtw with Philips RF section still does not receive any packets. Keep access-level (analog params > config[0123] registers > none) in sc_access. Add rtw_set_access for changing the access level. Make rtw_continuous_tx_enable and other subroutines use rtw_set_access instead of rtw_config0123_enable and rtw_anaparm_enable. Factor part of the chip-reset code into rtw_chip_reset1. Change the 'struct foo (*bar)[N]'-style arguments to 'struct foo *bar'-style arguments. Consolidate software/hardware Tx/Rx ring setup in rtw_hwring_setup, rtw_swring_setup. Add a new constant, SA2400_OPMODE_DEFAULTS, for the bits that we *always* set in the SA2400 OPMODE register. Factor some code out into rtw_sa2400_calibrate. (Inspired by the Linux driver.) [] When the receiver goes into underrun/overflow state, call a new subroutine, rtw_kick() that stops the Rx/Tx processes, resets the chip, reinitializes the Tx/Rx rings, and restarts Rx/Tx processes. (Inspired by the Linux driver.) [] In rtw_intr_rx, check for too-short packets before calling ieee80211_find_rxnode. I believe this will prevent a repeat of the MCHK exception I saw once on macppc. [] Use seconds-elapased as well as microseconds-elapsed to set the next "due date" for the timeout interrupt. This keeps the driver from programming the timeout to expire too early. [] In rtw_intr, read RTW_ISR at most 10 times, then get out. If the interface is not enabled (RTW_F_ENABLED), then get out. [] In rtw_stop, get out if the interface is not enabled (RTW_F_ENABLED). Block IPL_NET interrupts. Don't read/write any registers if the interface is invalid (RTW_F_INVALID). [] Call rtw_stop in rtw_detach.
2004-12-12 09:37:59 +03:00
{
- comment out impossible comparisons - remove superfluous if statement
2006-09-03 09:10:24 +04:00
KASSERT(/* naccess >= RTW_ACCESS_NONE && */
naccess <= RTW_ACCESS_ANAPARM);
KASSERT(/* regs->r_access >= RTW_ACCESS_NONE && */
regs->r_access <= RTW_ACCESS_ANAPARM);
Miscellaneous changes. Details below. Important changes flagged with []. Using the driver with my Linksys WPC11 ver. 4, it seems to be receiving packets for a change. The WPC11 ver. 4 has a Maxim RF section. My no-name rtw with Philips RF section still does not receive any packets. Keep access-level (analog params > config[0123] registers > none) in sc_access. Add rtw_set_access for changing the access level. Make rtw_continuous_tx_enable and other subroutines use rtw_set_access instead of rtw_config0123_enable and rtw_anaparm_enable. Factor part of the chip-reset code into rtw_chip_reset1. Change the 'struct foo (*bar)[N]'-style arguments to 'struct foo *bar'-style arguments. Consolidate software/hardware Tx/Rx ring setup in rtw_hwring_setup, rtw_swring_setup. Add a new constant, SA2400_OPMODE_DEFAULTS, for the bits that we *always* set in the SA2400 OPMODE register. Factor some code out into rtw_sa2400_calibrate. (Inspired by the Linux driver.) [] When the receiver goes into underrun/overflow state, call a new subroutine, rtw_kick() that stops the Rx/Tx processes, resets the chip, reinitializes the Tx/Rx rings, and restarts Rx/Tx processes. (Inspired by the Linux driver.) [] In rtw_intr_rx, check for too-short packets before calling ieee80211_find_rxnode. I believe this will prevent a repeat of the MCHK exception I saw once on macppc. [] Use seconds-elapased as well as microseconds-elapsed to set the next "due date" for the timeout interrupt. This keeps the driver from programming the timeout to expire too early. [] In rtw_intr, read RTW_ISR at most 10 times, then get out. If the interface is not enabled (RTW_F_ENABLED), then get out. [] In rtw_stop, get out if the interface is not enabled (RTW_F_ENABLED). Block IPL_NET interrupts. Don't read/write any registers if the interface is invalid (RTW_F_INVALID). [] Call rtw_stop in rtw_detach.
2004-12-12 09:37:59 +03:00
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
if (naccess == regs->r_access)
Miscellaneous changes. Details below. Important changes flagged with []. Using the driver with my Linksys WPC11 ver. 4, it seems to be receiving packets for a change. The WPC11 ver. 4 has a Maxim RF section. My no-name rtw with Philips RF section still does not receive any packets. Keep access-level (analog params > config[0123] registers > none) in sc_access. Add rtw_set_access for changing the access level. Make rtw_continuous_tx_enable and other subroutines use rtw_set_access instead of rtw_config0123_enable and rtw_anaparm_enable. Factor part of the chip-reset code into rtw_chip_reset1. Change the 'struct foo (*bar)[N]'-style arguments to 'struct foo *bar'-style arguments. Consolidate software/hardware Tx/Rx ring setup in rtw_hwring_setup, rtw_swring_setup. Add a new constant, SA2400_OPMODE_DEFAULTS, for the bits that we *always* set in the SA2400 OPMODE register. Factor some code out into rtw_sa2400_calibrate. (Inspired by the Linux driver.) [] When the receiver goes into underrun/overflow state, call a new subroutine, rtw_kick() that stops the Rx/Tx processes, resets the chip, reinitializes the Tx/Rx rings, and restarts Rx/Tx processes. (Inspired by the Linux driver.) [] In rtw_intr_rx, check for too-short packets before calling ieee80211_find_rxnode. I believe this will prevent a repeat of the MCHK exception I saw once on macppc. [] Use seconds-elapased as well as microseconds-elapsed to set the next "due date" for the timeout interrupt. This keeps the driver from programming the timeout to expire too early. [] In rtw_intr, read RTW_ISR at most 10 times, then get out. If the interface is not enabled (RTW_F_ENABLED), then get out. [] In rtw_stop, get out if the interface is not enabled (RTW_F_ENABLED). Block IPL_NET interrupts. Don't read/write any registers if the interface is invalid (RTW_F_INVALID). [] Call rtw_stop in rtw_detach.
2004-12-12 09:37:59 +03:00
return;
switch (naccess) {
case RTW_ACCESS_NONE:
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
switch (regs->r_access) {
Miscellaneous changes. Details below. Important changes flagged with []. Using the driver with my Linksys WPC11 ver. 4, it seems to be receiving packets for a change. The WPC11 ver. 4 has a Maxim RF section. My no-name rtw with Philips RF section still does not receive any packets. Keep access-level (analog params > config[0123] registers > none) in sc_access. Add rtw_set_access for changing the access level. Make rtw_continuous_tx_enable and other subroutines use rtw_set_access instead of rtw_config0123_enable and rtw_anaparm_enable. Factor part of the chip-reset code into rtw_chip_reset1. Change the 'struct foo (*bar)[N]'-style arguments to 'struct foo *bar'-style arguments. Consolidate software/hardware Tx/Rx ring setup in rtw_hwring_setup, rtw_swring_setup. Add a new constant, SA2400_OPMODE_DEFAULTS, for the bits that we *always* set in the SA2400 OPMODE register. Factor some code out into rtw_sa2400_calibrate. (Inspired by the Linux driver.) [] When the receiver goes into underrun/overflow state, call a new subroutine, rtw_kick() that stops the Rx/Tx processes, resets the chip, reinitializes the Tx/Rx rings, and restarts Rx/Tx processes. (Inspired by the Linux driver.) [] In rtw_intr_rx, check for too-short packets before calling ieee80211_find_rxnode. I believe this will prevent a repeat of the MCHK exception I saw once on macppc. [] Use seconds-elapased as well as microseconds-elapsed to set the next "due date" for the timeout interrupt. This keeps the driver from programming the timeout to expire too early. [] In rtw_intr, read RTW_ISR at most 10 times, then get out. If the interface is not enabled (RTW_F_ENABLED), then get out. [] In rtw_stop, get out if the interface is not enabled (RTW_F_ENABLED). Block IPL_NET interrupts. Don't read/write any registers if the interface is invalid (RTW_F_INVALID). [] Call rtw_stop in rtw_detach.
2004-12-12 09:37:59 +03:00
case RTW_ACCESS_ANAPARM:
rtw_anaparm_enable(regs, 0);
/*FALLTHROUGH*/
case RTW_ACCESS_CONFIG:
rtw_config0123_enable(regs, 0);
/*FALLTHROUGH*/
case RTW_ACCESS_NONE:
break;
}
break;
case RTW_ACCESS_CONFIG:
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
switch (regs->r_access) {
Miscellaneous changes. Details below. Important changes flagged with []. Using the driver with my Linksys WPC11 ver. 4, it seems to be receiving packets for a change. The WPC11 ver. 4 has a Maxim RF section. My no-name rtw with Philips RF section still does not receive any packets. Keep access-level (analog params > config[0123] registers > none) in sc_access. Add rtw_set_access for changing the access level. Make rtw_continuous_tx_enable and other subroutines use rtw_set_access instead of rtw_config0123_enable and rtw_anaparm_enable. Factor part of the chip-reset code into rtw_chip_reset1. Change the 'struct foo (*bar)[N]'-style arguments to 'struct foo *bar'-style arguments. Consolidate software/hardware Tx/Rx ring setup in rtw_hwring_setup, rtw_swring_setup. Add a new constant, SA2400_OPMODE_DEFAULTS, for the bits that we *always* set in the SA2400 OPMODE register. Factor some code out into rtw_sa2400_calibrate. (Inspired by the Linux driver.) [] When the receiver goes into underrun/overflow state, call a new subroutine, rtw_kick() that stops the Rx/Tx processes, resets the chip, reinitializes the Tx/Rx rings, and restarts Rx/Tx processes. (Inspired by the Linux driver.) [] In rtw_intr_rx, check for too-short packets before calling ieee80211_find_rxnode. I believe this will prevent a repeat of the MCHK exception I saw once on macppc. [] Use seconds-elapased as well as microseconds-elapsed to set the next "due date" for the timeout interrupt. This keeps the driver from programming the timeout to expire too early. [] In rtw_intr, read RTW_ISR at most 10 times, then get out. If the interface is not enabled (RTW_F_ENABLED), then get out. [] In rtw_stop, get out if the interface is not enabled (RTW_F_ENABLED). Block IPL_NET interrupts. Don't read/write any registers if the interface is invalid (RTW_F_INVALID). [] Call rtw_stop in rtw_detach.
2004-12-12 09:37:59 +03:00
case RTW_ACCESS_NONE:
rtw_config0123_enable(regs, 1);
/*FALLTHROUGH*/
case RTW_ACCESS_CONFIG:
break;
case RTW_ACCESS_ANAPARM:
rtw_anaparm_enable(regs, 0);
break;
}
break;
case RTW_ACCESS_ANAPARM:
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
switch (regs->r_access) {
Miscellaneous changes. Details below. Important changes flagged with []. Using the driver with my Linksys WPC11 ver. 4, it seems to be receiving packets for a change. The WPC11 ver. 4 has a Maxim RF section. My no-name rtw with Philips RF section still does not receive any packets. Keep access-level (analog params > config[0123] registers > none) in sc_access. Add rtw_set_access for changing the access level. Make rtw_continuous_tx_enable and other subroutines use rtw_set_access instead of rtw_config0123_enable and rtw_anaparm_enable. Factor part of the chip-reset code into rtw_chip_reset1. Change the 'struct foo (*bar)[N]'-style arguments to 'struct foo *bar'-style arguments. Consolidate software/hardware Tx/Rx ring setup in rtw_hwring_setup, rtw_swring_setup. Add a new constant, SA2400_OPMODE_DEFAULTS, for the bits that we *always* set in the SA2400 OPMODE register. Factor some code out into rtw_sa2400_calibrate. (Inspired by the Linux driver.) [] When the receiver goes into underrun/overflow state, call a new subroutine, rtw_kick() that stops the Rx/Tx processes, resets the chip, reinitializes the Tx/Rx rings, and restarts Rx/Tx processes. (Inspired by the Linux driver.) [] In rtw_intr_rx, check for too-short packets before calling ieee80211_find_rxnode. I believe this will prevent a repeat of the MCHK exception I saw once on macppc. [] Use seconds-elapased as well as microseconds-elapsed to set the next "due date" for the timeout interrupt. This keeps the driver from programming the timeout to expire too early. [] In rtw_intr, read RTW_ISR at most 10 times, then get out. If the interface is not enabled (RTW_F_ENABLED), then get out. [] In rtw_stop, get out if the interface is not enabled (RTW_F_ENABLED). Block IPL_NET interrupts. Don't read/write any registers if the interface is invalid (RTW_F_INVALID). [] Call rtw_stop in rtw_detach.
2004-12-12 09:37:59 +03:00
case RTW_ACCESS_NONE:
rtw_config0123_enable(regs, 1);
/*FALLTHROUGH*/
case RTW_ACCESS_CONFIG:
rtw_anaparm_enable(regs, 1);
/*FALLTHROUGH*/
case RTW_ACCESS_ANAPARM:
break;
}
break;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
}
}
Miscellaneous changes. Details below. Important changes flagged with []. Using the driver with my Linksys WPC11 ver. 4, it seems to be receiving packets for a change. The WPC11 ver. 4 has a Maxim RF section. My no-name rtw with Philips RF section still does not receive any packets. Keep access-level (analog params > config[0123] registers > none) in sc_access. Add rtw_set_access for changing the access level. Make rtw_continuous_tx_enable and other subroutines use rtw_set_access instead of rtw_config0123_enable and rtw_anaparm_enable. Factor part of the chip-reset code into rtw_chip_reset1. Change the 'struct foo (*bar)[N]'-style arguments to 'struct foo *bar'-style arguments. Consolidate software/hardware Tx/Rx ring setup in rtw_hwring_setup, rtw_swring_setup. Add a new constant, SA2400_OPMODE_DEFAULTS, for the bits that we *always* set in the SA2400 OPMODE register. Factor some code out into rtw_sa2400_calibrate. (Inspired by the Linux driver.) [] When the receiver goes into underrun/overflow state, call a new subroutine, rtw_kick() that stops the Rx/Tx processes, resets the chip, reinitializes the Tx/Rx rings, and restarts Rx/Tx processes. (Inspired by the Linux driver.) [] In rtw_intr_rx, check for too-short packets before calling ieee80211_find_rxnode. I believe this will prevent a repeat of the MCHK exception I saw once on macppc. [] Use seconds-elapased as well as microseconds-elapsed to set the next "due date" for the timeout interrupt. This keeps the driver from programming the timeout to expire too early. [] In rtw_intr, read RTW_ISR at most 10 times, then get out. If the interface is not enabled (RTW_F_ENABLED), then get out. [] In rtw_stop, get out if the interface is not enabled (RTW_F_ENABLED). Block IPL_NET interrupts. Don't read/write any registers if the interface is invalid (RTW_F_INVALID). [] Call rtw_stop in rtw_detach.
2004-12-12 09:37:59 +03:00
void
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
rtw_set_access(struct rtw_regs *regs, enum rtw_access access)
Miscellaneous changes. Details below. Important changes flagged with []. Using the driver with my Linksys WPC11 ver. 4, it seems to be receiving packets for a change. The WPC11 ver. 4 has a Maxim RF section. My no-name rtw with Philips RF section still does not receive any packets. Keep access-level (analog params > config[0123] registers > none) in sc_access. Add rtw_set_access for changing the access level. Make rtw_continuous_tx_enable and other subroutines use rtw_set_access instead of rtw_config0123_enable and rtw_anaparm_enable. Factor part of the chip-reset code into rtw_chip_reset1. Change the 'struct foo (*bar)[N]'-style arguments to 'struct foo *bar'-style arguments. Consolidate software/hardware Tx/Rx ring setup in rtw_hwring_setup, rtw_swring_setup. Add a new constant, SA2400_OPMODE_DEFAULTS, for the bits that we *always* set in the SA2400 OPMODE register. Factor some code out into rtw_sa2400_calibrate. (Inspired by the Linux driver.) [] When the receiver goes into underrun/overflow state, call a new subroutine, rtw_kick() that stops the Rx/Tx processes, resets the chip, reinitializes the Tx/Rx rings, and restarts Rx/Tx processes. (Inspired by the Linux driver.) [] In rtw_intr_rx, check for too-short packets before calling ieee80211_find_rxnode. I believe this will prevent a repeat of the MCHK exception I saw once on macppc. [] Use seconds-elapased as well as microseconds-elapsed to set the next "due date" for the timeout interrupt. This keeps the driver from programming the timeout to expire too early. [] In rtw_intr, read RTW_ISR at most 10 times, then get out. If the interface is not enabled (RTW_F_ENABLED), then get out. [] In rtw_stop, get out if the interface is not enabled (RTW_F_ENABLED). Block IPL_NET interrupts. Don't read/write any registers if the interface is invalid (RTW_F_INVALID). [] Call rtw_stop in rtw_detach.
2004-12-12 09:37:59 +03:00
{
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
rtw_set_access1(regs, access);
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
RTW_DPRINTF(RTW_DEBUG_ACCESS,
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
("%s: access %s -> %s\n", __func__,
rtw_access_string(regs->r_access),
Miscellaneous changes. Details below. Important changes flagged with []. Using the driver with my Linksys WPC11 ver. 4, it seems to be receiving packets for a change. The WPC11 ver. 4 has a Maxim RF section. My no-name rtw with Philips RF section still does not receive any packets. Keep access-level (analog params > config[0123] registers > none) in sc_access. Add rtw_set_access for changing the access level. Make rtw_continuous_tx_enable and other subroutines use rtw_set_access instead of rtw_config0123_enable and rtw_anaparm_enable. Factor part of the chip-reset code into rtw_chip_reset1. Change the 'struct foo (*bar)[N]'-style arguments to 'struct foo *bar'-style arguments. Consolidate software/hardware Tx/Rx ring setup in rtw_hwring_setup, rtw_swring_setup. Add a new constant, SA2400_OPMODE_DEFAULTS, for the bits that we *always* set in the SA2400 OPMODE register. Factor some code out into rtw_sa2400_calibrate. (Inspired by the Linux driver.) [] When the receiver goes into underrun/overflow state, call a new subroutine, rtw_kick() that stops the Rx/Tx processes, resets the chip, reinitializes the Tx/Rx rings, and restarts Rx/Tx processes. (Inspired by the Linux driver.) [] In rtw_intr_rx, check for too-short packets before calling ieee80211_find_rxnode. I believe this will prevent a repeat of the MCHK exception I saw once on macppc. [] Use seconds-elapased as well as microseconds-elapsed to set the next "due date" for the timeout interrupt. This keeps the driver from programming the timeout to expire too early. [] In rtw_intr, read RTW_ISR at most 10 times, then get out. If the interface is not enabled (RTW_F_ENABLED), then get out. [] In rtw_stop, get out if the interface is not enabled (RTW_F_ENABLED). Block IPL_NET interrupts. Don't read/write any registers if the interface is invalid (RTW_F_INVALID). [] Call rtw_stop in rtw_detach.
2004-12-12 09:37:59 +03:00
rtw_access_string(access)));
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
regs->r_access = access;
Miscellaneous changes. Details below. Important changes flagged with []. Using the driver with my Linksys WPC11 ver. 4, it seems to be receiving packets for a change. The WPC11 ver. 4 has a Maxim RF section. My no-name rtw with Philips RF section still does not receive any packets. Keep access-level (analog params > config[0123] registers > none) in sc_access. Add rtw_set_access for changing the access level. Make rtw_continuous_tx_enable and other subroutines use rtw_set_access instead of rtw_config0123_enable and rtw_anaparm_enable. Factor part of the chip-reset code into rtw_chip_reset1. Change the 'struct foo (*bar)[N]'-style arguments to 'struct foo *bar'-style arguments. Consolidate software/hardware Tx/Rx ring setup in rtw_hwring_setup, rtw_swring_setup. Add a new constant, SA2400_OPMODE_DEFAULTS, for the bits that we *always* set in the SA2400 OPMODE register. Factor some code out into rtw_sa2400_calibrate. (Inspired by the Linux driver.) [] When the receiver goes into underrun/overflow state, call a new subroutine, rtw_kick() that stops the Rx/Tx processes, resets the chip, reinitializes the Tx/Rx rings, and restarts Rx/Tx processes. (Inspired by the Linux driver.) [] In rtw_intr_rx, check for too-short packets before calling ieee80211_find_rxnode. I believe this will prevent a repeat of the MCHK exception I saw once on macppc. [] Use seconds-elapased as well as microseconds-elapsed to set the next "due date" for the timeout interrupt. This keeps the driver from programming the timeout to expire too early. [] In rtw_intr, read RTW_ISR at most 10 times, then get out. If the interface is not enabled (RTW_F_ENABLED), then get out. [] In rtw_stop, get out if the interface is not enabled (RTW_F_ENABLED). Block IPL_NET interrupts. Don't read/write any registers if the interface is invalid (RTW_F_INVALID). [] Call rtw_stop in rtw_detach.
2004-12-12 09:37:59 +03:00
}
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
/*
* Enable registers, switch register banks.
*/
void
rtw_config0123_enable(struct rtw_regs *regs, int enable)
{
Cosmetic: change u_int{8,16,32,64}_t to C99-style uint{8,16,32,64}_t, which is preferred.
2005-01-02 07:23:03 +03:00
uint8_t ecr;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
ecr = RTW_READ8(regs, RTW_9346CR);
ecr &= ~(RTW_9346CR_EEM_MASK | RTW_9346CR_EECS | RTW_9346CR_EESK);
if (enable)
ecr |= RTW_9346CR_EEM_CONFIG;
Add some bus barriers. Remark where they are needed, e.g., read_eeprom. Change some printfs to RTW_DPRINTF2s to quiet rtw(4) when debugging is disabled.
2004-12-20 04:13:45 +03:00
else {
RTW_WBW(regs, RTW_9346CR, MAX(RTW_CONFIG0, RTW_CONFIG3));
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
ecr |= RTW_9346CR_EEM_NORMAL;
Add some bus barriers. Remark where they are needed, e.g., read_eeprom. Change some printfs to RTW_DPRINTF2s to quiet rtw(4) when debugging is disabled.
2004-12-20 04:13:45 +03:00
}
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
RTW_WRITE8(regs, RTW_9346CR, ecr);
RTW_SYNC(regs, RTW_9346CR, RTW_9346CR);
}
/* requires rtw_config0123_enable(, 1) */
void
rtw_anaparm_enable(struct rtw_regs *regs, int enable)
{
Cosmetic: change u_int{8,16,32,64}_t to C99-style uint{8,16,32,64}_t, which is preferred.
2005-01-02 07:23:03 +03:00
uint8_t cfg3;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
cfg3 = RTW_READ8(regs, RTW_CONFIG3);
Miscellaneous changes. Details below. Important changes flagged with []. Using the driver with my Linksys WPC11 ver. 4, it seems to be receiving packets for a change. The WPC11 ver. 4 has a Maxim RF section. My no-name rtw with Philips RF section still does not receive any packets. Keep access-level (analog params > config[0123] registers > none) in sc_access. Add rtw_set_access for changing the access level. Make rtw_continuous_tx_enable and other subroutines use rtw_set_access instead of rtw_config0123_enable and rtw_anaparm_enable. Factor part of the chip-reset code into rtw_chip_reset1. Change the 'struct foo (*bar)[N]'-style arguments to 'struct foo *bar'-style arguments. Consolidate software/hardware Tx/Rx ring setup in rtw_hwring_setup, rtw_swring_setup. Add a new constant, SA2400_OPMODE_DEFAULTS, for the bits that we *always* set in the SA2400 OPMODE register. Factor some code out into rtw_sa2400_calibrate. (Inspired by the Linux driver.) [] When the receiver goes into underrun/overflow state, call a new subroutine, rtw_kick() that stops the Rx/Tx processes, resets the chip, reinitializes the Tx/Rx rings, and restarts Rx/Tx processes. (Inspired by the Linux driver.) [] In rtw_intr_rx, check for too-short packets before calling ieee80211_find_rxnode. I believe this will prevent a repeat of the MCHK exception I saw once on macppc. [] Use seconds-elapased as well as microseconds-elapsed to set the next "due date" for the timeout interrupt. This keeps the driver from programming the timeout to expire too early. [] In rtw_intr, read RTW_ISR at most 10 times, then get out. If the interface is not enabled (RTW_F_ENABLED), then get out. [] In rtw_stop, get out if the interface is not enabled (RTW_F_ENABLED). Block IPL_NET interrupts. Don't read/write any registers if the interface is invalid (RTW_F_INVALID). [] Call rtw_stop in rtw_detach.
2004-12-12 09:37:59 +03:00
cfg3 |= RTW_CONFIG3_CLKRUNEN;
if (enable)
cfg3 |= RTW_CONFIG3_PARMEN;
else
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
cfg3 &= ~RTW_CONFIG3_PARMEN;
RTW_WRITE8(regs, RTW_CONFIG3, cfg3);
RTW_SYNC(regs, RTW_CONFIG3, RTW_CONFIG3);
}
/* requires rtw_anaparm_enable(, 1) */
void
At last, I have rtw w/ Philips RF receiving packets. I added some sysctls to aid debugging: * hw.rtw.debug -- enable debugging * hw.rtw.flush_rfio -- Linux voodoo: possibly makes the MAC "flush" bits down the serial bus to the RF * hw.rtw.host_rfio: force the host to bang bits to the RF, instead of the MAC banging bits * hw.rtw.rfio_delay: after telling the MAC to bang bits to the RF front-end, delay rfio_delay microseconds. * hw.rtw.rfprog_fallback: there is this notion of the "RF programming method." I believe the choice influences the polarity/timing of the serial bus used to program the RF front-end. I know the correct choice for Intersil/RFMD/Philips front-ends, only. For all other front-ends, I "fallback" to rfprog_fallback. Make rtw_txdac_enable take an rtw_softc argument. I will probably revert this change. Add some Linux voodoo to rtw_continuous_tx_enable. I will probably revert this change. Important: add rtw_set_rfprog, which sets the correct RF programming method. This change and the following change are probably responsible for making the Philips RF work. Important: RTW_CONFIG1 is an 8-bit register, treat it that way! Important: RTW_BRSR is 16-bit, RTW_CRCOUNT, RTW_PHYDELAY, and RTW_MSR are 8-bit: treat them that way! Vastly simplify rtw_resume_ticks. Note to self: set the LED state to match the power state. Hedge against the possibility that RTW_MSR is protected as RTW_CONFIG[0123] are, meanwhile reworking that section of rtw_init a little. Add sc_anaparm, which isn't used, yet....
2004-12-13 03:48:02 +03:00
rtw_txdac_enable(struct rtw_softc *sc, int enable)
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
{
Cosmetic: change u_int{8,16,32,64}_t to C99-style uint{8,16,32,64}_t, which is preferred.
2005-01-02 07:23:03 +03:00
uint32_t anaparm;
At last, I have rtw w/ Philips RF receiving packets. I added some sysctls to aid debugging: * hw.rtw.debug -- enable debugging * hw.rtw.flush_rfio -- Linux voodoo: possibly makes the MAC "flush" bits down the serial bus to the RF * hw.rtw.host_rfio: force the host to bang bits to the RF, instead of the MAC banging bits * hw.rtw.rfio_delay: after telling the MAC to bang bits to the RF front-end, delay rfio_delay microseconds. * hw.rtw.rfprog_fallback: there is this notion of the "RF programming method." I believe the choice influences the polarity/timing of the serial bus used to program the RF front-end. I know the correct choice for Intersil/RFMD/Philips front-ends, only. For all other front-ends, I "fallback" to rfprog_fallback. Make rtw_txdac_enable take an rtw_softc argument. I will probably revert this change. Add some Linux voodoo to rtw_continuous_tx_enable. I will probably revert this change. Important: add rtw_set_rfprog, which sets the correct RF programming method. This change and the following change are probably responsible for making the Philips RF work. Important: RTW_CONFIG1 is an 8-bit register, treat it that way! Important: RTW_BRSR is 16-bit, RTW_CRCOUNT, RTW_PHYDELAY, and RTW_MSR are 8-bit: treat them that way! Vastly simplify rtw_resume_ticks. Note to self: set the LED state to match the power state. Hedge against the possibility that RTW_MSR is protected as RTW_CONFIG[0123] are, meanwhile reworking that section of rtw_init a little. Add sc_anaparm, which isn't used, yet....
2004-12-13 03:48:02 +03:00
struct rtw_regs *regs = &sc->sc_regs;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
anaparm = RTW_READ(regs, RTW_ANAPARM);
if (enable)
anaparm &= ~RTW_ANAPARM_TXDACOFF;
else
anaparm |= RTW_ANAPARM_TXDACOFF;
RTW_WRITE(regs, RTW_ANAPARM, anaparm);
RTW_SYNC(regs, RTW_ANAPARM, RTW_ANAPARM);
}
Remove leading __ from __(const|inline|signed|volatile) -- it is obsolete.
2005-12-24 23:27:29 +03:00
static inline int
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
rtw_chip_reset1(struct rtw_regs *regs, device_t dev)
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
{
Cosmetic: change u_int{8,16,32,64}_t to C99-style uint{8,16,32,64}_t, which is preferred.
2005-01-02 07:23:03 +03:00
uint8_t cr;
Miscellaneous changes. Details below. Important changes flagged with []. Using the driver with my Linksys WPC11 ver. 4, it seems to be receiving packets for a change. The WPC11 ver. 4 has a Maxim RF section. My no-name rtw with Philips RF section still does not receive any packets. Keep access-level (analog params > config[0123] registers > none) in sc_access. Add rtw_set_access for changing the access level. Make rtw_continuous_tx_enable and other subroutines use rtw_set_access instead of rtw_config0123_enable and rtw_anaparm_enable. Factor part of the chip-reset code into rtw_chip_reset1. Change the 'struct foo (*bar)[N]'-style arguments to 'struct foo *bar'-style arguments. Consolidate software/hardware Tx/Rx ring setup in rtw_hwring_setup, rtw_swring_setup. Add a new constant, SA2400_OPMODE_DEFAULTS, for the bits that we *always* set in the SA2400 OPMODE register. Factor some code out into rtw_sa2400_calibrate. (Inspired by the Linux driver.) [] When the receiver goes into underrun/overflow state, call a new subroutine, rtw_kick() that stops the Rx/Tx processes, resets the chip, reinitializes the Tx/Rx rings, and restarts Rx/Tx processes. (Inspired by the Linux driver.) [] In rtw_intr_rx, check for too-short packets before calling ieee80211_find_rxnode. I believe this will prevent a repeat of the MCHK exception I saw once on macppc. [] Use seconds-elapased as well as microseconds-elapsed to set the next "due date" for the timeout interrupt. This keeps the driver from programming the timeout to expire too early. [] In rtw_intr, read RTW_ISR at most 10 times, then get out. If the interface is not enabled (RTW_F_ENABLED), then get out. [] In rtw_stop, get out if the interface is not enabled (RTW_F_ENABLED). Block IPL_NET interrupts. Don't read/write any registers if the interface is invalid (RTW_F_INVALID). [] Call rtw_stop in rtw_detach.
2004-12-12 09:37:59 +03:00
int i;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
RTW_WRITE8(regs, RTW_CR, RTW_CR_RST);
RTW_WBR(regs, RTW_CR, RTW_CR);
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
for (i = 0; i < 1000; i++) {
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
if ((cr = RTW_READ8(regs, RTW_CR) & RTW_CR_RST) == 0) {
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
RTW_DPRINTF(RTW_DEBUG_RESET,
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
("%s: reset in %dus\n", device_xname(dev), i));
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
return 0;
}
RTW_RBR(regs, RTW_CR, RTW_CR);
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
DELAY(10); /* 10us */
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
}
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
aprint_error_dev(dev, "reset failed\n");
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
return ETIMEDOUT;
}
Remove leading __ from __(const|inline|signed|volatile) -- it is obsolete.
2005-12-24 23:27:29 +03:00
static inline int
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
rtw_chip_reset(struct rtw_regs *regs, device_t dev)
Miscellaneous changes. Details below. Important changes flagged with []. Using the driver with my Linksys WPC11 ver. 4, it seems to be receiving packets for a change. The WPC11 ver. 4 has a Maxim RF section. My no-name rtw with Philips RF section still does not receive any packets. Keep access-level (analog params > config[0123] registers > none) in sc_access. Add rtw_set_access for changing the access level. Make rtw_continuous_tx_enable and other subroutines use rtw_set_access instead of rtw_config0123_enable and rtw_anaparm_enable. Factor part of the chip-reset code into rtw_chip_reset1. Change the 'struct foo (*bar)[N]'-style arguments to 'struct foo *bar'-style arguments. Consolidate software/hardware Tx/Rx ring setup in rtw_hwring_setup, rtw_swring_setup. Add a new constant, SA2400_OPMODE_DEFAULTS, for the bits that we *always* set in the SA2400 OPMODE register. Factor some code out into rtw_sa2400_calibrate. (Inspired by the Linux driver.) [] When the receiver goes into underrun/overflow state, call a new subroutine, rtw_kick() that stops the Rx/Tx processes, resets the chip, reinitializes the Tx/Rx rings, and restarts Rx/Tx processes. (Inspired by the Linux driver.) [] In rtw_intr_rx, check for too-short packets before calling ieee80211_find_rxnode. I believe this will prevent a repeat of the MCHK exception I saw once on macppc. [] Use seconds-elapased as well as microseconds-elapsed to set the next "due date" for the timeout interrupt. This keeps the driver from programming the timeout to expire too early. [] In rtw_intr, read RTW_ISR at most 10 times, then get out. If the interface is not enabled (RTW_F_ENABLED), then get out. [] In rtw_stop, get out if the interface is not enabled (RTW_F_ENABLED). Block IPL_NET interrupts. Don't read/write any registers if the interface is invalid (RTW_F_INVALID). [] Call rtw_stop in rtw_detach.
2004-12-12 09:37:59 +03:00
{
uint32_t tcr;
/* from Linux driver */
tcr = RTW_TCR_CWMIN | RTW_TCR_MXDMA_2048 |
Per discussion on tech-kern and tech-userlevel, move the bit-twiddling macros, __BIT, __BITS, SHIFTIN, SHIFTOUT, and __arraycount() from lib/libkern/libkern.h to sys/cdefs.h. Add a __-prefix to SHIFTIN and SHIFTOUT, and add a manual page for the bit-twiddling macros, bits(3). Make the __BIT and __BITS macros "widthless," as best I can, by changing their type to uintmax_t from uint32_t. XXX The manual page lags this change by a bit. Define __PRIxBIT and __PRIxBITS printf(3) format strings.
2006-08-31 23:24:37 +04:00
__SHIFTIN(7, RTW_TCR_SRL_MASK) | __SHIFTIN(7, RTW_TCR_LRL_MASK);
Miscellaneous changes. Details below. Important changes flagged with []. Using the driver with my Linksys WPC11 ver. 4, it seems to be receiving packets for a change. The WPC11 ver. 4 has a Maxim RF section. My no-name rtw with Philips RF section still does not receive any packets. Keep access-level (analog params > config[0123] registers > none) in sc_access. Add rtw_set_access for changing the access level. Make rtw_continuous_tx_enable and other subroutines use rtw_set_access instead of rtw_config0123_enable and rtw_anaparm_enable. Factor part of the chip-reset code into rtw_chip_reset1. Change the 'struct foo (*bar)[N]'-style arguments to 'struct foo *bar'-style arguments. Consolidate software/hardware Tx/Rx ring setup in rtw_hwring_setup, rtw_swring_setup. Add a new constant, SA2400_OPMODE_DEFAULTS, for the bits that we *always* set in the SA2400 OPMODE register. Factor some code out into rtw_sa2400_calibrate. (Inspired by the Linux driver.) [] When the receiver goes into underrun/overflow state, call a new subroutine, rtw_kick() that stops the Rx/Tx processes, resets the chip, reinitializes the Tx/Rx rings, and restarts Rx/Tx processes. (Inspired by the Linux driver.) [] In rtw_intr_rx, check for too-short packets before calling ieee80211_find_rxnode. I believe this will prevent a repeat of the MCHK exception I saw once on macppc. [] Use seconds-elapased as well as microseconds-elapsed to set the next "due date" for the timeout interrupt. This keeps the driver from programming the timeout to expire too early. [] In rtw_intr, read RTW_ISR at most 10 times, then get out. If the interface is not enabled (RTW_F_ENABLED), then get out. [] In rtw_stop, get out if the interface is not enabled (RTW_F_ENABLED). Block IPL_NET interrupts. Don't read/write any registers if the interface is invalid (RTW_F_INVALID). [] Call rtw_stop in rtw_detach.
2004-12-12 09:37:59 +03:00
RTW_WRITE(regs, RTW_TCR, tcr);
RTW_WBW(regs, RTW_CR, RTW_TCR);
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
return rtw_chip_reset1(regs, dev);
Miscellaneous changes. Details below. Important changes flagged with []. Using the driver with my Linksys WPC11 ver. 4, it seems to be receiving packets for a change. The WPC11 ver. 4 has a Maxim RF section. My no-name rtw with Philips RF section still does not receive any packets. Keep access-level (analog params > config[0123] registers > none) in sc_access. Add rtw_set_access for changing the access level. Make rtw_continuous_tx_enable and other subroutines use rtw_set_access instead of rtw_config0123_enable and rtw_anaparm_enable. Factor part of the chip-reset code into rtw_chip_reset1. Change the 'struct foo (*bar)[N]'-style arguments to 'struct foo *bar'-style arguments. Consolidate software/hardware Tx/Rx ring setup in rtw_hwring_setup, rtw_swring_setup. Add a new constant, SA2400_OPMODE_DEFAULTS, for the bits that we *always* set in the SA2400 OPMODE register. Factor some code out into rtw_sa2400_calibrate. (Inspired by the Linux driver.) [] When the receiver goes into underrun/overflow state, call a new subroutine, rtw_kick() that stops the Rx/Tx processes, resets the chip, reinitializes the Tx/Rx rings, and restarts Rx/Tx processes. (Inspired by the Linux driver.) [] In rtw_intr_rx, check for too-short packets before calling ieee80211_find_rxnode. I believe this will prevent a repeat of the MCHK exception I saw once on macppc. [] Use seconds-elapased as well as microseconds-elapsed to set the next "due date" for the timeout interrupt. This keeps the driver from programming the timeout to expire too early. [] In rtw_intr, read RTW_ISR at most 10 times, then get out. If the interface is not enabled (RTW_F_ENABLED), then get out. [] In rtw_stop, get out if the interface is not enabled (RTW_F_ENABLED). Block IPL_NET interrupts. Don't read/write any registers if the interface is invalid (RTW_F_INVALID). [] Call rtw_stop in rtw_detach.
2004-12-12 09:37:59 +03:00
}
Here is a grotty hack that lets the RTL8180L handle WEP in the receive direction, while software handles WEP in the transmit direction. When net80211 calls rtw's rtw_key_set with a WEP key, I point the key's wk_cipher at our "fake" cipher, rtw_cipher_wep, which is alike to ieee80211_cipher_wep except it provides a different crypto-decapsulation routine, rtw_wep_decap. rtw_wep_decap copies the key passed to it by net80211, clears the key's SWCRYPT flag, and then calls wep_decap. Now wep_decap will decapsulate, but it will *not* re-decrypt. XXX I need to check whether the hardware supports 40-bit WEP, XXX 104-bit WEP, or both, and act accordingly.
2005-06-28 11:19:33 +04:00
static int
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
rtw_wep_decap(struct ieee80211_key *k, struct mbuf *m, int hdrlen)
Here is a grotty hack that lets the RTL8180L handle WEP in the receive direction, while software handles WEP in the transmit direction. When net80211 calls rtw's rtw_key_set with a WEP key, I point the key's wk_cipher at our "fake" cipher, rtw_cipher_wep, which is alike to ieee80211_cipher_wep except it provides a different crypto-decapsulation routine, rtw_wep_decap. rtw_wep_decap copies the key passed to it by net80211, clears the key's SWCRYPT flag, and then calls wep_decap. Now wep_decap will decapsulate, but it will *not* re-decrypt. XXX I need to check whether the hardware supports 40-bit WEP, XXX 104-bit WEP, or both, and act accordingly.
2005-06-28 11:19:33 +04:00
{
struct ieee80211_key keycopy;
RTW_DPRINTF(RTW_DEBUG_KEY, ("%s:\n", __func__));
keycopy = *k;
keycopy.wk_flags &= ~IEEE80211_KEY_SWCRYPT;
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
return (*ieee80211_cipher_wep.ic_decap)(&keycopy, m, hdrlen);
Here is a grotty hack that lets the RTL8180L handle WEP in the receive direction, while software handles WEP in the transmit direction. When net80211 calls rtw's rtw_key_set with a WEP key, I point the key's wk_cipher at our "fake" cipher, rtw_cipher_wep, which is alike to ieee80211_cipher_wep except it provides a different crypto-decapsulation routine, rtw_wep_decap. rtw_wep_decap copies the key passed to it by net80211, clears the key's SWCRYPT flag, and then calls wep_decap. Now wep_decap will decapsulate, but it will *not* re-decrypt. XXX I need to check whether the hardware supports 40-bit WEP, XXX 104-bit WEP, or both, and act accordingly.
2005-06-28 11:19:33 +04:00
}
rtw: Try to get hardware WEP to work. It seems to work in the RX direction, but not in the TX direction. The net80211 crypto framework doesn't seem to cope very well with the assymetry (I'm probably missing something), so I will use software WEP for now. net80211: In ieee80211_compute_duration, figure out whether to add the WEP header to the packet overhead by checking the WEP bit in the Frame Control field of the 802.11 header, instead of checking the IEEE80211_F_PRIVACY flag. Also, if the WEP bit is present, assume that the frame described by (wh, len) has already already been WEP encapsulated, and adjust the payload length accordingly. XXX that's a grotty hack that I will have to revisit, later.
2005-06-27 09:49:13 +04:00
static int
rtw_key_delete(struct ieee80211com *ic, const struct ieee80211_key *k)
{
struct rtw_softc *sc = ic->ic_ifp->if_softc;
Change tests for "impossible" keys to KASSERT()s. Get rid of a temporary variable. In rtw_key_delete(), do not invalidate the hardware keys unless a WEP key is deleted.
2008-03-15 02:04:42 +03:00
DPRINTF(sc, RTW_DEBUG_KEY, ("%s: delete key %u\n", __func__,
k->wk_keyix));
rtw: Try to get hardware WEP to work. It seems to work in the RX direction, but not in the TX direction. The net80211 crypto framework doesn't seem to cope very well with the assymetry (I'm probably missing something), so I will use software WEP for now. net80211: In ieee80211_compute_duration, figure out whether to add the WEP header to the packet overhead by checking the WEP bit in the Frame Control field of the 802.11 header, instead of checking the IEEE80211_F_PRIVACY flag. Also, if the WEP bit is present, assume that the frame described by (wh, len) has already already been WEP encapsulated, and adjust the payload length accordingly. XXX that's a grotty hack that I will have to revisit, later.
2005-06-27 09:49:13 +04:00
Change tests for "impossible" keys to KASSERT()s. Get rid of a temporary variable. In rtw_key_delete(), do not invalidate the hardware keys unless a WEP key is deleted.
2008-03-15 02:04:42 +03:00
KASSERT(k->wk_keyix < IEEE80211_WEP_NKID);
if (k->wk_keylen != 0 &&
k->wk_cipher->ic_cipher == IEEE80211_CIPHER_WEP)
rtw: Try to get hardware WEP to work. It seems to work in the RX direction, but not in the TX direction. The net80211 crypto framework doesn't seem to cope very well with the assymetry (I'm probably missing something), so I will use software WEP for now. net80211: In ieee80211_compute_duration, figure out whether to add the WEP header to the packet overhead by checking the WEP bit in the Frame Control field of the 802.11 header, instead of checking the IEEE80211_F_PRIVACY flag. Also, if the WEP bit is present, assume that the frame described by (wh, len) has already already been WEP encapsulated, and adjust the payload length accordingly. XXX that's a grotty hack that I will have to revisit, later.
2005-06-27 09:49:13 +04:00
sc->sc_flags &= ~RTW_F_DK_VALID;
return 1;
}
static int
rtw_key_set(struct ieee80211com *ic, const struct ieee80211_key *k,
__unused removal on arguments; approved by core.
2006-11-16 04:32:37 +03:00
const u_int8_t mac[IEEE80211_ADDR_LEN])
rtw: Try to get hardware WEP to work. It seems to work in the RX direction, but not in the TX direction. The net80211 crypto framework doesn't seem to cope very well with the assymetry (I'm probably missing something), so I will use software WEP for now. net80211: In ieee80211_compute_duration, figure out whether to add the WEP header to the packet overhead by checking the WEP bit in the Frame Control field of the 802.11 header, instead of checking the IEEE80211_F_PRIVACY flag. Also, if the WEP bit is present, assume that the frame described by (wh, len) has already already been WEP encapsulated, and adjust the payload length accordingly. XXX that's a grotty hack that I will have to revisit, later.
2005-06-27 09:49:13 +04:00
{
struct rtw_softc *sc = ic->ic_ifp->if_softc;
DPRINTF(sc, RTW_DEBUG_KEY, ("%s: set key %u\n", __func__, k->wk_keyix));
Change tests for "impossible" keys to KASSERT()s. Get rid of a temporary variable. In rtw_key_delete(), do not invalidate the hardware keys unless a WEP key is deleted.
2008-03-15 02:04:42 +03:00
KASSERT(k->wk_keyix < IEEE80211_WEP_NKID);
rtw: Try to get hardware WEP to work. It seems to work in the RX direction, but not in the TX direction. The net80211 crypto framework doesn't seem to cope very well with the assymetry (I'm probably missing something), so I will use software WEP for now. net80211: In ieee80211_compute_duration, figure out whether to add the WEP header to the packet overhead by checking the WEP bit in the Frame Control field of the 802.11 header, instead of checking the IEEE80211_F_PRIVACY flag. Also, if the WEP bit is present, assume that the frame described by (wh, len) has already already been WEP encapsulated, and adjust the payload length accordingly. XXX that's a grotty hack that I will have to revisit, later.
2005-06-27 09:49:13 +04:00
sc->sc_flags &= ~RTW_F_DK_VALID;
return 1;
}
static void
__unused removal on arguments; approved by core.
2006-11-16 04:32:37 +03:00
rtw_key_update_begin(struct ieee80211com *ic)
rtw: Try to get hardware WEP to work. It seems to work in the RX direction, but not in the TX direction. The net80211 crypto framework doesn't seem to cope very well with the assymetry (I'm probably missing something), so I will use software WEP for now. net80211: In ieee80211_compute_duration, figure out whether to add the WEP header to the packet overhead by checking the WEP bit in the Frame Control field of the 802.11 header, instead of checking the IEEE80211_F_PRIVACY flag. Also, if the WEP bit is present, assume that the frame described by (wh, len) has already already been WEP encapsulated, and adjust the payload length accordingly. XXX that's a grotty hack that I will have to revisit, later.
2005-06-27 09:49:13 +04:00
{
With the assistance of dyoung, fix the following problems: * some 8180L cards need a longer time for the srom to initialize; patch suggested via OpenBSD. * some cards give '0' for USA, some give '3'; for the meantime, accept both as being USA locale cards. * this is RTW, not ATW. :)
2005-09-07 09:25:41 +04:00
#ifdef RTW_DEBUG
rtw: Try to get hardware WEP to work. It seems to work in the RX direction, but not in the TX direction. The net80211 crypto framework doesn't seem to cope very well with the assymetry (I'm probably missing something), so I will use software WEP for now. net80211: In ieee80211_compute_duration, figure out whether to add the WEP header to the packet overhead by checking the WEP bit in the Frame Control field of the 802.11 header, instead of checking the IEEE80211_F_PRIVACY flag. Also, if the WEP bit is present, assume that the frame described by (wh, len) has already already been WEP encapsulated, and adjust the payload length accordingly. XXX that's a grotty hack that I will have to revisit, later.
2005-06-27 09:49:13 +04:00
struct ifnet *ifp = ic->ic_ifp;
struct rtw_softc *sc = ifp->if_softc;
#endif
DPRINTF(sc, RTW_DEBUG_KEY, ("%s:\n", __func__));
}
Extract subroutine rtw_tx_kick() from rtw_start(), and reuse it in rtw_io_enable() to restart the transmitter after resetting it.
2008-03-15 02:59:01 +03:00
static void
rtw_tx_kick(struct rtw_regs *regs, uint8_t ringsel)
{
uint8_t tppoll;
tppoll = RTW_READ8(regs, RTW_TPPOLL);
tppoll &= ~RTW_TPPOLL_SALL;
tppoll |= ringsel & RTW_TPPOLL_ALL;
RTW_WRITE8(regs, RTW_TPPOLL, tppoll);
RTW_SYNC(regs, RTW_TPPOLL, RTW_TPPOLL);
}
rtw: Try to get hardware WEP to work. It seems to work in the RX direction, but not in the TX direction. The net80211 crypto framework doesn't seem to cope very well with the assymetry (I'm probably missing something), so I will use software WEP for now. net80211: In ieee80211_compute_duration, figure out whether to add the WEP header to the packet overhead by checking the WEP bit in the Frame Control field of the 802.11 header, instead of checking the IEEE80211_F_PRIVACY flag. Also, if the WEP bit is present, assume that the frame described by (wh, len) has already already been WEP encapsulated, and adjust the payload length accordingly. XXX that's a grotty hack that I will have to revisit, later.
2005-06-27 09:49:13 +04:00
static void
rtw_key_update_end(struct ieee80211com *ic)
{
struct ifnet *ifp = ic->ic_ifp;
struct rtw_softc *sc = ifp->if_softc;
DPRINTF(sc, RTW_DEBUG_KEY, ("%s:\n", __func__));
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
if ((sc->sc_flags & RTW_F_DK_VALID) != 0 ||
Use device_t and its accessors throughout. Use aprint_*_dev(). Improve PMF-ability. Add a 'flags' argument to suspend/resume handlers and callers such as pmf_system_suspend(). Define a flag, PMF_F_SELF, which indicates to PMF that a device is suspending/resuming itself. Add helper routines, pmf_device_suspend_self(dev) and pmf_device_resume_self(dev), that call pmf_device_suspend(dev, PMF_F_SELF) and pmf_device_resume(dev, PMF_F_SELF), respectively. Use PMF_F_SELF to suspend/resume self in ath(4), audio(4), rtw(4), and sip(4). In ath(4) and in rtw(4), replace the icky sc_enable/sc_disable callbacks, provided by the bus front-end, with self-suspension/resumption. Also, clean up the bus front-ends. Make sure that the interrupt handler is disestablished during suspension. Get rid of driver-private flags (e.g., RTW_F_ENABLED, ath_softc->sc_invalid); use device_is_active()/device_has_power() calls, instead. In the network-class suspend handler, call if_stop(, 0) instead of if_stop(, 1), because the latter is superfluous (bus- and driver-suspension hooks will 'disable' the NIC), and it may cause recursion. In the network-class resume handler, prevent infinite recursion through if_init() by getting out early if we are self-suspending (PMF_F_SELF). rtw(4) improvements: Destroy rtw(4) callouts when we detach it. Make rtw at pci detachable. Print some more information with the "rx frame too long" warning. Remove activate() methods: Get rid of rtw_activate() and ath_activate(). The device activate() methods are not good for much these days. Make ath at cardbus resume with crypto functions intact: Introduce a boolean device property, "pmf-powerdown". If pmf-powerdown is present and false, it indicates that a bus back-end should not remove power from a device. Honor this property in cardbus_child_suspend(). Set this property to 'false' in ath_attach(), since removing power from an ath at cardbus seems to lobotomize the WPA crypto engine. XXX Should the pmf-powerdown property propagate toward the root of the device tree? Miscellaneous ath(4) changes: Warn if ath(4) tries to write crypto keys to suspended hardware. Reduce differences between FreeBSD and NetBSD in ath(4) multicast filter setup. Make ath_printrxbuf() print an rx descriptor's status & key index, to help debug crypto errors. Shorten a staircase in ath_ioctl(). Don't check for ieee80211_ioctl() return code ERESTART, it never happens.
2008-03-12 21:02:21 +03:00
!device_is_active(sc->sc_dev))
rtw: Try to get hardware WEP to work. It seems to work in the RX direction, but not in the TX direction. The net80211 crypto framework doesn't seem to cope very well with the assymetry (I'm probably missing something), so I will use software WEP for now. net80211: In ieee80211_compute_duration, figure out whether to add the WEP header to the packet overhead by checking the WEP bit in the Frame Control field of the 802.11 header, instead of checking the IEEE80211_F_PRIVACY flag. Also, if the WEP bit is present, assume that the frame described by (wh, len) has already already been WEP encapsulated, and adjust the payload length accordingly. XXX that's a grotty hack that I will have to revisit, later.
2005-06-27 09:49:13 +04:00
return;
I have made a small change that suppresses "recovery" from (spurious?) tx timeouts on the beacon queue. The driver's recovery process badly disrupts the MAC's receiver. This needs a closer look, later. Every other fix is by Charles M. Hannum. Charles writes: "With these, the device is quite stable under load, does not experience xmit timeouts, no longer mysteriously freezes, and does not spew a lot of garbage. It does, however, continue to get some "tx fifo underflow"s. global, SYSCTL_SETUP(): a) Removed unused sysctl variables. rtw_io_enable(): b) Change rtw_io_enable to take a softc, and make it call rtw_txring_fixup(). c) Enable the PCI multiple read/write bit. d) Always reset rdb_next to 0 when starting the receive engine. rtw_intr_rx(): e) When processing receive packets, store the next pointer in rdb_next and read it back on the next round, to fix a race condition vs. rtw_tune() during AP discovery and association. f) Remove the special "skip to 0th descriptor" code, because it is no longer needed. rtw_collect_txring(): g) In the "force" case, stop when we catch up to the hardware xmit pointer. rtw_intr_beacon(): h) Resetting tdb_next here looked fishy to me, in light of changes in rtw_collect_txring(). XXX However, this is the one part I haven't tested. rtw_txring_fixup(): i) Add function and line arguments, so we can print out information about the caller. j) In addition to reading the hardware pointer, skip over any additional packets with OWN set -- we may have queued more. rtw_rxring_fixup(): k) Remove this, because it's completely wrong. rtw_intr_ioerror(): l) Don't handle receive overruns here; handle them in the normal receive path. Also changed RTW_INTR_RX and RTW_INTR_IOERROR accordingly. m) Do not fiddle with CR; the transmit engine is still running just fine. n) Do not fiddle with IMR; it was unnecessary. o) Don't touch the receive engine at all; not necessary now that we're leaving CR alone. p) Remove the rtw_do_chip_reset case; it was dangerous and not actually used. rtw_intr(): q) Don't mask off ISR bits when we pass down the ISR value; it's not necessary, and sometimes the other bits useful in debugging output. rtw_seg_too_short(): r) Don't spew here; the caller will do it if necessary. (Stops spewage in an expected case.) rtw_rxdesc_blk: s) Minor change to make the elements have the same order as txdesc_blk. general: t) Added a bunch of debugging output, #ifdef RTW_DIAG, that was useful while investigating the ring management problems. Also moved rtw_txring_fixup() into RTW_DIAG, as I believe the transmit pointer is always kept in sync now, and the function has potential race conditions.
2007-01-06 11:27:53 +03:00
rtw_io_enable(sc, RTW_CR_RE | RTW_CR_TE, 0);
rtw: Try to get hardware WEP to work. It seems to work in the RX direction, but not in the TX direction. The net80211 crypto framework doesn't seem to cope very well with the assymetry (I'm probably missing something), so I will use software WEP for now. net80211: In ieee80211_compute_duration, figure out whether to add the WEP header to the packet overhead by checking the WEP bit in the Frame Control field of the 802.11 header, instead of checking the IEEE80211_F_PRIVACY flag. Also, if the WEP bit is present, assume that the frame described by (wh, len) has already already been WEP encapsulated, and adjust the payload length accordingly. XXX that's a grotty hack that I will have to revisit, later.
2005-06-27 09:49:13 +04:00
rtw_wep_setkeys(sc, ic->ic_nw_keys, ic->ic_def_txkey);
I have made a small change that suppresses "recovery" from (spurious?) tx timeouts on the beacon queue. The driver's recovery process badly disrupts the MAC's receiver. This needs a closer look, later. Every other fix is by Charles M. Hannum. Charles writes: "With these, the device is quite stable under load, does not experience xmit timeouts, no longer mysteriously freezes, and does not spew a lot of garbage. It does, however, continue to get some "tx fifo underflow"s. global, SYSCTL_SETUP(): a) Removed unused sysctl variables. rtw_io_enable(): b) Change rtw_io_enable to take a softc, and make it call rtw_txring_fixup(). c) Enable the PCI multiple read/write bit. d) Always reset rdb_next to 0 when starting the receive engine. rtw_intr_rx(): e) When processing receive packets, store the next pointer in rdb_next and read it back on the next round, to fix a race condition vs. rtw_tune() during AP discovery and association. f) Remove the special "skip to 0th descriptor" code, because it is no longer needed. rtw_collect_txring(): g) In the "force" case, stop when we catch up to the hardware xmit pointer. rtw_intr_beacon(): h) Resetting tdb_next here looked fishy to me, in light of changes in rtw_collect_txring(). XXX However, this is the one part I haven't tested. rtw_txring_fixup(): i) Add function and line arguments, so we can print out information about the caller. j) In addition to reading the hardware pointer, skip over any additional packets with OWN set -- we may have queued more. rtw_rxring_fixup(): k) Remove this, because it's completely wrong. rtw_intr_ioerror(): l) Don't handle receive overruns here; handle them in the normal receive path. Also changed RTW_INTR_RX and RTW_INTR_IOERROR accordingly. m) Do not fiddle with CR; the transmit engine is still running just fine. n) Do not fiddle with IMR; it was unnecessary. o) Don't touch the receive engine at all; not necessary now that we're leaving CR alone. p) Remove the rtw_do_chip_reset case; it was dangerous and not actually used. rtw_intr(): q) Don't mask off ISR bits when we pass down the ISR value; it's not necessary, and sometimes the other bits useful in debugging output. rtw_seg_too_short(): r) Don't spew here; the caller will do it if necessary. (Stops spewage in an expected case.) rtw_rxdesc_blk: s) Minor change to make the elements have the same order as txdesc_blk. general: t) Added a bunch of debugging output, #ifdef RTW_DIAG, that was useful while investigating the ring management problems. Also moved rtw_txring_fixup() into RTW_DIAG, as I believe the transmit pointer is always kept in sync now, and the function has potential race conditions.
2007-01-06 11:27:53 +03:00
rtw_io_enable(sc, RTW_CR_RE | RTW_CR_TE,
rtw: Try to get hardware WEP to work. It seems to work in the RX direction, but not in the TX direction. The net80211 crypto framework doesn't seem to cope very well with the assymetry (I'm probably missing something), so I will use software WEP for now. net80211: In ieee80211_compute_duration, figure out whether to add the WEP header to the packet overhead by checking the WEP bit in the Frame Control field of the 802.11 header, instead of checking the IEEE80211_F_PRIVACY flag. Also, if the WEP bit is present, assume that the frame described by (wh, len) has already already been WEP encapsulated, and adjust the payload length accordingly. XXX that's a grotty hack that I will have to revisit, later.
2005-06-27 09:49:13 +04:00
(ifp->if_flags & IFF_RUNNING) != 0);
}
Cosmetic: make rtw_hw_keysupp() return a bool instead of int. In rtw_wep_setkeys(), add a debug printf, and move the initialization of keylen from the for-loop to the declaration.
2008-03-15 03:21:12 +03:00
static bool
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
rtw_key_hwsupp(uint32_t flags, const struct ieee80211_key *k)
{
if (k->wk_cipher->ic_cipher != IEEE80211_CIPHER_WEP)
Cosmetic: make rtw_hw_keysupp() return a bool instead of int. In rtw_wep_setkeys(), add a debug printf, and move the initialization of keylen from the for-loop to the declaration.
2008-03-15 03:21:12 +03:00
return false;
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
return ((flags & RTW_C_RXWEP_40) != 0 && k->wk_keylen == 5) ||
((flags & RTW_C_RXWEP_104) != 0 && k->wk_keylen == 13);
}
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
static void
Resolve conflicts in importation of 18-May-2005 ath(4) / net80211(9) from FreeBSD. Introduce compatibility shims (sys/dev/ic/ath_netbsd.[ch], sys/net80211/ieee80211_netbsd.[ch]). Update drivers (an, atu, atw, awi, ipw, iwi, rtw, wi) for the new net80211(9) API.
2005-06-22 10:14:51 +04:00
rtw_wep_setkeys(struct rtw_softc *sc, struct ieee80211_key *wk, int txkey)
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
{
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
uint8_t psr, scr;
Cosmetic: make rtw_hw_keysupp() return a bool instead of int. In rtw_wep_setkeys(), add a debug printf, and move the initialization of keylen from the for-loop to the declaration.
2008-03-15 03:21:12 +03:00
int i, keylen = 0;
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
struct rtw_regs *regs;
union rtw_keys *rk;
regs = &sc->sc_regs;
rk = &sc->sc_keys;
Fix memset usage.
2011-07-04 20:06:17 +04:00
(void)memset(rk, 0, sizeof(*rk));
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
/* Temporarily use software crypto for all keys. */
for (i = 0; i < IEEE80211_WEP_NKID; i++) {
if (wk[i].wk_cipher == &rtw_cipher_wep)
wk[i].wk_cipher = &ieee80211_cipher_wep;
}
Here is a grotty hack that lets the RTL8180L handle WEP in the receive direction, while software handles WEP in the transmit direction. When net80211 calls rtw's rtw_key_set with a WEP key, I point the key's wk_cipher at our "fake" cipher, rtw_cipher_wep, which is alike to ieee80211_cipher_wep except it provides a different crypto-decapsulation routine, rtw_wep_decap. rtw_wep_decap copies the key passed to it by net80211, clears the key's SWCRYPT flag, and then calls wep_decap. Now wep_decap will decapsulate, but it will *not* re-decrypt. XXX I need to check whether the hardware supports 40-bit WEP, XXX 104-bit WEP, or both, and act accordingly.
2005-06-28 11:19:33 +04:00
rtw_set_access(regs, RTW_ACCESS_CONFIG);
psr = RTW_READ8(regs, RTW_PSR);
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
scr = RTW_READ8(regs, RTW_SCR);
scr &= ~(RTW_SCR_KM_MASK | RTW_SCR_TXSECON | RTW_SCR_RXSECON);
if ((sc->sc_ic.ic_flags & IEEE80211_F_PRIVACY) == 0)
goto out;
Cosmetic: make rtw_hw_keysupp() return a bool instead of int. In rtw_wep_setkeys(), add a debug printf, and move the initialization of keylen from the for-loop to the declaration.
2008-03-15 03:21:12 +03:00
for (i = 0; i < IEEE80211_WEP_NKID; i++) {
When rtw_enable is called, power may have been removed and re-applied, so invalidate the WEP key registers to force us to write the keys back to the hardware. Revamp key handling: use the hardware WEP engine for 40-bit and 104-bit keys, only; use software WEP for all other key lengths. Set the hardware key length to the length of the default Tx key, or if there is no default Tx key, use the length of the longest WEP key.
2006-02-19 11:10:06 +03:00
if (!rtw_key_hwsupp(sc->sc_flags, &wk[i]))
continue;
if (i == txkey) {
keylen = wk[i].wk_keylen;
break;
}
keylen = MAX(keylen, wk[i].wk_keylen);
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
}
When rtw_enable is called, power may have been removed and re-applied, so invalidate the WEP key registers to force us to write the keys back to the hardware. Revamp key handling: use the hardware WEP engine for 40-bit and 104-bit keys, only; use software WEP for all other key lengths. Set the hardware key length to the length of the default Tx key, or if there is no default Tx key, use the length of the longest WEP key.
2006-02-19 11:10:06 +03:00
if (keylen == 5)
scr |= RTW_SCR_KM_WEP40 | RTW_SCR_RXSECON;
else if (keylen == 13)
scr |= RTW_SCR_KM_WEP104 | RTW_SCR_RXSECON;
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
rtw: Try to get hardware WEP to work. It seems to work in the RX direction, but not in the TX direction. The net80211 crypto framework doesn't seem to cope very well with the assymetry (I'm probably missing something), so I will use software WEP for now. net80211: In ieee80211_compute_duration, figure out whether to add the WEP header to the packet overhead by checking the WEP bit in the Frame Control field of the 802.11 header, instead of checking the IEEE80211_F_PRIVACY flag. Also, if the WEP bit is present, assume that the frame described by (wh, len) has already already been WEP encapsulated, and adjust the payload length accordingly. XXX that's a grotty hack that I will have to revisit, later.
2005-06-27 09:49:13 +04:00
for (i = 0; i < IEEE80211_WEP_NKID; i++) {
When rtw_enable is called, power may have been removed and re-applied, so invalidate the WEP key registers to force us to write the keys back to the hardware. Revamp key handling: use the hardware WEP engine for 40-bit and 104-bit keys, only; use software WEP for all other key lengths. Set the hardware key length to the length of the default Tx key, or if there is no default Tx key, use the length of the longest WEP key.
2006-02-19 11:10:06 +03:00
if (wk[i].wk_keylen != keylen ||
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
wk[i].wk_cipher->ic_cipher != IEEE80211_CIPHER_WEP)
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
continue;
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
/* h/w will decrypt, s/w still strips headers */
wk[i].wk_cipher = &rtw_cipher_wep;
rtw: Try to get hardware WEP to work. It seems to work in the RX direction, but not in the TX direction. The net80211 crypto framework doesn't seem to cope very well with the assymetry (I'm probably missing something), so I will use software WEP for now. net80211: In ieee80211_compute_duration, figure out whether to add the WEP header to the packet overhead by checking the WEP bit in the Frame Control field of the 802.11 header, instead of checking the IEEE80211_F_PRIVACY flag. Also, if the WEP bit is present, assume that the frame described by (wh, len) has already already been WEP encapsulated, and adjust the payload length accordingly. XXX that's a grotty hack that I will have to revisit, later.
2005-06-27 09:49:13 +04:00
(void)memcpy(rk->rk_keys[i], wk[i].wk_key, wk[i].wk_keylen);
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
}
out:
Here is a grotty hack that lets the RTL8180L handle WEP in the receive direction, while software handles WEP in the transmit direction. When net80211 calls rtw's rtw_key_set with a WEP key, I point the key's wk_cipher at our "fake" cipher, rtw_cipher_wep, which is alike to ieee80211_cipher_wep except it provides a different crypto-decapsulation routine, rtw_wep_decap. rtw_wep_decap copies the key passed to it by net80211, clears the key's SWCRYPT flag, and then calls wep_decap. Now wep_decap will decapsulate, but it will *not* re-decrypt. XXX I need to check whether the hardware supports 40-bit WEP, XXX 104-bit WEP, or both, and act accordingly.
2005-06-28 11:19:33 +04:00
RTW_WRITE8(regs, RTW_PSR, psr & ~RTW_PSR_PSEN);
Fix rtw(4) on NetBSD/macppc: use bus_space_write_region_stream_4() instead of bus_space_write_region_4() to write WEP Rx keys to the NIC. Thanks macallan@ for detecting the bug and testing the fix.
2007-05-29 22:38:20 +04:00
bus_space_write_region_stream_4(regs->r_bt, regs->r_bh,
Cosmetic: simplify a memset() statement, use __arraycount(), and remove a gratuitous return statement at the end of a void subroutine.
2007-05-29 22:33:38 +04:00
RTW_DK0, rk->rk_words, __arraycount(rk->rk_words));
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
Here is a grotty hack that lets the RTL8180L handle WEP in the receive direction, while software handles WEP in the transmit direction. When net80211 calls rtw's rtw_key_set with a WEP key, I point the key's wk_cipher at our "fake" cipher, rtw_cipher_wep, which is alike to ieee80211_cipher_wep except it provides a different crypto-decapsulation routine, rtw_wep_decap. rtw_wep_decap copies the key passed to it by net80211, clears the key's SWCRYPT flag, and then calls wep_decap. Now wep_decap will decapsulate, but it will *not* re-decrypt. XXX I need to check whether the hardware supports 40-bit WEP, XXX 104-bit WEP, or both, and act accordingly.
2005-06-28 11:19:33 +04:00
bus_space_barrier(regs->r_bt, regs->r_bh, RTW_DK0, sizeof(rk->rk_words),
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
BUS_SPACE_BARRIER_SYNC);
Cosmetic: make rtw_hw_keysupp() return a bool instead of int. In rtw_wep_setkeys(), add a debug printf, and move the initialization of keylen from the for-loop to the declaration.
2008-03-15 03:21:12 +03:00
RTW_DPRINTF(RTW_DEBUG_KEY,
("%s.%d: scr %02" PRIx8 ", keylen %d\n", __func__, __LINE__, scr,
keylen));
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
RTW_WBW(regs, RTW_DK0, RTW_PSR);
Here is a grotty hack that lets the RTL8180L handle WEP in the receive direction, while software handles WEP in the transmit direction. When net80211 calls rtw's rtw_key_set with a WEP key, I point the key's wk_cipher at our "fake" cipher, rtw_cipher_wep, which is alike to ieee80211_cipher_wep except it provides a different crypto-decapsulation routine, rtw_wep_decap. rtw_wep_decap copies the key passed to it by net80211, clears the key's SWCRYPT flag, and then calls wep_decap. Now wep_decap will decapsulate, but it will *not* re-decrypt. XXX I need to check whether the hardware supports 40-bit WEP, XXX 104-bit WEP, or both, and act accordingly.
2005-06-28 11:19:33 +04:00
RTW_WRITE8(regs, RTW_PSR, psr);
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
RTW_WBW(regs, RTW_PSR, RTW_SCR);
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
RTW_WRITE8(regs, RTW_SCR, scr);
RTW_SYNC(regs, RTW_SCR, RTW_SCR);
rtw_set_access(regs, RTW_ACCESS_NONE);
rtw: Try to get hardware WEP to work. It seems to work in the RX direction, but not in the TX direction. The net80211 crypto framework doesn't seem to cope very well with the assymetry (I'm probably missing something), so I will use software WEP for now. net80211: In ieee80211_compute_duration, figure out whether to add the WEP header to the packet overhead by checking the WEP bit in the Frame Control field of the 802.11 header, instead of checking the IEEE80211_F_PRIVACY flag. Also, if the WEP bit is present, assume that the frame described by (wh, len) has already already been WEP encapsulated, and adjust the payload length accordingly. XXX that's a grotty hack that I will have to revisit, later.
2005-06-27 09:49:13 +04:00
sc->sc_flags |= RTW_F_DK_VALID;
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
}
Remove leading __ from __(const|inline|signed|volatile) -- it is obsolete.
2005-12-24 23:27:29 +03:00
static inline int
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
rtw_recall_eeprom(struct rtw_regs *regs, device_t dev)
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
{
int i;
Cosmetic: change u_int{8,16,32,64}_t to C99-style uint{8,16,32,64}_t, which is preferred.
2005-01-02 07:23:03 +03:00
uint8_t ecr;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
ecr = RTW_READ8(regs, RTW_9346CR);
ecr = (ecr & ~RTW_9346CR_EEM_MASK) | RTW_9346CR_EEM_AUTOLOAD;
RTW_WRITE8(regs, RTW_9346CR, ecr);
RTW_WBR(regs, RTW_9346CR, RTW_9346CR);
With the assistance of dyoung, fix the following problems: * some 8180L cards need a longer time for the srom to initialize; patch suggested via OpenBSD. * some cards give '0' for USA, some give '3'; for the meantime, accept both as being USA locale cards. * this is RTW, not ATW. :)
2005-09-07 09:25:41 +04:00
/* wait 25ms for completion */
for (i = 0; i < 250; i++) {
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
ecr = RTW_READ8(regs, RTW_9346CR);
if ((ecr & RTW_9346CR_EEM_MASK) == RTW_9346CR_EEM_NORMAL) {
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
RTW_DPRINTF(RTW_DEBUG_RESET,
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
("%s: recall EEPROM in %dus\n", device_xname(dev),
i * 100));
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
return 0;
}
RTW_RBR(regs, RTW_9346CR, RTW_9346CR);
DELAY(100);
}
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
aprint_error_dev(dev, "recall EEPROM failed\n");
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
return ETIMEDOUT;
}
Remove leading __ from __(const|inline|signed|volatile) -- it is obsolete.
2005-12-24 23:27:29 +03:00
static inline int
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
rtw_reset(struct rtw_softc *sc)
{
int rc;
At last, I have rtw w/ Philips RF receiving packets. I added some sysctls to aid debugging: * hw.rtw.debug -- enable debugging * hw.rtw.flush_rfio -- Linux voodoo: possibly makes the MAC "flush" bits down the serial bus to the RF * hw.rtw.host_rfio: force the host to bang bits to the RF, instead of the MAC banging bits * hw.rtw.rfio_delay: after telling the MAC to bang bits to the RF front-end, delay rfio_delay microseconds. * hw.rtw.rfprog_fallback: there is this notion of the "RF programming method." I believe the choice influences the polarity/timing of the serial bus used to program the RF front-end. I know the correct choice for Intersil/RFMD/Philips front-ends, only. For all other front-ends, I "fallback" to rfprog_fallback. Make rtw_txdac_enable take an rtw_softc argument. I will probably revert this change. Add some Linux voodoo to rtw_continuous_tx_enable. I will probably revert this change. Important: add rtw_set_rfprog, which sets the correct RF programming method. This change and the following change are probably responsible for making the Philips RF work. Important: RTW_CONFIG1 is an 8-bit register, treat it that way! Important: RTW_BRSR is 16-bit, RTW_CRCOUNT, RTW_PHYDELAY, and RTW_MSR are 8-bit: treat them that way! Vastly simplify rtw_resume_ticks. Note to self: set the LED state to match the power state. Hedge against the possibility that RTW_MSR is protected as RTW_CONFIG[0123] are, meanwhile reworking that section of rtw_init a little. Add sc_anaparm, which isn't used, yet....
2004-12-13 03:48:02 +03:00
uint8_t config1;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
rtw: Try to get hardware WEP to work. It seems to work in the RX direction, but not in the TX direction. The net80211 crypto framework doesn't seem to cope very well with the assymetry (I'm probably missing something), so I will use software WEP for now. net80211: In ieee80211_compute_duration, figure out whether to add the WEP header to the packet overhead by checking the WEP bit in the Frame Control field of the 802.11 header, instead of checking the IEEE80211_F_PRIVACY flag. Also, if the WEP bit is present, assume that the frame described by (wh, len) has already already been WEP encapsulated, and adjust the payload length accordingly. XXX that's a grotty hack that I will have to revisit, later.
2005-06-27 09:49:13 +04:00
sc->sc_flags &= ~RTW_F_DK_VALID;
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
if ((rc = rtw_chip_reset(&sc->sc_regs, sc->sc_dev)) != 0)
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
return rc;
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
rc = rtw_recall_eeprom(&sc->sc_regs, sc->sc_dev);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
At last, I have rtw w/ Philips RF receiving packets. I added some sysctls to aid debugging: * hw.rtw.debug -- enable debugging * hw.rtw.flush_rfio -- Linux voodoo: possibly makes the MAC "flush" bits down the serial bus to the RF * hw.rtw.host_rfio: force the host to bang bits to the RF, instead of the MAC banging bits * hw.rtw.rfio_delay: after telling the MAC to bang bits to the RF front-end, delay rfio_delay microseconds. * hw.rtw.rfprog_fallback: there is this notion of the "RF programming method." I believe the choice influences the polarity/timing of the serial bus used to program the RF front-end. I know the correct choice for Intersil/RFMD/Philips front-ends, only. For all other front-ends, I "fallback" to rfprog_fallback. Make rtw_txdac_enable take an rtw_softc argument. I will probably revert this change. Add some Linux voodoo to rtw_continuous_tx_enable. I will probably revert this change. Important: add rtw_set_rfprog, which sets the correct RF programming method. This change and the following change are probably responsible for making the Philips RF work. Important: RTW_CONFIG1 is an 8-bit register, treat it that way! Important: RTW_BRSR is 16-bit, RTW_CRCOUNT, RTW_PHYDELAY, and RTW_MSR are 8-bit: treat them that way! Vastly simplify rtw_resume_ticks. Note to self: set the LED state to match the power state. Hedge against the possibility that RTW_MSR is protected as RTW_CONFIG[0123] are, meanwhile reworking that section of rtw_init a little. Add sc_anaparm, which isn't used, yet....
2004-12-13 03:48:02 +03:00
config1 = RTW_READ8(&sc->sc_regs, RTW_CONFIG1);
RTW_WRITE8(&sc->sc_regs, RTW_CONFIG1, config1 & ~RTW_CONFIG1_PMEN);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
/* TBD turn off maximum power saving? */
return 0;
}
Remove leading __ from __(const|inline|signed|volatile) -- it is obsolete.
2005-12-24 23:27:29 +03:00
static inline int
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
rtw_txdesc_dmamaps_create(bus_dma_tag_t dmat, struct rtw_txsoft *descs,
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
u_int ndescs)
{
int i, rc = 0;
for (i = 0; i < ndescs; i++) {
rc = bus_dmamap_create(dmat, MCLBYTES, RTW_MAXPKTSEGS, MCLBYTES,
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
0, 0, &descs[i].ts_dmamap);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
if (rc != 0)
break;
}
return rc;
}
Remove leading __ from __(const|inline|signed|volatile) -- it is obsolete.
2005-12-24 23:27:29 +03:00
static inline int
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
rtw_rxdesc_dmamaps_create(bus_dma_tag_t dmat, struct rtw_rxsoft *descs,
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
u_int ndescs)
{
int i, rc = 0;
for (i = 0; i < ndescs; i++) {
rc = bus_dmamap_create(dmat, MCLBYTES, 1, MCLBYTES, 0, 0,
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
&descs[i].rs_dmamap);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
if (rc != 0)
break;
}
return rc;
}
Remove leading __ from __(const|inline|signed|volatile) -- it is obsolete.
2005-12-24 23:27:29 +03:00
static inline void
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
rtw_rxdesc_dmamaps_destroy(bus_dma_tag_t dmat, struct rtw_rxsoft *descs,
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
u_int ndescs)
{
int i;
for (i = 0; i < ndescs; i++) {
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
if (descs[i].rs_dmamap != NULL)
bus_dmamap_destroy(dmat, descs[i].rs_dmamap);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
}
}
Remove leading __ from __(const|inline|signed|volatile) -- it is obsolete.
2005-12-24 23:27:29 +03:00
static inline void
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
rtw_txdesc_dmamaps_destroy(bus_dma_tag_t dmat, struct rtw_txsoft *descs,
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
u_int ndescs)
{
int i;
for (i = 0; i < ndescs; i++) {
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
if (descs[i].ts_dmamap != NULL)
bus_dmamap_destroy(dmat, descs[i].ts_dmamap);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
}
}
Remove leading __ from __(const|inline|signed|volatile) -- it is obsolete.
2005-12-24 23:27:29 +03:00
static inline void
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
rtw_srom_free(struct rtw_srom *sr)
{
sr->sr_size = 0;
if (sr->sr_content == NULL)
return;
free(sr->sr_content, M_DEVBUF);
sr->sr_content = NULL;
}
static void
__unused removal on arguments; approved by core.
2006-11-16 04:32:37 +03:00
rtw_srom_defaults(struct rtw_srom *sr, uint32_t *flags,
- sprinkle __unused on function decls. - fix a couple of unused bugs - no more -Wno-unused for i386
2006-10-12 05:30:41 +04:00
uint8_t *cs_threshold, enum rtw_rfchipid *rfchipid, uint32_t *rcr)
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
{
*flags |= (RTW_F_DIGPHY|RTW_F_ANTDIV);
*cs_threshold = RTW_SR_ENERGYDETTHR_DEFAULT;
*rcr |= RTW_RCR_ENCS1;
*rfchipid = RTW_RFCHIPID_PHILIPS;
}
static int
Cosmetic: change u_int{8,16,32,64}_t to C99-style uint{8,16,32,64}_t, which is preferred.
2005-01-02 07:23:03 +03:00
rtw_srom_parse(struct rtw_srom *sr, uint32_t *flags, uint8_t *cs_threshold,
enum rtw_rfchipid *rfchipid, uint32_t *rcr, enum rtw_locale *locale,
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
device_t dev)
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
{
int i;
const char *rfname, *paname;
char scratch[sizeof("unknown 0xXX")];
Sprinkle __UNCONST() and rename variables to quiet uiet -Wcast-qual, -Wshadow complaints.
2005-06-10 06:35:34 +04:00
uint16_t srom_version;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
*flags &= ~(RTW_F_DIGPHY|RTW_F_DFLANTB|RTW_F_ANTDIV);
*rcr &= ~(RTW_RCR_ENCS1 | RTW_RCR_ENCS2);
Sprinkle __UNCONST() and rename variables to quiet uiet -Wcast-qual, -Wshadow complaints.
2005-06-10 06:35:34 +04:00
srom_version = RTW_SR_GET16(sr, RTW_SR_VERSION);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
Sprinkle __UNCONST() and rename variables to quiet uiet -Wcast-qual, -Wshadow complaints.
2005-06-10 06:35:34 +04:00
if (srom_version <= 0x0101) {
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
aprint_error_dev(dev,
"SROM version %d.%d is not understood, "
"limping along with defaults\n",
srom_version >> 8, srom_version & 0xff);
Get rid of the silly char (*dvname)[IFNAMSIZ] pattern. Replace it with const char *dvname. I'm not sure what I was thinking.
2004-12-20 03:28:02 +03:00
rtw_srom_defaults(sr, flags, cs_threshold, rfchipid, rcr);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
return 0;
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
} else {
Add the missing newline to a verbose autoconf message.
2010-01-31 21:12:51 +03:00
aprint_verbose_dev(dev, "SROM version %d.%d\n",
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
srom_version >> 8, srom_version & 0xff);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
}
- remove unused variables - move variables inside ifdef sections - ifdef notdef unused code - use __USE for debugging variables
2013-10-18 01:24:24 +04:00
uint8_t mac[IEEE80211_ADDR_LEN];
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
for (i = 0; i < IEEE80211_ADDR_LEN; i++)
mac[i] = RTW_SR_GET(sr, RTW_SR_MAC + i);
- remove unused variables - move variables inside ifdef sections - ifdef notdef unused code - use __USE for debugging variables
2013-10-18 01:24:24 +04:00
__USE(mac);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
RTW_DPRINTF(RTW_DEBUG_ATTACH,
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
("%s: EEPROM MAC %s\n", device_xname(dev), ether_sprintf(mac)));
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
*cs_threshold = RTW_SR_GET(sr, RTW_SR_ENERGYDETTHR);
if ((RTW_SR_GET(sr, RTW_SR_CONFIG2) & RTW_CONFIG2_ANT) != 0)
*flags |= RTW_F_ANTDIV;
Realtek sample sources are not consistent about the "Digital PHY" bit in the SROM. It seems as if it is set to 1 when the PHY is *analog*, not *digital*. Fix my sources. In rtw_intr_rx, use units of 500kb/s instead of 100kb/s for rate, to be consistent with net80211's expectations. Polish up some debugging ugly messages. Dump raw 802.11 packets if IFF_DEBUG|IFF_LINK2 and RTW_DEBUG is defined. Polish power-state (on/sleep/off) handling. Especially improve support for RFMD (totally untested) and Maxim. For Philips, take the Digital PHY property into account. Call the net80211 watchdog function from rtw_watchdog, so that we scan again if auth/assoc fails. Be a little more cautious about writing register[RTW_TPPOLL], since other drivers are.... Don't frob the high/low-priority queues right now, since I don't use them. Add rtw_join_bss which programs the card with the BSSID and other properties of a BSS. Use it on state transitions. Factor out rtw_set_nettype. Make rtw_recv_beacon call ieee80211_recv_mgmt instead of dropping beacons on the floor! TBD IBSS merges. Change some rtw_debug=2 printfs to rtw_debug=3 (RTW_DPRINTF3) printfs so the console doesn't get spammed so badly at rtw_debug=2. Change some debugging printfs to RTW_DPRINTFs. E.g., print the "RF programming method" only if debugging is enabled.
2004-12-21 02:05:41 +03:00
/* Note well: the sense of the RTW_SR_RFPARM_DIGPHY bit seems
* to be reversed.
*/
if ((RTW_SR_GET(sr, RTW_SR_RFPARM) & RTW_SR_RFPARM_DIGPHY) == 0)
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
*flags |= RTW_F_DIGPHY;
if ((RTW_SR_GET(sr, RTW_SR_RFPARM) & RTW_SR_RFPARM_DFLANTB) != 0)
*flags |= RTW_F_DFLANTB;
Per discussion on tech-kern and tech-userlevel, move the bit-twiddling macros, __BIT, __BITS, SHIFTIN, SHIFTOUT, and __arraycount() from lib/libkern/libkern.h to sys/cdefs.h. Add a __-prefix to SHIFTIN and SHIFTOUT, and add a manual page for the bit-twiddling macros, bits(3). Make the __BIT and __BITS macros "widthless," as best I can, by changing their type to uintmax_t from uint32_t. XXX The manual page lags this change by a bit. Define __PRIxBIT and __PRIxBITS printf(3) format strings.
2006-08-31 23:24:37 +04:00
*rcr |= __SHIFTIN(__SHIFTOUT(RTW_SR_GET(sr, RTW_SR_RFPARM),
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
RTW_SR_RFPARM_CS_MASK), RTW_RCR_ENCS1);
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
if ((RTW_SR_GET(sr, RTW_SR_CONFIG0) & RTW_CONFIG0_WEP104) != 0)
*flags |= RTW_C_RXWEP_104;
*flags |= RTW_C_RXWEP_40; /* XXX */
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
*rfchipid = RTW_SR_GET(sr, RTW_SR_RFCHIPID);
switch (*rfchipid) {
case RTW_RFCHIPID_GCT: /* this combo seen in the wild */
rfname = "GCT GRF5101";
paname = "Winspring WS9901";
break;
case RTW_RFCHIPID_MAXIM:
rfname = "Maxim MAX2820"; /* guess */
paname = "Maxim MAX2422"; /* guess */
break;
case RTW_RFCHIPID_INTERSIL:
rfname = "Intersil HFA3873"; /* guess */
paname = "Intersil <unknown>";
break;
case RTW_RFCHIPID_PHILIPS: /* this combo seen in the wild */
rfname = "Philips SA2400A";
paname = "Philips SA2411";
break;
case RTW_RFCHIPID_RFMD:
/* this is the same front-end as an atw(4)! */
rfname = "RFMD RF2948B, " /* mentioned in Realtek docs */
"LNA: RFMD RF2494, " /* mentioned in Realtek docs */
"SYN: Silicon Labs Si4126"; /* inferred from
* reference driver
*/
paname = "RFMD RF2189"; /* mentioned in Realtek docs */
break;
case RTW_RFCHIPID_RESERVED:
rfname = paname = "reserved";
break;
default:
snprintf(scratch, sizeof(scratch), "unknown 0x%02x", *rfchipid);
rfname = paname = scratch;
}
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
aprint_normal_dev(dev, "RF: %s, PA: %s\n", rfname, paname);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
switch (RTW_SR_GET(sr, RTW_SR_CONFIG0) & RTW_CONFIG0_GL_MASK) {
case RTW_CONFIG0_GL_USA:
With the assistance of dyoung, fix the following problems: * some 8180L cards need a longer time for the srom to initialize; patch suggested via OpenBSD. * some cards give '0' for USA, some give '3'; for the meantime, accept both as being USA locale cards. * this is RTW, not ATW. :)
2005-09-07 09:25:41 +04:00
case _RTW_CONFIG0_GL_USA:
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
*locale = RTW_LOCALE_USA;
break;
case RTW_CONFIG0_GL_EUROPE:
*locale = RTW_LOCALE_EUROPE;
break;
case RTW_CONFIG0_GL_JAPAN:
*locale = RTW_LOCALE_JAPAN;
break;
default:
*locale = RTW_LOCALE_UNKNOWN;
break;
}
return 0;
}
/* Returns -1 on failure. */
static int
Cosmetic: change u_int{8,16,32,64}_t to C99-style uint{8,16,32,64}_t, which is preferred.
2005-01-02 07:23:03 +03:00
rtw_srom_read(struct rtw_regs *regs, uint32_t flags, struct rtw_srom *sr,
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
device_t dev)
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
{
int rc;
struct seeprom_descriptor sd;
Cosmetic: change u_int{8,16,32,64}_t to C99-style uint{8,16,32,64}_t, which is preferred.
2005-01-02 07:23:03 +03:00
uint8_t ecr;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
(void)memset(&sd, 0, sizeof(sd));
ecr = RTW_READ8(regs, RTW_9346CR);
if ((flags & RTW_F_9356SROM) != 0) {
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
RTW_DPRINTF(RTW_DEBUG_ATTACH, ("%s: 93c56 SROM\n",
device_xname(dev)));
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
sr->sr_size = 256;
sd.sd_chip = C56_66;
} else {
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
RTW_DPRINTF(RTW_DEBUG_ATTACH, ("%s: 93c46 SROM\n",
device_xname(dev)));
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
sr->sr_size = 128;
sd.sd_chip = C46;
}
ecr &= ~(RTW_9346CR_EEDI | RTW_9346CR_EEDO | RTW_9346CR_EESK |
Set initial conditions to guarantee that when read_seeprom toggles Chip Select, it turns *on* rather than *off*.
2005-01-04 04:04:52 +03:00
RTW_9346CR_EEM_MASK | RTW_9346CR_EECS);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
ecr |= RTW_9346CR_EEM_PROGRAM;
RTW_WRITE8(regs, RTW_9346CR, ecr);
sr->sr_content = malloc(sr->sr_size, M_DEVBUF, M_NOWAIT);
if (sr->sr_content == NULL) {
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
aprint_error_dev(dev, "unable to allocate SROM buffer\n");
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
return ENOMEM;
}
(void)memset(sr->sr_content, 0, sr->sr_size);
/* RTL8180 has a single 8-bit register for controlling the
* 93cx6 SROM. There is no "ready" bit. The RTL8180
* input/output sense is the reverse of read_seeprom's.
*/
sd.sd_tag = regs->r_bt;
sd.sd_bsh = regs->r_bh;
sd.sd_regsize = 1;
sd.sd_control_offset = RTW_9346CR;
sd.sd_status_offset = RTW_9346CR;
sd.sd_dataout_offset = RTW_9346CR;
sd.sd_CK = RTW_9346CR_EESK;
sd.sd_CS = RTW_9346CR_EECS;
sd.sd_DI = RTW_9346CR_EEDO;
sd.sd_DO = RTW_9346CR_EEDI;
nuke trailing whitespace
2005-02-27 03:26:58 +03:00
/* make read_seeprom enter EEPROM read/write mode */
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
sd.sd_MS = ecr;
sd.sd_RDY = 0;
Add some bus barriers. Remark where they are needed, e.g., read_eeprom. Change some printfs to RTW_DPRINTF2s to quiet rtw(4) when debugging is disabled.
2004-12-20 04:13:45 +03:00
/* TBD bus barriers */
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
if (!read_seeprom(&sd, sr->sr_content, 0, sr->sr_size/2)) {
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
aprint_error_dev(dev, "could not read SROM\n");
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
free(sr->sr_content, M_DEVBUF);
sr->sr_content = NULL;
return -1; /* XXX */
}
nuke trailing whitespace
2005-02-27 03:26:58 +03:00
/* end EEPROM read/write mode */
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
RTW_WRITE8(regs, RTW_9346CR,
(ecr & ~RTW_9346CR_EEM_MASK) | RTW_9346CR_EEM_NORMAL);
RTW_WBRW(regs, RTW_9346CR, RTW_9346CR);
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
if ((rc = rtw_recall_eeprom(regs, dev)) != 0)
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
return rc;
#ifdef RTW_DEBUG
{
int i;
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
RTW_DPRINTF(RTW_DEBUG_ATTACH,
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
("\n%s: serial ROM:\n\t", device_xname(dev)));
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
for (i = 0; i < sr->sr_size/2; i++) {
if (((i % 8) == 0) && (i != 0))
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
RTW_DPRINTF(RTW_DEBUG_ATTACH, ("\n\t"));
RTW_DPRINTF(RTW_DEBUG_ATTACH,
(" %04x", sr->sr_content[i]));
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
}
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
RTW_DPRINTF(RTW_DEBUG_ATTACH, ("\n"));
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
}
#endif /* RTW_DEBUG */
return 0;
}
At last, I have rtw w/ Philips RF receiving packets. I added some sysctls to aid debugging: * hw.rtw.debug -- enable debugging * hw.rtw.flush_rfio -- Linux voodoo: possibly makes the MAC "flush" bits down the serial bus to the RF * hw.rtw.host_rfio: force the host to bang bits to the RF, instead of the MAC banging bits * hw.rtw.rfio_delay: after telling the MAC to bang bits to the RF front-end, delay rfio_delay microseconds. * hw.rtw.rfprog_fallback: there is this notion of the "RF programming method." I believe the choice influences the polarity/timing of the serial bus used to program the RF front-end. I know the correct choice for Intersil/RFMD/Philips front-ends, only. For all other front-ends, I "fallback" to rfprog_fallback. Make rtw_txdac_enable take an rtw_softc argument. I will probably revert this change. Add some Linux voodoo to rtw_continuous_tx_enable. I will probably revert this change. Important: add rtw_set_rfprog, which sets the correct RF programming method. This change and the following change are probably responsible for making the Philips RF work. Important: RTW_CONFIG1 is an 8-bit register, treat it that way! Important: RTW_BRSR is 16-bit, RTW_CRCOUNT, RTW_PHYDELAY, and RTW_MSR are 8-bit: treat them that way! Vastly simplify rtw_resume_ticks. Note to self: set the LED state to match the power state. Hedge against the possibility that RTW_MSR is protected as RTW_CONFIG[0123] are, meanwhile reworking that section of rtw_init a little. Add sc_anaparm, which isn't used, yet....
2004-12-13 03:48:02 +03:00
static void
rtw_set_rfprog(struct rtw_regs *regs, enum rtw_rfchipid rfchipid,
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
device_t dev)
At last, I have rtw w/ Philips RF receiving packets. I added some sysctls to aid debugging: * hw.rtw.debug -- enable debugging * hw.rtw.flush_rfio -- Linux voodoo: possibly makes the MAC "flush" bits down the serial bus to the RF * hw.rtw.host_rfio: force the host to bang bits to the RF, instead of the MAC banging bits * hw.rtw.rfio_delay: after telling the MAC to bang bits to the RF front-end, delay rfio_delay microseconds. * hw.rtw.rfprog_fallback: there is this notion of the "RF programming method." I believe the choice influences the polarity/timing of the serial bus used to program the RF front-end. I know the correct choice for Intersil/RFMD/Philips front-ends, only. For all other front-ends, I "fallback" to rfprog_fallback. Make rtw_txdac_enable take an rtw_softc argument. I will probably revert this change. Add some Linux voodoo to rtw_continuous_tx_enable. I will probably revert this change. Important: add rtw_set_rfprog, which sets the correct RF programming method. This change and the following change are probably responsible for making the Philips RF work. Important: RTW_CONFIG1 is an 8-bit register, treat it that way! Important: RTW_BRSR is 16-bit, RTW_CRCOUNT, RTW_PHYDELAY, and RTW_MSR are 8-bit: treat them that way! Vastly simplify rtw_resume_ticks. Note to self: set the LED state to match the power state. Hedge against the possibility that RTW_MSR is protected as RTW_CONFIG[0123] are, meanwhile reworking that section of rtw_init a little. Add sc_anaparm, which isn't used, yet....
2004-12-13 03:48:02 +03:00
{
Cosmetic: change u_int{8,16,32,64}_t to C99-style uint{8,16,32,64}_t, which is preferred.
2005-01-02 07:23:03 +03:00
uint8_t cfg4;
At last, I have rtw w/ Philips RF receiving packets. I added some sysctls to aid debugging: * hw.rtw.debug -- enable debugging * hw.rtw.flush_rfio -- Linux voodoo: possibly makes the MAC "flush" bits down the serial bus to the RF * hw.rtw.host_rfio: force the host to bang bits to the RF, instead of the MAC banging bits * hw.rtw.rfio_delay: after telling the MAC to bang bits to the RF front-end, delay rfio_delay microseconds. * hw.rtw.rfprog_fallback: there is this notion of the "RF programming method." I believe the choice influences the polarity/timing of the serial bus used to program the RF front-end. I know the correct choice for Intersil/RFMD/Philips front-ends, only. For all other front-ends, I "fallback" to rfprog_fallback. Make rtw_txdac_enable take an rtw_softc argument. I will probably revert this change. Add some Linux voodoo to rtw_continuous_tx_enable. I will probably revert this change. Important: add rtw_set_rfprog, which sets the correct RF programming method. This change and the following change are probably responsible for making the Philips RF work. Important: RTW_CONFIG1 is an 8-bit register, treat it that way! Important: RTW_BRSR is 16-bit, RTW_CRCOUNT, RTW_PHYDELAY, and RTW_MSR are 8-bit: treat them that way! Vastly simplify rtw_resume_ticks. Note to self: set the LED state to match the power state. Hedge against the possibility that RTW_MSR is protected as RTW_CONFIG[0123] are, meanwhile reworking that section of rtw_init a little. Add sc_anaparm, which isn't used, yet....
2004-12-13 03:48:02 +03:00
const char *method;
cfg4 = RTW_READ8(regs, RTW_CONFIG4) & ~RTW_CONFIG4_RFTYPE_MASK;
switch (rfchipid) {
default:
Per discussion on tech-kern and tech-userlevel, move the bit-twiddling macros, __BIT, __BITS, SHIFTIN, SHIFTOUT, and __arraycount() from lib/libkern/libkern.h to sys/cdefs.h. Add a __-prefix to SHIFTIN and SHIFTOUT, and add a manual page for the bit-twiddling macros, bits(3). Make the __BIT and __BITS macros "widthless," as best I can, by changing their type to uintmax_t from uint32_t. XXX The manual page lags this change by a bit. Define __PRIxBIT and __PRIxBITS printf(3) format strings.
2006-08-31 23:24:37 +04:00
cfg4 |= __SHIFTIN(rtw_rfprog_fallback, RTW_CONFIG4_RFTYPE_MASK);
At last, I have rtw w/ Philips RF receiving packets. I added some sysctls to aid debugging: * hw.rtw.debug -- enable debugging * hw.rtw.flush_rfio -- Linux voodoo: possibly makes the MAC "flush" bits down the serial bus to the RF * hw.rtw.host_rfio: force the host to bang bits to the RF, instead of the MAC banging bits * hw.rtw.rfio_delay: after telling the MAC to bang bits to the RF front-end, delay rfio_delay microseconds. * hw.rtw.rfprog_fallback: there is this notion of the "RF programming method." I believe the choice influences the polarity/timing of the serial bus used to program the RF front-end. I know the correct choice for Intersil/RFMD/Philips front-ends, only. For all other front-ends, I "fallback" to rfprog_fallback. Make rtw_txdac_enable take an rtw_softc argument. I will probably revert this change. Add some Linux voodoo to rtw_continuous_tx_enable. I will probably revert this change. Important: add rtw_set_rfprog, which sets the correct RF programming method. This change and the following change are probably responsible for making the Philips RF work. Important: RTW_CONFIG1 is an 8-bit register, treat it that way! Important: RTW_BRSR is 16-bit, RTW_CRCOUNT, RTW_PHYDELAY, and RTW_MSR are 8-bit: treat them that way! Vastly simplify rtw_resume_ticks. Note to self: set the LED state to match the power state. Hedge against the possibility that RTW_MSR is protected as RTW_CONFIG[0123] are, meanwhile reworking that section of rtw_init a little. Add sc_anaparm, which isn't used, yet....
2004-12-13 03:48:02 +03:00
method = "fallback";
break;
case RTW_RFCHIPID_INTERSIL:
cfg4 |= RTW_CONFIG4_RFTYPE_INTERSIL;
method = "Intersil";
break;
case RTW_RFCHIPID_PHILIPS:
cfg4 |= RTW_CONFIG4_RFTYPE_PHILIPS;
method = "Philips";
break;
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
case RTW_RFCHIPID_GCT: /* XXX a guess */
At last, I have rtw w/ Philips RF receiving packets. I added some sysctls to aid debugging: * hw.rtw.debug -- enable debugging * hw.rtw.flush_rfio -- Linux voodoo: possibly makes the MAC "flush" bits down the serial bus to the RF * hw.rtw.host_rfio: force the host to bang bits to the RF, instead of the MAC banging bits * hw.rtw.rfio_delay: after telling the MAC to bang bits to the RF front-end, delay rfio_delay microseconds. * hw.rtw.rfprog_fallback: there is this notion of the "RF programming method." I believe the choice influences the polarity/timing of the serial bus used to program the RF front-end. I know the correct choice for Intersil/RFMD/Philips front-ends, only. For all other front-ends, I "fallback" to rfprog_fallback. Make rtw_txdac_enable take an rtw_softc argument. I will probably revert this change. Add some Linux voodoo to rtw_continuous_tx_enable. I will probably revert this change. Important: add rtw_set_rfprog, which sets the correct RF programming method. This change and the following change are probably responsible for making the Philips RF work. Important: RTW_CONFIG1 is an 8-bit register, treat it that way! Important: RTW_BRSR is 16-bit, RTW_CRCOUNT, RTW_PHYDELAY, and RTW_MSR are 8-bit: treat them that way! Vastly simplify rtw_resume_ticks. Note to self: set the LED state to match the power state. Hedge against the possibility that RTW_MSR is protected as RTW_CONFIG[0123] are, meanwhile reworking that section of rtw_init a little. Add sc_anaparm, which isn't used, yet....
2004-12-13 03:48:02 +03:00
case RTW_RFCHIPID_RFMD:
cfg4 |= RTW_CONFIG4_RFTYPE_RFMD;
method = "RFMD";
break;
}
RTW_WRITE8(regs, RTW_CONFIG4, cfg4);
Add some bus barriers. Remark where they are needed, e.g., read_eeprom. Change some printfs to RTW_DPRINTF2s to quiet rtw(4) when debugging is disabled.
2004-12-20 04:13:45 +03:00
RTW_WBR(regs, RTW_CONFIG4, RTW_CONFIG4);
- remove unused variables - move variables inside ifdef sections - ifdef notdef unused code - use __USE for debugging variables
2013-10-18 01:24:24 +04:00
#ifdef RTW_DEBUG
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
RTW_DPRINTF(RTW_DEBUG_INIT,
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
("%s: %s RF programming method, %#02x\n", device_xname(dev), method,
Realtek sample sources are not consistent about the "Digital PHY" bit in the SROM. It seems as if it is set to 1 when the PHY is *analog*, not *digital*. Fix my sources. In rtw_intr_rx, use units of 500kb/s instead of 100kb/s for rate, to be consistent with net80211's expectations. Polish up some debugging ugly messages. Dump raw 802.11 packets if IFF_DEBUG|IFF_LINK2 and RTW_DEBUG is defined. Polish power-state (on/sleep/off) handling. Especially improve support for RFMD (totally untested) and Maxim. For Philips, take the Digital PHY property into account. Call the net80211 watchdog function from rtw_watchdog, so that we scan again if auth/assoc fails. Be a little more cautious about writing register[RTW_TPPOLL], since other drivers are.... Don't frob the high/low-priority queues right now, since I don't use them. Add rtw_join_bss which programs the card with the BSSID and other properties of a BSS. Use it on state transitions. Factor out rtw_set_nettype. Make rtw_recv_beacon call ieee80211_recv_mgmt instead of dropping beacons on the floor! TBD IBSS merges. Change some rtw_debug=2 printfs to rtw_debug=3 (RTW_DPRINTF3) printfs so the console doesn't get spammed so badly at rtw_debug=2. Change some debugging printfs to RTW_DPRINTFs. E.g., print the "RF programming method" only if debugging is enabled.
2004-12-21 02:05:41 +03:00
RTW_READ8(regs, RTW_CONFIG4)));
- remove unused variables - move variables inside ifdef sections - ifdef notdef unused code - use __USE for debugging variables
2013-10-18 01:24:24 +04:00
#else
__USE(method);
#endif
At last, I have rtw w/ Philips RF receiving packets. I added some sysctls to aid debugging: * hw.rtw.debug -- enable debugging * hw.rtw.flush_rfio -- Linux voodoo: possibly makes the MAC "flush" bits down the serial bus to the RF * hw.rtw.host_rfio: force the host to bang bits to the RF, instead of the MAC banging bits * hw.rtw.rfio_delay: after telling the MAC to bang bits to the RF front-end, delay rfio_delay microseconds. * hw.rtw.rfprog_fallback: there is this notion of the "RF programming method." I believe the choice influences the polarity/timing of the serial bus used to program the RF front-end. I know the correct choice for Intersil/RFMD/Philips front-ends, only. For all other front-ends, I "fallback" to rfprog_fallback. Make rtw_txdac_enable take an rtw_softc argument. I will probably revert this change. Add some Linux voodoo to rtw_continuous_tx_enable. I will probably revert this change. Important: add rtw_set_rfprog, which sets the correct RF programming method. This change and the following change are probably responsible for making the Philips RF work. Important: RTW_CONFIG1 is an 8-bit register, treat it that way! Important: RTW_BRSR is 16-bit, RTW_CRCOUNT, RTW_PHYDELAY, and RTW_MSR are 8-bit: treat them that way! Vastly simplify rtw_resume_ticks. Note to self: set the LED state to match the power state. Hedge against the possibility that RTW_MSR is protected as RTW_CONFIG[0123] are, meanwhile reworking that section of rtw_init a little. Add sc_anaparm, which isn't used, yet....
2004-12-13 03:48:02 +03:00
}
Remove leading __ from __(const|inline|signed|volatile) -- it is obsolete.
2005-12-24 23:27:29 +03:00
static inline void
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
rtw_init_channels(enum rtw_locale locale,
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
struct ieee80211_channel (*chans)[IEEE80211_CHAN_MAX+1], device_t dev)
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
{
int i;
const char *name = NULL;
#define ADD_CHANNEL(_chans, _chan) do { \
(*_chans)[_chan].ic_flags = IEEE80211_CHAN_B; \
(*_chans)[_chan].ic_freq = \
ieee80211_ieee2mhz(_chan, (*_chans)[_chan].ic_flags);\
} while (0)
switch (locale) {
case RTW_LOCALE_USA: /* 1-11 */
name = "USA";
for (i = 1; i <= 11; i++)
ADD_CHANNEL(chans, i);
break;
case RTW_LOCALE_JAPAN: /* 1-14 */
name = "Japan";
ADD_CHANNEL(chans, 14);
for (i = 1; i <= 14; i++)
ADD_CHANNEL(chans, i);
break;
case RTW_LOCALE_EUROPE: /* 1-13 */
name = "Europe";
for (i = 1; i <= 13; i++)
ADD_CHANNEL(chans, i);
break;
default: /* 10-11 allowed by most countries */
name = "<unknown>";
for (i = 10; i <= 11; i++)
ADD_CHANNEL(chans, i);
break;
}
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
aprint_normal_dev(dev, "Geographic Location %s\n", name);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
#undef ADD_CHANNEL
}
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
Remove leading __ from __(const|inline|signed|volatile) -- it is obsolete.
2005-12-24 23:27:29 +03:00
static inline void
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
rtw_identify_country(struct rtw_regs *regs, enum rtw_locale *locale)
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
{
Cosmetic: change u_int{8,16,32,64}_t to C99-style uint{8,16,32,64}_t, which is preferred.
2005-01-02 07:23:03 +03:00
uint8_t cfg0 = RTW_READ8(regs, RTW_CONFIG0);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
switch (cfg0 & RTW_CONFIG0_GL_MASK) {
case RTW_CONFIG0_GL_USA:
With the assistance of dyoung, fix the following problems: * some 8180L cards need a longer time for the srom to initialize; patch suggested via OpenBSD. * some cards give '0' for USA, some give '3'; for the meantime, accept both as being USA locale cards. * this is RTW, not ATW. :)
2005-09-07 09:25:41 +04:00
case _RTW_CONFIG0_GL_USA:
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
*locale = RTW_LOCALE_USA;
break;
case RTW_CONFIG0_GL_JAPAN:
*locale = RTW_LOCALE_JAPAN;
break;
case RTW_CONFIG0_GL_EUROPE:
*locale = RTW_LOCALE_EUROPE;
break;
default:
*locale = RTW_LOCALE_UNKNOWN;
break;
}
}
Remove leading __ from __(const|inline|signed|volatile) -- it is obsolete.
2005-12-24 23:27:29 +03:00
static inline int
Cosmetic: change u_int{8,16,32,64}_t to C99-style uint{8,16,32,64}_t, which is preferred.
2005-01-02 07:23:03 +03:00
rtw_identify_sta(struct rtw_regs *regs, uint8_t (*addr)[IEEE80211_ADDR_LEN],
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
device_t dev)
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
{
Cosmetic: change u_int{8,16,32,64}_t to C99-style uint{8,16,32,64}_t, which is preferred.
2005-01-02 07:23:03 +03:00
static const uint8_t empty_macaddr[IEEE80211_ADDR_LEN] = {
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
0x00, 0x00, 0x00, 0x00, 0x00, 0x00
};
Cosmetic: change u_int{8,16,32,64}_t to C99-style uint{8,16,32,64}_t, which is preferred.
2005-01-02 07:23:03 +03:00
uint32_t idr0 = RTW_READ(regs, RTW_IDR0),
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
idr1 = RTW_READ(regs, RTW_IDR1);
Per discussion on tech-kern and tech-userlevel, move the bit-twiddling macros, __BIT, __BITS, SHIFTIN, SHIFTOUT, and __arraycount() from lib/libkern/libkern.h to sys/cdefs.h. Add a __-prefix to SHIFTIN and SHIFTOUT, and add a manual page for the bit-twiddling macros, bits(3). Make the __BIT and __BITS macros "widthless," as best I can, by changing their type to uintmax_t from uint32_t. XXX The manual page lags this change by a bit. Define __PRIxBIT and __PRIxBITS printf(3) format strings.
2006-08-31 23:24:37 +04:00
(*addr)[0] = __SHIFTOUT(idr0, __BITS(0, 7));
(*addr)[1] = __SHIFTOUT(idr0, __BITS(8, 15));
(*addr)[2] = __SHIFTOUT(idr0, __BITS(16, 23));
(*addr)[3] = __SHIFTOUT(idr0, __BITS(24 ,31));
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
Per discussion on tech-kern and tech-userlevel, move the bit-twiddling macros, __BIT, __BITS, SHIFTIN, SHIFTOUT, and __arraycount() from lib/libkern/libkern.h to sys/cdefs.h. Add a __-prefix to SHIFTIN and SHIFTOUT, and add a manual page for the bit-twiddling macros, bits(3). Make the __BIT and __BITS macros "widthless," as best I can, by changing their type to uintmax_t from uint32_t. XXX The manual page lags this change by a bit. Define __PRIxBIT and __PRIxBITS printf(3) format strings.
2006-08-31 23:24:37 +04:00
(*addr)[4] = __SHIFTOUT(idr1, __BITS(0, 7));
(*addr)[5] = __SHIFTOUT(idr1, __BITS(8, 15));
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
if (IEEE80211_ADDR_EQ(addr, empty_macaddr)) {
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
aprint_error_dev(dev,
"could not get mac address, attach failed\n");
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
return ENXIO;
}
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
aprint_normal_dev(dev, "802.11 address %s\n", ether_sprintf(*addr));
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
return 0;
}
Cosmetic: change u_int{8,16,32,64}_t to C99-style uint{8,16,32,64}_t, which is preferred.
2005-01-02 07:23:03 +03:00
static uint8_t
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
rtw_chan2txpower(struct rtw_srom *sr, struct ieee80211com *ic,
struct ieee80211_channel *chan)
{
u_int idx = RTW_SR_TXPOWER1 + ieee80211_chan2ieee(ic, chan) - 1;
Remove shutdown hooks. Use KASSERT().
2007-12-20 21:47:23 +03:00
KASSERT(idx >= RTW_SR_TXPOWER1 && idx <= RTW_SR_TXPOWER14);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
return RTW_SR_GET(sr, idx);
}
static void
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
rtw_txdesc_blk_init_all(struct rtw_txdesc_blk *tdb)
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
{
int pri;
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
/* nfree: the number of free descriptors in each ring.
* The beacon ring is a special case: I do not let the
* driver use all of the descriptors on the beacon ring.
* The reasons are two-fold:
*
* (1) A BEACON descriptor's OWN bit is (apparently) not
* updated, so the driver cannot easily know if the descriptor
* belongs to it, or if it is racing the NIC. If the NIC
* does not OWN every descriptor, then the driver can safely
* update the descriptors when RTW_TBDA points at tdb_next.
*
* (2) I hope that the NIC will process more than one BEACON
* descriptor in a single beacon interval, since that will
* enable multiple-BSS support. Since the NIC does not
* clear the OWN bit, there is no natural place for it to
* stop processing BEACON desciptors. Maybe it will *not*
* stop processing them! I do not want to chance the NIC
* looping around and around a saturated beacon ring, so
* I will leave one descriptor unOWNed at all times.
*/
u_int nfree[RTW_NTXPRI] =
{RTW_NTXDESCLO, RTW_NTXDESCMD, RTW_NTXDESCHI,
RTW_NTXDESCBCN - 1};
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
for (pri = 0; pri < RTW_NTXPRI; pri++) {
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
tdb[pri].tdb_nfree = nfree[pri];
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
tdb[pri].tdb_next = 0;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
}
}
static int
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
rtw_txsoft_blk_init(struct rtw_txsoft_blk *tsb)
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
{
int i;
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
struct rtw_txsoft *ts;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
SIMPLEQ_INIT(&tsb->tsb_dirtyq);
SIMPLEQ_INIT(&tsb->tsb_freeq);
for (i = 0; i < tsb->tsb_ndesc; i++) {
ts = &tsb->tsb_desc[i];
ts->ts_mbuf = NULL;
SIMPLEQ_INSERT_TAIL(&tsb->tsb_freeq, ts, ts_q);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
}
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
tsb->tsb_tx_timer = 0;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
return 0;
}
static void
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
rtw_txsoft_blk_init_all(struct rtw_txsoft_blk *tsb)
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
{
int pri;
Miscellaneous changes. Details below. Important changes flagged with []. Using the driver with my Linksys WPC11 ver. 4, it seems to be receiving packets for a change. The WPC11 ver. 4 has a Maxim RF section. My no-name rtw with Philips RF section still does not receive any packets. Keep access-level (analog params > config[0123] registers > none) in sc_access. Add rtw_set_access for changing the access level. Make rtw_continuous_tx_enable and other subroutines use rtw_set_access instead of rtw_config0123_enable and rtw_anaparm_enable. Factor part of the chip-reset code into rtw_chip_reset1. Change the 'struct foo (*bar)[N]'-style arguments to 'struct foo *bar'-style arguments. Consolidate software/hardware Tx/Rx ring setup in rtw_hwring_setup, rtw_swring_setup. Add a new constant, SA2400_OPMODE_DEFAULTS, for the bits that we *always* set in the SA2400 OPMODE register. Factor some code out into rtw_sa2400_calibrate. (Inspired by the Linux driver.) [] When the receiver goes into underrun/overflow state, call a new subroutine, rtw_kick() that stops the Rx/Tx processes, resets the chip, reinitializes the Tx/Rx rings, and restarts Rx/Tx processes. (Inspired by the Linux driver.) [] In rtw_intr_rx, check for too-short packets before calling ieee80211_find_rxnode. I believe this will prevent a repeat of the MCHK exception I saw once on macppc. [] Use seconds-elapased as well as microseconds-elapsed to set the next "due date" for the timeout interrupt. This keeps the driver from programming the timeout to expire too early. [] In rtw_intr, read RTW_ISR at most 10 times, then get out. If the interface is not enabled (RTW_F_ENABLED), then get out. [] In rtw_stop, get out if the interface is not enabled (RTW_F_ENABLED). Block IPL_NET interrupts. Don't read/write any registers if the interface is invalid (RTW_F_INVALID). [] Call rtw_stop in rtw_detach.
2004-12-12 09:37:59 +03:00
for (pri = 0; pri < RTW_NTXPRI; pri++)
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
rtw_txsoft_blk_init(&tsb[pri]);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
}
Remove leading __ from __(const|inline|signed|volatile) -- it is obsolete.
2005-12-24 23:27:29 +03:00
static inline void
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
rtw_rxdescs_sync(struct rtw_rxdesc_blk *rdb, int desc0, int nsync, int ops)
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
{
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
KASSERT(nsync <= rdb->rdb_ndesc);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
/* sync to end of ring */
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
if (desc0 + nsync > rdb->rdb_ndesc) {
bus_dmamap_sync(rdb->rdb_dmat, rdb->rdb_dmamap,
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
offsetof(struct rtw_descs, hd_rx[desc0]),
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
sizeof(struct rtw_rxdesc) * (rdb->rdb_ndesc - desc0), ops);
nsync -= (rdb->rdb_ndesc - desc0);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
desc0 = 0;
}
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
KASSERT(desc0 < rdb->rdb_ndesc);
KASSERT(nsync <= rdb->rdb_ndesc);
KASSERT(desc0 + nsync <= rdb->rdb_ndesc);
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
/* sync what remains */
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
bus_dmamap_sync(rdb->rdb_dmat, rdb->rdb_dmamap,
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
offsetof(struct rtw_descs, hd_rx[desc0]),
sizeof(struct rtw_rxdesc) * nsync, ops);
}
static void
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
rtw_txdescs_sync(struct rtw_txdesc_blk *tdb, u_int desc0, u_int nsync, int ops)
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
{
/* sync to end of ring */
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
if (desc0 + nsync > tdb->tdb_ndesc) {
bus_dmamap_sync(tdb->tdb_dmat, tdb->tdb_dmamap,
tdb->tdb_ofs + sizeof(struct rtw_txdesc) * desc0,
sizeof(struct rtw_txdesc) * (tdb->tdb_ndesc - desc0),
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
ops);
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
nsync -= (tdb->tdb_ndesc - desc0);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
desc0 = 0;
}
/* sync what remains */
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
bus_dmamap_sync(tdb->tdb_dmat, tdb->tdb_dmamap,
tdb->tdb_ofs + sizeof(struct rtw_txdesc) * desc0,
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
sizeof(struct rtw_txdesc) * nsync, ops);
}
static void
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
rtw_txdescs_sync_all(struct rtw_txdesc_blk *tdb)
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
{
int pri;
for (pri = 0; pri < RTW_NTXPRI; pri++) {
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
rtw_txdescs_sync(&tdb[pri], 0, tdb[pri].tdb_ndesc,
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
BUS_DMASYNC_PREREAD|BUS_DMASYNC_PREWRITE);
}
}
static void
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
rtw_rxbufs_release(bus_dma_tag_t dmat, struct rtw_rxsoft *desc)
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
{
int i;
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
struct rtw_rxsoft *rs;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
for (i = 0; i < RTW_RXQLEN; i++) {
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
rs = &desc[i];
if (rs->rs_mbuf == NULL)
If defined(RTW_DEBUG), provide a sysctl (hw.rtw.rxbufs_limit) for limiting the number of rx buffers an rtw may allocate. Use this sysctl to test the code that copes with buffer exhaustion. Allocate at most RTW_RXQLEN rx buffers, stopping at the sysctl limit. Record in sc_nrxdesc how many were allocated, and put the end-of-ring indication on sc_rxdesc[sc_nrxdesc - 1]. In rtw_init, if no rx buffers could be allocated, log a complaint, clear IFF_RUNNING, and exit with an error. Many changes to accomodate a short rx ring, mainly of the "add a rx-ring length argument" variety. XXX I really should consolidate all of the rx ring variables in one struct and pass that to the rx-ring subroutines. Bug fix: after calling MCLGET, use the (m->m_flags & M_EXT) idiom to check for success, instead of m != NULL. Bug fix: at the top of rtw_start, if IFF_RUNNING is not set, or IFF_OACTIVE is, get out.
2004-12-29 01:21:15 +03:00
continue;
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
bus_dmamap_sync(dmat, rs->rs_dmamap, 0,
rs->rs_dmamap->dm_mapsize, BUS_DMASYNC_POSTREAD);
bus_dmamap_unload(dmat, rs->rs_dmamap);
m_freem(rs->rs_mbuf);
rs->rs_mbuf = NULL;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
}
}
Remove leading __ from __(const|inline|signed|volatile) -- it is obsolete.
2005-12-24 23:27:29 +03:00
static inline int
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
rtw_rxsoft_alloc(bus_dma_tag_t dmat, struct rtw_rxsoft *rs)
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
{
int rc;
struct mbuf *m;
nuke trailing whitespace
2005-02-27 03:26:58 +03:00
MGETHDR(m, M_DONTWAIT, MT_DATA);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
if (m == NULL)
Bug fix: in rtw_rxbuf_alloc, unload the previous mbuf chain before loading the new chain.
2004-12-23 09:03:09 +03:00
return ENOBUFS;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
nuke trailing whitespace
2005-02-27 03:26:58 +03:00
MCLGET(m, M_DONTWAIT);
If defined(RTW_DEBUG), provide a sysctl (hw.rtw.rxbufs_limit) for limiting the number of rx buffers an rtw may allocate. Use this sysctl to test the code that copes with buffer exhaustion. Allocate at most RTW_RXQLEN rx buffers, stopping at the sysctl limit. Record in sc_nrxdesc how many were allocated, and put the end-of-ring indication on sc_rxdesc[sc_nrxdesc - 1]. In rtw_init, if no rx buffers could be allocated, log a complaint, clear IFF_RUNNING, and exit with an error. Many changes to accomodate a short rx ring, mainly of the "add a rx-ring length argument" variety. XXX I really should consolidate all of the rx ring variables in one struct and pass that to the rx-ring subroutines. Bug fix: after calling MCLGET, use the (m->m_flags & M_EXT) idiom to check for success, instead of m != NULL. Bug fix: at the top of rtw_start, if IFF_RUNNING is not set, or IFF_OACTIVE is, get out.
2004-12-29 01:21:15 +03:00
if ((m->m_flags & M_EXT) == 0) {
m_freem(m);
Bug fix: in rtw_rxbuf_alloc, unload the previous mbuf chain before loading the new chain.
2004-12-23 09:03:09 +03:00
return ENOBUFS;
If defined(RTW_DEBUG), provide a sysctl (hw.rtw.rxbufs_limit) for limiting the number of rx buffers an rtw may allocate. Use this sysctl to test the code that copes with buffer exhaustion. Allocate at most RTW_RXQLEN rx buffers, stopping at the sysctl limit. Record in sc_nrxdesc how many were allocated, and put the end-of-ring indication on sc_rxdesc[sc_nrxdesc - 1]. In rtw_init, if no rx buffers could be allocated, log a complaint, clear IFF_RUNNING, and exit with an error. Many changes to accomodate a short rx ring, mainly of the "add a rx-ring length argument" variety. XXX I really should consolidate all of the rx ring variables in one struct and pass that to the rx-ring subroutines. Bug fix: after calling MCLGET, use the (m->m_flags & M_EXT) idiom to check for success, instead of m != NULL. Bug fix: at the top of rtw_start, if IFF_RUNNING is not set, or IFF_OACTIVE is, get out.
2004-12-29 01:21:15 +03:00
}
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
m->m_pkthdr.len = m->m_len = m->m_ext.ext_size;
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
if (rs->rs_mbuf != NULL)
bus_dmamap_unload(dmat, rs->rs_dmamap);
Bug fix: in rtw_rxbuf_alloc, unload the previous mbuf chain before loading the new chain.
2004-12-23 09:03:09 +03:00
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
rs->rs_mbuf = NULL;
Bug fix: in rtw_rxbuf_alloc, unload the previous mbuf chain before loading the new chain.
2004-12-23 09:03:09 +03:00
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
rc = bus_dmamap_load_mbuf(dmat, rs->rs_dmamap, m, BUS_DMA_NOWAIT);
Bug fix: in rtw_rxbuf_alloc, unload the previous mbuf chain before loading the new chain.
2004-12-23 09:03:09 +03:00
if (rc != 0) {
m_freem(m);
return -1;
}
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
rs->rs_mbuf = m;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
return 0;
}
static int
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
rtw_rxsoft_init_all(bus_dma_tag_t dmat, struct rtw_rxsoft *desc,
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
int *ndesc, device_t dev)
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
{
If defined(RTW_DEBUG), provide a sysctl (hw.rtw.rxbufs_limit) for limiting the number of rx buffers an rtw may allocate. Use this sysctl to test the code that copes with buffer exhaustion. Allocate at most RTW_RXQLEN rx buffers, stopping at the sysctl limit. Record in sc_nrxdesc how many were allocated, and put the end-of-ring indication on sc_rxdesc[sc_nrxdesc - 1]. In rtw_init, if no rx buffers could be allocated, log a complaint, clear IFF_RUNNING, and exit with an error. Many changes to accomodate a short rx ring, mainly of the "add a rx-ring length argument" variety. XXX I really should consolidate all of the rx ring variables in one struct and pass that to the rx-ring subroutines. Bug fix: after calling MCLGET, use the (m->m_flags & M_EXT) idiom to check for success, instead of m != NULL. Bug fix: at the top of rtw_start, if IFF_RUNNING is not set, or IFF_OACTIVE is, get out.
2004-12-29 01:21:15 +03:00
int i, rc = 0;
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
struct rtw_rxsoft *rs;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
for (i = 0; i < RTW_RXQLEN; i++) {
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
rs = &desc[i];
If defined(RTW_DEBUG), provide a sysctl (hw.rtw.rxbufs_limit) for limiting the number of rx buffers an rtw may allocate. Use this sysctl to test the code that copes with buffer exhaustion. Allocate at most RTW_RXQLEN rx buffers, stopping at the sysctl limit. Record in sc_nrxdesc how many were allocated, and put the end-of-ring indication on sc_rxdesc[sc_nrxdesc - 1]. In rtw_init, if no rx buffers could be allocated, log a complaint, clear IFF_RUNNING, and exit with an error. Many changes to accomodate a short rx ring, mainly of the "add a rx-ring length argument" variety. XXX I really should consolidate all of the rx ring variables in one struct and pass that to the rx-ring subroutines. Bug fix: after calling MCLGET, use the (m->m_flags & M_EXT) idiom to check for success, instead of m != NULL. Bug fix: at the top of rtw_start, if IFF_RUNNING is not set, or IFF_OACTIVE is, get out.
2004-12-29 01:21:15 +03:00
/* we're in rtw_init, so there should be no mbufs allocated */
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
KASSERT(rs->rs_mbuf == NULL);
If defined(RTW_DEBUG), provide a sysctl (hw.rtw.rxbufs_limit) for limiting the number of rx buffers an rtw may allocate. Use this sysctl to test the code that copes with buffer exhaustion. Allocate at most RTW_RXQLEN rx buffers, stopping at the sysctl limit. Record in sc_nrxdesc how many were allocated, and put the end-of-ring indication on sc_rxdesc[sc_nrxdesc - 1]. In rtw_init, if no rx buffers could be allocated, log a complaint, clear IFF_RUNNING, and exit with an error. Many changes to accomodate a short rx ring, mainly of the "add a rx-ring length argument" variety. XXX I really should consolidate all of the rx ring variables in one struct and pass that to the rx-ring subroutines. Bug fix: after calling MCLGET, use the (m->m_flags & M_EXT) idiom to check for success, instead of m != NULL. Bug fix: at the top of rtw_start, if IFF_RUNNING is not set, or IFF_OACTIVE is, get out.
2004-12-29 01:21:15 +03:00
#ifdef RTW_DEBUG
if (i == rtw_rxbufs_limit) {
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
aprint_error_dev(dev, "TEST hit %d-buffer limit\n", i);
If defined(RTW_DEBUG), provide a sysctl (hw.rtw.rxbufs_limit) for limiting the number of rx buffers an rtw may allocate. Use this sysctl to test the code that copes with buffer exhaustion. Allocate at most RTW_RXQLEN rx buffers, stopping at the sysctl limit. Record in sc_nrxdesc how many were allocated, and put the end-of-ring indication on sc_rxdesc[sc_nrxdesc - 1]. In rtw_init, if no rx buffers could be allocated, log a complaint, clear IFF_RUNNING, and exit with an error. Many changes to accomodate a short rx ring, mainly of the "add a rx-ring length argument" variety. XXX I really should consolidate all of the rx ring variables in one struct and pass that to the rx-ring subroutines. Bug fix: after calling MCLGET, use the (m->m_flags & M_EXT) idiom to check for success, instead of m != NULL. Bug fix: at the top of rtw_start, if IFF_RUNNING is not set, or IFF_OACTIVE is, get out.
2004-12-29 01:21:15 +03:00
rc = ENOBUFS;
break;
}
#endif /* RTW_DEBUG */
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
if ((rc = rtw_rxsoft_alloc(dmat, rs)) != 0) {
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
aprint_error_dev(dev,
"rtw_rxsoft_alloc failed, %d buffers, rc %d\n",
i, rc);
If defined(RTW_DEBUG), provide a sysctl (hw.rtw.rxbufs_limit) for limiting the number of rx buffers an rtw may allocate. Use this sysctl to test the code that copes with buffer exhaustion. Allocate at most RTW_RXQLEN rx buffers, stopping at the sysctl limit. Record in sc_nrxdesc how many were allocated, and put the end-of-ring indication on sc_rxdesc[sc_nrxdesc - 1]. In rtw_init, if no rx buffers could be allocated, log a complaint, clear IFF_RUNNING, and exit with an error. Many changes to accomodate a short rx ring, mainly of the "add a rx-ring length argument" variety. XXX I really should consolidate all of the rx ring variables in one struct and pass that to the rx-ring subroutines. Bug fix: after calling MCLGET, use the (m->m_flags & M_EXT) idiom to check for success, instead of m != NULL. Bug fix: at the top of rtw_start, if IFF_RUNNING is not set, or IFF_OACTIVE is, get out.
2004-12-29 01:21:15 +03:00
break;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
}
}
If defined(RTW_DEBUG), provide a sysctl (hw.rtw.rxbufs_limit) for limiting the number of rx buffers an rtw may allocate. Use this sysctl to test the code that copes with buffer exhaustion. Allocate at most RTW_RXQLEN rx buffers, stopping at the sysctl limit. Record in sc_nrxdesc how many were allocated, and put the end-of-ring indication on sc_rxdesc[sc_nrxdesc - 1]. In rtw_init, if no rx buffers could be allocated, log a complaint, clear IFF_RUNNING, and exit with an error. Many changes to accomodate a short rx ring, mainly of the "add a rx-ring length argument" variety. XXX I really should consolidate all of the rx ring variables in one struct and pass that to the rx-ring subroutines. Bug fix: after calling MCLGET, use the (m->m_flags & M_EXT) idiom to check for success, instead of m != NULL. Bug fix: at the top of rtw_start, if IFF_RUNNING is not set, or IFF_OACTIVE is, get out.
2004-12-29 01:21:15 +03:00
*ndesc = i;
return rc;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
}
Remove leading __ from __(const|inline|signed|volatile) -- it is obsolete.
2005-12-24 23:27:29 +03:00
static inline void
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
rtw_rxdesc_init(struct rtw_rxdesc_blk *rdb, struct rtw_rxsoft *rs,
__unused removal on arguments; approved by core.
2006-11-16 04:32:37 +03:00
int idx, int kick)
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
{
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
int is_last = (idx == rdb->rdb_ndesc - 1);
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
uint32_t ctl, octl, obuf;
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
struct rtw_rxdesc *rd = &rdb->rdb_desc[idx];
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
Fix order of operations: rtw_rxdesc_init: before initializing a Rx descriptor's control word, synchronize the Rx buffer rtw_start, be paranoid: set the "NIC owns" bit on a Tx descriptor dead last. Always reinitialize the "next Tx descriptor" pointer, just in case it got clobbered somehow. Let the compiler decide whether to inline rtw_collect_txring or not.
2007-11-16 01:55:50 +03:00
/* sync the mbuf before the descriptor */
bus_dmamap_sync(rdb->rdb_dmat, rs->rs_dmamap, 0,
rs->rs_dmamap->dm_mapsize, BUS_DMASYNC_PREREAD);
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
obuf = rd->rd_buf;
rd->rd_buf = htole32(rs->rs_dmamap->dm_segs[0].ds_addr);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
Per discussion on tech-kern and tech-userlevel, move the bit-twiddling macros, __BIT, __BITS, SHIFTIN, SHIFTOUT, and __arraycount() from lib/libkern/libkern.h to sys/cdefs.h. Add a __-prefix to SHIFTIN and SHIFTOUT, and add a manual page for the bit-twiddling macros, bits(3). Make the __BIT and __BITS macros "widthless," as best I can, by changing their type to uintmax_t from uint32_t. XXX The manual page lags this change by a bit. Define __PRIxBIT and __PRIxBITS printf(3) format strings.
2006-08-31 23:24:37 +04:00
ctl = __SHIFTIN(rs->rs_mbuf->m_len, RTW_RXCTL_LENGTH_MASK) |
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
RTW_RXCTL_OWN | RTW_RXCTL_FS | RTW_RXCTL_LS;
if (is_last)
ctl |= RTW_RXCTL_EOR;
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
octl = rd->rd_ctl;
rd->rd_ctl = htole32(ctl);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
- remove unused variables - move variables inside ifdef sections - ifdef notdef unused code - use __USE for debugging variables
2013-10-18 01:24:24 +04:00
#ifdef RTW_DEBUG
Make rtw(4) build w/o RTW_DEBUG. Thanks to Charles Hannum for pointing out my oversight.
2004-12-27 01:59:41 +03:00
RTW_DPRINTF(
kick ? (RTW_DEBUG_RECV_DESC | RTW_DEBUG_IO_KICK)
: RTW_DEBUG_RECV_DESC,
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
("%s: rd %p buf %08x -> %08x ctl %08x -> %08x\n", __func__, rd,
le32toh(obuf), le32toh(rd->rd_buf), le32toh(octl),
le32toh(rd->rd_ctl)));
- remove unused variables - move variables inside ifdef sections - ifdef notdef unused code - use __USE for debugging variables
2013-10-18 01:24:24 +04:00
#else
__USE(octl);
__USE(obuf);
#endif
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
/* sync the descriptor */
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
bus_dmamap_sync(rdb->rdb_dmat, rdb->rdb_dmamap,
Consolidate variables related to the rx ring in sc_rxdesc_blk, which is a struct rtw_rxdesc_blk. Put a copy of the DMA tag and the DMA map into the rx- and tx-ring blocks so that I don't have to pass them to subroutines all of the time.
2004-12-29 03:21:37 +03:00
RTW_DESC_OFFSET(hd_rx, idx), sizeof(struct rtw_rxdesc),
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
BUS_DMASYNC_PREREAD|BUS_DMASYNC_PREWRITE);
}
static void
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
rtw_rxdesc_init_all(struct rtw_rxdesc_blk *rdb, struct rtw_rxsoft *ctl, int kick)
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
{
int i;
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
struct rtw_rxsoft *rs;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
for (i = 0; i < rdb->rdb_ndesc; i++) {
rs = &ctl[i];
rtw_rxdesc_init(rdb, rs, i, kick);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
}
}
static void
I have made a small change that suppresses "recovery" from (spurious?) tx timeouts on the beacon queue. The driver's recovery process badly disrupts the MAC's receiver. This needs a closer look, later. Every other fix is by Charles M. Hannum. Charles writes: "With these, the device is quite stable under load, does not experience xmit timeouts, no longer mysteriously freezes, and does not spew a lot of garbage. It does, however, continue to get some "tx fifo underflow"s. global, SYSCTL_SETUP(): a) Removed unused sysctl variables. rtw_io_enable(): b) Change rtw_io_enable to take a softc, and make it call rtw_txring_fixup(). c) Enable the PCI multiple read/write bit. d) Always reset rdb_next to 0 when starting the receive engine. rtw_intr_rx(): e) When processing receive packets, store the next pointer in rdb_next and read it back on the next round, to fix a race condition vs. rtw_tune() during AP discovery and association. f) Remove the special "skip to 0th descriptor" code, because it is no longer needed. rtw_collect_txring(): g) In the "force" case, stop when we catch up to the hardware xmit pointer. rtw_intr_beacon(): h) Resetting tdb_next here looked fishy to me, in light of changes in rtw_collect_txring(). XXX However, this is the one part I haven't tested. rtw_txring_fixup(): i) Add function and line arguments, so we can print out information about the caller. j) In addition to reading the hardware pointer, skip over any additional packets with OWN set -- we may have queued more. rtw_rxring_fixup(): k) Remove this, because it's completely wrong. rtw_intr_ioerror(): l) Don't handle receive overruns here; handle them in the normal receive path. Also changed RTW_INTR_RX and RTW_INTR_IOERROR accordingly. m) Do not fiddle with CR; the transmit engine is still running just fine. n) Do not fiddle with IMR; it was unnecessary. o) Don't touch the receive engine at all; not necessary now that we're leaving CR alone. p) Remove the rtw_do_chip_reset case; it was dangerous and not actually used. rtw_intr(): q) Don't mask off ISR bits when we pass down the ISR value; it's not necessary, and sometimes the other bits useful in debugging output. rtw_seg_too_short(): r) Don't spew here; the caller will do it if necessary. (Stops spewage in an expected case.) rtw_rxdesc_blk: s) Minor change to make the elements have the same order as txdesc_blk. general: t) Added a bunch of debugging output, #ifdef RTW_DIAG, that was useful while investigating the ring management problems. Also moved rtw_txring_fixup() into RTW_DIAG, as I believe the transmit pointer is always kept in sync now, and the function has potential race conditions.
2007-01-06 11:27:53 +03:00
rtw_io_enable(struct rtw_softc *sc, uint8_t flags, int enable)
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
{
I have made a small change that suppresses "recovery" from (spurious?) tx timeouts on the beacon queue. The driver's recovery process badly disrupts the MAC's receiver. This needs a closer look, later. Every other fix is by Charles M. Hannum. Charles writes: "With these, the device is quite stable under load, does not experience xmit timeouts, no longer mysteriously freezes, and does not spew a lot of garbage. It does, however, continue to get some "tx fifo underflow"s. global, SYSCTL_SETUP(): a) Removed unused sysctl variables. rtw_io_enable(): b) Change rtw_io_enable to take a softc, and make it call rtw_txring_fixup(). c) Enable the PCI multiple read/write bit. d) Always reset rdb_next to 0 when starting the receive engine. rtw_intr_rx(): e) When processing receive packets, store the next pointer in rdb_next and read it back on the next round, to fix a race condition vs. rtw_tune() during AP discovery and association. f) Remove the special "skip to 0th descriptor" code, because it is no longer needed. rtw_collect_txring(): g) In the "force" case, stop when we catch up to the hardware xmit pointer. rtw_intr_beacon(): h) Resetting tdb_next here looked fishy to me, in light of changes in rtw_collect_txring(). XXX However, this is the one part I haven't tested. rtw_txring_fixup(): i) Add function and line arguments, so we can print out information about the caller. j) In addition to reading the hardware pointer, skip over any additional packets with OWN set -- we may have queued more. rtw_rxring_fixup(): k) Remove this, because it's completely wrong. rtw_intr_ioerror(): l) Don't handle receive overruns here; handle them in the normal receive path. Also changed RTW_INTR_RX and RTW_INTR_IOERROR accordingly. m) Do not fiddle with CR; the transmit engine is still running just fine. n) Do not fiddle with IMR; it was unnecessary. o) Don't touch the receive engine at all; not necessary now that we're leaving CR alone. p) Remove the rtw_do_chip_reset case; it was dangerous and not actually used. rtw_intr(): q) Don't mask off ISR bits when we pass down the ISR value; it's not necessary, and sometimes the other bits useful in debugging output. rtw_seg_too_short(): r) Don't spew here; the caller will do it if necessary. (Stops spewage in an expected case.) rtw_rxdesc_blk: s) Minor change to make the elements have the same order as txdesc_blk. general: t) Added a bunch of debugging output, #ifdef RTW_DIAG, that was useful while investigating the ring management problems. Also moved rtw_txring_fixup() into RTW_DIAG, as I believe the transmit pointer is always kept in sync now, and the function has potential race conditions.
2007-01-06 11:27:53 +03:00
struct rtw_regs *regs = &sc->sc_regs;
Cosmetic: change u_int{8,16,32,64}_t to C99-style uint{8,16,32,64}_t, which is preferred.
2005-01-02 07:23:03 +03:00
uint8_t cr;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
RTW_DPRINTF(RTW_DEBUG_IOSTATE, ("%s: %s 0x%02x\n", __func__,
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
enable ? "enable" : "disable", flags));
cr = RTW_READ8(regs, RTW_CR);
/* XXX reference source does not enable MULRW */
/* enable PCI Read/Write Multiple */
cr |= RTW_CR_MULRW;
I have made a small change that suppresses "recovery" from (spurious?) tx timeouts on the beacon queue. The driver's recovery process badly disrupts the MAC's receiver. This needs a closer look, later. Every other fix is by Charles M. Hannum. Charles writes: "With these, the device is quite stable under load, does not experience xmit timeouts, no longer mysteriously freezes, and does not spew a lot of garbage. It does, however, continue to get some "tx fifo underflow"s. global, SYSCTL_SETUP(): a) Removed unused sysctl variables. rtw_io_enable(): b) Change rtw_io_enable to take a softc, and make it call rtw_txring_fixup(). c) Enable the PCI multiple read/write bit. d) Always reset rdb_next to 0 when starting the receive engine. rtw_intr_rx(): e) When processing receive packets, store the next pointer in rdb_next and read it back on the next round, to fix a race condition vs. rtw_tune() during AP discovery and association. f) Remove the special "skip to 0th descriptor" code, because it is no longer needed. rtw_collect_txring(): g) In the "force" case, stop when we catch up to the hardware xmit pointer. rtw_intr_beacon(): h) Resetting tdb_next here looked fishy to me, in light of changes in rtw_collect_txring(). XXX However, this is the one part I haven't tested. rtw_txring_fixup(): i) Add function and line arguments, so we can print out information about the caller. j) In addition to reading the hardware pointer, skip over any additional packets with OWN set -- we may have queued more. rtw_rxring_fixup(): k) Remove this, because it's completely wrong. rtw_intr_ioerror(): l) Don't handle receive overruns here; handle them in the normal receive path. Also changed RTW_INTR_RX and RTW_INTR_IOERROR accordingly. m) Do not fiddle with CR; the transmit engine is still running just fine. n) Do not fiddle with IMR; it was unnecessary. o) Don't touch the receive engine at all; not necessary now that we're leaving CR alone. p) Remove the rtw_do_chip_reset case; it was dangerous and not actually used. rtw_intr(): q) Don't mask off ISR bits when we pass down the ISR value; it's not necessary, and sometimes the other bits useful in debugging output. rtw_seg_too_short(): r) Don't spew here; the caller will do it if necessary. (Stops spewage in an expected case.) rtw_rxdesc_blk: s) Minor change to make the elements have the same order as txdesc_blk. general: t) Added a bunch of debugging output, #ifdef RTW_DIAG, that was useful while investigating the ring management problems. Also moved rtw_txring_fixup() into RTW_DIAG, as I believe the transmit pointer is always kept in sync now, and the function has potential race conditions.
2007-01-06 11:27:53 +03:00
/* The receive engine will always start at RDSAR. */
if (enable && (flags & ~cr & RTW_CR_RE)) {
struct rtw_rxdesc_blk *rdb;
rdb = &sc->sc_rxdesc_blk;
rdb->rdb_next = 0;
}
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
RTW_RBW(regs, RTW_CR, RTW_CR); /* XXX paranoia? */
if (enable)
cr |= flags;
else
cr &= ~flags;
RTW_WRITE8(regs, RTW_CR, cr);
RTW_SYNC(regs, RTW_CR, RTW_CR);
I have made a small change that suppresses "recovery" from (spurious?) tx timeouts on the beacon queue. The driver's recovery process badly disrupts the MAC's receiver. This needs a closer look, later. Every other fix is by Charles M. Hannum. Charles writes: "With these, the device is quite stable under load, does not experience xmit timeouts, no longer mysteriously freezes, and does not spew a lot of garbage. It does, however, continue to get some "tx fifo underflow"s. global, SYSCTL_SETUP(): a) Removed unused sysctl variables. rtw_io_enable(): b) Change rtw_io_enable to take a softc, and make it call rtw_txring_fixup(). c) Enable the PCI multiple read/write bit. d) Always reset rdb_next to 0 when starting the receive engine. rtw_intr_rx(): e) When processing receive packets, store the next pointer in rdb_next and read it back on the next round, to fix a race condition vs. rtw_tune() during AP discovery and association. f) Remove the special "skip to 0th descriptor" code, because it is no longer needed. rtw_collect_txring(): g) In the "force" case, stop when we catch up to the hardware xmit pointer. rtw_intr_beacon(): h) Resetting tdb_next here looked fishy to me, in light of changes in rtw_collect_txring(). XXX However, this is the one part I haven't tested. rtw_txring_fixup(): i) Add function and line arguments, so we can print out information about the caller. j) In addition to reading the hardware pointer, skip over any additional packets with OWN set -- we may have queued more. rtw_rxring_fixup(): k) Remove this, because it's completely wrong. rtw_intr_ioerror(): l) Don't handle receive overruns here; handle them in the normal receive path. Also changed RTW_INTR_RX and RTW_INTR_IOERROR accordingly. m) Do not fiddle with CR; the transmit engine is still running just fine. n) Do not fiddle with IMR; it was unnecessary. o) Don't touch the receive engine at all; not necessary now that we're leaving CR alone. p) Remove the rtw_do_chip_reset case; it was dangerous and not actually used. rtw_intr(): q) Don't mask off ISR bits when we pass down the ISR value; it's not necessary, and sometimes the other bits useful in debugging output. rtw_seg_too_short(): r) Don't spew here; the caller will do it if necessary. (Stops spewage in an expected case.) rtw_rxdesc_blk: s) Minor change to make the elements have the same order as txdesc_blk. general: t) Added a bunch of debugging output, #ifdef RTW_DIAG, that was useful while investigating the ring management problems. Also moved rtw_txring_fixup() into RTW_DIAG, as I believe the transmit pointer is always kept in sync now, and the function has potential race conditions.
2007-01-06 11:27:53 +03:00
#ifdef RTW_DIAG
if (cr & RTW_CR_TE)
rtw_txring_fixup(sc, __func__, __LINE__);
#endif
Extract subroutine rtw_tx_kick() from rtw_start(), and reuse it in rtw_io_enable() to restart the transmitter after resetting it.
2008-03-15 02:59:01 +03:00
if (cr & RTW_CR_TE) {
rtw_tx_kick(&sc->sc_regs,
RTW_TPPOLL_HPQ | RTW_TPPOLL_NPQ | RTW_TPPOLL_LPQ);
}
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
}
static void
__unused removal on arguments; approved by core.
2006-11-16 04:32:37 +03:00
rtw_intr_rx(struct rtw_softc *sc, uint16_t isr)
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
{
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
#define IS_BEACON(__fc0) \
((__fc0 & (IEEE80211_FC0_TYPE_MASK | IEEE80211_FC0_SUBTYPE_MASK)) ==\
(IEEE80211_FC0_TYPE_MGT | IEEE80211_FC0_SUBTYPE_BEACON))
Use a table to convert the hardware's bit rates (0, 1, 2, 3) to net80211 bit rates (2, 4, 11, 22). Count input errors (if_ierrors++) for "DMA error/FIFO overflow" and unknown bit rate.
2004-12-27 23:04:45 +03:00
static const int ratetbl[4] = {2, 4, 11, 22}; /* convert rates:
* hardware -> net80211
*/
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
u_int next, nproc = 0;
Add bpf(4) packet taps to rtw(4). Now you can tap DLT_EN10MB (Ethernet II), DLT_IEEE802_11 (802.11), and DLT_IEEE802_11_RADIO (radiotap + 802.11).
2004-12-29 01:30:07 +03:00
int hwrate, len, rate, rssi, sq;
Cosmetic: change u_int{8,16,32,64}_t to C99-style uint{8,16,32,64}_t, which is preferred.
2005-01-02 07:23:03 +03:00
uint32_t hrssi, hstat, htsfth, htsftl;
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
struct rtw_rxdesc *rd;
struct rtw_rxsoft *rs;
struct rtw_rxdesc_blk *rdb;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
struct mbuf *m;
Resolve conflicts in importation of 18-May-2005 ath(4) / net80211(9) from FreeBSD. Introduce compatibility shims (sys/dev/ic/ath_netbsd.[ch], sys/net80211/ieee80211_netbsd.[ch]). Update drivers (an, atu, atw, awi, ipw, iwi, rtw, wi) for the new net80211(9) API.
2005-06-22 10:14:51 +04:00
struct ifnet *ifp = &sc->sc_if;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
struct ieee80211_node *ni;
Resolve conflicts in importation of 18-May-2005 ath(4) / net80211(9) from FreeBSD. Introduce compatibility shims (sys/dev/ic/ath_netbsd.[ch], sys/net80211/ieee80211_netbsd.[ch]). Update drivers (an, atu, atw, awi, ipw, iwi, rtw, wi) for the new net80211(9) API.
2005-06-22 10:14:51 +04:00
struct ieee80211_frame_min *wh;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
rdb = &sc->sc_rxdesc_blk;
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
I have made a small change that suppresses "recovery" from (spurious?) tx timeouts on the beacon queue. The driver's recovery process badly disrupts the MAC's receiver. This needs a closer look, later. Every other fix is by Charles M. Hannum. Charles writes: "With these, the device is quite stable under load, does not experience xmit timeouts, no longer mysteriously freezes, and does not spew a lot of garbage. It does, however, continue to get some "tx fifo underflow"s. global, SYSCTL_SETUP(): a) Removed unused sysctl variables. rtw_io_enable(): b) Change rtw_io_enable to take a softc, and make it call rtw_txring_fixup(). c) Enable the PCI multiple read/write bit. d) Always reset rdb_next to 0 when starting the receive engine. rtw_intr_rx(): e) When processing receive packets, store the next pointer in rdb_next and read it back on the next round, to fix a race condition vs. rtw_tune() during AP discovery and association. f) Remove the special "skip to 0th descriptor" code, because it is no longer needed. rtw_collect_txring(): g) In the "force" case, stop when we catch up to the hardware xmit pointer. rtw_intr_beacon(): h) Resetting tdb_next here looked fishy to me, in light of changes in rtw_collect_txring(). XXX However, this is the one part I haven't tested. rtw_txring_fixup(): i) Add function and line arguments, so we can print out information about the caller. j) In addition to reading the hardware pointer, skip over any additional packets with OWN set -- we may have queued more. rtw_rxring_fixup(): k) Remove this, because it's completely wrong. rtw_intr_ioerror(): l) Don't handle receive overruns here; handle them in the normal receive path. Also changed RTW_INTR_RX and RTW_INTR_IOERROR accordingly. m) Do not fiddle with CR; the transmit engine is still running just fine. n) Do not fiddle with IMR; it was unnecessary. o) Don't touch the receive engine at all; not necessary now that we're leaving CR alone. p) Remove the rtw_do_chip_reset case; it was dangerous and not actually used. rtw_intr(): q) Don't mask off ISR bits when we pass down the ISR value; it's not necessary, and sometimes the other bits useful in debugging output. rtw_seg_too_short(): r) Don't spew here; the caller will do it if necessary. (Stops spewage in an expected case.) rtw_rxdesc_blk: s) Minor change to make the elements have the same order as txdesc_blk. general: t) Added a bunch of debugging output, #ifdef RTW_DIAG, that was useful while investigating the ring management problems. Also moved rtw_txring_fixup() into RTW_DIAG, as I believe the transmit pointer is always kept in sync now, and the function has potential race conditions.
2007-01-06 11:27:53 +03:00
for (next = rdb->rdb_next; ; next = rdb->rdb_next) {
KASSERT(next < rdb->rdb_ndesc);
Accommodate the variety of RF front-ends when radiotapping frames: tap only RSSI when there is a Philips RF front-end. Tap both Barker lock quality and RSSI when there is any other RF front-end. Provide radiotap listeners a more complete picture of channel activity: in promiscuous mode, tap frames who do not pass the CRC32 check. Flag packets that were received with a short preamble. Ask the NIC to pass us 802.11 Control frames. Pass frames to radiotap listeners before stripping the FCS. Re-order operations in rtw_intr_rx() in order to accomplish all of this, taking care not to pass a broken packet to net80211! Do not provide a flags field when tapping xmitted frames. Assert sane Rx packet lengths. Really should check and drop instead of KASSERTing. I will revisit this, soon. Update copyright.
2007-11-17 02:35:19 +03:00
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
rtw_rxdescs_sync(rdb, next, 1,
If defined(RTW_DEBUG), provide a sysctl (hw.rtw.rxbufs_limit) for limiting the number of rx buffers an rtw may allocate. Use this sysctl to test the code that copes with buffer exhaustion. Allocate at most RTW_RXQLEN rx buffers, stopping at the sysctl limit. Record in sc_nrxdesc how many were allocated, and put the end-of-ring indication on sc_rxdesc[sc_nrxdesc - 1]. In rtw_init, if no rx buffers could be allocated, log a complaint, clear IFF_RUNNING, and exit with an error. Many changes to accomodate a short rx ring, mainly of the "add a rx-ring length argument" variety. XXX I really should consolidate all of the rx ring variables in one struct and pass that to the rx-ring subroutines. Bug fix: after calling MCLGET, use the (m->m_flags & M_EXT) idiom to check for success, instead of m != NULL. Bug fix: at the top of rtw_start, if IFF_RUNNING is not set, or IFF_OACTIVE is, get out.
2004-12-29 01:21:15 +03:00
BUS_DMASYNC_POSTREAD|BUS_DMASYNC_POSTWRITE);
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
rd = &rdb->rdb_desc[next];
rs = &sc->sc_rxsoft[next];
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
hstat = le32toh(rd->rd_stat);
hrssi = le32toh(rd->rd_rssi);
htsfth = le32toh(rd->rd_tsfth);
htsftl = le32toh(rd->rd_tsftl);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
RTW_DPRINTF(RTW_DEBUG_RECV_DESC,
("%s: rxdesc[%d] hstat %08x hrssi %08x htsft %08x%08x\n",
__func__, next, hstat, hrssi, htsfth, htsftl));
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
++nproc;
/* still belongs to NIC */
if ((hstat & RTW_RXSTAT_OWN) != 0) {
After we examine a rx/tx descriptor and find that it's still owned by the NIC, re-sync the descriptor so that a cacheline will not "cover" the NIC's modifications to the descriptor when we next examine it. This compiles, but I have not tested it, yet.
2007-01-06 09:07:55 +03:00
rtw_rxdescs_sync(rdb, next, 1, BUS_DMASYNC_PREREAD);
I have made a small change that suppresses "recovery" from (spurious?) tx timeouts on the beacon queue. The driver's recovery process badly disrupts the MAC's receiver. This needs a closer look, later. Every other fix is by Charles M. Hannum. Charles writes: "With these, the device is quite stable under load, does not experience xmit timeouts, no longer mysteriously freezes, and does not spew a lot of garbage. It does, however, continue to get some "tx fifo underflow"s. global, SYSCTL_SETUP(): a) Removed unused sysctl variables. rtw_io_enable(): b) Change rtw_io_enable to take a softc, and make it call rtw_txring_fixup(). c) Enable the PCI multiple read/write bit. d) Always reset rdb_next to 0 when starting the receive engine. rtw_intr_rx(): e) When processing receive packets, store the next pointer in rdb_next and read it back on the next round, to fix a race condition vs. rtw_tune() during AP discovery and association. f) Remove the special "skip to 0th descriptor" code, because it is no longer needed. rtw_collect_txring(): g) In the "force" case, stop when we catch up to the hardware xmit pointer. rtw_intr_beacon(): h) Resetting tdb_next here looked fishy to me, in light of changes in rtw_collect_txring(). XXX However, this is the one part I haven't tested. rtw_txring_fixup(): i) Add function and line arguments, so we can print out information about the caller. j) In addition to reading the hardware pointer, skip over any additional packets with OWN set -- we may have queued more. rtw_rxring_fixup(): k) Remove this, because it's completely wrong. rtw_intr_ioerror(): l) Don't handle receive overruns here; handle them in the normal receive path. Also changed RTW_INTR_RX and RTW_INTR_IOERROR accordingly. m) Do not fiddle with CR; the transmit engine is still running just fine. n) Do not fiddle with IMR; it was unnecessary. o) Don't touch the receive engine at all; not necessary now that we're leaving CR alone. p) Remove the rtw_do_chip_reset case; it was dangerous and not actually used. rtw_intr(): q) Don't mask off ISR bits when we pass down the ISR value; it's not necessary, and sometimes the other bits useful in debugging output. rtw_seg_too_short(): r) Don't spew here; the caller will do it if necessary. (Stops spewage in an expected case.) rtw_rxdesc_blk: s) Minor change to make the elements have the same order as txdesc_blk. general: t) Added a bunch of debugging output, #ifdef RTW_DIAG, that was useful while investigating the ring management problems. Also moved rtw_txring_fixup() into RTW_DIAG, as I believe the transmit pointer is always kept in sync now, and the function has potential race conditions.
2007-01-06 11:27:53 +03:00
break;
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
}
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
I have made a small change that suppresses "recovery" from (spurious?) tx timeouts on the beacon queue. The driver's recovery process badly disrupts the MAC's receiver. This needs a closer look, later. Every other fix is by Charles M. Hannum. Charles writes: "With these, the device is quite stable under load, does not experience xmit timeouts, no longer mysteriously freezes, and does not spew a lot of garbage. It does, however, continue to get some "tx fifo underflow"s. global, SYSCTL_SETUP(): a) Removed unused sysctl variables. rtw_io_enable(): b) Change rtw_io_enable to take a softc, and make it call rtw_txring_fixup(). c) Enable the PCI multiple read/write bit. d) Always reset rdb_next to 0 when starting the receive engine. rtw_intr_rx(): e) When processing receive packets, store the next pointer in rdb_next and read it back on the next round, to fix a race condition vs. rtw_tune() during AP discovery and association. f) Remove the special "skip to 0th descriptor" code, because it is no longer needed. rtw_collect_txring(): g) In the "force" case, stop when we catch up to the hardware xmit pointer. rtw_intr_beacon(): h) Resetting tdb_next here looked fishy to me, in light of changes in rtw_collect_txring(). XXX However, this is the one part I haven't tested. rtw_txring_fixup(): i) Add function and line arguments, so we can print out information about the caller. j) In addition to reading the hardware pointer, skip over any additional packets with OWN set -- we may have queued more. rtw_rxring_fixup(): k) Remove this, because it's completely wrong. rtw_intr_ioerror(): l) Don't handle receive overruns here; handle them in the normal receive path. Also changed RTW_INTR_RX and RTW_INTR_IOERROR accordingly. m) Do not fiddle with CR; the transmit engine is still running just fine. n) Do not fiddle with IMR; it was unnecessary. o) Don't touch the receive engine at all; not necessary now that we're leaving CR alone. p) Remove the rtw_do_chip_reset case; it was dangerous and not actually used. rtw_intr(): q) Don't mask off ISR bits when we pass down the ISR value; it's not necessary, and sometimes the other bits useful in debugging output. rtw_seg_too_short(): r) Don't spew here; the caller will do it if necessary. (Stops spewage in an expected case.) rtw_rxdesc_blk: s) Minor change to make the elements have the same order as txdesc_blk. general: t) Added a bunch of debugging output, #ifdef RTW_DIAG, that was useful while investigating the ring management problems. Also moved rtw_txring_fixup() into RTW_DIAG, as I believe the transmit pointer is always kept in sync now, and the function has potential race conditions.
2007-01-06 11:27:53 +03:00
/* ieee80211_input() might reset the receive engine
* (e.g. by indirectly calling rtw_tune()), so save
* the next pointer here and retrieve it again on
* the next round.
*/
rdb->rdb_next = (next + 1) % rdb->rdb_ndesc;
Misc. changes to improve ad hoc mode and to enable hostap mode: Fix the work-around for the NIC bug where it skips to rx descriptor 0. The driver used to skip to rx descriptor 1. Hopefully this stops the out-of-order packet reception that Charles Hannum saw. When debugging is enabled, print rx-descriptor status flags before printing the rx bit rate. Add a debug message for when a beacon tx buffer reclamation. Reset IFF_OACTIVE when we reset the transmitter. Pass the consolidated LED state, a struct rtw_led_state, to rtw_led_attach. Choose the bit-rate for management frames (1Mb/s) at the same place we choose for all other frames. Do not use the NIC's short preamble or RTS options for management frames. Label beacons for the NIC. Following a Linux driver, take care not to zero arbitrary bits in the TPPOLL register. Use the new idiom for IBSS merges: disable transmitter, kick the state machine. Add a second descriptor to the beacon ring. The NIC seems to like this much better.
2005-03-02 08:20:43 +03:00
#ifdef RTW_DEBUG
#define PRINTSTAT(flag) do { \
if ((hstat & flag) != 0) { \
printf("%s" #flag, delim); \
delim = ","; \
} \
} while (0)
if ((rtw_debug & RTW_DEBUG_RECV_DESC) != 0) {
const char *delim = "<";
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
printf("%s: ", device_xname(sc->sc_dev));
Misc. changes to improve ad hoc mode and to enable hostap mode: Fix the work-around for the NIC bug where it skips to rx descriptor 0. The driver used to skip to rx descriptor 1. Hopefully this stops the out-of-order packet reception that Charles Hannum saw. When debugging is enabled, print rx-descriptor status flags before printing the rx bit rate. Add a debug message for when a beacon tx buffer reclamation. Reset IFF_OACTIVE when we reset the transmitter. Pass the consolidated LED state, a struct rtw_led_state, to rtw_led_attach. Choose the bit-rate for management frames (1Mb/s) at the same place we choose for all other frames. Do not use the NIC's short preamble or RTS options for management frames. Label beacons for the NIC. Following a Linux driver, take care not to zero arbitrary bits in the TPPOLL register. Use the new idiom for IBSS merges: disable transmitter, kick the state machine. Add a second descriptor to the beacon ring. The NIC seems to like this much better.
2005-03-02 08:20:43 +03:00
if ((hstat & RTW_RXSTAT_DEBUG) != 0) {
printf("status %08x", hstat);
PRINTSTAT(RTW_RXSTAT_SPLCP);
PRINTSTAT(RTW_RXSTAT_MAR);
PRINTSTAT(RTW_RXSTAT_PAR);
PRINTSTAT(RTW_RXSTAT_BAR);
PRINTSTAT(RTW_RXSTAT_PWRMGT);
PRINTSTAT(RTW_RXSTAT_CRC32);
PRINTSTAT(RTW_RXSTAT_ICV);
printf(">, ");
}
}
#endif /* RTW_DEBUG */
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
if ((hstat & RTW_RXSTAT_IOERROR) != 0) {
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
aprint_error_dev(sc->sc_dev,
"DMA error/FIFO overflow %08" PRIx32 ", "
"rx descriptor %d\n", hstat, next);
Resolve conflicts in importation of 18-May-2005 ath(4) / net80211(9) from FreeBSD. Introduce compatibility shims (sys/dev/ic/ath_netbsd.[ch], sys/net80211/ieee80211_netbsd.[ch]). Update drivers (an, atu, atw, awi, ipw, iwi, rtw, wi) for the new net80211(9) API.
2005-06-22 10:14:51 +04:00
ifp->if_ierrors++;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
goto next;
}
Per discussion on tech-kern and tech-userlevel, move the bit-twiddling macros, __BIT, __BITS, SHIFTIN, SHIFTOUT, and __arraycount() from lib/libkern/libkern.h to sys/cdefs.h. Add a __-prefix to SHIFTIN and SHIFTOUT, and add a manual page for the bit-twiddling macros, bits(3). Make the __BIT and __BITS macros "widthless," as best I can, by changing their type to uintmax_t from uint32_t. XXX The manual page lags this change by a bit. Define __PRIxBIT and __PRIxBITS printf(3) format strings.
2006-08-31 23:24:37 +04:00
len = __SHIFTOUT(hstat, RTW_RXSTAT_LENGTH_MASK);
IMPORTANT: set m->m_pkthdr.rcvif on rx'd packets. Illegal memory references abound without this fix. In rtw_intr_rx, skip to the next rx packet earlier if the rx packet length is too short. Also in rtw_intr_rx, if a packet is rx'd with unknown rate, printf a warning, drop the packet, recycle the rx buffer, skip to next rx packet.
2004-12-25 10:24:17 +03:00
if (len < IEEE80211_MIN_LEN) {
sc->sc_ic.ic_stats.is_rx_tooshort++;
goto next;
}
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
if (len > rs->rs_mbuf->m_len) {
Use device_t and its accessors throughout. Use aprint_*_dev(). Improve PMF-ability. Add a 'flags' argument to suspend/resume handlers and callers such as pmf_system_suspend(). Define a flag, PMF_F_SELF, which indicates to PMF that a device is suspending/resuming itself. Add helper routines, pmf_device_suspend_self(dev) and pmf_device_resume_self(dev), that call pmf_device_suspend(dev, PMF_F_SELF) and pmf_device_resume(dev, PMF_F_SELF), respectively. Use PMF_F_SELF to suspend/resume self in ath(4), audio(4), rtw(4), and sip(4). In ath(4) and in rtw(4), replace the icky sc_enable/sc_disable callbacks, provided by the bus front-end, with self-suspension/resumption. Also, clean up the bus front-ends. Make sure that the interrupt handler is disestablished during suspension. Get rid of driver-private flags (e.g., RTW_F_ENABLED, ath_softc->sc_invalid); use device_is_active()/device_has_power() calls, instead. In the network-class suspend handler, call if_stop(, 0) instead of if_stop(, 1), because the latter is superfluous (bus- and driver-suspension hooks will 'disable' the NIC), and it may cause recursion. In the network-class resume handler, prevent infinite recursion through if_init() by getting out early if we are self-suspending (PMF_F_SELF). rtw(4) improvements: Destroy rtw(4) callouts when we detach it. Make rtw at pci detachable. Print some more information with the "rx frame too long" warning. Remove activate() methods: Get rid of rtw_activate() and ath_activate(). The device activate() methods are not good for much these days. Make ath at cardbus resume with crypto functions intact: Introduce a boolean device property, "pmf-powerdown". If pmf-powerdown is present and false, it indicates that a bus back-end should not remove power from a device. Honor this property in cardbus_child_suspend(). Set this property to 'false' in ath_attach(), since removing power from an ath at cardbus seems to lobotomize the WPA crypto engine. XXX Should the pmf-powerdown property propagate toward the root of the device tree? Miscellaneous ath(4) changes: Warn if ath(4) tries to write crypto keys to suspended hardware. Reduce differences between FreeBSD and NetBSD in ath(4) multicast filter setup. Make ath_printrxbuf() print an rx descriptor's status & key index, to help debug crypto errors. Shorten a staircase in ath_ioctl(). Don't check for ieee80211_ioctl() return code ERESTART, it never happens.
2008-03-12 21:02:21 +03:00
aprint_error_dev(sc->sc_dev,
"rx frame too long, %d > %d, %08" PRIx32
", desc %d\n",
len, rs->rs_mbuf->m_len, hstat, next);
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
ifp->if_ierrors++;
goto next;
}
Eliminate use of M_HASFCS.
2005-01-31 05:51:17 +03:00
Per discussion on tech-kern and tech-userlevel, move the bit-twiddling macros, __BIT, __BITS, SHIFTIN, SHIFTOUT, and __arraycount() from lib/libkern/libkern.h to sys/cdefs.h. Add a __-prefix to SHIFTIN and SHIFTOUT, and add a manual page for the bit-twiddling macros, bits(3). Make the __BIT and __BITS macros "widthless," as best I can, by changing their type to uintmax_t from uint32_t. XXX The manual page lags this change by a bit. Define __PRIxBIT and __PRIxBITS printf(3) format strings.
2006-08-31 23:24:37 +04:00
hwrate = __SHIFTOUT(hstat, RTW_RXSTAT_RATE_MASK);
Cosmetic: simplify a memset() statement, use __arraycount(), and remove a gratuitous return statement at the end of a void subroutine.
2007-05-29 22:33:38 +04:00
if (hwrate >= __arraycount(ratetbl)) {
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
aprint_error_dev(sc->sc_dev,
"unknown rate #%" __PRIuBITS "\n",
Per discussion on tech-kern and tech-userlevel, move the bit-twiddling macros, __BIT, __BITS, SHIFTIN, SHIFTOUT, and __arraycount() from lib/libkern/libkern.h to sys/cdefs.h. Add a __-prefix to SHIFTIN and SHIFTOUT, and add a manual page for the bit-twiddling macros, bits(3). Make the __BIT and __BITS macros "widthless," as best I can, by changing their type to uintmax_t from uint32_t. XXX The manual page lags this change by a bit. Define __PRIxBIT and __PRIxBITS printf(3) format strings.
2006-08-31 23:24:37 +04:00
__SHIFTOUT(hstat, RTW_RXSTAT_RATE_MASK));
Resolve conflicts in importation of 18-May-2005 ath(4) / net80211(9) from FreeBSD. Introduce compatibility shims (sys/dev/ic/ath_netbsd.[ch], sys/net80211/ieee80211_netbsd.[ch]). Update drivers (an, atu, atw, awi, ipw, iwi, rtw, wi) for the new net80211(9) API.
2005-06-22 10:14:51 +04:00
ifp->if_ierrors++;
IMPORTANT: set m->m_pkthdr.rcvif on rx'd packets. Illegal memory references abound without this fix. In rtw_intr_rx, skip to the next rx packet earlier if the rx packet length is too short. Also in rtw_intr_rx, if a packet is rx'd with unknown rate, printf a warning, drop the packet, recycle the rx buffer, skip to next rx packet.
2004-12-25 10:24:17 +03:00
goto next;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
}
Use a table to convert the hardware's bit rates (0, 1, 2, 3) to net80211 bit rates (2, 4, 11, 22). Count input errors (if_ierrors++) for "DMA error/FIFO overflow" and unknown bit rate.
2004-12-27 23:04:45 +03:00
rate = ratetbl[hwrate];
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
#ifdef RTW_DEBUG
Misc. changes to improve ad hoc mode and to enable hostap mode: Fix the work-around for the NIC bug where it skips to rx descriptor 0. The driver used to skip to rx descriptor 1. Hopefully this stops the out-of-order packet reception that Charles Hannum saw. When debugging is enabled, print rx-descriptor status flags before printing the rx bit rate. Add a debug message for when a beacon tx buffer reclamation. Reset IFF_OACTIVE when we reset the transmitter. Pass the consolidated LED state, a struct rtw_led_state, to rtw_led_attach. Choose the bit-rate for management frames (1Mb/s) at the same place we choose for all other frames. Do not use the NIC's short preamble or RTS options for management frames. Label beacons for the NIC. Following a Linux driver, take care not to zero arbitrary bits in the TPPOLL register. Use the new idiom for IBSS merges: disable transmitter, kick the state machine. Add a second descriptor to the beacon ring. The NIC seems to like this much better.
2005-03-02 08:20:43 +03:00
RTW_DPRINTF(RTW_DEBUG_RECV_DESC,
("rate %d.%d Mb/s, time %08x%08x\n", (rate * 5) / 10,
(rate * 5) % 10, htsfth, htsftl));
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
#endif /* RTW_DEBUG */
/* if bad flags, skip descriptor */
if ((hstat & RTW_RXSTAT_ONESEG) != RTW_RXSTAT_ONESEG) {
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
aprint_error_dev(sc->sc_dev, "too many rx segments, "
"next=%d, %08" PRIx32 "\n", next, hstat);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
goto next;
}
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
bus_dmamap_sync(sc->sc_dmat, rs->rs_dmamap, 0,
rs->rs_dmamap->dm_mapsize, BUS_DMASYNC_POSTREAD);
Bug fix: in rtw_rxbuf_alloc, unload the previous mbuf chain before loading the new chain.
2004-12-23 09:03:09 +03:00
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
m = rs->rs_mbuf;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
/* if temporarily out of memory, re-use mbuf */
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
switch (rtw_rxsoft_alloc(sc->sc_dmat, rs)) {
Bug fix: in rtw_rxbuf_alloc, unload the previous mbuf chain before loading the new chain.
2004-12-23 09:03:09 +03:00
case 0:
break;
case ENOBUFS:
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
aprint_error_dev(sc->sc_dev,
"rtw_rxsoft_alloc(, %d) failed, dropping packet\n",
next);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
goto next;
Bug fix: in rtw_rxbuf_alloc, unload the previous mbuf chain before loading the new chain.
2004-12-23 09:03:09 +03:00
default:
/* XXX shorten rx ring, instead? */
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
aprint_error_dev(sc->sc_dev,
"could not load DMA map\n");
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
}
Accommodate the variety of RF front-ends when radiotapping frames: tap only RSSI when there is a Philips RF front-end. Tap both Barker lock quality and RSSI when there is any other RF front-end. Provide radiotap listeners a more complete picture of channel activity: in promiscuous mode, tap frames who do not pass the CRC32 check. Flag packets that were received with a short preamble. Ask the NIC to pass us 802.11 Control frames. Pass frames to radiotap listeners before stripping the FCS. Re-order operations in rtw_intr_rx() in order to accomplish all of this, taking care not to pass a broken packet to net80211! Do not provide a flags field when tapping xmitted frames. Assert sane Rx packet lengths. Really should check and drop instead of KASSERTing. I will revisit this, soon. Update copyright.
2007-11-17 02:35:19 +03:00
sq = __SHIFTOUT(hrssi, RTW_RXRSSI_SQ);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
if (sc->sc_rfchipid == RTW_RFCHIPID_PHILIPS)
Accommodate the variety of RF front-ends when radiotapping frames: tap only RSSI when there is a Philips RF front-end. Tap both Barker lock quality and RSSI when there is any other RF front-end. Provide radiotap listeners a more complete picture of channel activity: in promiscuous mode, tap frames who do not pass the CRC32 check. Flag packets that were received with a short preamble. Ask the NIC to pass us 802.11 Control frames. Pass frames to radiotap listeners before stripping the FCS. Re-order operations in rtw_intr_rx() in order to accomplish all of this, taking care not to pass a broken packet to net80211! Do not provide a flags field when tapping xmitted frames. Assert sane Rx packet lengths. Really should check and drop instead of KASSERTing. I will revisit this, soon. Update copyright.
2007-11-17 02:35:19 +03:00
rssi = UINT8_MAX - sq;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
else {
Per discussion on tech-kern and tech-userlevel, move the bit-twiddling macros, __BIT, __BITS, SHIFTIN, SHIFTOUT, and __arraycount() from lib/libkern/libkern.h to sys/cdefs.h. Add a __-prefix to SHIFTIN and SHIFTOUT, and add a manual page for the bit-twiddling macros, bits(3). Make the __BIT and __BITS macros "widthless," as best I can, by changing their type to uintmax_t from uint32_t. XXX The manual page lags this change by a bit. Define __PRIxBIT and __PRIxBITS printf(3) format strings.
2006-08-31 23:24:37 +04:00
rssi = __SHIFTOUT(hrssi, RTW_RXRSSI_IMR_RSSI);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
/* TBD find out each front-end's LNA gain in the
* front-end's units
*/
if ((hrssi & RTW_RXRSSI_IMR_LNA) == 0)
rssi |= 0x80;
}
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
/* Note well: now we cannot recycle the rs_mbuf unless
Add bpf(4) packet taps to rtw(4). Now you can tap DLT_EN10MB (Ethernet II), DLT_IEEE802_11 (802.11), and DLT_IEEE802_11_RADIO (radiotap + 802.11).
2004-12-29 01:30:07 +03:00
* we restore its original length.
*/
Resolve conflicts in importation of 18-May-2005 ath(4) / net80211(9) from FreeBSD. Introduce compatibility shims (sys/dev/ic/ath_netbsd.[ch], sys/net80211/ieee80211_netbsd.[ch]). Update drivers (an, atu, atw, awi, ipw, iwi, rtw, wi) for the new net80211(9) API.
2005-06-22 10:14:51 +04:00
m->m_pkthdr.rcvif = ifp;
IMPORTANT: set m->m_pkthdr.rcvif on rx'd packets. Illegal memory references abound without this fix. In rtw_intr_rx, skip to the next rx packet earlier if the rx packet length is too short. Also in rtw_intr_rx, if a packet is rx'd with unknown rate, printf a warning, drop the packet, recycle the rx buffer, skip to next rx packet.
2004-12-25 10:24:17 +03:00
m->m_pkthdr.len = m->m_len = len;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
Resolve conflicts in importation of 18-May-2005 ath(4) / net80211(9) from FreeBSD. Introduce compatibility shims (sys/dev/ic/ath_netbsd.[ch], sys/net80211/ieee80211_netbsd.[ch]). Update drivers (an, atu, atw, awi, ipw, iwi, rtw, wi) for the new net80211(9) API.
2005-06-22 10:14:51 +04:00
wh = mtod(m, struct ieee80211_frame_min *);
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
if (!IS_BEACON(wh->i_fc[0]))
sc->sc_led_state.ls_event |= RTW_LED_S_RX;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
sc->sc_tsfth = htsfth;
Realtek sample sources are not consistent about the "Digital PHY" bit in the SROM. It seems as if it is set to 1 when the PHY is *analog*, not *digital*. Fix my sources. In rtw_intr_rx, use units of 500kb/s instead of 100kb/s for rate, to be consistent with net80211's expectations. Polish up some debugging ugly messages. Dump raw 802.11 packets if IFF_DEBUG|IFF_LINK2 and RTW_DEBUG is defined. Polish power-state (on/sleep/off) handling. Especially improve support for RFMD (totally untested) and Maxim. For Philips, take the Digital PHY property into account. Call the net80211 watchdog function from rtw_watchdog, so that we scan again if auth/assoc fails. Be a little more cautious about writing register[RTW_TPPOLL], since other drivers are.... Don't frob the high/low-priority queues right now, since I don't use them. Add rtw_join_bss which programs the card with the BSSID and other properties of a BSS. Use it on state transitions. Factor out rtw_set_nettype. Make rtw_recv_beacon call ieee80211_recv_mgmt instead of dropping beacons on the floor! TBD IBSS merges. Change some rtw_debug=2 printfs to rtw_debug=3 (RTW_DPRINTF3) printfs so the console doesn't get spammed so badly at rtw_debug=2. Change some debugging printfs to RTW_DPRINTFs. E.g., print the "RF programming method" only if debugging is enabled.
2004-12-21 02:05:41 +03:00
#ifdef RTW_DEBUG
Resolve conflicts in importation of 18-May-2005 ath(4) / net80211(9) from FreeBSD. Introduce compatibility shims (sys/dev/ic/ath_netbsd.[ch], sys/net80211/ieee80211_netbsd.[ch]). Update drivers (an, atu, atw, awi, ipw, iwi, rtw, wi) for the new net80211(9) API.
2005-06-22 10:14:51 +04:00
if ((ifp->if_flags & (IFF_DEBUG|IFF_LINK2)) ==
Realtek sample sources are not consistent about the "Digital PHY" bit in the SROM. It seems as if it is set to 1 when the PHY is *analog*, not *digital*. Fix my sources. In rtw_intr_rx, use units of 500kb/s instead of 100kb/s for rate, to be consistent with net80211's expectations. Polish up some debugging ugly messages. Dump raw 802.11 packets if IFF_DEBUG|IFF_LINK2 and RTW_DEBUG is defined. Polish power-state (on/sleep/off) handling. Especially improve support for RFMD (totally untested) and Maxim. For Philips, take the Digital PHY property into account. Call the net80211 watchdog function from rtw_watchdog, so that we scan again if auth/assoc fails. Be a little more cautious about writing register[RTW_TPPOLL], since other drivers are.... Don't frob the high/low-priority queues right now, since I don't use them. Add rtw_join_bss which programs the card with the BSSID and other properties of a BSS. Use it on state transitions. Factor out rtw_set_nettype. Make rtw_recv_beacon call ieee80211_recv_mgmt instead of dropping beacons on the floor! TBD IBSS merges. Change some rtw_debug=2 printfs to rtw_debug=3 (RTW_DPRINTF3) printfs so the console doesn't get spammed so badly at rtw_debug=2. Change some debugging printfs to RTW_DPRINTFs. E.g., print the "RF programming method" only if debugging is enabled.
2004-12-21 02:05:41 +03:00
(IFF_DEBUG|IFF_LINK2)) {
ieee80211_dump_pkt(mtod(m, uint8_t *), m->m_pkthdr.len,
rate, rssi);
}
#endif /* RTW_DEBUG */
Add bpf(4) packet taps to rtw(4). Now you can tap DLT_EN10MB (Ethernet II), DLT_IEEE802_11 (802.11), and DLT_IEEE802_11_RADIO (radiotap + 802.11).
2004-12-29 01:30:07 +03:00
if (sc->sc_radiobpf != NULL) {
struct rtw_rx_radiotap_header *rr = &sc->sc_rxtap;
rr->rr_tsft =
htole64(((uint64_t)htsfth << 32) | htsftl);
Accommodate the variety of RF front-ends when radiotapping frames: tap only RSSI when there is a Philips RF front-end. Tap both Barker lock quality and RSSI when there is any other RF front-end. Provide radiotap listeners a more complete picture of channel activity: in promiscuous mode, tap frames who do not pass the CRC32 check. Flag packets that were received with a short preamble. Ask the NIC to pass us 802.11 Control frames. Pass frames to radiotap listeners before stripping the FCS. Re-order operations in rtw_intr_rx() in order to accomplish all of this, taking care not to pass a broken packet to net80211! Do not provide a flags field when tapping xmitted frames. Assert sane Rx packet lengths. Really should check and drop instead of KASSERTing. I will revisit this, soon. Update copyright.
2007-11-17 02:35:19 +03:00
rr->rr_flags = IEEE80211_RADIOTAP_F_FCS;
Add bpf(4) packet taps to rtw(4). Now you can tap DLT_EN10MB (Ethernet II), DLT_IEEE802_11 (802.11), and DLT_IEEE802_11_RADIO (radiotap + 802.11).
2004-12-29 01:30:07 +03:00
if ((hstat & RTW_RXSTAT_SPLCP) != 0)
Accommodate the variety of RF front-ends when radiotapping frames: tap only RSSI when there is a Philips RF front-end. Tap both Barker lock quality and RSSI when there is any other RF front-end. Provide radiotap listeners a more complete picture of channel activity: in promiscuous mode, tap frames who do not pass the CRC32 check. Flag packets that were received with a short preamble. Ask the NIC to pass us 802.11 Control frames. Pass frames to radiotap listeners before stripping the FCS. Re-order operations in rtw_intr_rx() in order to accomplish all of this, taking care not to pass a broken packet to net80211! Do not provide a flags field when tapping xmitted frames. Assert sane Rx packet lengths. Really should check and drop instead of KASSERTing. I will revisit this, soon. Update copyright.
2007-11-17 02:35:19 +03:00
rr->rr_flags |= IEEE80211_RADIOTAP_F_SHORTPRE;
if ((hstat & RTW_RXSTAT_CRC32) != 0)
rr->rr_flags |= IEEE80211_RADIOTAP_F_BADFCS;
Add bpf(4) packet taps to rtw(4). Now you can tap DLT_EN10MB (Ethernet II), DLT_IEEE802_11 (802.11), and DLT_IEEE802_11_RADIO (radiotap + 802.11).
2004-12-29 01:30:07 +03:00
rr->rr_rate = rate;
Accommodate the variety of RF front-ends when radiotapping frames: tap only RSSI when there is a Philips RF front-end. Tap both Barker lock quality and RSSI when there is any other RF front-end. Provide radiotap listeners a more complete picture of channel activity: in promiscuous mode, tap frames who do not pass the CRC32 check. Flag packets that were received with a short preamble. Ask the NIC to pass us 802.11 Control frames. Pass frames to radiotap listeners before stripping the FCS. Re-order operations in rtw_intr_rx() in order to accomplish all of this, taking care not to pass a broken packet to net80211! Do not provide a flags field when tapping xmitted frames. Assert sane Rx packet lengths. Really should check and drop instead of KASSERTing. I will revisit this, soon. Update copyright.
2007-11-17 02:35:19 +03:00
if (sc->sc_rfchipid == RTW_RFCHIPID_PHILIPS)
rr->rr_u.u_philips.p_antsignal = rssi;
else {
rr->rr_u.u_other.o_antsignal = rssi;
rr->rr_u.u_other.o_barker_lock =
htole16(UINT8_MAX - sq);
}
Push the bpf_ops usage back into bpf.h. Push the common ifp->if_bpf check into the inline functions as well the fourth argument for bpf_attach.
2010-04-05 11:19:28 +04:00
bpf_mtap2(sc->sc_radiobpf,
Redefine bpf linkage through an always present op vector, i.e. #if NBPFILTER is no longer required in the client. This change doesn't yet add support for loading bpf as a module, since drivers can register before bpf is attached. However, callers of bpf can now be modularized. Dynamically loadable bpf could probably be done fairly easily with coordination from the stub driver and the real driver by registering attachments in the stub before the real driver is loaded and doing a handoff. ... and I'm not going to ponder the depths of unload here. Tested with i386/MONOLITHIC, modified MONOLITHIC without bpf and rump.
2010-01-20 01:06:18 +03:00
rr, sizeof(sc->sc_rxtapu), m);
Add bpf(4) packet taps to rtw(4). Now you can tap DLT_EN10MB (Ethernet II), DLT_IEEE802_11 (802.11), and DLT_IEEE802_11_RADIO (radiotap + 802.11).
2004-12-29 01:30:07 +03:00
}
Accommodate the variety of RF front-ends when radiotapping frames: tap only RSSI when there is a Philips RF front-end. Tap both Barker lock quality and RSSI when there is any other RF front-end. Provide radiotap listeners a more complete picture of channel activity: in promiscuous mode, tap frames who do not pass the CRC32 check. Flag packets that were received with a short preamble. Ask the NIC to pass us 802.11 Control frames. Pass frames to radiotap listeners before stripping the FCS. Re-order operations in rtw_intr_rx() in order to accomplish all of this, taking care not to pass a broken packet to net80211! Do not provide a flags field when tapping xmitted frames. Assert sane Rx packet lengths. Really should check and drop instead of KASSERTing. I will revisit this, soon. Update copyright.
2007-11-17 02:35:19 +03:00
if ((hstat & RTW_RXSTAT_RES) != 0) {
m_freem(m);
goto next;
}
/* CRC is included with the packet; trim it off. */
m_adj(m, -IEEE80211_CRC_LEN);
/* TBD use _MAR, _BAR, _PAR flags as hints to _find_rxnode? */
ni = ieee80211_find_rxnode(&sc->sc_ic, wh);
Resolve conflicts in importation of 18-May-2005 ath(4) / net80211(9) from FreeBSD. Introduce compatibility shims (sys/dev/ic/ath_netbsd.[ch], sys/net80211/ieee80211_netbsd.[ch]). Update drivers (an, atu, atw, awi, ipw, iwi, rtw, wi) for the new net80211(9) API.
2005-06-22 10:14:51 +04:00
ieee80211_input(&sc->sc_ic, m, ni, rssi, htsftl);
ieee80211_free_node(ni);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
next:
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
rtw_rxdesc_init(rdb, rs, next, 0);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
}
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
#undef IS_BEACON
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
}
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
static void
__unused removal on arguments; approved by core.
2006-11-16 04:32:37 +03:00
rtw_txsoft_release(bus_dma_tag_t dmat, struct ieee80211com *ic,
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
struct rtw_txsoft *ts)
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
{
struct mbuf *m;
struct ieee80211_node *ni;
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
m = ts->ts_mbuf;
ni = ts->ts_ni;
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
KASSERT(m != NULL);
KASSERT(ni != NULL);
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
ts->ts_mbuf = NULL;
ts->ts_ni = NULL;
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
bus_dmamap_sync(dmat, ts->ts_dmamap, 0, ts->ts_dmamap->dm_mapsize,
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
BUS_DMASYNC_POSTWRITE);
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
bus_dmamap_unload(dmat, ts->ts_dmamap);
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
m_freem(m);
Resolve conflicts in importation of 18-May-2005 ath(4) / net80211(9) from FreeBSD. Introduce compatibility shims (sys/dev/ic/ath_netbsd.[ch], sys/net80211/ieee80211_netbsd.[ch]). Update drivers (an, atu, atw, awi, ipw, iwi, rtw, wi) for the new net80211(9) API.
2005-06-22 10:14:51 +04:00
ieee80211_free_node(ni);
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
}
static void
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
rtw_txsofts_release(bus_dma_tag_t dmat, struct ieee80211com *ic,
struct rtw_txsoft_blk *tsb)
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
{
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
struct rtw_txsoft *ts;
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
while ((ts = SIMPLEQ_FIRST(&tsb->tsb_dirtyq)) != NULL) {
rtw_txsoft_release(dmat, ic, ts);
SIMPLEQ_REMOVE_HEAD(&tsb->tsb_dirtyq, ts_q);
SIMPLEQ_INSERT_TAIL(&tsb->tsb_freeq, ts, ts_q);
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
}
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
tsb->tsb_tx_timer = 0;
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
}
Remove leading __ from __(const|inline|signed|volatile) -- it is obsolete.
2005-12-24 23:27:29 +03:00
static inline void
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
rtw_collect_txpkt(struct rtw_softc *sc, struct rtw_txdesc_blk *tdb,
struct rtw_txsoft *ts, int ndesc)
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
{
Extract the transmission status from a tx descriptor, and convert to host byte-order, once.
2004-12-23 08:44:39 +03:00
uint32_t hstat;
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
int data_retry, rts_retry;
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
struct rtw_txdesc *tdn;
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
const char *condstring;
Resolve conflicts in importation of 18-May-2005 ath(4) / net80211(9) from FreeBSD. Introduce compatibility shims (sys/dev/ic/ath_netbsd.[ch], sys/net80211/ieee80211_netbsd.[ch]). Update drivers (an, atu, atw, awi, ipw, iwi, rtw, wi) for the new net80211(9) API.
2005-06-22 10:14:51 +04:00
struct ifnet *ifp = &sc->sc_if;
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
rtw_txsoft_release(sc->sc_dmat, &sc->sc_ic, ts);
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
tdb->tdb_nfree += ndesc;
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
tdn = &tdb->tdb_desc[ts->ts_last];
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
hstat = le32toh(tdn->td_stat);
Per discussion on tech-kern and tech-userlevel, move the bit-twiddling macros, __BIT, __BITS, SHIFTIN, SHIFTOUT, and __arraycount() from lib/libkern/libkern.h to sys/cdefs.h. Add a __-prefix to SHIFTIN and SHIFTOUT, and add a manual page for the bit-twiddling macros, bits(3). Make the __BIT and __BITS macros "widthless," as best I can, by changing their type to uintmax_t from uint32_t. XXX The manual page lags this change by a bit. Define __PRIxBIT and __PRIxBITS printf(3) format strings.
2006-08-31 23:24:37 +04:00
rts_retry = __SHIFTOUT(hstat, RTW_TXSTAT_RTSRETRY_MASK);
data_retry = __SHIFTOUT(hstat, RTW_TXSTAT_DRC_MASK);
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
Resolve conflicts in importation of 18-May-2005 ath(4) / net80211(9) from FreeBSD. Introduce compatibility shims (sys/dev/ic/ath_netbsd.[ch], sys/net80211/ieee80211_netbsd.[ch]). Update drivers (an, atu, atw, awi, ipw, iwi, rtw, wi) for the new net80211(9) API.
2005-06-22 10:14:51 +04:00
ifp->if_collisions += rts_retry + data_retry;
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
Extract the transmission status from a tx descriptor, and convert to host byte-order, once.
2004-12-23 08:44:39 +03:00
if ((hstat & RTW_TXSTAT_TOK) != 0)
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
condstring = "ok";
else {
Resolve conflicts in importation of 18-May-2005 ath(4) / net80211(9) from FreeBSD. Introduce compatibility shims (sys/dev/ic/ath_netbsd.[ch], sys/net80211/ieee80211_netbsd.[ch]). Update drivers (an, atu, atw, awi, ipw, iwi, rtw, wi) for the new net80211(9) API.
2005-06-22 10:14:51 +04:00
ifp->if_oerrors++;
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
condstring = "error";
}
- remove unused variables - move variables inside ifdef sections - ifdef notdef unused code - use __USE for debugging variables
2013-10-18 01:24:24 +04:00
#ifdef RTW_DEBUG
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
DPRINTF(sc, RTW_DEBUG_XMIT_DESC,
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
("%s: ts %p txdesc[%d, %d] %s tries rts %u data %u\n",
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
device_xname(sc->sc_dev), ts, ts->ts_first, ts->ts_last,
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
condstring, rts_retry, data_retry));
- remove unused variables - move variables inside ifdef sections - ifdef notdef unused code - use __USE for debugging variables
2013-10-18 01:24:24 +04:00
#else
__USE(condstring);
#endif
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
}
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
static void
rtw_reset_oactive(struct rtw_softc *sc)
{
short oflags;
int pri;
struct rtw_txsoft_blk *tsb;
struct rtw_txdesc_blk *tdb;
oflags = sc->sc_if.if_flags;
for (pri = 0; pri < RTW_NTXPRI; pri++) {
tsb = &sc->sc_txsoft_blk[pri];
tdb = &sc->sc_txdesc_blk[pri];
if (!SIMPLEQ_EMPTY(&tsb->tsb_freeq) && tdb->tdb_nfree > 0)
sc->sc_if.if_flags &= ~IFF_OACTIVE;
}
if (oflags != sc->sc_if.if_flags) {
DPRINTF(sc, RTW_DEBUG_OACTIVE,
("%s: reset OACTIVE\n", __func__));
}
}
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
/* Collect transmitted packets. */
Add the WEP capability to our net80211 capability flags, since we do sort of support hardware WEP. Only change the WEP key index in a Tx descriptor from 0 if the type of our transmit key is _WEP, not _TKIP: i.e., only if we're really doing WEP crypto in the hardware. Ignore a watchdog timeout on any Tx ring if we can collect some packets from that ring. Restart both the receiver and the transmitter when a watchdog timeout occurs instead of restarting only the transmitter.
2009-04-02 04:57:20 +04:00
static bool
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
rtw_collect_txring(struct rtw_softc *sc, struct rtw_txsoft_blk *tsb,
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
struct rtw_txdesc_blk *tdb, int force)
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
{
Add the WEP capability to our net80211 capability flags, since we do sort of support hardware WEP. Only change the WEP key index in a Tx descriptor from 0 if the type of our transmit key is _WEP, not _TKIP: i.e., only if we're really doing WEP crypto in the hardware. Ignore a watchdog timeout on any Tx ring if we can collect some packets from that ring. Restart both the receiver and the transmitter when a watchdog timeout occurs instead of restarting only the transmitter.
2009-04-02 04:57:20 +04:00
bool collected = false;
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
int ndesc;
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
struct rtw_txsoft *ts;
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
I have made a small change that suppresses "recovery" from (spurious?) tx timeouts on the beacon queue. The driver's recovery process badly disrupts the MAC's receiver. This needs a closer look, later. Every other fix is by Charles M. Hannum. Charles writes: "With these, the device is quite stable under load, does not experience xmit timeouts, no longer mysteriously freezes, and does not spew a lot of garbage. It does, however, continue to get some "tx fifo underflow"s. global, SYSCTL_SETUP(): a) Removed unused sysctl variables. rtw_io_enable(): b) Change rtw_io_enable to take a softc, and make it call rtw_txring_fixup(). c) Enable the PCI multiple read/write bit. d) Always reset rdb_next to 0 when starting the receive engine. rtw_intr_rx(): e) When processing receive packets, store the next pointer in rdb_next and read it back on the next round, to fix a race condition vs. rtw_tune() during AP discovery and association. f) Remove the special "skip to 0th descriptor" code, because it is no longer needed. rtw_collect_txring(): g) In the "force" case, stop when we catch up to the hardware xmit pointer. rtw_intr_beacon(): h) Resetting tdb_next here looked fishy to me, in light of changes in rtw_collect_txring(). XXX However, this is the one part I haven't tested. rtw_txring_fixup(): i) Add function and line arguments, so we can print out information about the caller. j) In addition to reading the hardware pointer, skip over any additional packets with OWN set -- we may have queued more. rtw_rxring_fixup(): k) Remove this, because it's completely wrong. rtw_intr_ioerror(): l) Don't handle receive overruns here; handle them in the normal receive path. Also changed RTW_INTR_RX and RTW_INTR_IOERROR accordingly. m) Do not fiddle with CR; the transmit engine is still running just fine. n) Do not fiddle with IMR; it was unnecessary. o) Don't touch the receive engine at all; not necessary now that we're leaving CR alone. p) Remove the rtw_do_chip_reset case; it was dangerous and not actually used. rtw_intr(): q) Don't mask off ISR bits when we pass down the ISR value; it's not necessary, and sometimes the other bits useful in debugging output. rtw_seg_too_short(): r) Don't spew here; the caller will do it if necessary. (Stops spewage in an expected case.) rtw_rxdesc_blk: s) Minor change to make the elements have the same order as txdesc_blk. general: t) Added a bunch of debugging output, #ifdef RTW_DIAG, that was useful while investigating the ring management problems. Also moved rtw_txring_fixup() into RTW_DIAG, as I believe the transmit pointer is always kept in sync now, and the function has potential race conditions.
2007-01-06 11:27:53 +03:00
#ifdef RTW_DEBUG
rtw_dump_rings(sc);
#endif
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
while ((ts = SIMPLEQ_FIRST(&tsb->tsb_dirtyq)) != NULL) {
I have made a small change that suppresses "recovery" from (spurious?) tx timeouts on the beacon queue. The driver's recovery process badly disrupts the MAC's receiver. This needs a closer look, later. Every other fix is by Charles M. Hannum. Charles writes: "With these, the device is quite stable under load, does not experience xmit timeouts, no longer mysteriously freezes, and does not spew a lot of garbage. It does, however, continue to get some "tx fifo underflow"s. global, SYSCTL_SETUP(): a) Removed unused sysctl variables. rtw_io_enable(): b) Change rtw_io_enable to take a softc, and make it call rtw_txring_fixup(). c) Enable the PCI multiple read/write bit. d) Always reset rdb_next to 0 when starting the receive engine. rtw_intr_rx(): e) When processing receive packets, store the next pointer in rdb_next and read it back on the next round, to fix a race condition vs. rtw_tune() during AP discovery and association. f) Remove the special "skip to 0th descriptor" code, because it is no longer needed. rtw_collect_txring(): g) In the "force" case, stop when we catch up to the hardware xmit pointer. rtw_intr_beacon(): h) Resetting tdb_next here looked fishy to me, in light of changes in rtw_collect_txring(). XXX However, this is the one part I haven't tested. rtw_txring_fixup(): i) Add function and line arguments, so we can print out information about the caller. j) In addition to reading the hardware pointer, skip over any additional packets with OWN set -- we may have queued more. rtw_rxring_fixup(): k) Remove this, because it's completely wrong. rtw_intr_ioerror(): l) Don't handle receive overruns here; handle them in the normal receive path. Also changed RTW_INTR_RX and RTW_INTR_IOERROR accordingly. m) Do not fiddle with CR; the transmit engine is still running just fine. n) Do not fiddle with IMR; it was unnecessary. o) Don't touch the receive engine at all; not necessary now that we're leaving CR alone. p) Remove the rtw_do_chip_reset case; it was dangerous and not actually used. rtw_intr(): q) Don't mask off ISR bits when we pass down the ISR value; it's not necessary, and sometimes the other bits useful in debugging output. rtw_seg_too_short(): r) Don't spew here; the caller will do it if necessary. (Stops spewage in an expected case.) rtw_rxdesc_blk: s) Minor change to make the elements have the same order as txdesc_blk. general: t) Added a bunch of debugging output, #ifdef RTW_DIAG, that was useful while investigating the ring management problems. Also moved rtw_txring_fixup() into RTW_DIAG, as I believe the transmit pointer is always kept in sync now, and the function has potential race conditions.
2007-01-06 11:27:53 +03:00
/* If we're clearing a failed transmission, only clear
up to the last packet the hardware has processed. */
if (ts->ts_first == rtw_txring_next(&sc->sc_regs, tdb))
break;
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
ndesc = 1 + ts->ts_last - ts->ts_first;
if (ts->ts_last < ts->ts_first)
ndesc += tdb->tdb_ndesc;
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
Assert that I'm DMA-syncing a non-negative number of transmit descriptors. As I free dirty transmit buffers, move them to the free list, not back to the dirty list! Add a 5-second transmit timeout.
2004-12-20 03:16:21 +03:00
KASSERT(ndesc > 0);
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
rtw_txdescs_sync(tdb, ts->ts_first, ndesc,
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
BUS_DMASYNC_POSTREAD|BUS_DMASYNC_POSTWRITE);
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
if (force) {
I have made a small change that suppresses "recovery" from (spurious?) tx timeouts on the beacon queue. The driver's recovery process badly disrupts the MAC's receiver. This needs a closer look, later. Every other fix is by Charles M. Hannum. Charles writes: "With these, the device is quite stable under load, does not experience xmit timeouts, no longer mysteriously freezes, and does not spew a lot of garbage. It does, however, continue to get some "tx fifo underflow"s. global, SYSCTL_SETUP(): a) Removed unused sysctl variables. rtw_io_enable(): b) Change rtw_io_enable to take a softc, and make it call rtw_txring_fixup(). c) Enable the PCI multiple read/write bit. d) Always reset rdb_next to 0 when starting the receive engine. rtw_intr_rx(): e) When processing receive packets, store the next pointer in rdb_next and read it back on the next round, to fix a race condition vs. rtw_tune() during AP discovery and association. f) Remove the special "skip to 0th descriptor" code, because it is no longer needed. rtw_collect_txring(): g) In the "force" case, stop when we catch up to the hardware xmit pointer. rtw_intr_beacon(): h) Resetting tdb_next here looked fishy to me, in light of changes in rtw_collect_txring(). XXX However, this is the one part I haven't tested. rtw_txring_fixup(): i) Add function and line arguments, so we can print out information about the caller. j) In addition to reading the hardware pointer, skip over any additional packets with OWN set -- we may have queued more. rtw_rxring_fixup(): k) Remove this, because it's completely wrong. rtw_intr_ioerror(): l) Don't handle receive overruns here; handle them in the normal receive path. Also changed RTW_INTR_RX and RTW_INTR_IOERROR accordingly. m) Do not fiddle with CR; the transmit engine is still running just fine. n) Do not fiddle with IMR; it was unnecessary. o) Don't touch the receive engine at all; not necessary now that we're leaving CR alone. p) Remove the rtw_do_chip_reset case; it was dangerous and not actually used. rtw_intr(): q) Don't mask off ISR bits when we pass down the ISR value; it's not necessary, and sometimes the other bits useful in debugging output. rtw_seg_too_short(): r) Don't spew here; the caller will do it if necessary. (Stops spewage in an expected case.) rtw_rxdesc_blk: s) Minor change to make the elements have the same order as txdesc_blk. general: t) Added a bunch of debugging output, #ifdef RTW_DIAG, that was useful while investigating the ring management problems. Also moved rtw_txring_fixup() into RTW_DIAG, as I believe the transmit pointer is always kept in sync now, and the function has potential race conditions.
2007-01-06 11:27:53 +03:00
int next;
#ifdef RTW_DIAG
printf("%s: clearing packet, stats", __func__);
#endif
for (next = ts->ts_first; ;
next = RTW_NEXT_IDX(tdb, next)) {
#ifdef RTW_DIAG
printf(" %" PRIx32 "/%" PRIx32 "/%" PRIx32 "/%" PRIu32 "/%" PRIx32, le32toh(tdb->tdb_desc[next].td_stat), le32toh(tdb->tdb_desc[next].td_ctl1), le32toh(tdb->tdb_desc[next].td_buf), le32toh(tdb->tdb_desc[next].td_len), le32toh(tdb->tdb_desc[next].td_next));
#endif
tdb->tdb_desc[next].td_stat &=
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
~htole32(RTW_TXSTAT_OWN);
I have made a small change that suppresses "recovery" from (spurious?) tx timeouts on the beacon queue. The driver's recovery process badly disrupts the MAC's receiver. This needs a closer look, later. Every other fix is by Charles M. Hannum. Charles writes: "With these, the device is quite stable under load, does not experience xmit timeouts, no longer mysteriously freezes, and does not spew a lot of garbage. It does, however, continue to get some "tx fifo underflow"s. global, SYSCTL_SETUP(): a) Removed unused sysctl variables. rtw_io_enable(): b) Change rtw_io_enable to take a softc, and make it call rtw_txring_fixup(). c) Enable the PCI multiple read/write bit. d) Always reset rdb_next to 0 when starting the receive engine. rtw_intr_rx(): e) When processing receive packets, store the next pointer in rdb_next and read it back on the next round, to fix a race condition vs. rtw_tune() during AP discovery and association. f) Remove the special "skip to 0th descriptor" code, because it is no longer needed. rtw_collect_txring(): g) In the "force" case, stop when we catch up to the hardware xmit pointer. rtw_intr_beacon(): h) Resetting tdb_next here looked fishy to me, in light of changes in rtw_collect_txring(). XXX However, this is the one part I haven't tested. rtw_txring_fixup(): i) Add function and line arguments, so we can print out information about the caller. j) In addition to reading the hardware pointer, skip over any additional packets with OWN set -- we may have queued more. rtw_rxring_fixup(): k) Remove this, because it's completely wrong. rtw_intr_ioerror(): l) Don't handle receive overruns here; handle them in the normal receive path. Also changed RTW_INTR_RX and RTW_INTR_IOERROR accordingly. m) Do not fiddle with CR; the transmit engine is still running just fine. n) Do not fiddle with IMR; it was unnecessary. o) Don't touch the receive engine at all; not necessary now that we're leaving CR alone. p) Remove the rtw_do_chip_reset case; it was dangerous and not actually used. rtw_intr(): q) Don't mask off ISR bits when we pass down the ISR value; it's not necessary, and sometimes the other bits useful in debugging output. rtw_seg_too_short(): r) Don't spew here; the caller will do it if necessary. (Stops spewage in an expected case.) rtw_rxdesc_blk: s) Minor change to make the elements have the same order as txdesc_blk. general: t) Added a bunch of debugging output, #ifdef RTW_DIAG, that was useful while investigating the ring management problems. Also moved rtw_txring_fixup() into RTW_DIAG, as I believe the transmit pointer is always kept in sync now, and the function has potential race conditions.
2007-01-06 11:27:53 +03:00
if (next == ts->ts_last)
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
break;
}
rtw_txdescs_sync(tdb, ts->ts_first, ndesc,
BUS_DMASYNC_PREREAD|BUS_DMASYNC_PREWRITE);
I have made a small change that suppresses "recovery" from (spurious?) tx timeouts on the beacon queue. The driver's recovery process badly disrupts the MAC's receiver. This needs a closer look, later. Every other fix is by Charles M. Hannum. Charles writes: "With these, the device is quite stable under load, does not experience xmit timeouts, no longer mysteriously freezes, and does not spew a lot of garbage. It does, however, continue to get some "tx fifo underflow"s. global, SYSCTL_SETUP(): a) Removed unused sysctl variables. rtw_io_enable(): b) Change rtw_io_enable to take a softc, and make it call rtw_txring_fixup(). c) Enable the PCI multiple read/write bit. d) Always reset rdb_next to 0 when starting the receive engine. rtw_intr_rx(): e) When processing receive packets, store the next pointer in rdb_next and read it back on the next round, to fix a race condition vs. rtw_tune() during AP discovery and association. f) Remove the special "skip to 0th descriptor" code, because it is no longer needed. rtw_collect_txring(): g) In the "force" case, stop when we catch up to the hardware xmit pointer. rtw_intr_beacon(): h) Resetting tdb_next here looked fishy to me, in light of changes in rtw_collect_txring(). XXX However, this is the one part I haven't tested. rtw_txring_fixup(): i) Add function and line arguments, so we can print out information about the caller. j) In addition to reading the hardware pointer, skip over any additional packets with OWN set -- we may have queued more. rtw_rxring_fixup(): k) Remove this, because it's completely wrong. rtw_intr_ioerror(): l) Don't handle receive overruns here; handle them in the normal receive path. Also changed RTW_INTR_RX and RTW_INTR_IOERROR accordingly. m) Do not fiddle with CR; the transmit engine is still running just fine. n) Do not fiddle with IMR; it was unnecessary. o) Don't touch the receive engine at all; not necessary now that we're leaving CR alone. p) Remove the rtw_do_chip_reset case; it was dangerous and not actually used. rtw_intr(): q) Don't mask off ISR bits when we pass down the ISR value; it's not necessary, and sometimes the other bits useful in debugging output. rtw_seg_too_short(): r) Don't spew here; the caller will do it if necessary. (Stops spewage in an expected case.) rtw_rxdesc_blk: s) Minor change to make the elements have the same order as txdesc_blk. general: t) Added a bunch of debugging output, #ifdef RTW_DIAG, that was useful while investigating the ring management problems. Also moved rtw_txring_fixup() into RTW_DIAG, as I believe the transmit pointer is always kept in sync now, and the function has potential race conditions.
2007-01-06 11:27:53 +03:00
#ifdef RTW_DIAG
next = RTW_NEXT_IDX(tdb, next);
printf(" -> end %u stat %" PRIx32 ", was %u\n", next,
le32toh(tdb->tdb_desc[next].td_stat),
rtw_txring_next(&sc->sc_regs, tdb));
#endif
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
} else if ((tdb->tdb_desc[ts->ts_last].td_stat &
After we examine a rx/tx descriptor and find that it's still owned by the NIC, re-sync the descriptor so that a cacheline will not "cover" the NIC's modifications to the descriptor when we next examine it. This compiles, but I have not tested it, yet.
2007-01-06 09:07:55 +03:00
htole32(RTW_TXSTAT_OWN)) != 0) {
rtw_txdescs_sync(tdb, ts->ts_last, 1,
BUS_DMASYNC_PREREAD);
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
break;
After we examine a rx/tx descriptor and find that it's still owned by the NIC, re-sync the descriptor so that a cacheline will not "cover" the NIC's modifications to the descriptor when we next examine it. This compiles, but I have not tested it, yet.
2007-01-06 09:07:55 +03:00
}
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
Add the WEP capability to our net80211 capability flags, since we do sort of support hardware WEP. Only change the WEP key index in a Tx descriptor from 0 if the type of our transmit key is _WEP, not _TKIP: i.e., only if we're really doing WEP crypto in the hardware. Ignore a watchdog timeout on any Tx ring if we can collect some packets from that ring. Restart both the receiver and the transmitter when a watchdog timeout occurs instead of restarting only the transmitter.
2009-04-02 04:57:20 +04:00
collected = true;
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
rtw_collect_txpkt(sc, tdb, ts, ndesc);
SIMPLEQ_REMOVE_HEAD(&tsb->tsb_dirtyq, ts_q);
SIMPLEQ_INSERT_TAIL(&tsb->tsb_freeq, ts, ts_q);
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
}
I have made a small change that suppresses "recovery" from (spurious?) tx timeouts on the beacon queue. The driver's recovery process badly disrupts the MAC's receiver. This needs a closer look, later. Every other fix is by Charles M. Hannum. Charles writes: "With these, the device is quite stable under load, does not experience xmit timeouts, no longer mysteriously freezes, and does not spew a lot of garbage. It does, however, continue to get some "tx fifo underflow"s. global, SYSCTL_SETUP(): a) Removed unused sysctl variables. rtw_io_enable(): b) Change rtw_io_enable to take a softc, and make it call rtw_txring_fixup(). c) Enable the PCI multiple read/write bit. d) Always reset rdb_next to 0 when starting the receive engine. rtw_intr_rx(): e) When processing receive packets, store the next pointer in rdb_next and read it back on the next round, to fix a race condition vs. rtw_tune() during AP discovery and association. f) Remove the special "skip to 0th descriptor" code, because it is no longer needed. rtw_collect_txring(): g) In the "force" case, stop when we catch up to the hardware xmit pointer. rtw_intr_beacon(): h) Resetting tdb_next here looked fishy to me, in light of changes in rtw_collect_txring(). XXX However, this is the one part I haven't tested. rtw_txring_fixup(): i) Add function and line arguments, so we can print out information about the caller. j) In addition to reading the hardware pointer, skip over any additional packets with OWN set -- we may have queued more. rtw_rxring_fixup(): k) Remove this, because it's completely wrong. rtw_intr_ioerror(): l) Don't handle receive overruns here; handle them in the normal receive path. Also changed RTW_INTR_RX and RTW_INTR_IOERROR accordingly. m) Do not fiddle with CR; the transmit engine is still running just fine. n) Do not fiddle with IMR; it was unnecessary. o) Don't touch the receive engine at all; not necessary now that we're leaving CR alone. p) Remove the rtw_do_chip_reset case; it was dangerous and not actually used. rtw_intr(): q) Don't mask off ISR bits when we pass down the ISR value; it's not necessary, and sometimes the other bits useful in debugging output. rtw_seg_too_short(): r) Don't spew here; the caller will do it if necessary. (Stops spewage in an expected case.) rtw_rxdesc_blk: s) Minor change to make the elements have the same order as txdesc_blk. general: t) Added a bunch of debugging output, #ifdef RTW_DIAG, that was useful while investigating the ring management problems. Also moved rtw_txring_fixup() into RTW_DIAG, as I believe the transmit pointer is always kept in sync now, and the function has potential race conditions.
2007-01-06 11:27:53 +03:00
Clean up bogus whitespace
2006-05-26 16:52:45 +04:00
/* no more pending transmissions, cancel watchdog */
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
if (ts == NULL)
tsb->tsb_tx_timer = 0;
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
rtw_reset_oactive(sc);
Add the WEP capability to our net80211 capability flags, since we do sort of support hardware WEP. Only change the WEP key index in a Tx descriptor from 0 if the type of our transmit key is _WEP, not _TKIP: i.e., only if we're really doing WEP crypto in the hardware. Ignore a watchdog timeout on any Tx ring if we can collect some packets from that ring. Restart both the receiver and the transmitter when a watchdog timeout occurs instead of restarting only the transmitter.
2009-04-02 04:57:20 +04:00
return collected;
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
}
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
static void
Cosmetic: change u_int{8,16,32,64}_t to C99-style uint{8,16,32,64}_t, which is preferred.
2005-01-02 07:23:03 +03:00
rtw_intr_tx(struct rtw_softc *sc, uint16_t isr)
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
{
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
int pri;
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
struct rtw_txsoft_blk *tsb;
struct rtw_txdesc_blk *tdb;
Resolve conflicts in importation of 18-May-2005 ath(4) / net80211(9) from FreeBSD. Introduce compatibility shims (sys/dev/ic/ath_netbsd.[ch], sys/net80211/ieee80211_netbsd.[ch]). Update drivers (an, atu, atw, awi, ipw, iwi, rtw, wi) for the new net80211(9) API.
2005-06-22 10:14:51 +04:00
struct ifnet *ifp = &sc->sc_if;
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
for (pri = 0; pri < RTW_NTXPRI; pri++) {
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
tsb = &sc->sc_txsoft_blk[pri];
tdb = &sc->sc_txdesc_blk[pri];
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
rtw_collect_txring(sc, tsb, tdb, 0);
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
}
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
if ((isr & RTW_INTR_TX) != 0)
rtw_start(ifp);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
return;
}
static void
Cosmetic: change u_int{8,16,32,64}_t to C99-style uint{8,16,32,64}_t, which is preferred.
2005-01-02 07:23:03 +03:00
rtw_intr_beacon(struct rtw_softc *sc, uint16_t isr)
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
{
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
u_int next;
uint32_t tsfth, tsftl;
struct ieee80211com *ic;
struct rtw_txdesc_blk *tdb = &sc->sc_txdesc_blk[RTW_TXPRIBCN];
struct rtw_txsoft_blk *tsb = &sc->sc_txsoft_blk[RTW_TXPRIBCN];
struct mbuf *m;
tsfth = RTW_READ(&sc->sc_regs, RTW_TSFTRH);
tsftl = RTW_READ(&sc->sc_regs, RTW_TSFTRL);
if ((isr & (RTW_INTR_TBDOK|RTW_INTR_TBDER)) != 0) {
next = rtw_txring_next(&sc->sc_regs, tdb);
- remove unused variables - move variables inside ifdef sections - ifdef notdef unused code - use __USE for debugging variables
2013-10-18 01:24:24 +04:00
#ifdef RTW_DEBUG
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
RTW_DPRINTF(RTW_DEBUG_BEACON,
("%s: beacon ring %sprocessed, isr = %#04" PRIx16
", next %u expected %u, %" PRIu64 "\n", __func__,
(next == tdb->tdb_next) ? "" : "un", isr, next,
tdb->tdb_next, (uint64_t)tsfth << 32 | tsftl));
- remove unused variables - move variables inside ifdef sections - ifdef notdef unused code - use __USE for debugging variables
2013-10-18 01:24:24 +04:00
#else
__USE(next);
__USE(tsfth);
__USE(tsftl);
#endif
I have made a small change that suppresses "recovery" from (spurious?) tx timeouts on the beacon queue. The driver's recovery process badly disrupts the MAC's receiver. This needs a closer look, later. Every other fix is by Charles M. Hannum. Charles writes: "With these, the device is quite stable under load, does not experience xmit timeouts, no longer mysteriously freezes, and does not spew a lot of garbage. It does, however, continue to get some "tx fifo underflow"s. global, SYSCTL_SETUP(): a) Removed unused sysctl variables. rtw_io_enable(): b) Change rtw_io_enable to take a softc, and make it call rtw_txring_fixup(). c) Enable the PCI multiple read/write bit. d) Always reset rdb_next to 0 when starting the receive engine. rtw_intr_rx(): e) When processing receive packets, store the next pointer in rdb_next and read it back on the next round, to fix a race condition vs. rtw_tune() during AP discovery and association. f) Remove the special "skip to 0th descriptor" code, because it is no longer needed. rtw_collect_txring(): g) In the "force" case, stop when we catch up to the hardware xmit pointer. rtw_intr_beacon(): h) Resetting tdb_next here looked fishy to me, in light of changes in rtw_collect_txring(). XXX However, this is the one part I haven't tested. rtw_txring_fixup(): i) Add function and line arguments, so we can print out information about the caller. j) In addition to reading the hardware pointer, skip over any additional packets with OWN set -- we may have queued more. rtw_rxring_fixup(): k) Remove this, because it's completely wrong. rtw_intr_ioerror(): l) Don't handle receive overruns here; handle them in the normal receive path. Also changed RTW_INTR_RX and RTW_INTR_IOERROR accordingly. m) Do not fiddle with CR; the transmit engine is still running just fine. n) Do not fiddle with IMR; it was unnecessary. o) Don't touch the receive engine at all; not necessary now that we're leaving CR alone. p) Remove the rtw_do_chip_reset case; it was dangerous and not actually used. rtw_intr(): q) Don't mask off ISR bits when we pass down the ISR value; it's not necessary, and sometimes the other bits useful in debugging output. rtw_seg_too_short(): r) Don't spew here; the caller will do it if necessary. (Stops spewage in an expected case.) rtw_rxdesc_blk: s) Minor change to make the elements have the same order as txdesc_blk. general: t) Added a bunch of debugging output, #ifdef RTW_DIAG, that was useful while investigating the ring management problems. Also moved rtw_txring_fixup() into RTW_DIAG, as I believe the transmit pointer is always kept in sync now, and the function has potential race conditions.
2007-01-06 11:27:53 +03:00
if ((RTW_READ8(&sc->sc_regs, RTW_TPPOLL) & RTW_TPPOLL_BQ) == 0)
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
rtw_collect_txring(sc, tsb, tdb, 1);
}
/* Start beacon transmission. */
if ((isr & RTW_INTR_BCNINT) != 0 &&
sc->sc_ic.ic_state == IEEE80211_S_RUN &&
SIMPLEQ_EMPTY(&tsb->tsb_dirtyq)) {
RTW_DPRINTF(RTW_DEBUG_BEACON,
("%s: beacon prep. time, isr = %#04" PRIx16
", %16" PRIu64 "\n", __func__, isr,
(uint64_t)tsfth << 32 | tsftl));
ic = &sc->sc_ic;
m = rtw_beacon_alloc(sc, ic->ic_bss);
if (m == NULL) {
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
aprint_error_dev(sc->sc_dev,
"could not allocate beacon\n");
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
return;
}
m->m_pkthdr.rcvif = (void *)ieee80211_ref_node(ic->ic_bss);
IF_ENQUEUE(&sc->sc_beaconq, m);
rtw_start(&sc->sc_if);
}
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
}
static void
__unused removal on arguments; approved by core.
2006-11-16 04:32:37 +03:00
rtw_intr_atim(struct rtw_softc *sc)
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
{
/* TBD */
return;
}
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
#ifdef RTW_DEBUG
static void
rtw_dump_rings(struct rtw_softc *sc)
{
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
struct rtw_txdesc_blk *tdb;
struct rtw_rxdesc *rd;
struct rtw_rxdesc_blk *rdb;
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
int desc, pri;
if ((rtw_debug & RTW_DEBUG_IO_KICK) == 0)
return;
for (pri = 0; pri < RTW_NTXPRI; pri++) {
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
tdb = &sc->sc_txdesc_blk[pri];
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
printf("%s: txpri %d ndesc %d nfree %d\n", __func__, pri,
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
tdb->tdb_ndesc, tdb->tdb_nfree);
for (desc = 0; desc < tdb->tdb_ndesc; desc++)
rtw_print_txdesc(sc, ".", NULL, tdb, desc);
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
}
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
rdb = &sc->sc_rxdesc_blk;
Consolidate variables related to the rx ring in sc_rxdesc_blk, which is a struct rtw_rxdesc_blk. Put a copy of the DMA tag and the DMA map into the rx- and tx-ring blocks so that I don't have to pass them to subroutines all of the time.
2004-12-29 03:21:37 +03:00
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
for (desc = 0; desc < RTW_RXQLEN; desc++) {
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
rd = &rdb->rdb_desc[desc];
If defined(RTW_DEBUG), provide a sysctl (hw.rtw.rxbufs_limit) for limiting the number of rx buffers an rtw may allocate. Use this sysctl to test the code that copes with buffer exhaustion. Allocate at most RTW_RXQLEN rx buffers, stopping at the sysctl limit. Record in sc_nrxdesc how many were allocated, and put the end-of-ring indication on sc_rxdesc[sc_nrxdesc - 1]. In rtw_init, if no rx buffers could be allocated, log a complaint, clear IFF_RUNNING, and exit with an error. Many changes to accomodate a short rx ring, mainly of the "add a rx-ring length argument" variety. XXX I really should consolidate all of the rx ring variables in one struct and pass that to the rx-ring subroutines. Bug fix: after calling MCLGET, use the (m->m_flags & M_EXT) idiom to check for success, instead of m != NULL. Bug fix: at the top of rtw_start, if IFF_RUNNING is not set, or IFF_OACTIVE is, get out.
2004-12-29 01:21:15 +03:00
printf("%s: %sctl %08x rsvd0/rssi %08x buf/tsftl %08x "
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
"rsvd1/tsfth %08x\n", __func__,
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
(desc >= rdb->rdb_ndesc) ? "UNUSED " : "",
le32toh(rd->rd_ctl), le32toh(rd->rd_rssi),
le32toh(rd->rd_buf), le32toh(rd->rd_tsfth));
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
}
}
#endif /* RTW_DEBUG */
Miscellaneous changes. Details below. Important changes flagged with []. Using the driver with my Linksys WPC11 ver. 4, it seems to be receiving packets for a change. The WPC11 ver. 4 has a Maxim RF section. My no-name rtw with Philips RF section still does not receive any packets. Keep access-level (analog params > config[0123] registers > none) in sc_access. Add rtw_set_access for changing the access level. Make rtw_continuous_tx_enable and other subroutines use rtw_set_access instead of rtw_config0123_enable and rtw_anaparm_enable. Factor part of the chip-reset code into rtw_chip_reset1. Change the 'struct foo (*bar)[N]'-style arguments to 'struct foo *bar'-style arguments. Consolidate software/hardware Tx/Rx ring setup in rtw_hwring_setup, rtw_swring_setup. Add a new constant, SA2400_OPMODE_DEFAULTS, for the bits that we *always* set in the SA2400 OPMODE register. Factor some code out into rtw_sa2400_calibrate. (Inspired by the Linux driver.) [] When the receiver goes into underrun/overflow state, call a new subroutine, rtw_kick() that stops the Rx/Tx processes, resets the chip, reinitializes the Tx/Rx rings, and restarts Rx/Tx processes. (Inspired by the Linux driver.) [] In rtw_intr_rx, check for too-short packets before calling ieee80211_find_rxnode. I believe this will prevent a repeat of the MCHK exception I saw once on macppc. [] Use seconds-elapased as well as microseconds-elapsed to set the next "due date" for the timeout interrupt. This keeps the driver from programming the timeout to expire too early. [] In rtw_intr, read RTW_ISR at most 10 times, then get out. If the interface is not enabled (RTW_F_ENABLED), then get out. [] In rtw_stop, get out if the interface is not enabled (RTW_F_ENABLED). Block IPL_NET interrupts. Don't read/write any registers if the interface is invalid (RTW_F_INVALID). [] Call rtw_stop in rtw_detach.
2004-12-12 09:37:59 +03:00
static void
rtw_hwring_setup(struct rtw_softc *sc)
{
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
int pri;
Miscellaneous changes. Details below. Important changes flagged with []. Using the driver with my Linksys WPC11 ver. 4, it seems to be receiving packets for a change. The WPC11 ver. 4 has a Maxim RF section. My no-name rtw with Philips RF section still does not receive any packets. Keep access-level (analog params > config[0123] registers > none) in sc_access. Add rtw_set_access for changing the access level. Make rtw_continuous_tx_enable and other subroutines use rtw_set_access instead of rtw_config0123_enable and rtw_anaparm_enable. Factor part of the chip-reset code into rtw_chip_reset1. Change the 'struct foo (*bar)[N]'-style arguments to 'struct foo *bar'-style arguments. Consolidate software/hardware Tx/Rx ring setup in rtw_hwring_setup, rtw_swring_setup. Add a new constant, SA2400_OPMODE_DEFAULTS, for the bits that we *always* set in the SA2400 OPMODE register. Factor some code out into rtw_sa2400_calibrate. (Inspired by the Linux driver.) [] When the receiver goes into underrun/overflow state, call a new subroutine, rtw_kick() that stops the Rx/Tx processes, resets the chip, reinitializes the Tx/Rx rings, and restarts Rx/Tx processes. (Inspired by the Linux driver.) [] In rtw_intr_rx, check for too-short packets before calling ieee80211_find_rxnode. I believe this will prevent a repeat of the MCHK exception I saw once on macppc. [] Use seconds-elapased as well as microseconds-elapsed to set the next "due date" for the timeout interrupt. This keeps the driver from programming the timeout to expire too early. [] In rtw_intr, read RTW_ISR at most 10 times, then get out. If the interface is not enabled (RTW_F_ENABLED), then get out. [] In rtw_stop, get out if the interface is not enabled (RTW_F_ENABLED). Block IPL_NET interrupts. Don't read/write any registers if the interface is invalid (RTW_F_INVALID). [] Call rtw_stop in rtw_detach.
2004-12-12 09:37:59 +03:00
struct rtw_regs *regs = &sc->sc_regs;
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
struct rtw_txdesc_blk *tdb;
sc->sc_txdesc_blk[RTW_TXPRILO].tdb_basereg = RTW_TLPDA;
sc->sc_txdesc_blk[RTW_TXPRILO].tdb_base = RTW_RING_BASE(sc, hd_txlo);
sc->sc_txdesc_blk[RTW_TXPRIMD].tdb_basereg = RTW_TNPDA;
sc->sc_txdesc_blk[RTW_TXPRIMD].tdb_base = RTW_RING_BASE(sc, hd_txmd);
sc->sc_txdesc_blk[RTW_TXPRIHI].tdb_basereg = RTW_THPDA;
sc->sc_txdesc_blk[RTW_TXPRIHI].tdb_base = RTW_RING_BASE(sc, hd_txhi);
sc->sc_txdesc_blk[RTW_TXPRIBCN].tdb_basereg = RTW_TBDA;
sc->sc_txdesc_blk[RTW_TXPRIBCN].tdb_base = RTW_RING_BASE(sc, hd_bcn);
for (pri = 0; pri < RTW_NTXPRI; pri++) {
tdb = &sc->sc_txdesc_blk[pri];
RTW_WRITE(regs, tdb->tdb_basereg, tdb->tdb_base);
RTW_DPRINTF(RTW_DEBUG_XMIT_DESC,
("%s: reg[tdb->tdb_basereg] <- %" PRIxPTR "\n", __func__,
(uintptr_t)tdb->tdb_base));
}
Miscellaneous changes. Details below. Important changes flagged with []. Using the driver with my Linksys WPC11 ver. 4, it seems to be receiving packets for a change. The WPC11 ver. 4 has a Maxim RF section. My no-name rtw with Philips RF section still does not receive any packets. Keep access-level (analog params > config[0123] registers > none) in sc_access. Add rtw_set_access for changing the access level. Make rtw_continuous_tx_enable and other subroutines use rtw_set_access instead of rtw_config0123_enable and rtw_anaparm_enable. Factor part of the chip-reset code into rtw_chip_reset1. Change the 'struct foo (*bar)[N]'-style arguments to 'struct foo *bar'-style arguments. Consolidate software/hardware Tx/Rx ring setup in rtw_hwring_setup, rtw_swring_setup. Add a new constant, SA2400_OPMODE_DEFAULTS, for the bits that we *always* set in the SA2400 OPMODE register. Factor some code out into rtw_sa2400_calibrate. (Inspired by the Linux driver.) [] When the receiver goes into underrun/overflow state, call a new subroutine, rtw_kick() that stops the Rx/Tx processes, resets the chip, reinitializes the Tx/Rx rings, and restarts Rx/Tx processes. (Inspired by the Linux driver.) [] In rtw_intr_rx, check for too-short packets before calling ieee80211_find_rxnode. I believe this will prevent a repeat of the MCHK exception I saw once on macppc. [] Use seconds-elapased as well as microseconds-elapsed to set the next "due date" for the timeout interrupt. This keeps the driver from programming the timeout to expire too early. [] In rtw_intr, read RTW_ISR at most 10 times, then get out. If the interface is not enabled (RTW_F_ENABLED), then get out. [] In rtw_stop, get out if the interface is not enabled (RTW_F_ENABLED). Block IPL_NET interrupts. Don't read/write any registers if the interface is invalid (RTW_F_INVALID). [] Call rtw_stop in rtw_detach.
2004-12-12 09:37:59 +03:00
RTW_WRITE(regs, RTW_RDSAR, RTW_RING_BASE(sc, hd_rx));
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
RTW_DPRINTF(RTW_DEBUG_RECV_DESC,
("%s: reg[RDSAR] <- %" PRIxPTR "\n", __func__,
(uintptr_t)RTW_RING_BASE(sc, hd_rx)));
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
RTW_SYNC(regs, RTW_TLPDA, RTW_RDSAR);
Miscellaneous changes. Details below. Important changes flagged with []. Using the driver with my Linksys WPC11 ver. 4, it seems to be receiving packets for a change. The WPC11 ver. 4 has a Maxim RF section. My no-name rtw with Philips RF section still does not receive any packets. Keep access-level (analog params > config[0123] registers > none) in sc_access. Add rtw_set_access for changing the access level. Make rtw_continuous_tx_enable and other subroutines use rtw_set_access instead of rtw_config0123_enable and rtw_anaparm_enable. Factor part of the chip-reset code into rtw_chip_reset1. Change the 'struct foo (*bar)[N]'-style arguments to 'struct foo *bar'-style arguments. Consolidate software/hardware Tx/Rx ring setup in rtw_hwring_setup, rtw_swring_setup. Add a new constant, SA2400_OPMODE_DEFAULTS, for the bits that we *always* set in the SA2400 OPMODE register. Factor some code out into rtw_sa2400_calibrate. (Inspired by the Linux driver.) [] When the receiver goes into underrun/overflow state, call a new subroutine, rtw_kick() that stops the Rx/Tx processes, resets the chip, reinitializes the Tx/Rx rings, and restarts Rx/Tx processes. (Inspired by the Linux driver.) [] In rtw_intr_rx, check for too-short packets before calling ieee80211_find_rxnode. I believe this will prevent a repeat of the MCHK exception I saw once on macppc. [] Use seconds-elapased as well as microseconds-elapsed to set the next "due date" for the timeout interrupt. This keeps the driver from programming the timeout to expire too early. [] In rtw_intr, read RTW_ISR at most 10 times, then get out. If the interface is not enabled (RTW_F_ENABLED), then get out. [] In rtw_stop, get out if the interface is not enabled (RTW_F_ENABLED). Block IPL_NET interrupts. Don't read/write any registers if the interface is invalid (RTW_F_INVALID). [] Call rtw_stop in rtw_detach.
2004-12-12 09:37:59 +03:00
}
If defined(RTW_DEBUG), provide a sysctl (hw.rtw.rxbufs_limit) for limiting the number of rx buffers an rtw may allocate. Use this sysctl to test the code that copes with buffer exhaustion. Allocate at most RTW_RXQLEN rx buffers, stopping at the sysctl limit. Record in sc_nrxdesc how many were allocated, and put the end-of-ring indication on sc_rxdesc[sc_nrxdesc - 1]. In rtw_init, if no rx buffers could be allocated, log a complaint, clear IFF_RUNNING, and exit with an error. Many changes to accomodate a short rx ring, mainly of the "add a rx-ring length argument" variety. XXX I really should consolidate all of the rx ring variables in one struct and pass that to the rx-ring subroutines. Bug fix: after calling MCLGET, use the (m->m_flags & M_EXT) idiom to check for success, instead of m != NULL. Bug fix: at the top of rtw_start, if IFF_RUNNING is not set, or IFF_OACTIVE is, get out.
2004-12-29 01:21:15 +03:00
static int
Miscellaneous changes. Details below. Important changes flagged with []. Using the driver with my Linksys WPC11 ver. 4, it seems to be receiving packets for a change. The WPC11 ver. 4 has a Maxim RF section. My no-name rtw with Philips RF section still does not receive any packets. Keep access-level (analog params > config[0123] registers > none) in sc_access. Add rtw_set_access for changing the access level. Make rtw_continuous_tx_enable and other subroutines use rtw_set_access instead of rtw_config0123_enable and rtw_anaparm_enable. Factor part of the chip-reset code into rtw_chip_reset1. Change the 'struct foo (*bar)[N]'-style arguments to 'struct foo *bar'-style arguments. Consolidate software/hardware Tx/Rx ring setup in rtw_hwring_setup, rtw_swring_setup. Add a new constant, SA2400_OPMODE_DEFAULTS, for the bits that we *always* set in the SA2400 OPMODE register. Factor some code out into rtw_sa2400_calibrate. (Inspired by the Linux driver.) [] When the receiver goes into underrun/overflow state, call a new subroutine, rtw_kick() that stops the Rx/Tx processes, resets the chip, reinitializes the Tx/Rx rings, and restarts Rx/Tx processes. (Inspired by the Linux driver.) [] In rtw_intr_rx, check for too-short packets before calling ieee80211_find_rxnode. I believe this will prevent a repeat of the MCHK exception I saw once on macppc. [] Use seconds-elapased as well as microseconds-elapsed to set the next "due date" for the timeout interrupt. This keeps the driver from programming the timeout to expire too early. [] In rtw_intr, read RTW_ISR at most 10 times, then get out. If the interface is not enabled (RTW_F_ENABLED), then get out. [] In rtw_stop, get out if the interface is not enabled (RTW_F_ENABLED). Block IPL_NET interrupts. Don't read/write any registers if the interface is invalid (RTW_F_INVALID). [] Call rtw_stop in rtw_detach.
2004-12-12 09:37:59 +03:00
rtw_swring_setup(struct rtw_softc *sc)
{
If defined(RTW_DEBUG), provide a sysctl (hw.rtw.rxbufs_limit) for limiting the number of rx buffers an rtw may allocate. Use this sysctl to test the code that copes with buffer exhaustion. Allocate at most RTW_RXQLEN rx buffers, stopping at the sysctl limit. Record in sc_nrxdesc how many were allocated, and put the end-of-ring indication on sc_rxdesc[sc_nrxdesc - 1]. In rtw_init, if no rx buffers could be allocated, log a complaint, clear IFF_RUNNING, and exit with an error. Many changes to accomodate a short rx ring, mainly of the "add a rx-ring length argument" variety. XXX I really should consolidate all of the rx ring variables in one struct and pass that to the rx-ring subroutines. Bug fix: after calling MCLGET, use the (m->m_flags & M_EXT) idiom to check for success, instead of m != NULL. Bug fix: at the top of rtw_start, if IFF_RUNNING is not set, or IFF_OACTIVE is, get out.
2004-12-29 01:21:15 +03:00
int rc;
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
struct rtw_rxdesc_blk *rdb;
Consolidate variables related to the rx ring in sc_rxdesc_blk, which is a struct rtw_rxdesc_blk. Put a copy of the DMA tag and the DMA map into the rx- and tx-ring blocks so that I don't have to pass them to subroutines all of the time.
2004-12-29 03:21:37 +03:00
Miscellaneous changes. Details below. Important changes flagged with []. Using the driver with my Linksys WPC11 ver. 4, it seems to be receiving packets for a change. The WPC11 ver. 4 has a Maxim RF section. My no-name rtw with Philips RF section still does not receive any packets. Keep access-level (analog params > config[0123] registers > none) in sc_access. Add rtw_set_access for changing the access level. Make rtw_continuous_tx_enable and other subroutines use rtw_set_access instead of rtw_config0123_enable and rtw_anaparm_enable. Factor part of the chip-reset code into rtw_chip_reset1. Change the 'struct foo (*bar)[N]'-style arguments to 'struct foo *bar'-style arguments. Consolidate software/hardware Tx/Rx ring setup in rtw_hwring_setup, rtw_swring_setup. Add a new constant, SA2400_OPMODE_DEFAULTS, for the bits that we *always* set in the SA2400 OPMODE register. Factor some code out into rtw_sa2400_calibrate. (Inspired by the Linux driver.) [] When the receiver goes into underrun/overflow state, call a new subroutine, rtw_kick() that stops the Rx/Tx processes, resets the chip, reinitializes the Tx/Rx rings, and restarts Rx/Tx processes. (Inspired by the Linux driver.) [] In rtw_intr_rx, check for too-short packets before calling ieee80211_find_rxnode. I believe this will prevent a repeat of the MCHK exception I saw once on macppc. [] Use seconds-elapased as well as microseconds-elapsed to set the next "due date" for the timeout interrupt. This keeps the driver from programming the timeout to expire too early. [] In rtw_intr, read RTW_ISR at most 10 times, then get out. If the interface is not enabled (RTW_F_ENABLED), then get out. [] In rtw_stop, get out if the interface is not enabled (RTW_F_ENABLED). Block IPL_NET interrupts. Don't read/write any registers if the interface is invalid (RTW_F_INVALID). [] Call rtw_stop in rtw_detach.
2004-12-12 09:37:59 +03:00
rtw_txdesc_blk_init_all(&sc->sc_txdesc_blk[0]);
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
rtw_txsoft_blk_init_all(&sc->sc_txsoft_blk[0]);
Miscellaneous changes. Details below. Important changes flagged with []. Using the driver with my Linksys WPC11 ver. 4, it seems to be receiving packets for a change. The WPC11 ver. 4 has a Maxim RF section. My no-name rtw with Philips RF section still does not receive any packets. Keep access-level (analog params > config[0123] registers > none) in sc_access. Add rtw_set_access for changing the access level. Make rtw_continuous_tx_enable and other subroutines use rtw_set_access instead of rtw_config0123_enable and rtw_anaparm_enable. Factor part of the chip-reset code into rtw_chip_reset1. Change the 'struct foo (*bar)[N]'-style arguments to 'struct foo *bar'-style arguments. Consolidate software/hardware Tx/Rx ring setup in rtw_hwring_setup, rtw_swring_setup. Add a new constant, SA2400_OPMODE_DEFAULTS, for the bits that we *always* set in the SA2400 OPMODE register. Factor some code out into rtw_sa2400_calibrate. (Inspired by the Linux driver.) [] When the receiver goes into underrun/overflow state, call a new subroutine, rtw_kick() that stops the Rx/Tx processes, resets the chip, reinitializes the Tx/Rx rings, and restarts Rx/Tx processes. (Inspired by the Linux driver.) [] In rtw_intr_rx, check for too-short packets before calling ieee80211_find_rxnode. I believe this will prevent a repeat of the MCHK exception I saw once on macppc. [] Use seconds-elapased as well as microseconds-elapsed to set the next "due date" for the timeout interrupt. This keeps the driver from programming the timeout to expire too early. [] In rtw_intr, read RTW_ISR at most 10 times, then get out. If the interface is not enabled (RTW_F_ENABLED), then get out. [] In rtw_stop, get out if the interface is not enabled (RTW_F_ENABLED). Block IPL_NET interrupts. Don't read/write any registers if the interface is invalid (RTW_F_INVALID). [] Call rtw_stop in rtw_detach.
2004-12-12 09:37:59 +03:00
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
rdb = &sc->sc_rxdesc_blk;
if ((rc = rtw_rxsoft_init_all(sc->sc_dmat, sc->sc_rxsoft, &rdb->rdb_ndesc,
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
sc->sc_dev)) != 0 && rdb->rdb_ndesc == 0) {
aprint_error_dev(sc->sc_dev, "could not allocate rx buffers\n");
If defined(RTW_DEBUG), provide a sysctl (hw.rtw.rxbufs_limit) for limiting the number of rx buffers an rtw may allocate. Use this sysctl to test the code that copes with buffer exhaustion. Allocate at most RTW_RXQLEN rx buffers, stopping at the sysctl limit. Record in sc_nrxdesc how many were allocated, and put the end-of-ring indication on sc_rxdesc[sc_nrxdesc - 1]. In rtw_init, if no rx buffers could be allocated, log a complaint, clear IFF_RUNNING, and exit with an error. Many changes to accomodate a short rx ring, mainly of the "add a rx-ring length argument" variety. XXX I really should consolidate all of the rx ring variables in one struct and pass that to the rx-ring subroutines. Bug fix: after calling MCLGET, use the (m->m_flags & M_EXT) idiom to check for success, instead of m != NULL. Bug fix: at the top of rtw_start, if IFF_RUNNING is not set, or IFF_OACTIVE is, get out.
2004-12-29 01:21:15 +03:00
return rc;
}
nuke trailing whitespace
2005-02-27 03:26:58 +03:00
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
rdb = &sc->sc_rxdesc_blk;
rtw_rxdescs_sync(rdb, 0, rdb->rdb_ndesc,
If defined(RTW_DEBUG), provide a sysctl (hw.rtw.rxbufs_limit) for limiting the number of rx buffers an rtw may allocate. Use this sysctl to test the code that copes with buffer exhaustion. Allocate at most RTW_RXQLEN rx buffers, stopping at the sysctl limit. Record in sc_nrxdesc how many were allocated, and put the end-of-ring indication on sc_rxdesc[sc_nrxdesc - 1]. In rtw_init, if no rx buffers could be allocated, log a complaint, clear IFF_RUNNING, and exit with an error. Many changes to accomodate a short rx ring, mainly of the "add a rx-ring length argument" variety. XXX I really should consolidate all of the rx ring variables in one struct and pass that to the rx-ring subroutines. Bug fix: after calling MCLGET, use the (m->m_flags & M_EXT) idiom to check for success, instead of m != NULL. Bug fix: at the top of rtw_start, if IFF_RUNNING is not set, or IFF_OACTIVE is, get out.
2004-12-29 01:21:15 +03:00
BUS_DMASYNC_POSTREAD|BUS_DMASYNC_POSTWRITE);
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
rtw_rxdesc_init_all(rdb, sc->sc_rxsoft, 1);
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
rdb->rdb_next = 0;
Miscellaneous changes. Details below. Important changes flagged with []. Using the driver with my Linksys WPC11 ver. 4, it seems to be receiving packets for a change. The WPC11 ver. 4 has a Maxim RF section. My no-name rtw with Philips RF section still does not receive any packets. Keep access-level (analog params > config[0123] registers > none) in sc_access. Add rtw_set_access for changing the access level. Make rtw_continuous_tx_enable and other subroutines use rtw_set_access instead of rtw_config0123_enable and rtw_anaparm_enable. Factor part of the chip-reset code into rtw_chip_reset1. Change the 'struct foo (*bar)[N]'-style arguments to 'struct foo *bar'-style arguments. Consolidate software/hardware Tx/Rx ring setup in rtw_hwring_setup, rtw_swring_setup. Add a new constant, SA2400_OPMODE_DEFAULTS, for the bits that we *always* set in the SA2400 OPMODE register. Factor some code out into rtw_sa2400_calibrate. (Inspired by the Linux driver.) [] When the receiver goes into underrun/overflow state, call a new subroutine, rtw_kick() that stops the Rx/Tx processes, resets the chip, reinitializes the Tx/Rx rings, and restarts Rx/Tx processes. (Inspired by the Linux driver.) [] In rtw_intr_rx, check for too-short packets before calling ieee80211_find_rxnode. I believe this will prevent a repeat of the MCHK exception I saw once on macppc. [] Use seconds-elapased as well as microseconds-elapsed to set the next "due date" for the timeout interrupt. This keeps the driver from programming the timeout to expire too early. [] In rtw_intr, read RTW_ISR at most 10 times, then get out. If the interface is not enabled (RTW_F_ENABLED), then get out. [] In rtw_stop, get out if the interface is not enabled (RTW_F_ENABLED). Block IPL_NET interrupts. Don't read/write any registers if the interface is invalid (RTW_F_INVALID). [] Call rtw_stop in rtw_detach.
2004-12-12 09:37:59 +03:00
Consolidate variables related to the rx ring in sc_rxdesc_blk, which is a struct rtw_rxdesc_blk. Put a copy of the DMA tag and the DMA map into the rx- and tx-ring blocks so that I don't have to pass them to subroutines all of the time.
2004-12-29 03:21:37 +03:00
rtw_txdescs_sync_all(&sc->sc_txdesc_blk[0]);
If defined(RTW_DEBUG), provide a sysctl (hw.rtw.rxbufs_limit) for limiting the number of rx buffers an rtw may allocate. Use this sysctl to test the code that copes with buffer exhaustion. Allocate at most RTW_RXQLEN rx buffers, stopping at the sysctl limit. Record in sc_nrxdesc how many were allocated, and put the end-of-ring indication on sc_rxdesc[sc_nrxdesc - 1]. In rtw_init, if no rx buffers could be allocated, log a complaint, clear IFF_RUNNING, and exit with an error. Many changes to accomodate a short rx ring, mainly of the "add a rx-ring length argument" variety. XXX I really should consolidate all of the rx ring variables in one struct and pass that to the rx-ring subroutines. Bug fix: after calling MCLGET, use the (m->m_flags & M_EXT) idiom to check for success, instead of m != NULL. Bug fix: at the top of rtw_start, if IFF_RUNNING is not set, or IFF_OACTIVE is, get out.
2004-12-29 01:21:15 +03:00
return 0;
Miscellaneous changes. Details below. Important changes flagged with []. Using the driver with my Linksys WPC11 ver. 4, it seems to be receiving packets for a change. The WPC11 ver. 4 has a Maxim RF section. My no-name rtw with Philips RF section still does not receive any packets. Keep access-level (analog params > config[0123] registers > none) in sc_access. Add rtw_set_access for changing the access level. Make rtw_continuous_tx_enable and other subroutines use rtw_set_access instead of rtw_config0123_enable and rtw_anaparm_enable. Factor part of the chip-reset code into rtw_chip_reset1. Change the 'struct foo (*bar)[N]'-style arguments to 'struct foo *bar'-style arguments. Consolidate software/hardware Tx/Rx ring setup in rtw_hwring_setup, rtw_swring_setup. Add a new constant, SA2400_OPMODE_DEFAULTS, for the bits that we *always* set in the SA2400 OPMODE register. Factor some code out into rtw_sa2400_calibrate. (Inspired by the Linux driver.) [] When the receiver goes into underrun/overflow state, call a new subroutine, rtw_kick() that stops the Rx/Tx processes, resets the chip, reinitializes the Tx/Rx rings, and restarts Rx/Tx processes. (Inspired by the Linux driver.) [] In rtw_intr_rx, check for too-short packets before calling ieee80211_find_rxnode. I believe this will prevent a repeat of the MCHK exception I saw once on macppc. [] Use seconds-elapased as well as microseconds-elapsed to set the next "due date" for the timeout interrupt. This keeps the driver from programming the timeout to expire too early. [] In rtw_intr, read RTW_ISR at most 10 times, then get out. If the interface is not enabled (RTW_F_ENABLED), then get out. [] In rtw_stop, get out if the interface is not enabled (RTW_F_ENABLED). Block IPL_NET interrupts. Don't read/write any registers if the interface is invalid (RTW_F_INVALID). [] Call rtw_stop in rtw_detach.
2004-12-12 09:37:59 +03:00
}
static void
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
rtw_txdesc_blk_init(struct rtw_txdesc_blk *tdb)
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
{
int i;
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
(void)memset(tdb->tdb_desc, 0,
sizeof(tdb->tdb_desc[0]) * tdb->tdb_ndesc);
for (i = 0; i < tdb->tdb_ndesc; i++)
tdb->tdb_desc[i].td_next = htole32(RTW_NEXT_DESC(tdb, i));
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
}
static u_int
rtw_txring_next(struct rtw_regs *regs, struct rtw_txdesc_blk *tdb)
{
return (le32toh(RTW_READ(regs, tdb->tdb_basereg)) - tdb->tdb_base) /
sizeof(struct rtw_txdesc);
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
}
I have made a small change that suppresses "recovery" from (spurious?) tx timeouts on the beacon queue. The driver's recovery process badly disrupts the MAC's receiver. This needs a closer look, later. Every other fix is by Charles M. Hannum. Charles writes: "With these, the device is quite stable under load, does not experience xmit timeouts, no longer mysteriously freezes, and does not spew a lot of garbage. It does, however, continue to get some "tx fifo underflow"s. global, SYSCTL_SETUP(): a) Removed unused sysctl variables. rtw_io_enable(): b) Change rtw_io_enable to take a softc, and make it call rtw_txring_fixup(). c) Enable the PCI multiple read/write bit. d) Always reset rdb_next to 0 when starting the receive engine. rtw_intr_rx(): e) When processing receive packets, store the next pointer in rdb_next and read it back on the next round, to fix a race condition vs. rtw_tune() during AP discovery and association. f) Remove the special "skip to 0th descriptor" code, because it is no longer needed. rtw_collect_txring(): g) In the "force" case, stop when we catch up to the hardware xmit pointer. rtw_intr_beacon(): h) Resetting tdb_next here looked fishy to me, in light of changes in rtw_collect_txring(). XXX However, this is the one part I haven't tested. rtw_txring_fixup(): i) Add function and line arguments, so we can print out information about the caller. j) In addition to reading the hardware pointer, skip over any additional packets with OWN set -- we may have queued more. rtw_rxring_fixup(): k) Remove this, because it's completely wrong. rtw_intr_ioerror(): l) Don't handle receive overruns here; handle them in the normal receive path. Also changed RTW_INTR_RX and RTW_INTR_IOERROR accordingly. m) Do not fiddle with CR; the transmit engine is still running just fine. n) Do not fiddle with IMR; it was unnecessary. o) Don't touch the receive engine at all; not necessary now that we're leaving CR alone. p) Remove the rtw_do_chip_reset case; it was dangerous and not actually used. rtw_intr(): q) Don't mask off ISR bits when we pass down the ISR value; it's not necessary, and sometimes the other bits useful in debugging output. rtw_seg_too_short(): r) Don't spew here; the caller will do it if necessary. (Stops spewage in an expected case.) rtw_rxdesc_blk: s) Minor change to make the elements have the same order as txdesc_blk. general: t) Added a bunch of debugging output, #ifdef RTW_DIAG, that was useful while investigating the ring management problems. Also moved rtw_txring_fixup() into RTW_DIAG, as I believe the transmit pointer is always kept in sync now, and the function has potential race conditions.
2007-01-06 11:27:53 +03:00
#ifdef RTW_DIAG
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
static void
I have made a small change that suppresses "recovery" from (spurious?) tx timeouts on the beacon queue. The driver's recovery process badly disrupts the MAC's receiver. This needs a closer look, later. Every other fix is by Charles M. Hannum. Charles writes: "With these, the device is quite stable under load, does not experience xmit timeouts, no longer mysteriously freezes, and does not spew a lot of garbage. It does, however, continue to get some "tx fifo underflow"s. global, SYSCTL_SETUP(): a) Removed unused sysctl variables. rtw_io_enable(): b) Change rtw_io_enable to take a softc, and make it call rtw_txring_fixup(). c) Enable the PCI multiple read/write bit. d) Always reset rdb_next to 0 when starting the receive engine. rtw_intr_rx(): e) When processing receive packets, store the next pointer in rdb_next and read it back on the next round, to fix a race condition vs. rtw_tune() during AP discovery and association. f) Remove the special "skip to 0th descriptor" code, because it is no longer needed. rtw_collect_txring(): g) In the "force" case, stop when we catch up to the hardware xmit pointer. rtw_intr_beacon(): h) Resetting tdb_next here looked fishy to me, in light of changes in rtw_collect_txring(). XXX However, this is the one part I haven't tested. rtw_txring_fixup(): i) Add function and line arguments, so we can print out information about the caller. j) In addition to reading the hardware pointer, skip over any additional packets with OWN set -- we may have queued more. rtw_rxring_fixup(): k) Remove this, because it's completely wrong. rtw_intr_ioerror(): l) Don't handle receive overruns here; handle them in the normal receive path. Also changed RTW_INTR_RX and RTW_INTR_IOERROR accordingly. m) Do not fiddle with CR; the transmit engine is still running just fine. n) Do not fiddle with IMR; it was unnecessary. o) Don't touch the receive engine at all; not necessary now that we're leaving CR alone. p) Remove the rtw_do_chip_reset case; it was dangerous and not actually used. rtw_intr(): q) Don't mask off ISR bits when we pass down the ISR value; it's not necessary, and sometimes the other bits useful in debugging output. rtw_seg_too_short(): r) Don't spew here; the caller will do it if necessary. (Stops spewage in an expected case.) rtw_rxdesc_blk: s) Minor change to make the elements have the same order as txdesc_blk. general: t) Added a bunch of debugging output, #ifdef RTW_DIAG, that was useful while investigating the ring management problems. Also moved rtw_txring_fixup() into RTW_DIAG, as I believe the transmit pointer is always kept in sync now, and the function has potential race conditions.
2007-01-06 11:27:53 +03:00
rtw_txring_fixup(struct rtw_softc *sc, const char *fn, int ln)
Miscellaneous changes. Details below. Important changes flagged with []. Using the driver with my Linksys WPC11 ver. 4, it seems to be receiving packets for a change. The WPC11 ver. 4 has a Maxim RF section. My no-name rtw with Philips RF section still does not receive any packets. Keep access-level (analog params > config[0123] registers > none) in sc_access. Add rtw_set_access for changing the access level. Make rtw_continuous_tx_enable and other subroutines use rtw_set_access instead of rtw_config0123_enable and rtw_anaparm_enable. Factor part of the chip-reset code into rtw_chip_reset1. Change the 'struct foo (*bar)[N]'-style arguments to 'struct foo *bar'-style arguments. Consolidate software/hardware Tx/Rx ring setup in rtw_hwring_setup, rtw_swring_setup. Add a new constant, SA2400_OPMODE_DEFAULTS, for the bits that we *always* set in the SA2400 OPMODE register. Factor some code out into rtw_sa2400_calibrate. (Inspired by the Linux driver.) [] When the receiver goes into underrun/overflow state, call a new subroutine, rtw_kick() that stops the Rx/Tx processes, resets the chip, reinitializes the Tx/Rx rings, and restarts Rx/Tx processes. (Inspired by the Linux driver.) [] In rtw_intr_rx, check for too-short packets before calling ieee80211_find_rxnode. I believe this will prevent a repeat of the MCHK exception I saw once on macppc. [] Use seconds-elapased as well as microseconds-elapsed to set the next "due date" for the timeout interrupt. This keeps the driver from programming the timeout to expire too early. [] In rtw_intr, read RTW_ISR at most 10 times, then get out. If the interface is not enabled (RTW_F_ENABLED), then get out. [] In rtw_stop, get out if the interface is not enabled (RTW_F_ENABLED). Block IPL_NET interrupts. Don't read/write any registers if the interface is invalid (RTW_F_INVALID). [] Call rtw_stop in rtw_detach.
2004-12-12 09:37:59 +03:00
{
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
int pri;
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
u_int next;
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
struct rtw_txdesc_blk *tdb;
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
struct rtw_regs *regs = &sc->sc_regs;
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
for (pri = 0; pri < RTW_NTXPRI; pri++) {
I have made a small change that suppresses "recovery" from (spurious?) tx timeouts on the beacon queue. The driver's recovery process badly disrupts the MAC's receiver. This needs a closer look, later. Every other fix is by Charles M. Hannum. Charles writes: "With these, the device is quite stable under load, does not experience xmit timeouts, no longer mysteriously freezes, and does not spew a lot of garbage. It does, however, continue to get some "tx fifo underflow"s. global, SYSCTL_SETUP(): a) Removed unused sysctl variables. rtw_io_enable(): b) Change rtw_io_enable to take a softc, and make it call rtw_txring_fixup(). c) Enable the PCI multiple read/write bit. d) Always reset rdb_next to 0 when starting the receive engine. rtw_intr_rx(): e) When processing receive packets, store the next pointer in rdb_next and read it back on the next round, to fix a race condition vs. rtw_tune() during AP discovery and association. f) Remove the special "skip to 0th descriptor" code, because it is no longer needed. rtw_collect_txring(): g) In the "force" case, stop when we catch up to the hardware xmit pointer. rtw_intr_beacon(): h) Resetting tdb_next here looked fishy to me, in light of changes in rtw_collect_txring(). XXX However, this is the one part I haven't tested. rtw_txring_fixup(): i) Add function and line arguments, so we can print out information about the caller. j) In addition to reading the hardware pointer, skip over any additional packets with OWN set -- we may have queued more. rtw_rxring_fixup(): k) Remove this, because it's completely wrong. rtw_intr_ioerror(): l) Don't handle receive overruns here; handle them in the normal receive path. Also changed RTW_INTR_RX and RTW_INTR_IOERROR accordingly. m) Do not fiddle with CR; the transmit engine is still running just fine. n) Do not fiddle with IMR; it was unnecessary. o) Don't touch the receive engine at all; not necessary now that we're leaving CR alone. p) Remove the rtw_do_chip_reset case; it was dangerous and not actually used. rtw_intr(): q) Don't mask off ISR bits when we pass down the ISR value; it's not necessary, and sometimes the other bits useful in debugging output. rtw_seg_too_short(): r) Don't spew here; the caller will do it if necessary. (Stops spewage in an expected case.) rtw_rxdesc_blk: s) Minor change to make the elements have the same order as txdesc_blk. general: t) Added a bunch of debugging output, #ifdef RTW_DIAG, that was useful while investigating the ring management problems. Also moved rtw_txring_fixup() into RTW_DIAG, as I believe the transmit pointer is always kept in sync now, and the function has potential race conditions.
2007-01-06 11:27:53 +03:00
int i;
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
tdb = &sc->sc_txdesc_blk[pri];
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
next = rtw_txring_next(regs, tdb);
if (tdb->tdb_next == next)
continue;
I have made a small change that suppresses "recovery" from (spurious?) tx timeouts on the beacon queue. The driver's recovery process badly disrupts the MAC's receiver. This needs a closer look, later. Every other fix is by Charles M. Hannum. Charles writes: "With these, the device is quite stable under load, does not experience xmit timeouts, no longer mysteriously freezes, and does not spew a lot of garbage. It does, however, continue to get some "tx fifo underflow"s. global, SYSCTL_SETUP(): a) Removed unused sysctl variables. rtw_io_enable(): b) Change rtw_io_enable to take a softc, and make it call rtw_txring_fixup(). c) Enable the PCI multiple read/write bit. d) Always reset rdb_next to 0 when starting the receive engine. rtw_intr_rx(): e) When processing receive packets, store the next pointer in rdb_next and read it back on the next round, to fix a race condition vs. rtw_tune() during AP discovery and association. f) Remove the special "skip to 0th descriptor" code, because it is no longer needed. rtw_collect_txring(): g) In the "force" case, stop when we catch up to the hardware xmit pointer. rtw_intr_beacon(): h) Resetting tdb_next here looked fishy to me, in light of changes in rtw_collect_txring(). XXX However, this is the one part I haven't tested. rtw_txring_fixup(): i) Add function and line arguments, so we can print out information about the caller. j) In addition to reading the hardware pointer, skip over any additional packets with OWN set -- we may have queued more. rtw_rxring_fixup(): k) Remove this, because it's completely wrong. rtw_intr_ioerror(): l) Don't handle receive overruns here; handle them in the normal receive path. Also changed RTW_INTR_RX and RTW_INTR_IOERROR accordingly. m) Do not fiddle with CR; the transmit engine is still running just fine. n) Do not fiddle with IMR; it was unnecessary. o) Don't touch the receive engine at all; not necessary now that we're leaving CR alone. p) Remove the rtw_do_chip_reset case; it was dangerous and not actually used. rtw_intr(): q) Don't mask off ISR bits when we pass down the ISR value; it's not necessary, and sometimes the other bits useful in debugging output. rtw_seg_too_short(): r) Don't spew here; the caller will do it if necessary. (Stops spewage in an expected case.) rtw_rxdesc_blk: s) Minor change to make the elements have the same order as txdesc_blk. general: t) Added a bunch of debugging output, #ifdef RTW_DIAG, that was useful while investigating the ring management problems. Also moved rtw_txring_fixup() into RTW_DIAG, as I believe the transmit pointer is always kept in sync now, and the function has potential race conditions.
2007-01-06 11:27:53 +03:00
for (i = 0; next != tdb->tdb_next;
next = RTW_NEXT_IDX(tdb, next), i++) {
if ((tdb->tdb_desc[next].td_stat & htole32(RTW_TXSTAT_OWN)) == 0)
break;
}
printf("%s:%d: tx-ring %d expected next %u, read %u+%d -> %s\n", fn,
ln, pri, tdb->tdb_next, next, i, tdb->tdb_next == next ? "okay" : "BAD");
if (tdb->tdb_next == next)
continue;
In rtw_{rx,tx}ring_fixup, do not set the tx/rx ring index to something out of bounds if we read an invalid pointer form the hardware. Prevents a panic when an rtw@cardbus is ejected.
2006-02-05 02:22:27 +03:00
tdb->tdb_next = MIN(next, tdb->tdb_ndesc - 1);
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
}
}
I have made a small change that suppresses "recovery" from (spurious?) tx timeouts on the beacon queue. The driver's recovery process badly disrupts the MAC's receiver. This needs a closer look, later. Every other fix is by Charles M. Hannum. Charles writes: "With these, the device is quite stable under load, does not experience xmit timeouts, no longer mysteriously freezes, and does not spew a lot of garbage. It does, however, continue to get some "tx fifo underflow"s. global, SYSCTL_SETUP(): a) Removed unused sysctl variables. rtw_io_enable(): b) Change rtw_io_enable to take a softc, and make it call rtw_txring_fixup(). c) Enable the PCI multiple read/write bit. d) Always reset rdb_next to 0 when starting the receive engine. rtw_intr_rx(): e) When processing receive packets, store the next pointer in rdb_next and read it back on the next round, to fix a race condition vs. rtw_tune() during AP discovery and association. f) Remove the special "skip to 0th descriptor" code, because it is no longer needed. rtw_collect_txring(): g) In the "force" case, stop when we catch up to the hardware xmit pointer. rtw_intr_beacon(): h) Resetting tdb_next here looked fishy to me, in light of changes in rtw_collect_txring(). XXX However, this is the one part I haven't tested. rtw_txring_fixup(): i) Add function and line arguments, so we can print out information about the caller. j) In addition to reading the hardware pointer, skip over any additional packets with OWN set -- we may have queued more. rtw_rxring_fixup(): k) Remove this, because it's completely wrong. rtw_intr_ioerror(): l) Don't handle receive overruns here; handle them in the normal receive path. Also changed RTW_INTR_RX and RTW_INTR_IOERROR accordingly. m) Do not fiddle with CR; the transmit engine is still running just fine. n) Do not fiddle with IMR; it was unnecessary. o) Don't touch the receive engine at all; not necessary now that we're leaving CR alone. p) Remove the rtw_do_chip_reset case; it was dangerous and not actually used. rtw_intr(): q) Don't mask off ISR bits when we pass down the ISR value; it's not necessary, and sometimes the other bits useful in debugging output. rtw_seg_too_short(): r) Don't spew here; the caller will do it if necessary. (Stops spewage in an expected case.) rtw_rxdesc_blk: s) Minor change to make the elements have the same order as txdesc_blk. general: t) Added a bunch of debugging output, #ifdef RTW_DIAG, that was useful while investigating the ring management problems. Also moved rtw_txring_fixup() into RTW_DIAG, as I believe the transmit pointer is always kept in sync now, and the function has potential race conditions.
2007-01-06 11:27:53 +03:00
#endif
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
static void
rtw_txdescs_reset(struct rtw_softc *sc)
{
int pri;
I have made a small change that suppresses "recovery" from (spurious?) tx timeouts on the beacon queue. The driver's recovery process badly disrupts the MAC's receiver. This needs a closer look, later. Every other fix is by Charles M. Hannum. Charles writes: "With these, the device is quite stable under load, does not experience xmit timeouts, no longer mysteriously freezes, and does not spew a lot of garbage. It does, however, continue to get some "tx fifo underflow"s. global, SYSCTL_SETUP(): a) Removed unused sysctl variables. rtw_io_enable(): b) Change rtw_io_enable to take a softc, and make it call rtw_txring_fixup(). c) Enable the PCI multiple read/write bit. d) Always reset rdb_next to 0 when starting the receive engine. rtw_intr_rx(): e) When processing receive packets, store the next pointer in rdb_next and read it back on the next round, to fix a race condition vs. rtw_tune() during AP discovery and association. f) Remove the special "skip to 0th descriptor" code, because it is no longer needed. rtw_collect_txring(): g) In the "force" case, stop when we catch up to the hardware xmit pointer. rtw_intr_beacon(): h) Resetting tdb_next here looked fishy to me, in light of changes in rtw_collect_txring(). XXX However, this is the one part I haven't tested. rtw_txring_fixup(): i) Add function and line arguments, so we can print out information about the caller. j) In addition to reading the hardware pointer, skip over any additional packets with OWN set -- we may have queued more. rtw_rxring_fixup(): k) Remove this, because it's completely wrong. rtw_intr_ioerror(): l) Don't handle receive overruns here; handle them in the normal receive path. Also changed RTW_INTR_RX and RTW_INTR_IOERROR accordingly. m) Do not fiddle with CR; the transmit engine is still running just fine. n) Do not fiddle with IMR; it was unnecessary. o) Don't touch the receive engine at all; not necessary now that we're leaving CR alone. p) Remove the rtw_do_chip_reset case; it was dangerous and not actually used. rtw_intr(): q) Don't mask off ISR bits when we pass down the ISR value; it's not necessary, and sometimes the other bits useful in debugging output. rtw_seg_too_short(): r) Don't spew here; the caller will do it if necessary. (Stops spewage in an expected case.) rtw_rxdesc_blk: s) Minor change to make the elements have the same order as txdesc_blk. general: t) Added a bunch of debugging output, #ifdef RTW_DIAG, that was useful while investigating the ring management problems. Also moved rtw_txring_fixup() into RTW_DIAG, as I believe the transmit pointer is always kept in sync now, and the function has potential race conditions.
2007-01-06 11:27:53 +03:00
struct rtw_txsoft_blk *tsb;
struct rtw_txdesc_blk *tdb;
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
for (pri = 0; pri < RTW_NTXPRI; pri++) {
I have made a small change that suppresses "recovery" from (spurious?) tx timeouts on the beacon queue. The driver's recovery process badly disrupts the MAC's receiver. This needs a closer look, later. Every other fix is by Charles M. Hannum. Charles writes: "With these, the device is quite stable under load, does not experience xmit timeouts, no longer mysteriously freezes, and does not spew a lot of garbage. It does, however, continue to get some "tx fifo underflow"s. global, SYSCTL_SETUP(): a) Removed unused sysctl variables. rtw_io_enable(): b) Change rtw_io_enable to take a softc, and make it call rtw_txring_fixup(). c) Enable the PCI multiple read/write bit. d) Always reset rdb_next to 0 when starting the receive engine. rtw_intr_rx(): e) When processing receive packets, store the next pointer in rdb_next and read it back on the next round, to fix a race condition vs. rtw_tune() during AP discovery and association. f) Remove the special "skip to 0th descriptor" code, because it is no longer needed. rtw_collect_txring(): g) In the "force" case, stop when we catch up to the hardware xmit pointer. rtw_intr_beacon(): h) Resetting tdb_next here looked fishy to me, in light of changes in rtw_collect_txring(). XXX However, this is the one part I haven't tested. rtw_txring_fixup(): i) Add function and line arguments, so we can print out information about the caller. j) In addition to reading the hardware pointer, skip over any additional packets with OWN set -- we may have queued more. rtw_rxring_fixup(): k) Remove this, because it's completely wrong. rtw_intr_ioerror(): l) Don't handle receive overruns here; handle them in the normal receive path. Also changed RTW_INTR_RX and RTW_INTR_IOERROR accordingly. m) Do not fiddle with CR; the transmit engine is still running just fine. n) Do not fiddle with IMR; it was unnecessary. o) Don't touch the receive engine at all; not necessary now that we're leaving CR alone. p) Remove the rtw_do_chip_reset case; it was dangerous and not actually used. rtw_intr(): q) Don't mask off ISR bits when we pass down the ISR value; it's not necessary, and sometimes the other bits useful in debugging output. rtw_seg_too_short(): r) Don't spew here; the caller will do it if necessary. (Stops spewage in an expected case.) rtw_rxdesc_blk: s) Minor change to make the elements have the same order as txdesc_blk. general: t) Added a bunch of debugging output, #ifdef RTW_DIAG, that was useful while investigating the ring management problems. Also moved rtw_txring_fixup() into RTW_DIAG, as I believe the transmit pointer is always kept in sync now, and the function has potential race conditions.
2007-01-06 11:27:53 +03:00
tsb = &sc->sc_txsoft_blk[pri];
tdb = &sc->sc_txdesc_blk[pri];
rtw_collect_txring(sc, tsb, tdb, 1);
#ifdef RTW_DIAG
if (!SIMPLEQ_EMPTY(&tsb->tsb_dirtyq))
printf("%s: packets left in ring %d\n", __func__, pri);
#endif
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
}
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
}
static void
rtw_intr_ioerror(struct rtw_softc *sc, uint16_t isr)
{
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
aprint_error_dev(sc->sc_dev, "tx fifo underflow\n");
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
I have made a small change that suppresses "recovery" from (spurious?) tx timeouts on the beacon queue. The driver's recovery process badly disrupts the MAC's receiver. This needs a closer look, later. Every other fix is by Charles M. Hannum. Charles writes: "With these, the device is quite stable under load, does not experience xmit timeouts, no longer mysteriously freezes, and does not spew a lot of garbage. It does, however, continue to get some "tx fifo underflow"s. global, SYSCTL_SETUP(): a) Removed unused sysctl variables. rtw_io_enable(): b) Change rtw_io_enable to take a softc, and make it call rtw_txring_fixup(). c) Enable the PCI multiple read/write bit. d) Always reset rdb_next to 0 when starting the receive engine. rtw_intr_rx(): e) When processing receive packets, store the next pointer in rdb_next and read it back on the next round, to fix a race condition vs. rtw_tune() during AP discovery and association. f) Remove the special "skip to 0th descriptor" code, because it is no longer needed. rtw_collect_txring(): g) In the "force" case, stop when we catch up to the hardware xmit pointer. rtw_intr_beacon(): h) Resetting tdb_next here looked fishy to me, in light of changes in rtw_collect_txring(). XXX However, this is the one part I haven't tested. rtw_txring_fixup(): i) Add function and line arguments, so we can print out information about the caller. j) In addition to reading the hardware pointer, skip over any additional packets with OWN set -- we may have queued more. rtw_rxring_fixup(): k) Remove this, because it's completely wrong. rtw_intr_ioerror(): l) Don't handle receive overruns here; handle them in the normal receive path. Also changed RTW_INTR_RX and RTW_INTR_IOERROR accordingly. m) Do not fiddle with CR; the transmit engine is still running just fine. n) Do not fiddle with IMR; it was unnecessary. o) Don't touch the receive engine at all; not necessary now that we're leaving CR alone. p) Remove the rtw_do_chip_reset case; it was dangerous and not actually used. rtw_intr(): q) Don't mask off ISR bits when we pass down the ISR value; it's not necessary, and sometimes the other bits useful in debugging output. rtw_seg_too_short(): r) Don't spew here; the caller will do it if necessary. (Stops spewage in an expected case.) rtw_rxdesc_blk: s) Minor change to make the elements have the same order as txdesc_blk. general: t) Added a bunch of debugging output, #ifdef RTW_DIAG, that was useful while investigating the ring management problems. Also moved rtw_txring_fixup() into RTW_DIAG, as I believe the transmit pointer is always kept in sync now, and the function has potential race conditions.
2007-01-06 11:27:53 +03:00
RTW_DPRINTF(RTW_DEBUG_BUGS, ("%s: cleaning up xmit, isr %" PRIx16
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
"\n", device_xname(sc->sc_dev), isr));
Disable interrupts and stop transmit DMA before disabling the transmitter and receiver.
2004-12-23 08:54:54 +03:00
Make rtw(4) build w/o RTW_DEBUG. Thanks to Charles Hannum for pointing out my oversight.
2004-12-27 01:59:41 +03:00
#ifdef RTW_DEBUG
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
rtw_dump_rings(sc);
Make rtw(4) build w/o RTW_DEBUG. Thanks to Charles Hannum for pointing out my oversight.
2004-12-27 01:59:41 +03:00
#endif /* RTW_DEBUG */
Disable interrupts and stop transmit DMA before disabling the transmitter and receiver.
2004-12-23 08:54:54 +03:00
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
/* Collect tx'd packets. XXX let's hope this stops the transmit
* timeouts.
*/
I have made a small change that suppresses "recovery" from (spurious?) tx timeouts on the beacon queue. The driver's recovery process badly disrupts the MAC's receiver. This needs a closer look, later. Every other fix is by Charles M. Hannum. Charles writes: "With these, the device is quite stable under load, does not experience xmit timeouts, no longer mysteriously freezes, and does not spew a lot of garbage. It does, however, continue to get some "tx fifo underflow"s. global, SYSCTL_SETUP(): a) Removed unused sysctl variables. rtw_io_enable(): b) Change rtw_io_enable to take a softc, and make it call rtw_txring_fixup(). c) Enable the PCI multiple read/write bit. d) Always reset rdb_next to 0 when starting the receive engine. rtw_intr_rx(): e) When processing receive packets, store the next pointer in rdb_next and read it back on the next round, to fix a race condition vs. rtw_tune() during AP discovery and association. f) Remove the special "skip to 0th descriptor" code, because it is no longer needed. rtw_collect_txring(): g) In the "force" case, stop when we catch up to the hardware xmit pointer. rtw_intr_beacon(): h) Resetting tdb_next here looked fishy to me, in light of changes in rtw_collect_txring(). XXX However, this is the one part I haven't tested. rtw_txring_fixup(): i) Add function and line arguments, so we can print out information about the caller. j) In addition to reading the hardware pointer, skip over any additional packets with OWN set -- we may have queued more. rtw_rxring_fixup(): k) Remove this, because it's completely wrong. rtw_intr_ioerror(): l) Don't handle receive overruns here; handle them in the normal receive path. Also changed RTW_INTR_RX and RTW_INTR_IOERROR accordingly. m) Do not fiddle with CR; the transmit engine is still running just fine. n) Do not fiddle with IMR; it was unnecessary. o) Don't touch the receive engine at all; not necessary now that we're leaving CR alone. p) Remove the rtw_do_chip_reset case; it was dangerous and not actually used. rtw_intr(): q) Don't mask off ISR bits when we pass down the ISR value; it's not necessary, and sometimes the other bits useful in debugging output. rtw_seg_too_short(): r) Don't spew here; the caller will do it if necessary. (Stops spewage in an expected case.) rtw_rxdesc_blk: s) Minor change to make the elements have the same order as txdesc_blk. general: t) Added a bunch of debugging output, #ifdef RTW_DIAG, that was useful while investigating the ring management problems. Also moved rtw_txring_fixup() into RTW_DIAG, as I believe the transmit pointer is always kept in sync now, and the function has potential race conditions.
2007-01-06 11:27:53 +03:00
rtw_txdescs_reset(sc);
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
Make rtw(4) build w/o RTW_DEBUG. Thanks to Charles Hannum for pointing out my oversight.
2004-12-27 01:59:41 +03:00
#ifdef RTW_DEBUG
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
rtw_dump_rings(sc);
Make rtw(4) build w/o RTW_DEBUG. Thanks to Charles Hannum for pointing out my oversight.
2004-12-27 01:59:41 +03:00
#endif /* RTW_DEBUG */
Miscellaneous changes. Details below. Important changes flagged with []. Using the driver with my Linksys WPC11 ver. 4, it seems to be receiving packets for a change. The WPC11 ver. 4 has a Maxim RF section. My no-name rtw with Philips RF section still does not receive any packets. Keep access-level (analog params > config[0123] registers > none) in sc_access. Add rtw_set_access for changing the access level. Make rtw_continuous_tx_enable and other subroutines use rtw_set_access instead of rtw_config0123_enable and rtw_anaparm_enable. Factor part of the chip-reset code into rtw_chip_reset1. Change the 'struct foo (*bar)[N]'-style arguments to 'struct foo *bar'-style arguments. Consolidate software/hardware Tx/Rx ring setup in rtw_hwring_setup, rtw_swring_setup. Add a new constant, SA2400_OPMODE_DEFAULTS, for the bits that we *always* set in the SA2400 OPMODE register. Factor some code out into rtw_sa2400_calibrate. (Inspired by the Linux driver.) [] When the receiver goes into underrun/overflow state, call a new subroutine, rtw_kick() that stops the Rx/Tx processes, resets the chip, reinitializes the Tx/Rx rings, and restarts Rx/Tx processes. (Inspired by the Linux driver.) [] In rtw_intr_rx, check for too-short packets before calling ieee80211_find_rxnode. I believe this will prevent a repeat of the MCHK exception I saw once on macppc. [] Use seconds-elapased as well as microseconds-elapsed to set the next "due date" for the timeout interrupt. This keeps the driver from programming the timeout to expire too early. [] In rtw_intr, read RTW_ISR at most 10 times, then get out. If the interface is not enabled (RTW_F_ENABLED), then get out. [] In rtw_stop, get out if the interface is not enabled (RTW_F_ENABLED). Block IPL_NET interrupts. Don't read/write any registers if the interface is invalid (RTW_F_INVALID). [] Call rtw_stop in rtw_detach.
2004-12-12 09:37:59 +03:00
}
Remove leading __ from __(const|inline|signed|volatile) -- it is obsolete.
2005-12-24 23:27:29 +03:00
static inline void
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
rtw_suspend_ticks(struct rtw_softc *sc)
{
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
RTW_DPRINTF(RTW_DEBUG_TIMEOUT,
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
("%s: suspending ticks\n", device_xname(sc->sc_dev)));
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
sc->sc_do_tick = 0;
}
Remove leading __ from __(const|inline|signed|volatile) -- it is obsolete.
2005-12-24 23:27:29 +03:00
static inline void
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
rtw_resume_ticks(struct rtw_softc *sc)
{
Make this compile on sparc64: 1 s/next_tick/next_tint/ because next_tick shadows a global on sparc64. 2 Fix the bus-barrier helper routines, which had the tag & handle swapped (!!). Move the helpers from rtwreg.h to rtwvar.h, and change RTW_BARRIER() into an inline subroutine called rtw_barrier().
2010-03-16 02:21:08 +03:00
uint32_t tsftrl0, tsftrl1, next_tint;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
tsftrl0 = RTW_READ(&sc->sc_regs, RTW_TSFTRL);
tsftrl1 = RTW_READ(&sc->sc_regs, RTW_TSFTRL);
Make this compile on sparc64: 1 s/next_tick/next_tint/ because next_tick shadows a global on sparc64. 2 Fix the bus-barrier helper routines, which had the tag & handle swapped (!!). Move the helpers from rtwreg.h to rtwvar.h, and change RTW_BARRIER() into an inline subroutine called rtw_barrier().
2010-03-16 02:21:08 +03:00
next_tint = tsftrl1 + 1000000;
RTW_WRITE(&sc->sc_regs, RTW_TINT, next_tint);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
sc->sc_do_tick = 1;
- remove unused variables - move variables inside ifdef sections - ifdef notdef unused code - use __USE for debugging variables
2013-10-18 01:24:24 +04:00
#ifdef RTW_DEBUG
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
RTW_DPRINTF(RTW_DEBUG_TIMEOUT,
("%s: resume ticks delta %#08x now %#08x next %#08x\n",
Make this compile on sparc64: 1 s/next_tick/next_tint/ because next_tick shadows a global on sparc64. 2 Fix the bus-barrier helper routines, which had the tag & handle swapped (!!). Move the helpers from rtwreg.h to rtwvar.h, and change RTW_BARRIER() into an inline subroutine called rtw_barrier().
2010-03-16 02:21:08 +03:00
device_xname(sc->sc_dev), tsftrl1 - tsftrl0, tsftrl1, next_tint));
- remove unused variables - move variables inside ifdef sections - ifdef notdef unused code - use __USE for debugging variables
2013-10-18 01:24:24 +04:00
#else
__USE(tsftrl0);
#endif
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
}
static void
rtw_intr_timeout(struct rtw_softc *sc)
{
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
RTW_DPRINTF(RTW_DEBUG_TIMEOUT, ("%s: timeout\n", device_xname(sc->sc_dev)));
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
if (sc->sc_do_tick)
rtw_resume_ticks(sc);
return;
}
int
rtw_intr(void *arg)
{
Miscellaneous changes. Details below. Important changes flagged with []. Using the driver with my Linksys WPC11 ver. 4, it seems to be receiving packets for a change. The WPC11 ver. 4 has a Maxim RF section. My no-name rtw with Philips RF section still does not receive any packets. Keep access-level (analog params > config[0123] registers > none) in sc_access. Add rtw_set_access for changing the access level. Make rtw_continuous_tx_enable and other subroutines use rtw_set_access instead of rtw_config0123_enable and rtw_anaparm_enable. Factor part of the chip-reset code into rtw_chip_reset1. Change the 'struct foo (*bar)[N]'-style arguments to 'struct foo *bar'-style arguments. Consolidate software/hardware Tx/Rx ring setup in rtw_hwring_setup, rtw_swring_setup. Add a new constant, SA2400_OPMODE_DEFAULTS, for the bits that we *always* set in the SA2400 OPMODE register. Factor some code out into rtw_sa2400_calibrate. (Inspired by the Linux driver.) [] When the receiver goes into underrun/overflow state, call a new subroutine, rtw_kick() that stops the Rx/Tx processes, resets the chip, reinitializes the Tx/Rx rings, and restarts Rx/Tx processes. (Inspired by the Linux driver.) [] In rtw_intr_rx, check for too-short packets before calling ieee80211_find_rxnode. I believe this will prevent a repeat of the MCHK exception I saw once on macppc. [] Use seconds-elapased as well as microseconds-elapsed to set the next "due date" for the timeout interrupt. This keeps the driver from programming the timeout to expire too early. [] In rtw_intr, read RTW_ISR at most 10 times, then get out. If the interface is not enabled (RTW_F_ENABLED), then get out. [] In rtw_stop, get out if the interface is not enabled (RTW_F_ENABLED). Block IPL_NET interrupts. Don't read/write any registers if the interface is invalid (RTW_F_INVALID). [] Call rtw_stop in rtw_detach.
2004-12-12 09:37:59 +03:00
int i;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
struct rtw_softc *sc = arg;
struct rtw_regs *regs = &sc->sc_regs;
Cosmetic: change u_int{8,16,32,64}_t to C99-style uint{8,16,32,64}_t, which is preferred.
2005-01-02 07:23:03 +03:00
uint16_t isr;
Resolve conflicts in importation of 18-May-2005 ath(4) / net80211(9) from FreeBSD. Introduce compatibility shims (sys/dev/ic/ath_netbsd.[ch], sys/net80211/ieee80211_netbsd.[ch]). Update drivers (an, atu, atw, awi, ipw, iwi, rtw, wi) for the new net80211(9) API.
2005-06-22 10:14:51 +04:00
struct ifnet *ifp = &sc->sc_if;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
/*
* If the interface isn't running, the interrupt couldn't
* possibly have come from us.
*/
Use device_t and its accessors throughout. Use aprint_*_dev(). Improve PMF-ability. Add a 'flags' argument to suspend/resume handlers and callers such as pmf_system_suspend(). Define a flag, PMF_F_SELF, which indicates to PMF that a device is suspending/resuming itself. Add helper routines, pmf_device_suspend_self(dev) and pmf_device_resume_self(dev), that call pmf_device_suspend(dev, PMF_F_SELF) and pmf_device_resume(dev, PMF_F_SELF), respectively. Use PMF_F_SELF to suspend/resume self in ath(4), audio(4), rtw(4), and sip(4). In ath(4) and in rtw(4), replace the icky sc_enable/sc_disable callbacks, provided by the bus front-end, with self-suspension/resumption. Also, clean up the bus front-ends. Make sure that the interrupt handler is disestablished during suspension. Get rid of driver-private flags (e.g., RTW_F_ENABLED, ath_softc->sc_invalid); use device_is_active()/device_has_power() calls, instead. In the network-class suspend handler, call if_stop(, 0) instead of if_stop(, 1), because the latter is superfluous (bus- and driver-suspension hooks will 'disable' the NIC), and it may cause recursion. In the network-class resume handler, prevent infinite recursion through if_init() by getting out early if we are self-suspending (PMF_F_SELF). rtw(4) improvements: Destroy rtw(4) callouts when we detach it. Make rtw at pci detachable. Print some more information with the "rx frame too long" warning. Remove activate() methods: Get rid of rtw_activate() and ath_activate(). The device activate() methods are not good for much these days. Make ath at cardbus resume with crypto functions intact: Introduce a boolean device property, "pmf-powerdown". If pmf-powerdown is present and false, it indicates that a bus back-end should not remove power from a device. Honor this property in cardbus_child_suspend(). Set this property to 'false' in ath_attach(), since removing power from an ath at cardbus seems to lobotomize the WPA crypto engine. XXX Should the pmf-powerdown property propagate toward the root of the device tree? Miscellaneous ath(4) changes: Warn if ath(4) tries to write crypto keys to suspended hardware. Reduce differences between FreeBSD and NetBSD in ath(4) multicast filter setup. Make ath_printrxbuf() print an rx descriptor's status & key index, to help debug crypto errors. Shorten a staircase in ath_ioctl(). Don't check for ieee80211_ioctl() return code ERESTART, it never happens.
2008-03-12 21:02:21 +03:00
if ((ifp->if_flags & IFF_RUNNING) == 0 ||
In pmf(9), improve the implementation of device self-suspension and make suspension by self, by drvctl(8), and by ACPI system sleep play nice together. Start solidifying some temporary API changes. 1. Extract a new header file, <sys/device_if.h>, from <sys/device.h> and #include it from <sys/pmf.h> instead of <sys/device.h> to break the circular dependency between <sys/device.h> and <sys/pmf.h>. 2. Introduce pmf_qual_t, an aggregate of qualifications on a PMF suspend/resume call. Start to replace instances of PMF_FN_PROTO, PMF_FN_ARGS, et cetera, with a pmf_qual_t. 3. Introduce the notion of a "suspensor," an entity that holds a device in suspension. More than one suspensor may hold a device at once. A device stays suspended as long as at least one suspensor holds it. A device resumes when the last suspensor releases it. Currently, the kernel defines three suspensors, 3a the system-suspensor: for system suspension, initiated by 'sysctl -w machdep.sleep_state=3', by lid closure, by power-button press, et cetera, 3b the drvctl-suspensor: for device suspension by /dev/drvctl ioctl, e.g., drvctl -S sip0. 3c the system self-suspensor: for device drivers that suspend themselves and their children. Several drivers for network interfaces put the network device to sleep while it is not administratively up, that is, after the kernel calls if_stop(, 1). The self-suspensor should not be used directly. See the description of suspensor delegates, below. A suspensor can have one or more "delegates". A suspensor can release devices that its delegates hold suspended. Right now, only the system self-suspensor has delegates. For each device that a self-suspending driver attaches, it creates the device's self-suspensor, a delegate of the system self-suspensor. Suspensors stop a system-wide suspend/resume cycle from waking devices that the operator put to sleep with drvctl before the cycle. They also help self-suspension to work more simply, safely, and in accord with expectations. 4. Add the notion of device activation level, devact_level_t, and a routine for checking the current activation level, device_activation(). Current activation levels are DEVACT_LEVEL_BUS, DEVACT_LEVEL_DRIVER, and DEVACT_LEVEL_CLASS, which respectively indicate that the device's bus is active, that the bus and device are active, and that the bus, device, and the functions of the device's class (network, audio) are active. Suspend/resume calls can be qualified with a devact_level_t. The power-management framework treats a devact_level_t that qualifies a device suspension as the device's current activation level; it only runs hooks to reduce the activation level from the presumed current level to the fully suspended state. The framework treats a devact_level_t qualifying device resumption as the target activation level; it only runs hooks to raise the activation level to the target. 5. Use pmf_qual_t, devact_level_t, and self-suspensors in several drivers. 6. Temporarily add an unused power-management workqueue that I will remove or replace, soon.
2009-09-16 20:34:49 +04:00
!device_activation(sc->sc_dev, DEVACT_LEVEL_DRIVER)) {
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
RTW_DPRINTF(RTW_DEBUG_INTR, ("%s: stray interrupt\n",
device_xname(sc->sc_dev)));
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
return (0);
}
Miscellaneous changes. Details below. Important changes flagged with []. Using the driver with my Linksys WPC11 ver. 4, it seems to be receiving packets for a change. The WPC11 ver. 4 has a Maxim RF section. My no-name rtw with Philips RF section still does not receive any packets. Keep access-level (analog params > config[0123] registers > none) in sc_access. Add rtw_set_access for changing the access level. Make rtw_continuous_tx_enable and other subroutines use rtw_set_access instead of rtw_config0123_enable and rtw_anaparm_enable. Factor part of the chip-reset code into rtw_chip_reset1. Change the 'struct foo (*bar)[N]'-style arguments to 'struct foo *bar'-style arguments. Consolidate software/hardware Tx/Rx ring setup in rtw_hwring_setup, rtw_swring_setup. Add a new constant, SA2400_OPMODE_DEFAULTS, for the bits that we *always* set in the SA2400 OPMODE register. Factor some code out into rtw_sa2400_calibrate. (Inspired by the Linux driver.) [] When the receiver goes into underrun/overflow state, call a new subroutine, rtw_kick() that stops the Rx/Tx processes, resets the chip, reinitializes the Tx/Rx rings, and restarts Rx/Tx processes. (Inspired by the Linux driver.) [] In rtw_intr_rx, check for too-short packets before calling ieee80211_find_rxnode. I believe this will prevent a repeat of the MCHK exception I saw once on macppc. [] Use seconds-elapased as well as microseconds-elapsed to set the next "due date" for the timeout interrupt. This keeps the driver from programming the timeout to expire too early. [] In rtw_intr, read RTW_ISR at most 10 times, then get out. If the interface is not enabled (RTW_F_ENABLED), then get out. [] In rtw_stop, get out if the interface is not enabled (RTW_F_ENABLED). Block IPL_NET interrupts. Don't read/write any registers if the interface is invalid (RTW_F_INVALID). [] Call rtw_stop in rtw_detach.
2004-12-12 09:37:59 +03:00
for (i = 0; i < 10; i++) {
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
isr = RTW_READ16(regs, RTW_ISR);
RTW_WRITE16(regs, RTW_ISR, isr);
Add some bus barriers. Remark where they are needed, e.g., read_eeprom. Change some printfs to RTW_DPRINTF2s to quiet rtw(4) when debugging is disabled.
2004-12-20 04:13:45 +03:00
RTW_WBR(regs, RTW_ISR, RTW_ISR);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
if (sc->sc_intr_ack != NULL)
(*sc->sc_intr_ack)(regs);
if (isr == 0)
break;
#ifdef RTW_DEBUG
#define PRINTINTR(flag) do { \
if ((isr & flag) != 0) { \
printf("%s" #flag, delim); \
delim = ","; \
} \
} while (0)
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
if ((rtw_debug & RTW_DEBUG_INTR) != 0 && isr != 0) {
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
const char *delim = "<";
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
printf("%s: reg[ISR] = %x", device_xname(sc->sc_dev),
isr);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
PRINTINTR(RTW_INTR_TXFOVW);
PRINTINTR(RTW_INTR_TIMEOUT);
PRINTINTR(RTW_INTR_BCNINT);
PRINTINTR(RTW_INTR_ATIMINT);
PRINTINTR(RTW_INTR_TBDER);
PRINTINTR(RTW_INTR_TBDOK);
PRINTINTR(RTW_INTR_THPDER);
PRINTINTR(RTW_INTR_THPDOK);
PRINTINTR(RTW_INTR_TNPDER);
PRINTINTR(RTW_INTR_TNPDOK);
PRINTINTR(RTW_INTR_RXFOVW);
PRINTINTR(RTW_INTR_RDU);
PRINTINTR(RTW_INTR_TLPDER);
PRINTINTR(RTW_INTR_TLPDOK);
PRINTINTR(RTW_INTR_RER);
PRINTINTR(RTW_INTR_ROK);
printf(">\n");
}
#undef PRINTINTR
#endif /* RTW_DEBUG */
if ((isr & RTW_INTR_RX) != 0)
I have made a small change that suppresses "recovery" from (spurious?) tx timeouts on the beacon queue. The driver's recovery process badly disrupts the MAC's receiver. This needs a closer look, later. Every other fix is by Charles M. Hannum. Charles writes: "With these, the device is quite stable under load, does not experience xmit timeouts, no longer mysteriously freezes, and does not spew a lot of garbage. It does, however, continue to get some "tx fifo underflow"s. global, SYSCTL_SETUP(): a) Removed unused sysctl variables. rtw_io_enable(): b) Change rtw_io_enable to take a softc, and make it call rtw_txring_fixup(). c) Enable the PCI multiple read/write bit. d) Always reset rdb_next to 0 when starting the receive engine. rtw_intr_rx(): e) When processing receive packets, store the next pointer in rdb_next and read it back on the next round, to fix a race condition vs. rtw_tune() during AP discovery and association. f) Remove the special "skip to 0th descriptor" code, because it is no longer needed. rtw_collect_txring(): g) In the "force" case, stop when we catch up to the hardware xmit pointer. rtw_intr_beacon(): h) Resetting tdb_next here looked fishy to me, in light of changes in rtw_collect_txring(). XXX However, this is the one part I haven't tested. rtw_txring_fixup(): i) Add function and line arguments, so we can print out information about the caller. j) In addition to reading the hardware pointer, skip over any additional packets with OWN set -- we may have queued more. rtw_rxring_fixup(): k) Remove this, because it's completely wrong. rtw_intr_ioerror(): l) Don't handle receive overruns here; handle them in the normal receive path. Also changed RTW_INTR_RX and RTW_INTR_IOERROR accordingly. m) Do not fiddle with CR; the transmit engine is still running just fine. n) Do not fiddle with IMR; it was unnecessary. o) Don't touch the receive engine at all; not necessary now that we're leaving CR alone. p) Remove the rtw_do_chip_reset case; it was dangerous and not actually used. rtw_intr(): q) Don't mask off ISR bits when we pass down the ISR value; it's not necessary, and sometimes the other bits useful in debugging output. rtw_seg_too_short(): r) Don't spew here; the caller will do it if necessary. (Stops spewage in an expected case.) rtw_rxdesc_blk: s) Minor change to make the elements have the same order as txdesc_blk. general: t) Added a bunch of debugging output, #ifdef RTW_DIAG, that was useful while investigating the ring management problems. Also moved rtw_txring_fixup() into RTW_DIAG, as I believe the transmit pointer is always kept in sync now, and the function has potential race conditions.
2007-01-06 11:27:53 +03:00
rtw_intr_rx(sc, isr);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
if ((isr & RTW_INTR_TX) != 0)
I have made a small change that suppresses "recovery" from (spurious?) tx timeouts on the beacon queue. The driver's recovery process badly disrupts the MAC's receiver. This needs a closer look, later. Every other fix is by Charles M. Hannum. Charles writes: "With these, the device is quite stable under load, does not experience xmit timeouts, no longer mysteriously freezes, and does not spew a lot of garbage. It does, however, continue to get some "tx fifo underflow"s. global, SYSCTL_SETUP(): a) Removed unused sysctl variables. rtw_io_enable(): b) Change rtw_io_enable to take a softc, and make it call rtw_txring_fixup(). c) Enable the PCI multiple read/write bit. d) Always reset rdb_next to 0 when starting the receive engine. rtw_intr_rx(): e) When processing receive packets, store the next pointer in rdb_next and read it back on the next round, to fix a race condition vs. rtw_tune() during AP discovery and association. f) Remove the special "skip to 0th descriptor" code, because it is no longer needed. rtw_collect_txring(): g) In the "force" case, stop when we catch up to the hardware xmit pointer. rtw_intr_beacon(): h) Resetting tdb_next here looked fishy to me, in light of changes in rtw_collect_txring(). XXX However, this is the one part I haven't tested. rtw_txring_fixup(): i) Add function and line arguments, so we can print out information about the caller. j) In addition to reading the hardware pointer, skip over any additional packets with OWN set -- we may have queued more. rtw_rxring_fixup(): k) Remove this, because it's completely wrong. rtw_intr_ioerror(): l) Don't handle receive overruns here; handle them in the normal receive path. Also changed RTW_INTR_RX and RTW_INTR_IOERROR accordingly. m) Do not fiddle with CR; the transmit engine is still running just fine. n) Do not fiddle with IMR; it was unnecessary. o) Don't touch the receive engine at all; not necessary now that we're leaving CR alone. p) Remove the rtw_do_chip_reset case; it was dangerous and not actually used. rtw_intr(): q) Don't mask off ISR bits when we pass down the ISR value; it's not necessary, and sometimes the other bits useful in debugging output. rtw_seg_too_short(): r) Don't spew here; the caller will do it if necessary. (Stops spewage in an expected case.) rtw_rxdesc_blk: s) Minor change to make the elements have the same order as txdesc_blk. general: t) Added a bunch of debugging output, #ifdef RTW_DIAG, that was useful while investigating the ring management problems. Also moved rtw_txring_fixup() into RTW_DIAG, as I believe the transmit pointer is always kept in sync now, and the function has potential race conditions.
2007-01-06 11:27:53 +03:00
rtw_intr_tx(sc, isr);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
if ((isr & RTW_INTR_BEACON) != 0)
I have made a small change that suppresses "recovery" from (spurious?) tx timeouts on the beacon queue. The driver's recovery process badly disrupts the MAC's receiver. This needs a closer look, later. Every other fix is by Charles M. Hannum. Charles writes: "With these, the device is quite stable under load, does not experience xmit timeouts, no longer mysteriously freezes, and does not spew a lot of garbage. It does, however, continue to get some "tx fifo underflow"s. global, SYSCTL_SETUP(): a) Removed unused sysctl variables. rtw_io_enable(): b) Change rtw_io_enable to take a softc, and make it call rtw_txring_fixup(). c) Enable the PCI multiple read/write bit. d) Always reset rdb_next to 0 when starting the receive engine. rtw_intr_rx(): e) When processing receive packets, store the next pointer in rdb_next and read it back on the next round, to fix a race condition vs. rtw_tune() during AP discovery and association. f) Remove the special "skip to 0th descriptor" code, because it is no longer needed. rtw_collect_txring(): g) In the "force" case, stop when we catch up to the hardware xmit pointer. rtw_intr_beacon(): h) Resetting tdb_next here looked fishy to me, in light of changes in rtw_collect_txring(). XXX However, this is the one part I haven't tested. rtw_txring_fixup(): i) Add function and line arguments, so we can print out information about the caller. j) In addition to reading the hardware pointer, skip over any additional packets with OWN set -- we may have queued more. rtw_rxring_fixup(): k) Remove this, because it's completely wrong. rtw_intr_ioerror(): l) Don't handle receive overruns here; handle them in the normal receive path. Also changed RTW_INTR_RX and RTW_INTR_IOERROR accordingly. m) Do not fiddle with CR; the transmit engine is still running just fine. n) Do not fiddle with IMR; it was unnecessary. o) Don't touch the receive engine at all; not necessary now that we're leaving CR alone. p) Remove the rtw_do_chip_reset case; it was dangerous and not actually used. rtw_intr(): q) Don't mask off ISR bits when we pass down the ISR value; it's not necessary, and sometimes the other bits useful in debugging output. rtw_seg_too_short(): r) Don't spew here; the caller will do it if necessary. (Stops spewage in an expected case.) rtw_rxdesc_blk: s) Minor change to make the elements have the same order as txdesc_blk. general: t) Added a bunch of debugging output, #ifdef RTW_DIAG, that was useful while investigating the ring management problems. Also moved rtw_txring_fixup() into RTW_DIAG, as I believe the transmit pointer is always kept in sync now, and the function has potential race conditions.
2007-01-06 11:27:53 +03:00
rtw_intr_beacon(sc, isr);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
if ((isr & RTW_INTR_ATIMINT) != 0)
rtw_intr_atim(sc);
if ((isr & RTW_INTR_IOERROR) != 0)
I have made a small change that suppresses "recovery" from (spurious?) tx timeouts on the beacon queue. The driver's recovery process badly disrupts the MAC's receiver. This needs a closer look, later. Every other fix is by Charles M. Hannum. Charles writes: "With these, the device is quite stable under load, does not experience xmit timeouts, no longer mysteriously freezes, and does not spew a lot of garbage. It does, however, continue to get some "tx fifo underflow"s. global, SYSCTL_SETUP(): a) Removed unused sysctl variables. rtw_io_enable(): b) Change rtw_io_enable to take a softc, and make it call rtw_txring_fixup(). c) Enable the PCI multiple read/write bit. d) Always reset rdb_next to 0 when starting the receive engine. rtw_intr_rx(): e) When processing receive packets, store the next pointer in rdb_next and read it back on the next round, to fix a race condition vs. rtw_tune() during AP discovery and association. f) Remove the special "skip to 0th descriptor" code, because it is no longer needed. rtw_collect_txring(): g) In the "force" case, stop when we catch up to the hardware xmit pointer. rtw_intr_beacon(): h) Resetting tdb_next here looked fishy to me, in light of changes in rtw_collect_txring(). XXX However, this is the one part I haven't tested. rtw_txring_fixup(): i) Add function and line arguments, so we can print out information about the caller. j) In addition to reading the hardware pointer, skip over any additional packets with OWN set -- we may have queued more. rtw_rxring_fixup(): k) Remove this, because it's completely wrong. rtw_intr_ioerror(): l) Don't handle receive overruns here; handle them in the normal receive path. Also changed RTW_INTR_RX and RTW_INTR_IOERROR accordingly. m) Do not fiddle with CR; the transmit engine is still running just fine. n) Do not fiddle with IMR; it was unnecessary. o) Don't touch the receive engine at all; not necessary now that we're leaving CR alone. p) Remove the rtw_do_chip_reset case; it was dangerous and not actually used. rtw_intr(): q) Don't mask off ISR bits when we pass down the ISR value; it's not necessary, and sometimes the other bits useful in debugging output. rtw_seg_too_short(): r) Don't spew here; the caller will do it if necessary. (Stops spewage in an expected case.) rtw_rxdesc_blk: s) Minor change to make the elements have the same order as txdesc_blk. general: t) Added a bunch of debugging output, #ifdef RTW_DIAG, that was useful while investigating the ring management problems. Also moved rtw_txring_fixup() into RTW_DIAG, as I believe the transmit pointer is always kept in sync now, and the function has potential race conditions.
2007-01-06 11:27:53 +03:00
rtw_intr_ioerror(sc, isr);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
if ((isr & RTW_INTR_TIMEOUT) != 0)
rtw_intr_timeout(sc);
}
return 1;
}
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
/* Must be called at splnet. */
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
static void
rtw_stop(struct ifnet *ifp, int disable)
{
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
int pri;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
struct rtw_softc *sc = (struct rtw_softc *)ifp->if_softc;
struct ieee80211com *ic = &sc->sc_ic;
struct rtw_regs *regs = &sc->sc_regs;
rtw_suspend_ticks(sc);
ieee80211_new_state(ic, IEEE80211_S_INIT, -1);
Use device_t and its accessors throughout. Use aprint_*_dev(). Improve PMF-ability. Add a 'flags' argument to suspend/resume handlers and callers such as pmf_system_suspend(). Define a flag, PMF_F_SELF, which indicates to PMF that a device is suspending/resuming itself. Add helper routines, pmf_device_suspend_self(dev) and pmf_device_resume_self(dev), that call pmf_device_suspend(dev, PMF_F_SELF) and pmf_device_resume(dev, PMF_F_SELF), respectively. Use PMF_F_SELF to suspend/resume self in ath(4), audio(4), rtw(4), and sip(4). In ath(4) and in rtw(4), replace the icky sc_enable/sc_disable callbacks, provided by the bus front-end, with self-suspension/resumption. Also, clean up the bus front-ends. Make sure that the interrupt handler is disestablished during suspension. Get rid of driver-private flags (e.g., RTW_F_ENABLED, ath_softc->sc_invalid); use device_is_active()/device_has_power() calls, instead. In the network-class suspend handler, call if_stop(, 0) instead of if_stop(, 1), because the latter is superfluous (bus- and driver-suspension hooks will 'disable' the NIC), and it may cause recursion. In the network-class resume handler, prevent infinite recursion through if_init() by getting out early if we are self-suspending (PMF_F_SELF). rtw(4) improvements: Destroy rtw(4) callouts when we detach it. Make rtw at pci detachable. Print some more information with the "rx frame too long" warning. Remove activate() methods: Get rid of rtw_activate() and ath_activate(). The device activate() methods are not good for much these days. Make ath at cardbus resume with crypto functions intact: Introduce a boolean device property, "pmf-powerdown". If pmf-powerdown is present and false, it indicates that a bus back-end should not remove power from a device. Honor this property in cardbus_child_suspend(). Set this property to 'false' in ath_attach(), since removing power from an ath at cardbus seems to lobotomize the WPA crypto engine. XXX Should the pmf-powerdown property propagate toward the root of the device tree? Miscellaneous ath(4) changes: Warn if ath(4) tries to write crypto keys to suspended hardware. Reduce differences between FreeBSD and NetBSD in ath(4) multicast filter setup. Make ath_printrxbuf() print an rx descriptor's status & key index, to help debug crypto errors. Shorten a staircase in ath_ioctl(). Don't check for ieee80211_ioctl() return code ERESTART, it never happens.
2008-03-12 21:02:21 +03:00
if (device_has_power(sc->sc_dev)) {
Miscellaneous changes. Details below. Important changes flagged with []. Using the driver with my Linksys WPC11 ver. 4, it seems to be receiving packets for a change. The WPC11 ver. 4 has a Maxim RF section. My no-name rtw with Philips RF section still does not receive any packets. Keep access-level (analog params > config[0123] registers > none) in sc_access. Add rtw_set_access for changing the access level. Make rtw_continuous_tx_enable and other subroutines use rtw_set_access instead of rtw_config0123_enable and rtw_anaparm_enable. Factor part of the chip-reset code into rtw_chip_reset1. Change the 'struct foo (*bar)[N]'-style arguments to 'struct foo *bar'-style arguments. Consolidate software/hardware Tx/Rx ring setup in rtw_hwring_setup, rtw_swring_setup. Add a new constant, SA2400_OPMODE_DEFAULTS, for the bits that we *always* set in the SA2400 OPMODE register. Factor some code out into rtw_sa2400_calibrate. (Inspired by the Linux driver.) [] When the receiver goes into underrun/overflow state, call a new subroutine, rtw_kick() that stops the Rx/Tx processes, resets the chip, reinitializes the Tx/Rx rings, and restarts Rx/Tx processes. (Inspired by the Linux driver.) [] In rtw_intr_rx, check for too-short packets before calling ieee80211_find_rxnode. I believe this will prevent a repeat of the MCHK exception I saw once on macppc. [] Use seconds-elapased as well as microseconds-elapsed to set the next "due date" for the timeout interrupt. This keeps the driver from programming the timeout to expire too early. [] In rtw_intr, read RTW_ISR at most 10 times, then get out. If the interface is not enabled (RTW_F_ENABLED), then get out. [] In rtw_stop, get out if the interface is not enabled (RTW_F_ENABLED). Block IPL_NET interrupts. Don't read/write any registers if the interface is invalid (RTW_F_INVALID). [] Call rtw_stop in rtw_detach.
2004-12-12 09:37:59 +03:00
/* Disable interrupts. */
RTW_WRITE16(regs, RTW_IMR, 0);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
Add some bus barriers. Remark where they are needed, e.g., read_eeprom. Change some printfs to RTW_DPRINTF2s to quiet rtw(4) when debugging is disabled.
2004-12-20 04:13:45 +03:00
RTW_WBW(regs, RTW_TPPOLL, RTW_IMR);
Miscellaneous changes. Details below. Important changes flagged with []. Using the driver with my Linksys WPC11 ver. 4, it seems to be receiving packets for a change. The WPC11 ver. 4 has a Maxim RF section. My no-name rtw with Philips RF section still does not receive any packets. Keep access-level (analog params > config[0123] registers > none) in sc_access. Add rtw_set_access for changing the access level. Make rtw_continuous_tx_enable and other subroutines use rtw_set_access instead of rtw_config0123_enable and rtw_anaparm_enable. Factor part of the chip-reset code into rtw_chip_reset1. Change the 'struct foo (*bar)[N]'-style arguments to 'struct foo *bar'-style arguments. Consolidate software/hardware Tx/Rx ring setup in rtw_hwring_setup, rtw_swring_setup. Add a new constant, SA2400_OPMODE_DEFAULTS, for the bits that we *always* set in the SA2400 OPMODE register. Factor some code out into rtw_sa2400_calibrate. (Inspired by the Linux driver.) [] When the receiver goes into underrun/overflow state, call a new subroutine, rtw_kick() that stops the Rx/Tx processes, resets the chip, reinitializes the Tx/Rx rings, and restarts Rx/Tx processes. (Inspired by the Linux driver.) [] In rtw_intr_rx, check for too-short packets before calling ieee80211_find_rxnode. I believe this will prevent a repeat of the MCHK exception I saw once on macppc. [] Use seconds-elapased as well as microseconds-elapsed to set the next "due date" for the timeout interrupt. This keeps the driver from programming the timeout to expire too early. [] In rtw_intr, read RTW_ISR at most 10 times, then get out. If the interface is not enabled (RTW_F_ENABLED), then get out. [] In rtw_stop, get out if the interface is not enabled (RTW_F_ENABLED). Block IPL_NET interrupts. Don't read/write any registers if the interface is invalid (RTW_F_INVALID). [] Call rtw_stop in rtw_detach.
2004-12-12 09:37:59 +03:00
/* Stop the transmit and receive processes. First stop DMA,
* then disable receiver and transmitter.
*/
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
RTW_WRITE8(regs, RTW_TPPOLL, RTW_TPPOLL_SALL);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
Add some bus barriers. Remark where they are needed, e.g., read_eeprom. Change some printfs to RTW_DPRINTF2s to quiet rtw(4) when debugging is disabled.
2004-12-20 04:13:45 +03:00
RTW_SYNC(regs, RTW_TPPOLL, RTW_IMR);
I have made a small change that suppresses "recovery" from (spurious?) tx timeouts on the beacon queue. The driver's recovery process badly disrupts the MAC's receiver. This needs a closer look, later. Every other fix is by Charles M. Hannum. Charles writes: "With these, the device is quite stable under load, does not experience xmit timeouts, no longer mysteriously freezes, and does not spew a lot of garbage. It does, however, continue to get some "tx fifo underflow"s. global, SYSCTL_SETUP(): a) Removed unused sysctl variables. rtw_io_enable(): b) Change rtw_io_enable to take a softc, and make it call rtw_txring_fixup(). c) Enable the PCI multiple read/write bit. d) Always reset rdb_next to 0 when starting the receive engine. rtw_intr_rx(): e) When processing receive packets, store the next pointer in rdb_next and read it back on the next round, to fix a race condition vs. rtw_tune() during AP discovery and association. f) Remove the special "skip to 0th descriptor" code, because it is no longer needed. rtw_collect_txring(): g) In the "force" case, stop when we catch up to the hardware xmit pointer. rtw_intr_beacon(): h) Resetting tdb_next here looked fishy to me, in light of changes in rtw_collect_txring(). XXX However, this is the one part I haven't tested. rtw_txring_fixup(): i) Add function and line arguments, so we can print out information about the caller. j) In addition to reading the hardware pointer, skip over any additional packets with OWN set -- we may have queued more. rtw_rxring_fixup(): k) Remove this, because it's completely wrong. rtw_intr_ioerror(): l) Don't handle receive overruns here; handle them in the normal receive path. Also changed RTW_INTR_RX and RTW_INTR_IOERROR accordingly. m) Do not fiddle with CR; the transmit engine is still running just fine. n) Do not fiddle with IMR; it was unnecessary. o) Don't touch the receive engine at all; not necessary now that we're leaving CR alone. p) Remove the rtw_do_chip_reset case; it was dangerous and not actually used. rtw_intr(): q) Don't mask off ISR bits when we pass down the ISR value; it's not necessary, and sometimes the other bits useful in debugging output. rtw_seg_too_short(): r) Don't spew here; the caller will do it if necessary. (Stops spewage in an expected case.) rtw_rxdesc_blk: s) Minor change to make the elements have the same order as txdesc_blk. general: t) Added a bunch of debugging output, #ifdef RTW_DIAG, that was useful while investigating the ring management problems. Also moved rtw_txring_fixup() into RTW_DIAG, as I believe the transmit pointer is always kept in sync now, and the function has potential race conditions.
2007-01-06 11:27:53 +03:00
rtw_io_enable(sc, RTW_CR_RE | RTW_CR_TE, 0);
Miscellaneous changes. Details below. Important changes flagged with []. Using the driver with my Linksys WPC11 ver. 4, it seems to be receiving packets for a change. The WPC11 ver. 4 has a Maxim RF section. My no-name rtw with Philips RF section still does not receive any packets. Keep access-level (analog params > config[0123] registers > none) in sc_access. Add rtw_set_access for changing the access level. Make rtw_continuous_tx_enable and other subroutines use rtw_set_access instead of rtw_config0123_enable and rtw_anaparm_enable. Factor part of the chip-reset code into rtw_chip_reset1. Change the 'struct foo (*bar)[N]'-style arguments to 'struct foo *bar'-style arguments. Consolidate software/hardware Tx/Rx ring setup in rtw_hwring_setup, rtw_swring_setup. Add a new constant, SA2400_OPMODE_DEFAULTS, for the bits that we *always* set in the SA2400 OPMODE register. Factor some code out into rtw_sa2400_calibrate. (Inspired by the Linux driver.) [] When the receiver goes into underrun/overflow state, call a new subroutine, rtw_kick() that stops the Rx/Tx processes, resets the chip, reinitializes the Tx/Rx rings, and restarts Rx/Tx processes. (Inspired by the Linux driver.) [] In rtw_intr_rx, check for too-short packets before calling ieee80211_find_rxnode. I believe this will prevent a repeat of the MCHK exception I saw once on macppc. [] Use seconds-elapased as well as microseconds-elapsed to set the next "due date" for the timeout interrupt. This keeps the driver from programming the timeout to expire too early. [] In rtw_intr, read RTW_ISR at most 10 times, then get out. If the interface is not enabled (RTW_F_ENABLED), then get out. [] In rtw_stop, get out if the interface is not enabled (RTW_F_ENABLED). Block IPL_NET interrupts. Don't read/write any registers if the interface is invalid (RTW_F_INVALID). [] Call rtw_stop in rtw_detach.
2004-12-12 09:37:59 +03:00
}
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
for (pri = 0; pri < RTW_NTXPRI; pri++) {
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
rtw_txsofts_release(sc->sc_dmat, &sc->sc_ic,
&sc->sc_txsoft_blk[pri]);
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
}
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
rtw_rxbufs_release(sc->sc_dmat, &sc->sc_rxsoft[0]);
If defined(RTW_DEBUG), provide a sysctl (hw.rtw.rxbufs_limit) for limiting the number of rx buffers an rtw may allocate. Use this sysctl to test the code that copes with buffer exhaustion. Allocate at most RTW_RXQLEN rx buffers, stopping at the sysctl limit. Record in sc_nrxdesc how many were allocated, and put the end-of-ring indication on sc_rxdesc[sc_nrxdesc - 1]. In rtw_init, if no rx buffers could be allocated, log a complaint, clear IFF_RUNNING, and exit with an error. Many changes to accomodate a short rx ring, mainly of the "add a rx-ring length argument" variety. XXX I really should consolidate all of the rx ring variables in one struct and pass that to the rx-ring subroutines. Bug fix: after calling MCLGET, use the (m->m_flags & M_EXT) idiom to check for success, instead of m != NULL. Bug fix: at the top of rtw_start, if IFF_RUNNING is not set, or IFF_OACTIVE is, get out.
2004-12-29 01:21:15 +03:00
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
/* Mark the interface as not running. Cancel the watchdog timer. */
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
ifp->if_flags &= ~(IFF_RUNNING | IFF_OACTIVE);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
ifp->if_timer = 0;
Miscellaneous changes. Details below. Important changes flagged with []. Using the driver with my Linksys WPC11 ver. 4, it seems to be receiving packets for a change. The WPC11 ver. 4 has a Maxim RF section. My no-name rtw with Philips RF section still does not receive any packets. Keep access-level (analog params > config[0123] registers > none) in sc_access. Add rtw_set_access for changing the access level. Make rtw_continuous_tx_enable and other subroutines use rtw_set_access instead of rtw_config0123_enable and rtw_anaparm_enable. Factor part of the chip-reset code into rtw_chip_reset1. Change the 'struct foo (*bar)[N]'-style arguments to 'struct foo *bar'-style arguments. Consolidate software/hardware Tx/Rx ring setup in rtw_hwring_setup, rtw_swring_setup. Add a new constant, SA2400_OPMODE_DEFAULTS, for the bits that we *always* set in the SA2400 OPMODE register. Factor some code out into rtw_sa2400_calibrate. (Inspired by the Linux driver.) [] When the receiver goes into underrun/overflow state, call a new subroutine, rtw_kick() that stops the Rx/Tx processes, resets the chip, reinitializes the Tx/Rx rings, and restarts Rx/Tx processes. (Inspired by the Linux driver.) [] In rtw_intr_rx, check for too-short packets before calling ieee80211_find_rxnode. I believe this will prevent a repeat of the MCHK exception I saw once on macppc. [] Use seconds-elapased as well as microseconds-elapsed to set the next "due date" for the timeout interrupt. This keeps the driver from programming the timeout to expire too early. [] In rtw_intr, read RTW_ISR at most 10 times, then get out. If the interface is not enabled (RTW_F_ENABLED), then get out. [] In rtw_stop, get out if the interface is not enabled (RTW_F_ENABLED). Block IPL_NET interrupts. Don't read/write any registers if the interface is invalid (RTW_F_INVALID). [] Call rtw_stop in rtw_detach.
2004-12-12 09:37:59 +03:00
Use device_t and its accessors throughout. Use aprint_*_dev(). Improve PMF-ability. Add a 'flags' argument to suspend/resume handlers and callers such as pmf_system_suspend(). Define a flag, PMF_F_SELF, which indicates to PMF that a device is suspending/resuming itself. Add helper routines, pmf_device_suspend_self(dev) and pmf_device_resume_self(dev), that call pmf_device_suspend(dev, PMF_F_SELF) and pmf_device_resume(dev, PMF_F_SELF), respectively. Use PMF_F_SELF to suspend/resume self in ath(4), audio(4), rtw(4), and sip(4). In ath(4) and in rtw(4), replace the icky sc_enable/sc_disable callbacks, provided by the bus front-end, with self-suspension/resumption. Also, clean up the bus front-ends. Make sure that the interrupt handler is disestablished during suspension. Get rid of driver-private flags (e.g., RTW_F_ENABLED, ath_softc->sc_invalid); use device_is_active()/device_has_power() calls, instead. In the network-class suspend handler, call if_stop(, 0) instead of if_stop(, 1), because the latter is superfluous (bus- and driver-suspension hooks will 'disable' the NIC), and it may cause recursion. In the network-class resume handler, prevent infinite recursion through if_init() by getting out early if we are self-suspending (PMF_F_SELF). rtw(4) improvements: Destroy rtw(4) callouts when we detach it. Make rtw at pci detachable. Print some more information with the "rx frame too long" warning. Remove activate() methods: Get rid of rtw_activate() and ath_activate(). The device activate() methods are not good for much these days. Make ath at cardbus resume with crypto functions intact: Introduce a boolean device property, "pmf-powerdown". If pmf-powerdown is present and false, it indicates that a bus back-end should not remove power from a device. Honor this property in cardbus_child_suspend(). Set this property to 'false' in ath_attach(), since removing power from an ath at cardbus seems to lobotomize the WPA crypto engine. XXX Should the pmf-powerdown property propagate toward the root of the device tree? Miscellaneous ath(4) changes: Warn if ath(4) tries to write crypto keys to suspended hardware. Reduce differences between FreeBSD and NetBSD in ath(4) multicast filter setup. Make ath_printrxbuf() print an rx descriptor's status & key index, to help debug crypto errors. Shorten a staircase in ath_ioctl(). Don't check for ieee80211_ioctl() return code ERESTART, it never happens.
2008-03-12 21:02:21 +03:00
if (disable)
In pmf(9), improve the implementation of device self-suspension and make suspension by self, by drvctl(8), and by ACPI system sleep play nice together. Start solidifying some temporary API changes. 1. Extract a new header file, <sys/device_if.h>, from <sys/device.h> and #include it from <sys/pmf.h> instead of <sys/device.h> to break the circular dependency between <sys/device.h> and <sys/pmf.h>. 2. Introduce pmf_qual_t, an aggregate of qualifications on a PMF suspend/resume call. Start to replace instances of PMF_FN_PROTO, PMF_FN_ARGS, et cetera, with a pmf_qual_t. 3. Introduce the notion of a "suspensor," an entity that holds a device in suspension. More than one suspensor may hold a device at once. A device stays suspended as long as at least one suspensor holds it. A device resumes when the last suspensor releases it. Currently, the kernel defines three suspensors, 3a the system-suspensor: for system suspension, initiated by 'sysctl -w machdep.sleep_state=3', by lid closure, by power-button press, et cetera, 3b the drvctl-suspensor: for device suspension by /dev/drvctl ioctl, e.g., drvctl -S sip0. 3c the system self-suspensor: for device drivers that suspend themselves and their children. Several drivers for network interfaces put the network device to sleep while it is not administratively up, that is, after the kernel calls if_stop(, 1). The self-suspensor should not be used directly. See the description of suspensor delegates, below. A suspensor can have one or more "delegates". A suspensor can release devices that its delegates hold suspended. Right now, only the system self-suspensor has delegates. For each device that a self-suspending driver attaches, it creates the device's self-suspensor, a delegate of the system self-suspensor. Suspensors stop a system-wide suspend/resume cycle from waking devices that the operator put to sleep with drvctl before the cycle. They also help self-suspension to work more simply, safely, and in accord with expectations. 4. Add the notion of device activation level, devact_level_t, and a routine for checking the current activation level, device_activation(). Current activation levels are DEVACT_LEVEL_BUS, DEVACT_LEVEL_DRIVER, and DEVACT_LEVEL_CLASS, which respectively indicate that the device's bus is active, that the bus and device are active, and that the bus, device, and the functions of the device's class (network, audio) are active. Suspend/resume calls can be qualified with a devact_level_t. The power-management framework treats a devact_level_t that qualifies a device suspension as the device's current activation level; it only runs hooks to reduce the activation level from the presumed current level to the fully suspended state. The framework treats a devact_level_t qualifying device resumption as the target activation level; it only runs hooks to raise the activation level to the target. 5. Use pmf_qual_t, devact_level_t, and self-suspensors in several drivers. 6. Temporarily add an unused power-management workqueue that I will remove or replace, soon.
2009-09-16 20:34:49 +04:00
pmf_device_suspend(sc->sc_dev, &sc->sc_qual);
Use device_t and its accessors throughout. Use aprint_*_dev(). Improve PMF-ability. Add a 'flags' argument to suspend/resume handlers and callers such as pmf_system_suspend(). Define a flag, PMF_F_SELF, which indicates to PMF that a device is suspending/resuming itself. Add helper routines, pmf_device_suspend_self(dev) and pmf_device_resume_self(dev), that call pmf_device_suspend(dev, PMF_F_SELF) and pmf_device_resume(dev, PMF_F_SELF), respectively. Use PMF_F_SELF to suspend/resume self in ath(4), audio(4), rtw(4), and sip(4). In ath(4) and in rtw(4), replace the icky sc_enable/sc_disable callbacks, provided by the bus front-end, with self-suspension/resumption. Also, clean up the bus front-ends. Make sure that the interrupt handler is disestablished during suspension. Get rid of driver-private flags (e.g., RTW_F_ENABLED, ath_softc->sc_invalid); use device_is_active()/device_has_power() calls, instead. In the network-class suspend handler, call if_stop(, 0) instead of if_stop(, 1), because the latter is superfluous (bus- and driver-suspension hooks will 'disable' the NIC), and it may cause recursion. In the network-class resume handler, prevent infinite recursion through if_init() by getting out early if we are self-suspending (PMF_F_SELF). rtw(4) improvements: Destroy rtw(4) callouts when we detach it. Make rtw at pci detachable. Print some more information with the "rx frame too long" warning. Remove activate() methods: Get rid of rtw_activate() and ath_activate(). The device activate() methods are not good for much these days. Make ath at cardbus resume with crypto functions intact: Introduce a boolean device property, "pmf-powerdown". If pmf-powerdown is present and false, it indicates that a bus back-end should not remove power from a device. Honor this property in cardbus_child_suspend(). Set this property to 'false' in ath_attach(), since removing power from an ath at cardbus seems to lobotomize the WPA crypto engine. XXX Should the pmf-powerdown property propagate toward the root of the device tree? Miscellaneous ath(4) changes: Warn if ath(4) tries to write crypto keys to suspended hardware. Reduce differences between FreeBSD and NetBSD in ath(4) multicast filter setup. Make ath_printrxbuf() print an rx descriptor's status & key index, to help debug crypto errors. Shorten a staircase in ath_ioctl(). Don't check for ieee80211_ioctl() return code ERESTART, it never happens.
2008-03-12 21:02:21 +03:00
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
return;
}
const char *
rtw_pwrstate_string(enum rtw_pwrstate power)
{
switch (power) {
case RTW_ON:
return "on";
case RTW_SLEEP:
return "sleep";
case RTW_OFF:
return "off";
default:
return "unknown";
}
}
Realtek sample sources are not consistent about the "Digital PHY" bit in the SROM. It seems as if it is set to 1 when the PHY is *analog*, not *digital*. Fix my sources. In rtw_intr_rx, use units of 500kb/s instead of 100kb/s for rate, to be consistent with net80211's expectations. Polish up some debugging ugly messages. Dump raw 802.11 packets if IFF_DEBUG|IFF_LINK2 and RTW_DEBUG is defined. Polish power-state (on/sleep/off) handling. Especially improve support for RFMD (totally untested) and Maxim. For Philips, take the Digital PHY property into account. Call the net80211 watchdog function from rtw_watchdog, so that we scan again if auth/assoc fails. Be a little more cautious about writing register[RTW_TPPOLL], since other drivers are.... Don't frob the high/low-priority queues right now, since I don't use them. Add rtw_join_bss which programs the card with the BSSID and other properties of a BSS. Use it on state transitions. Factor out rtw_set_nettype. Make rtw_recv_beacon call ieee80211_recv_mgmt instead of dropping beacons on the floor! TBD IBSS merges. Change some rtw_debug=2 printfs to rtw_debug=3 (RTW_DPRINTF3) printfs so the console doesn't get spammed so badly at rtw_debug=2. Change some debugging printfs to RTW_DPRINTFs. E.g., print the "RF programming method" only if debugging is enabled.
2004-12-21 02:05:41 +03:00
/* XXX For Maxim, I am using the RFMD settings gleaned from the
* reference driver, plus a magic Maxim "ON" value that comes from
* the Realtek document "Windows PG for Rtl8180."
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
*/
static void
rtw_maxim_pwrstate(struct rtw_regs *regs, enum rtw_pwrstate power,
__unused removal on arguments; approved by core.
2006-11-16 04:32:37 +03:00
int before_rf, int digphy)
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
{
Cosmetic: change u_int{8,16,32,64}_t to C99-style uint{8,16,32,64}_t, which is preferred.
2005-01-02 07:23:03 +03:00
uint32_t anaparm;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
Realtek sample sources are not consistent about the "Digital PHY" bit in the SROM. It seems as if it is set to 1 when the PHY is *analog*, not *digital*. Fix my sources. In rtw_intr_rx, use units of 500kb/s instead of 100kb/s for rate, to be consistent with net80211's expectations. Polish up some debugging ugly messages. Dump raw 802.11 packets if IFF_DEBUG|IFF_LINK2 and RTW_DEBUG is defined. Polish power-state (on/sleep/off) handling. Especially improve support for RFMD (totally untested) and Maxim. For Philips, take the Digital PHY property into account. Call the net80211 watchdog function from rtw_watchdog, so that we scan again if auth/assoc fails. Be a little more cautious about writing register[RTW_TPPOLL], since other drivers are.... Don't frob the high/low-priority queues right now, since I don't use them. Add rtw_join_bss which programs the card with the BSSID and other properties of a BSS. Use it on state transitions. Factor out rtw_set_nettype. Make rtw_recv_beacon call ieee80211_recv_mgmt instead of dropping beacons on the floor! TBD IBSS merges. Change some rtw_debug=2 printfs to rtw_debug=3 (RTW_DPRINTF3) printfs so the console doesn't get spammed so badly at rtw_debug=2. Change some debugging printfs to RTW_DPRINTFs. E.g., print the "RF programming method" only if debugging is enabled.
2004-12-21 02:05:41 +03:00
anaparm = RTW_READ(regs, RTW_ANAPARM);
anaparm &= ~(RTW_ANAPARM_RFPOW_MASK | RTW_ANAPARM_TXDACOFF);
switch (power) {
case RTW_OFF:
if (before_rf)
return;
anaparm |= RTW_ANAPARM_RFPOW_MAXIM_OFF;
anaparm |= RTW_ANAPARM_TXDACOFF;
break;
case RTW_SLEEP:
if (!before_rf)
return;
anaparm |= RTW_ANAPARM_RFPOW_MAXIM_SLEEP;
anaparm |= RTW_ANAPARM_TXDACOFF;
break;
case RTW_ON:
if (!before_rf)
return;
anaparm |= RTW_ANAPARM_RFPOW_MAXIM_ON;
break;
}
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
RTW_DPRINTF(RTW_DEBUG_PWR,
("%s: power state %s, %s RF, reg[ANAPARM] <- %08x\n",
Realtek sample sources are not consistent about the "Digital PHY" bit in the SROM. It seems as if it is set to 1 when the PHY is *analog*, not *digital*. Fix my sources. In rtw_intr_rx, use units of 500kb/s instead of 100kb/s for rate, to be consistent with net80211's expectations. Polish up some debugging ugly messages. Dump raw 802.11 packets if IFF_DEBUG|IFF_LINK2 and RTW_DEBUG is defined. Polish power-state (on/sleep/off) handling. Especially improve support for RFMD (totally untested) and Maxim. For Philips, take the Digital PHY property into account. Call the net80211 watchdog function from rtw_watchdog, so that we scan again if auth/assoc fails. Be a little more cautious about writing register[RTW_TPPOLL], since other drivers are.... Don't frob the high/low-priority queues right now, since I don't use them. Add rtw_join_bss which programs the card with the BSSID and other properties of a BSS. Use it on state transitions. Factor out rtw_set_nettype. Make rtw_recv_beacon call ieee80211_recv_mgmt instead of dropping beacons on the floor! TBD IBSS merges. Change some rtw_debug=2 printfs to rtw_debug=3 (RTW_DPRINTF3) printfs so the console doesn't get spammed so badly at rtw_debug=2. Change some debugging printfs to RTW_DPRINTFs. E.g., print the "RF programming method" only if debugging is enabled.
2004-12-21 02:05:41 +03:00
__func__, rtw_pwrstate_string(power),
(before_rf) ? "before" : "after", anaparm));
RTW_WRITE(regs, RTW_ANAPARM, anaparm);
RTW_SYNC(regs, RTW_ANAPARM, RTW_ANAPARM);
}
/* XXX I am using the RFMD settings gleaned from the reference
nuke trailing whitespace
2005-02-27 03:26:58 +03:00
* driver. They agree
Realtek sample sources are not consistent about the "Digital PHY" bit in the SROM. It seems as if it is set to 1 when the PHY is *analog*, not *digital*. Fix my sources. In rtw_intr_rx, use units of 500kb/s instead of 100kb/s for rate, to be consistent with net80211's expectations. Polish up some debugging ugly messages. Dump raw 802.11 packets if IFF_DEBUG|IFF_LINK2 and RTW_DEBUG is defined. Polish power-state (on/sleep/off) handling. Especially improve support for RFMD (totally untested) and Maxim. For Philips, take the Digital PHY property into account. Call the net80211 watchdog function from rtw_watchdog, so that we scan again if auth/assoc fails. Be a little more cautious about writing register[RTW_TPPOLL], since other drivers are.... Don't frob the high/low-priority queues right now, since I don't use them. Add rtw_join_bss which programs the card with the BSSID and other properties of a BSS. Use it on state transitions. Factor out rtw_set_nettype. Make rtw_recv_beacon call ieee80211_recv_mgmt instead of dropping beacons on the floor! TBD IBSS merges. Change some rtw_debug=2 printfs to rtw_debug=3 (RTW_DPRINTF3) printfs so the console doesn't get spammed so badly at rtw_debug=2. Change some debugging printfs to RTW_DPRINTFs. E.g., print the "RF programming method" only if debugging is enabled.
2004-12-21 02:05:41 +03:00
*/
static void
rtw_rfmd_pwrstate(struct rtw_regs *regs, enum rtw_pwrstate power,
__unused removal on arguments; approved by core.
2006-11-16 04:32:37 +03:00
int before_rf, int digphy)
Realtek sample sources are not consistent about the "Digital PHY" bit in the SROM. It seems as if it is set to 1 when the PHY is *analog*, not *digital*. Fix my sources. In rtw_intr_rx, use units of 500kb/s instead of 100kb/s for rate, to be consistent with net80211's expectations. Polish up some debugging ugly messages. Dump raw 802.11 packets if IFF_DEBUG|IFF_LINK2 and RTW_DEBUG is defined. Polish power-state (on/sleep/off) handling. Especially improve support for RFMD (totally untested) and Maxim. For Philips, take the Digital PHY property into account. Call the net80211 watchdog function from rtw_watchdog, so that we scan again if auth/assoc fails. Be a little more cautious about writing register[RTW_TPPOLL], since other drivers are.... Don't frob the high/low-priority queues right now, since I don't use them. Add rtw_join_bss which programs the card with the BSSID and other properties of a BSS. Use it on state transitions. Factor out rtw_set_nettype. Make rtw_recv_beacon call ieee80211_recv_mgmt instead of dropping beacons on the floor! TBD IBSS merges. Change some rtw_debug=2 printfs to rtw_debug=3 (RTW_DPRINTF3) printfs so the console doesn't get spammed so badly at rtw_debug=2. Change some debugging printfs to RTW_DPRINTFs. E.g., print the "RF programming method" only if debugging is enabled.
2004-12-21 02:05:41 +03:00
{
Cosmetic: change u_int{8,16,32,64}_t to C99-style uint{8,16,32,64}_t, which is preferred.
2005-01-02 07:23:03 +03:00
uint32_t anaparm;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
anaparm = RTW_READ(regs, RTW_ANAPARM);
Realtek sample sources are not consistent about the "Digital PHY" bit in the SROM. It seems as if it is set to 1 when the PHY is *analog*, not *digital*. Fix my sources. In rtw_intr_rx, use units of 500kb/s instead of 100kb/s for rate, to be consistent with net80211's expectations. Polish up some debugging ugly messages. Dump raw 802.11 packets if IFF_DEBUG|IFF_LINK2 and RTW_DEBUG is defined. Polish power-state (on/sleep/off) handling. Especially improve support for RFMD (totally untested) and Maxim. For Philips, take the Digital PHY property into account. Call the net80211 watchdog function from rtw_watchdog, so that we scan again if auth/assoc fails. Be a little more cautious about writing register[RTW_TPPOLL], since other drivers are.... Don't frob the high/low-priority queues right now, since I don't use them. Add rtw_join_bss which programs the card with the BSSID and other properties of a BSS. Use it on state transitions. Factor out rtw_set_nettype. Make rtw_recv_beacon call ieee80211_recv_mgmt instead of dropping beacons on the floor! TBD IBSS merges. Change some rtw_debug=2 printfs to rtw_debug=3 (RTW_DPRINTF3) printfs so the console doesn't get spammed so badly at rtw_debug=2. Change some debugging printfs to RTW_DPRINTFs. E.g., print the "RF programming method" only if debugging is enabled.
2004-12-21 02:05:41 +03:00
anaparm &= ~(RTW_ANAPARM_RFPOW_MASK | RTW_ANAPARM_TXDACOFF);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
switch (power) {
case RTW_OFF:
if (before_rf)
return;
Realtek sample sources are not consistent about the "Digital PHY" bit in the SROM. It seems as if it is set to 1 when the PHY is *analog*, not *digital*. Fix my sources. In rtw_intr_rx, use units of 500kb/s instead of 100kb/s for rate, to be consistent with net80211's expectations. Polish up some debugging ugly messages. Dump raw 802.11 packets if IFF_DEBUG|IFF_LINK2 and RTW_DEBUG is defined. Polish power-state (on/sleep/off) handling. Especially improve support for RFMD (totally untested) and Maxim. For Philips, take the Digital PHY property into account. Call the net80211 watchdog function from rtw_watchdog, so that we scan again if auth/assoc fails. Be a little more cautious about writing register[RTW_TPPOLL], since other drivers are.... Don't frob the high/low-priority queues right now, since I don't use them. Add rtw_join_bss which programs the card with the BSSID and other properties of a BSS. Use it on state transitions. Factor out rtw_set_nettype. Make rtw_recv_beacon call ieee80211_recv_mgmt instead of dropping beacons on the floor! TBD IBSS merges. Change some rtw_debug=2 printfs to rtw_debug=3 (RTW_DPRINTF3) printfs so the console doesn't get spammed so badly at rtw_debug=2. Change some debugging printfs to RTW_DPRINTFs. E.g., print the "RF programming method" only if debugging is enabled.
2004-12-21 02:05:41 +03:00
anaparm |= RTW_ANAPARM_RFPOW_RFMD_OFF;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
anaparm |= RTW_ANAPARM_TXDACOFF;
break;
case RTW_SLEEP:
if (!before_rf)
return;
Realtek sample sources are not consistent about the "Digital PHY" bit in the SROM. It seems as if it is set to 1 when the PHY is *analog*, not *digital*. Fix my sources. In rtw_intr_rx, use units of 500kb/s instead of 100kb/s for rate, to be consistent with net80211's expectations. Polish up some debugging ugly messages. Dump raw 802.11 packets if IFF_DEBUG|IFF_LINK2 and RTW_DEBUG is defined. Polish power-state (on/sleep/off) handling. Especially improve support for RFMD (totally untested) and Maxim. For Philips, take the Digital PHY property into account. Call the net80211 watchdog function from rtw_watchdog, so that we scan again if auth/assoc fails. Be a little more cautious about writing register[RTW_TPPOLL], since other drivers are.... Don't frob the high/low-priority queues right now, since I don't use them. Add rtw_join_bss which programs the card with the BSSID and other properties of a BSS. Use it on state transitions. Factor out rtw_set_nettype. Make rtw_recv_beacon call ieee80211_recv_mgmt instead of dropping beacons on the floor! TBD IBSS merges. Change some rtw_debug=2 printfs to rtw_debug=3 (RTW_DPRINTF3) printfs so the console doesn't get spammed so badly at rtw_debug=2. Change some debugging printfs to RTW_DPRINTFs. E.g., print the "RF programming method" only if debugging is enabled.
2004-12-21 02:05:41 +03:00
anaparm |= RTW_ANAPARM_RFPOW_RFMD_SLEEP;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
anaparm |= RTW_ANAPARM_TXDACOFF;
break;
case RTW_ON:
if (!before_rf)
return;
Realtek sample sources are not consistent about the "Digital PHY" bit in the SROM. It seems as if it is set to 1 when the PHY is *analog*, not *digital*. Fix my sources. In rtw_intr_rx, use units of 500kb/s instead of 100kb/s for rate, to be consistent with net80211's expectations. Polish up some debugging ugly messages. Dump raw 802.11 packets if IFF_DEBUG|IFF_LINK2 and RTW_DEBUG is defined. Polish power-state (on/sleep/off) handling. Especially improve support for RFMD (totally untested) and Maxim. For Philips, take the Digital PHY property into account. Call the net80211 watchdog function from rtw_watchdog, so that we scan again if auth/assoc fails. Be a little more cautious about writing register[RTW_TPPOLL], since other drivers are.... Don't frob the high/low-priority queues right now, since I don't use them. Add rtw_join_bss which programs the card with the BSSID and other properties of a BSS. Use it on state transitions. Factor out rtw_set_nettype. Make rtw_recv_beacon call ieee80211_recv_mgmt instead of dropping beacons on the floor! TBD IBSS merges. Change some rtw_debug=2 printfs to rtw_debug=3 (RTW_DPRINTF3) printfs so the console doesn't get spammed so badly at rtw_debug=2. Change some debugging printfs to RTW_DPRINTFs. E.g., print the "RF programming method" only if debugging is enabled.
2004-12-21 02:05:41 +03:00
anaparm |= RTW_ANAPARM_RFPOW_RFMD_ON;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
break;
}
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
RTW_DPRINTF(RTW_DEBUG_PWR,
("%s: power state %s, %s RF, reg[ANAPARM] <- %08x\n",
Realtek sample sources are not consistent about the "Digital PHY" bit in the SROM. It seems as if it is set to 1 when the PHY is *analog*, not *digital*. Fix my sources. In rtw_intr_rx, use units of 500kb/s instead of 100kb/s for rate, to be consistent with net80211's expectations. Polish up some debugging ugly messages. Dump raw 802.11 packets if IFF_DEBUG|IFF_LINK2 and RTW_DEBUG is defined. Polish power-state (on/sleep/off) handling. Especially improve support for RFMD (totally untested) and Maxim. For Philips, take the Digital PHY property into account. Call the net80211 watchdog function from rtw_watchdog, so that we scan again if auth/assoc fails. Be a little more cautious about writing register[RTW_TPPOLL], since other drivers are.... Don't frob the high/low-priority queues right now, since I don't use them. Add rtw_join_bss which programs the card with the BSSID and other properties of a BSS. Use it on state transitions. Factor out rtw_set_nettype. Make rtw_recv_beacon call ieee80211_recv_mgmt instead of dropping beacons on the floor! TBD IBSS merges. Change some rtw_debug=2 printfs to rtw_debug=3 (RTW_DPRINTF3) printfs so the console doesn't get spammed so badly at rtw_debug=2. Change some debugging printfs to RTW_DPRINTFs. E.g., print the "RF programming method" only if debugging is enabled.
2004-12-21 02:05:41 +03:00
__func__, rtw_pwrstate_string(power),
(before_rf) ? "before" : "after", anaparm));
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
RTW_WRITE(regs, RTW_ANAPARM, anaparm);
RTW_SYNC(regs, RTW_ANAPARM, RTW_ANAPARM);
}
static void
rtw_philips_pwrstate(struct rtw_regs *regs, enum rtw_pwrstate power,
Realtek sample sources are not consistent about the "Digital PHY" bit in the SROM. It seems as if it is set to 1 when the PHY is *analog*, not *digital*. Fix my sources. In rtw_intr_rx, use units of 500kb/s instead of 100kb/s for rate, to be consistent with net80211's expectations. Polish up some debugging ugly messages. Dump raw 802.11 packets if IFF_DEBUG|IFF_LINK2 and RTW_DEBUG is defined. Polish power-state (on/sleep/off) handling. Especially improve support for RFMD (totally untested) and Maxim. For Philips, take the Digital PHY property into account. Call the net80211 watchdog function from rtw_watchdog, so that we scan again if auth/assoc fails. Be a little more cautious about writing register[RTW_TPPOLL], since other drivers are.... Don't frob the high/low-priority queues right now, since I don't use them. Add rtw_join_bss which programs the card with the BSSID and other properties of a BSS. Use it on state transitions. Factor out rtw_set_nettype. Make rtw_recv_beacon call ieee80211_recv_mgmt instead of dropping beacons on the floor! TBD IBSS merges. Change some rtw_debug=2 printfs to rtw_debug=3 (RTW_DPRINTF3) printfs so the console doesn't get spammed so badly at rtw_debug=2. Change some debugging printfs to RTW_DPRINTFs. E.g., print the "RF programming method" only if debugging is enabled.
2004-12-21 02:05:41 +03:00
int before_rf, int digphy)
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
{
Cosmetic: change u_int{8,16,32,64}_t to C99-style uint{8,16,32,64}_t, which is preferred.
2005-01-02 07:23:03 +03:00
uint32_t anaparm;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
anaparm = RTW_READ(regs, RTW_ANAPARM);
Realtek sample sources are not consistent about the "Digital PHY" bit in the SROM. It seems as if it is set to 1 when the PHY is *analog*, not *digital*. Fix my sources. In rtw_intr_rx, use units of 500kb/s instead of 100kb/s for rate, to be consistent with net80211's expectations. Polish up some debugging ugly messages. Dump raw 802.11 packets if IFF_DEBUG|IFF_LINK2 and RTW_DEBUG is defined. Polish power-state (on/sleep/off) handling. Especially improve support for RFMD (totally untested) and Maxim. For Philips, take the Digital PHY property into account. Call the net80211 watchdog function from rtw_watchdog, so that we scan again if auth/assoc fails. Be a little more cautious about writing register[RTW_TPPOLL], since other drivers are.... Don't frob the high/low-priority queues right now, since I don't use them. Add rtw_join_bss which programs the card with the BSSID and other properties of a BSS. Use it on state transitions. Factor out rtw_set_nettype. Make rtw_recv_beacon call ieee80211_recv_mgmt instead of dropping beacons on the floor! TBD IBSS merges. Change some rtw_debug=2 printfs to rtw_debug=3 (RTW_DPRINTF3) printfs so the console doesn't get spammed so badly at rtw_debug=2. Change some debugging printfs to RTW_DPRINTFs. E.g., print the "RF programming method" only if debugging is enabled.
2004-12-21 02:05:41 +03:00
anaparm &= ~(RTW_ANAPARM_RFPOW_MASK | RTW_ANAPARM_TXDACOFF);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
switch (power) {
case RTW_OFF:
if (before_rf)
return;
Realtek sample sources are not consistent about the "Digital PHY" bit in the SROM. It seems as if it is set to 1 when the PHY is *analog*, not *digital*. Fix my sources. In rtw_intr_rx, use units of 500kb/s instead of 100kb/s for rate, to be consistent with net80211's expectations. Polish up some debugging ugly messages. Dump raw 802.11 packets if IFF_DEBUG|IFF_LINK2 and RTW_DEBUG is defined. Polish power-state (on/sleep/off) handling. Especially improve support for RFMD (totally untested) and Maxim. For Philips, take the Digital PHY property into account. Call the net80211 watchdog function from rtw_watchdog, so that we scan again if auth/assoc fails. Be a little more cautious about writing register[RTW_TPPOLL], since other drivers are.... Don't frob the high/low-priority queues right now, since I don't use them. Add rtw_join_bss which programs the card with the BSSID and other properties of a BSS. Use it on state transitions. Factor out rtw_set_nettype. Make rtw_recv_beacon call ieee80211_recv_mgmt instead of dropping beacons on the floor! TBD IBSS merges. Change some rtw_debug=2 printfs to rtw_debug=3 (RTW_DPRINTF3) printfs so the console doesn't get spammed so badly at rtw_debug=2. Change some debugging printfs to RTW_DPRINTFs. E.g., print the "RF programming method" only if debugging is enabled.
2004-12-21 02:05:41 +03:00
anaparm |= RTW_ANAPARM_RFPOW_PHILIPS_OFF;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
anaparm |= RTW_ANAPARM_TXDACOFF;
break;
case RTW_SLEEP:
if (!before_rf)
return;
Realtek sample sources are not consistent about the "Digital PHY" bit in the SROM. It seems as if it is set to 1 when the PHY is *analog*, not *digital*. Fix my sources. In rtw_intr_rx, use units of 500kb/s instead of 100kb/s for rate, to be consistent with net80211's expectations. Polish up some debugging ugly messages. Dump raw 802.11 packets if IFF_DEBUG|IFF_LINK2 and RTW_DEBUG is defined. Polish power-state (on/sleep/off) handling. Especially improve support for RFMD (totally untested) and Maxim. For Philips, take the Digital PHY property into account. Call the net80211 watchdog function from rtw_watchdog, so that we scan again if auth/assoc fails. Be a little more cautious about writing register[RTW_TPPOLL], since other drivers are.... Don't frob the high/low-priority queues right now, since I don't use them. Add rtw_join_bss which programs the card with the BSSID and other properties of a BSS. Use it on state transitions. Factor out rtw_set_nettype. Make rtw_recv_beacon call ieee80211_recv_mgmt instead of dropping beacons on the floor! TBD IBSS merges. Change some rtw_debug=2 printfs to rtw_debug=3 (RTW_DPRINTF3) printfs so the console doesn't get spammed so badly at rtw_debug=2. Change some debugging printfs to RTW_DPRINTFs. E.g., print the "RF programming method" only if debugging is enabled.
2004-12-21 02:05:41 +03:00
anaparm |= RTW_ANAPARM_RFPOW_PHILIPS_SLEEP;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
anaparm |= RTW_ANAPARM_TXDACOFF;
break;
case RTW_ON:
if (!before_rf)
return;
Realtek sample sources are not consistent about the "Digital PHY" bit in the SROM. It seems as if it is set to 1 when the PHY is *analog*, not *digital*. Fix my sources. In rtw_intr_rx, use units of 500kb/s instead of 100kb/s for rate, to be consistent with net80211's expectations. Polish up some debugging ugly messages. Dump raw 802.11 packets if IFF_DEBUG|IFF_LINK2 and RTW_DEBUG is defined. Polish power-state (on/sleep/off) handling. Especially improve support for RFMD (totally untested) and Maxim. For Philips, take the Digital PHY property into account. Call the net80211 watchdog function from rtw_watchdog, so that we scan again if auth/assoc fails. Be a little more cautious about writing register[RTW_TPPOLL], since other drivers are.... Don't frob the high/low-priority queues right now, since I don't use them. Add rtw_join_bss which programs the card with the BSSID and other properties of a BSS. Use it on state transitions. Factor out rtw_set_nettype. Make rtw_recv_beacon call ieee80211_recv_mgmt instead of dropping beacons on the floor! TBD IBSS merges. Change some rtw_debug=2 printfs to rtw_debug=3 (RTW_DPRINTF3) printfs so the console doesn't get spammed so badly at rtw_debug=2. Change some debugging printfs to RTW_DPRINTFs. E.g., print the "RF programming method" only if debugging is enabled.
2004-12-21 02:05:41 +03:00
if (digphy) {
anaparm |= RTW_ANAPARM_RFPOW_DIG_PHILIPS_ON;
/* XXX guess */
anaparm |= RTW_ANAPARM_TXDACOFF;
} else
anaparm |= RTW_ANAPARM_RFPOW_ANA_PHILIPS_ON;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
break;
}
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
RTW_DPRINTF(RTW_DEBUG_PWR,
("%s: power state %s, %s RF, reg[ANAPARM] <- %08x\n",
Realtek sample sources are not consistent about the "Digital PHY" bit in the SROM. It seems as if it is set to 1 when the PHY is *analog*, not *digital*. Fix my sources. In rtw_intr_rx, use units of 500kb/s instead of 100kb/s for rate, to be consistent with net80211's expectations. Polish up some debugging ugly messages. Dump raw 802.11 packets if IFF_DEBUG|IFF_LINK2 and RTW_DEBUG is defined. Polish power-state (on/sleep/off) handling. Especially improve support for RFMD (totally untested) and Maxim. For Philips, take the Digital PHY property into account. Call the net80211 watchdog function from rtw_watchdog, so that we scan again if auth/assoc fails. Be a little more cautious about writing register[RTW_TPPOLL], since other drivers are.... Don't frob the high/low-priority queues right now, since I don't use them. Add rtw_join_bss which programs the card with the BSSID and other properties of a BSS. Use it on state transitions. Factor out rtw_set_nettype. Make rtw_recv_beacon call ieee80211_recv_mgmt instead of dropping beacons on the floor! TBD IBSS merges. Change some rtw_debug=2 printfs to rtw_debug=3 (RTW_DPRINTF3) printfs so the console doesn't get spammed so badly at rtw_debug=2. Change some debugging printfs to RTW_DPRINTFs. E.g., print the "RF programming method" only if debugging is enabled.
2004-12-21 02:05:41 +03:00
__func__, rtw_pwrstate_string(power),
(before_rf) ? "before" : "after", anaparm));
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
RTW_WRITE(regs, RTW_ANAPARM, anaparm);
RTW_SYNC(regs, RTW_ANAPARM, RTW_ANAPARM);
}
static void
Realtek sample sources are not consistent about the "Digital PHY" bit in the SROM. It seems as if it is set to 1 when the PHY is *analog*, not *digital*. Fix my sources. In rtw_intr_rx, use units of 500kb/s instead of 100kb/s for rate, to be consistent with net80211's expectations. Polish up some debugging ugly messages. Dump raw 802.11 packets if IFF_DEBUG|IFF_LINK2 and RTW_DEBUG is defined. Polish power-state (on/sleep/off) handling. Especially improve support for RFMD (totally untested) and Maxim. For Philips, take the Digital PHY property into account. Call the net80211 watchdog function from rtw_watchdog, so that we scan again if auth/assoc fails. Be a little more cautious about writing register[RTW_TPPOLL], since other drivers are.... Don't frob the high/low-priority queues right now, since I don't use them. Add rtw_join_bss which programs the card with the BSSID and other properties of a BSS. Use it on state transitions. Factor out rtw_set_nettype. Make rtw_recv_beacon call ieee80211_recv_mgmt instead of dropping beacons on the floor! TBD IBSS merges. Change some rtw_debug=2 printfs to rtw_debug=3 (RTW_DPRINTF3) printfs so the console doesn't get spammed so badly at rtw_debug=2. Change some debugging printfs to RTW_DPRINTFs. E.g., print the "RF programming method" only if debugging is enabled.
2004-12-21 02:05:41 +03:00
rtw_pwrstate0(struct rtw_softc *sc, enum rtw_pwrstate power, int before_rf,
int digphy)
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
{
struct rtw_regs *regs = &sc->sc_regs;
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
rtw_set_access(regs, RTW_ACCESS_ANAPARM);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
Realtek sample sources are not consistent about the "Digital PHY" bit in the SROM. It seems as if it is set to 1 when the PHY is *analog*, not *digital*. Fix my sources. In rtw_intr_rx, use units of 500kb/s instead of 100kb/s for rate, to be consistent with net80211's expectations. Polish up some debugging ugly messages. Dump raw 802.11 packets if IFF_DEBUG|IFF_LINK2 and RTW_DEBUG is defined. Polish power-state (on/sleep/off) handling. Especially improve support for RFMD (totally untested) and Maxim. For Philips, take the Digital PHY property into account. Call the net80211 watchdog function from rtw_watchdog, so that we scan again if auth/assoc fails. Be a little more cautious about writing register[RTW_TPPOLL], since other drivers are.... Don't frob the high/low-priority queues right now, since I don't use them. Add rtw_join_bss which programs the card with the BSSID and other properties of a BSS. Use it on state transitions. Factor out rtw_set_nettype. Make rtw_recv_beacon call ieee80211_recv_mgmt instead of dropping beacons on the floor! TBD IBSS merges. Change some rtw_debug=2 printfs to rtw_debug=3 (RTW_DPRINTF3) printfs so the console doesn't get spammed so badly at rtw_debug=2. Change some debugging printfs to RTW_DPRINTFs. E.g., print the "RF programming method" only if debugging is enabled.
2004-12-21 02:05:41 +03:00
(*sc->sc_pwrstate_cb)(regs, power, before_rf, digphy);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
rtw_set_access(regs, RTW_ACCESS_NONE);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
return;
}
static int
rtw_pwrstate(struct rtw_softc *sc, enum rtw_pwrstate power)
{
int rc;
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
RTW_DPRINTF(RTW_DEBUG_PWR,
("%s: %s->%s\n", __func__,
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
rtw_pwrstate_string(sc->sc_pwrstate), rtw_pwrstate_string(power)));
if (sc->sc_pwrstate == power)
return 0;
Realtek sample sources are not consistent about the "Digital PHY" bit in the SROM. It seems as if it is set to 1 when the PHY is *analog*, not *digital*. Fix my sources. In rtw_intr_rx, use units of 500kb/s instead of 100kb/s for rate, to be consistent with net80211's expectations. Polish up some debugging ugly messages. Dump raw 802.11 packets if IFF_DEBUG|IFF_LINK2 and RTW_DEBUG is defined. Polish power-state (on/sleep/off) handling. Especially improve support for RFMD (totally untested) and Maxim. For Philips, take the Digital PHY property into account. Call the net80211 watchdog function from rtw_watchdog, so that we scan again if auth/assoc fails. Be a little more cautious about writing register[RTW_TPPOLL], since other drivers are.... Don't frob the high/low-priority queues right now, since I don't use them. Add rtw_join_bss which programs the card with the BSSID and other properties of a BSS. Use it on state transitions. Factor out rtw_set_nettype. Make rtw_recv_beacon call ieee80211_recv_mgmt instead of dropping beacons on the floor! TBD IBSS merges. Change some rtw_debug=2 printfs to rtw_debug=3 (RTW_DPRINTF3) printfs so the console doesn't get spammed so badly at rtw_debug=2. Change some debugging printfs to RTW_DPRINTFs. E.g., print the "RF programming method" only if debugging is enabled.
2004-12-21 02:05:41 +03:00
rtw_pwrstate0(sc, power, 1, sc->sc_flags & RTW_F_DIGPHY);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
rc = rtw_rf_pwrstate(sc->sc_rf, power);
Realtek sample sources are not consistent about the "Digital PHY" bit in the SROM. It seems as if it is set to 1 when the PHY is *analog*, not *digital*. Fix my sources. In rtw_intr_rx, use units of 500kb/s instead of 100kb/s for rate, to be consistent with net80211's expectations. Polish up some debugging ugly messages. Dump raw 802.11 packets if IFF_DEBUG|IFF_LINK2 and RTW_DEBUG is defined. Polish power-state (on/sleep/off) handling. Especially improve support for RFMD (totally untested) and Maxim. For Philips, take the Digital PHY property into account. Call the net80211 watchdog function from rtw_watchdog, so that we scan again if auth/assoc fails. Be a little more cautious about writing register[RTW_TPPOLL], since other drivers are.... Don't frob the high/low-priority queues right now, since I don't use them. Add rtw_join_bss which programs the card with the BSSID and other properties of a BSS. Use it on state transitions. Factor out rtw_set_nettype. Make rtw_recv_beacon call ieee80211_recv_mgmt instead of dropping beacons on the floor! TBD IBSS merges. Change some rtw_debug=2 printfs to rtw_debug=3 (RTW_DPRINTF3) printfs so the console doesn't get spammed so badly at rtw_debug=2. Change some debugging printfs to RTW_DPRINTFs. E.g., print the "RF programming method" only if debugging is enabled.
2004-12-21 02:05:41 +03:00
rtw_pwrstate0(sc, power, 0, sc->sc_flags & RTW_F_DIGPHY);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
switch (power) {
case RTW_ON:
At last, I have rtw w/ Philips RF receiving packets. I added some sysctls to aid debugging: * hw.rtw.debug -- enable debugging * hw.rtw.flush_rfio -- Linux voodoo: possibly makes the MAC "flush" bits down the serial bus to the RF * hw.rtw.host_rfio: force the host to bang bits to the RF, instead of the MAC banging bits * hw.rtw.rfio_delay: after telling the MAC to bang bits to the RF front-end, delay rfio_delay microseconds. * hw.rtw.rfprog_fallback: there is this notion of the "RF programming method." I believe the choice influences the polarity/timing of the serial bus used to program the RF front-end. I know the correct choice for Intersil/RFMD/Philips front-ends, only. For all other front-ends, I "fallback" to rfprog_fallback. Make rtw_txdac_enable take an rtw_softc argument. I will probably revert this change. Add some Linux voodoo to rtw_continuous_tx_enable. I will probably revert this change. Important: add rtw_set_rfprog, which sets the correct RF programming method. This change and the following change are probably responsible for making the Philips RF work. Important: RTW_CONFIG1 is an 8-bit register, treat it that way! Important: RTW_BRSR is 16-bit, RTW_CRCOUNT, RTW_PHYDELAY, and RTW_MSR are 8-bit: treat them that way! Vastly simplify rtw_resume_ticks. Note to self: set the LED state to match the power state. Hedge against the possibility that RTW_MSR is protected as RTW_CONFIG[0123] are, meanwhile reworking that section of rtw_init a little. Add sc_anaparm, which isn't used, yet....
2004-12-13 03:48:02 +03:00
/* TBD set LEDs */
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
break;
case RTW_SLEEP:
/* TBD */
break;
case RTW_OFF:
/* TBD */
break;
}
if (rc == 0)
sc->sc_pwrstate = power;
else
sc->sc_pwrstate = RTW_OFF;
return rc;
}
static int
rtw_tune(struct rtw_softc *sc)
{
struct ieee80211com *ic = &sc->sc_ic;
Set the channel frequency and flags in the radiotap headers in rtw_tune instead of in rtw_start, rtw_intr_rx.
2006-03-28 02:03:36 +04:00
struct rtw_tx_radiotap_header *rt = &sc->sc_txtap;
struct rtw_rx_radiotap_header *rr = &sc->sc_rxtap;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
u_int chan;
int rc;
int antdiv = sc->sc_flags & RTW_F_ANTDIV,
dflantb = sc->sc_flags & RTW_F_DFLANTB;
Adapt drivers to the new net80211(9). Most of this is from dyoung@. Thanks!
2005-11-18 19:53:56 +03:00
chan = ieee80211_chan2ieee(ic, ic->ic_curchan);
Remove shutdown hooks. Use KASSERT().
2007-12-20 21:47:23 +03:00
KASSERT(chan != IEEE80211_CHAN_ANY);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
Set the channel frequency and flags in the radiotap headers in rtw_tune instead of in rtw_start, rtw_intr_rx.
2006-03-28 02:03:36 +04:00
rt->rt_chan_freq = htole16(ic->ic_curchan->ic_freq);
rt->rt_chan_flags = htole16(ic->ic_curchan->ic_flags);
rr->rr_chan_freq = htole16(ic->ic_curchan->ic_freq);
rr->rr_chan_flags = htole16(ic->ic_curchan->ic_flags);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
if (chan == sc->sc_cur_chan) {
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
RTW_DPRINTF(RTW_DEBUG_TUNE,
nuke trailing whitespace
2005-02-27 03:26:58 +03:00
("%s: already tuned chan #%d\n", __func__, chan));
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
return 0;
}
rtw_suspend_ticks(sc);
I have made a small change that suppresses "recovery" from (spurious?) tx timeouts on the beacon queue. The driver's recovery process badly disrupts the MAC's receiver. This needs a closer look, later. Every other fix is by Charles M. Hannum. Charles writes: "With these, the device is quite stable under load, does not experience xmit timeouts, no longer mysteriously freezes, and does not spew a lot of garbage. It does, however, continue to get some "tx fifo underflow"s. global, SYSCTL_SETUP(): a) Removed unused sysctl variables. rtw_io_enable(): b) Change rtw_io_enable to take a softc, and make it call rtw_txring_fixup(). c) Enable the PCI multiple read/write bit. d) Always reset rdb_next to 0 when starting the receive engine. rtw_intr_rx(): e) When processing receive packets, store the next pointer in rdb_next and read it back on the next round, to fix a race condition vs. rtw_tune() during AP discovery and association. f) Remove the special "skip to 0th descriptor" code, because it is no longer needed. rtw_collect_txring(): g) In the "force" case, stop when we catch up to the hardware xmit pointer. rtw_intr_beacon(): h) Resetting tdb_next here looked fishy to me, in light of changes in rtw_collect_txring(). XXX However, this is the one part I haven't tested. rtw_txring_fixup(): i) Add function and line arguments, so we can print out information about the caller. j) In addition to reading the hardware pointer, skip over any additional packets with OWN set -- we may have queued more. rtw_rxring_fixup(): k) Remove this, because it's completely wrong. rtw_intr_ioerror(): l) Don't handle receive overruns here; handle them in the normal receive path. Also changed RTW_INTR_RX and RTW_INTR_IOERROR accordingly. m) Do not fiddle with CR; the transmit engine is still running just fine. n) Do not fiddle with IMR; it was unnecessary. o) Don't touch the receive engine at all; not necessary now that we're leaving CR alone. p) Remove the rtw_do_chip_reset case; it was dangerous and not actually used. rtw_intr(): q) Don't mask off ISR bits when we pass down the ISR value; it's not necessary, and sometimes the other bits useful in debugging output. rtw_seg_too_short(): r) Don't spew here; the caller will do it if necessary. (Stops spewage in an expected case.) rtw_rxdesc_blk: s) Minor change to make the elements have the same order as txdesc_blk. general: t) Added a bunch of debugging output, #ifdef RTW_DIAG, that was useful while investigating the ring management problems. Also moved rtw_txring_fixup() into RTW_DIAG, as I believe the transmit pointer is always kept in sync now, and the function has potential race conditions.
2007-01-06 11:27:53 +03:00
rtw_io_enable(sc, RTW_CR_RE | RTW_CR_TE, 0);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
/* TBD wait for Tx to complete */
Use device_t and its accessors throughout. Use aprint_*_dev(). Improve PMF-ability. Add a 'flags' argument to suspend/resume handlers and callers such as pmf_system_suspend(). Define a flag, PMF_F_SELF, which indicates to PMF that a device is suspending/resuming itself. Add helper routines, pmf_device_suspend_self(dev) and pmf_device_resume_self(dev), that call pmf_device_suspend(dev, PMF_F_SELF) and pmf_device_resume(dev, PMF_F_SELF), respectively. Use PMF_F_SELF to suspend/resume self in ath(4), audio(4), rtw(4), and sip(4). In ath(4) and in rtw(4), replace the icky sc_enable/sc_disable callbacks, provided by the bus front-end, with self-suspension/resumption. Also, clean up the bus front-ends. Make sure that the interrupt handler is disestablished during suspension. Get rid of driver-private flags (e.g., RTW_F_ENABLED, ath_softc->sc_invalid); use device_is_active()/device_has_power() calls, instead. In the network-class suspend handler, call if_stop(, 0) instead of if_stop(, 1), because the latter is superfluous (bus- and driver-suspension hooks will 'disable' the NIC), and it may cause recursion. In the network-class resume handler, prevent infinite recursion through if_init() by getting out early if we are self-suspending (PMF_F_SELF). rtw(4) improvements: Destroy rtw(4) callouts when we detach it. Make rtw at pci detachable. Print some more information with the "rx frame too long" warning. Remove activate() methods: Get rid of rtw_activate() and ath_activate(). The device activate() methods are not good for much these days. Make ath at cardbus resume with crypto functions intact: Introduce a boolean device property, "pmf-powerdown". If pmf-powerdown is present and false, it indicates that a bus back-end should not remove power from a device. Honor this property in cardbus_child_suspend(). Set this property to 'false' in ath_attach(), since removing power from an ath at cardbus seems to lobotomize the WPA crypto engine. XXX Should the pmf-powerdown property propagate toward the root of the device tree? Miscellaneous ath(4) changes: Warn if ath(4) tries to write crypto keys to suspended hardware. Reduce differences between FreeBSD and NetBSD in ath(4) multicast filter setup. Make ath_printrxbuf() print an rx descriptor's status & key index, to help debug crypto errors. Shorten a staircase in ath_ioctl(). Don't check for ieee80211_ioctl() return code ERESTART, it never happens.
2008-03-12 21:02:21 +03:00
KASSERT(device_has_power(sc->sc_dev));
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
if ((rc = rtw_phy_init(&sc->sc_regs, sc->sc_rf,
Adapt drivers to the new net80211(9). Most of this is from dyoung@. Thanks!
2005-11-18 19:53:56 +03:00
rtw_chan2txpower(&sc->sc_srom, ic, ic->ic_curchan), sc->sc_csthr,
ic->ic_curchan->ic_freq, antdiv, dflantb, RTW_ON)) != 0) {
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
/* XXX condition on powersaving */
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
aprint_error_dev(sc->sc_dev, "phy init failed\n");
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
}
sc->sc_cur_chan = chan;
I have made a small change that suppresses "recovery" from (spurious?) tx timeouts on the beacon queue. The driver's recovery process badly disrupts the MAC's receiver. This needs a closer look, later. Every other fix is by Charles M. Hannum. Charles writes: "With these, the device is quite stable under load, does not experience xmit timeouts, no longer mysteriously freezes, and does not spew a lot of garbage. It does, however, continue to get some "tx fifo underflow"s. global, SYSCTL_SETUP(): a) Removed unused sysctl variables. rtw_io_enable(): b) Change rtw_io_enable to take a softc, and make it call rtw_txring_fixup(). c) Enable the PCI multiple read/write bit. d) Always reset rdb_next to 0 when starting the receive engine. rtw_intr_rx(): e) When processing receive packets, store the next pointer in rdb_next and read it back on the next round, to fix a race condition vs. rtw_tune() during AP discovery and association. f) Remove the special "skip to 0th descriptor" code, because it is no longer needed. rtw_collect_txring(): g) In the "force" case, stop when we catch up to the hardware xmit pointer. rtw_intr_beacon(): h) Resetting tdb_next here looked fishy to me, in light of changes in rtw_collect_txring(). XXX However, this is the one part I haven't tested. rtw_txring_fixup(): i) Add function and line arguments, so we can print out information about the caller. j) In addition to reading the hardware pointer, skip over any additional packets with OWN set -- we may have queued more. rtw_rxring_fixup(): k) Remove this, because it's completely wrong. rtw_intr_ioerror(): l) Don't handle receive overruns here; handle them in the normal receive path. Also changed RTW_INTR_RX and RTW_INTR_IOERROR accordingly. m) Do not fiddle with CR; the transmit engine is still running just fine. n) Do not fiddle with IMR; it was unnecessary. o) Don't touch the receive engine at all; not necessary now that we're leaving CR alone. p) Remove the rtw_do_chip_reset case; it was dangerous and not actually used. rtw_intr(): q) Don't mask off ISR bits when we pass down the ISR value; it's not necessary, and sometimes the other bits useful in debugging output. rtw_seg_too_short(): r) Don't spew here; the caller will do it if necessary. (Stops spewage in an expected case.) rtw_rxdesc_blk: s) Minor change to make the elements have the same order as txdesc_blk. general: t) Added a bunch of debugging output, #ifdef RTW_DIAG, that was useful while investigating the ring management problems. Also moved rtw_txring_fixup() into RTW_DIAG, as I believe the transmit pointer is always kept in sync now, and the function has potential race conditions.
2007-01-06 11:27:53 +03:00
rtw_io_enable(sc, RTW_CR_RE | RTW_CR_TE, 1);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
rtw_resume_ticks(sc);
return rc;
}
Use device_t and its accessors throughout. Use aprint_*_dev(). Improve PMF-ability. Add a 'flags' argument to suspend/resume handlers and callers such as pmf_system_suspend(). Define a flag, PMF_F_SELF, which indicates to PMF that a device is suspending/resuming itself. Add helper routines, pmf_device_suspend_self(dev) and pmf_device_resume_self(dev), that call pmf_device_suspend(dev, PMF_F_SELF) and pmf_device_resume(dev, PMF_F_SELF), respectively. Use PMF_F_SELF to suspend/resume self in ath(4), audio(4), rtw(4), and sip(4). In ath(4) and in rtw(4), replace the icky sc_enable/sc_disable callbacks, provided by the bus front-end, with self-suspension/resumption. Also, clean up the bus front-ends. Make sure that the interrupt handler is disestablished during suspension. Get rid of driver-private flags (e.g., RTW_F_ENABLED, ath_softc->sc_invalid); use device_is_active()/device_has_power() calls, instead. In the network-class suspend handler, call if_stop(, 0) instead of if_stop(, 1), because the latter is superfluous (bus- and driver-suspension hooks will 'disable' the NIC), and it may cause recursion. In the network-class resume handler, prevent infinite recursion through if_init() by getting out early if we are self-suspending (PMF_F_SELF). rtw(4) improvements: Destroy rtw(4) callouts when we detach it. Make rtw at pci detachable. Print some more information with the "rx frame too long" warning. Remove activate() methods: Get rid of rtw_activate() and ath_activate(). The device activate() methods are not good for much these days. Make ath at cardbus resume with crypto functions intact: Introduce a boolean device property, "pmf-powerdown". If pmf-powerdown is present and false, it indicates that a bus back-end should not remove power from a device. Honor this property in cardbus_child_suspend(). Set this property to 'false' in ath_attach(), since removing power from an ath at cardbus seems to lobotomize the WPA crypto engine. XXX Should the pmf-powerdown property propagate toward the root of the device tree? Miscellaneous ath(4) changes: Warn if ath(4) tries to write crypto keys to suspended hardware. Reduce differences between FreeBSD and NetBSD in ath(4) multicast filter setup. Make ath_printrxbuf() print an rx descriptor's status & key index, to help debug crypto errors. Shorten a staircase in ath_ioctl(). Don't check for ieee80211_ioctl() return code ERESTART, it never happens.
2008-03-12 21:02:21 +03:00
bool
A pointer typedef entails trading too much flexibility to declare const and non-const types, and the kernel uses both const and non-const PMF qualifiers and device suspensors, so change the pmf_qual_t and device_suspensor_t typedefs from "pointers to const" to non-pointer, non-const types.
2010-02-25 01:37:54 +03:00
rtw_suspend(device_t self, const pmf_qual_t *qual)
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
{
int rc;
Use device_t and its accessors throughout. Use aprint_*_dev(). Improve PMF-ability. Add a 'flags' argument to suspend/resume handlers and callers such as pmf_system_suspend(). Define a flag, PMF_F_SELF, which indicates to PMF that a device is suspending/resuming itself. Add helper routines, pmf_device_suspend_self(dev) and pmf_device_resume_self(dev), that call pmf_device_suspend(dev, PMF_F_SELF) and pmf_device_resume(dev, PMF_F_SELF), respectively. Use PMF_F_SELF to suspend/resume self in ath(4), audio(4), rtw(4), and sip(4). In ath(4) and in rtw(4), replace the icky sc_enable/sc_disable callbacks, provided by the bus front-end, with self-suspension/resumption. Also, clean up the bus front-ends. Make sure that the interrupt handler is disestablished during suspension. Get rid of driver-private flags (e.g., RTW_F_ENABLED, ath_softc->sc_invalid); use device_is_active()/device_has_power() calls, instead. In the network-class suspend handler, call if_stop(, 0) instead of if_stop(, 1), because the latter is superfluous (bus- and driver-suspension hooks will 'disable' the NIC), and it may cause recursion. In the network-class resume handler, prevent infinite recursion through if_init() by getting out early if we are self-suspending (PMF_F_SELF). rtw(4) improvements: Destroy rtw(4) callouts when we detach it. Make rtw at pci detachable. Print some more information with the "rx frame too long" warning. Remove activate() methods: Get rid of rtw_activate() and ath_activate(). The device activate() methods are not good for much these days. Make ath at cardbus resume with crypto functions intact: Introduce a boolean device property, "pmf-powerdown". If pmf-powerdown is present and false, it indicates that a bus back-end should not remove power from a device. Honor this property in cardbus_child_suspend(). Set this property to 'false' in ath_attach(), since removing power from an ath at cardbus seems to lobotomize the WPA crypto engine. XXX Should the pmf-powerdown property propagate toward the root of the device tree? Miscellaneous ath(4) changes: Warn if ath(4) tries to write crypto keys to suspended hardware. Reduce differences between FreeBSD and NetBSD in ath(4) multicast filter setup. Make ath_printrxbuf() print an rx descriptor's status & key index, to help debug crypto errors. Shorten a staircase in ath_ioctl(). Don't check for ieee80211_ioctl() return code ERESTART, it never happens.
2008-03-12 21:02:21 +03:00
struct rtw_softc *sc = device_private(self);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
Use device_t and its accessors throughout. Use aprint_*_dev(). Improve PMF-ability. Add a 'flags' argument to suspend/resume handlers and callers such as pmf_system_suspend(). Define a flag, PMF_F_SELF, which indicates to PMF that a device is suspending/resuming itself. Add helper routines, pmf_device_suspend_self(dev) and pmf_device_resume_self(dev), that call pmf_device_suspend(dev, PMF_F_SELF) and pmf_device_resume(dev, PMF_F_SELF), respectively. Use PMF_F_SELF to suspend/resume self in ath(4), audio(4), rtw(4), and sip(4). In ath(4) and in rtw(4), replace the icky sc_enable/sc_disable callbacks, provided by the bus front-end, with self-suspension/resumption. Also, clean up the bus front-ends. Make sure that the interrupt handler is disestablished during suspension. Get rid of driver-private flags (e.g., RTW_F_ENABLED, ath_softc->sc_invalid); use device_is_active()/device_has_power() calls, instead. In the network-class suspend handler, call if_stop(, 0) instead of if_stop(, 1), because the latter is superfluous (bus- and driver-suspension hooks will 'disable' the NIC), and it may cause recursion. In the network-class resume handler, prevent infinite recursion through if_init() by getting out early if we are self-suspending (PMF_F_SELF). rtw(4) improvements: Destroy rtw(4) callouts when we detach it. Make rtw at pci detachable. Print some more information with the "rx frame too long" warning. Remove activate() methods: Get rid of rtw_activate() and ath_activate(). The device activate() methods are not good for much these days. Make ath at cardbus resume with crypto functions intact: Introduce a boolean device property, "pmf-powerdown". If pmf-powerdown is present and false, it indicates that a bus back-end should not remove power from a device. Honor this property in cardbus_child_suspend(). Set this property to 'false' in ath_attach(), since removing power from an ath at cardbus seems to lobotomize the WPA crypto engine. XXX Should the pmf-powerdown property propagate toward the root of the device tree? Miscellaneous ath(4) changes: Warn if ath(4) tries to write crypto keys to suspended hardware. Reduce differences between FreeBSD and NetBSD in ath(4) multicast filter setup. Make ath_printrxbuf() print an rx descriptor's status & key index, to help debug crypto errors. Shorten a staircase in ath_ioctl(). Don't check for ieee80211_ioctl() return code ERESTART, it never happens.
2008-03-12 21:02:21 +03:00
sc->sc_flags &= ~RTW_F_DK_VALID;
if (!device_has_power(self))
return false;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
/* turn off PHY */
Use device_t and its accessors throughout. Use aprint_*_dev(). Improve PMF-ability. Add a 'flags' argument to suspend/resume handlers and callers such as pmf_system_suspend(). Define a flag, PMF_F_SELF, which indicates to PMF that a device is suspending/resuming itself. Add helper routines, pmf_device_suspend_self(dev) and pmf_device_resume_self(dev), that call pmf_device_suspend(dev, PMF_F_SELF) and pmf_device_resume(dev, PMF_F_SELF), respectively. Use PMF_F_SELF to suspend/resume self in ath(4), audio(4), rtw(4), and sip(4). In ath(4) and in rtw(4), replace the icky sc_enable/sc_disable callbacks, provided by the bus front-end, with self-suspension/resumption. Also, clean up the bus front-ends. Make sure that the interrupt handler is disestablished during suspension. Get rid of driver-private flags (e.g., RTW_F_ENABLED, ath_softc->sc_invalid); use device_is_active()/device_has_power() calls, instead. In the network-class suspend handler, call if_stop(, 0) instead of if_stop(, 1), because the latter is superfluous (bus- and driver-suspension hooks will 'disable' the NIC), and it may cause recursion. In the network-class resume handler, prevent infinite recursion through if_init() by getting out early if we are self-suspending (PMF_F_SELF). rtw(4) improvements: Destroy rtw(4) callouts when we detach it. Make rtw at pci detachable. Print some more information with the "rx frame too long" warning. Remove activate() methods: Get rid of rtw_activate() and ath_activate(). The device activate() methods are not good for much these days. Make ath at cardbus resume with crypto functions intact: Introduce a boolean device property, "pmf-powerdown". If pmf-powerdown is present and false, it indicates that a bus back-end should not remove power from a device. Honor this property in cardbus_child_suspend(). Set this property to 'false' in ath_attach(), since removing power from an ath at cardbus seems to lobotomize the WPA crypto engine. XXX Should the pmf-powerdown property propagate toward the root of the device tree? Miscellaneous ath(4) changes: Warn if ath(4) tries to write crypto keys to suspended hardware. Reduce differences between FreeBSD and NetBSD in ath(4) multicast filter setup. Make ath_printrxbuf() print an rx descriptor's status & key index, to help debug crypto errors. Shorten a staircase in ath_ioctl(). Don't check for ieee80211_ioctl() return code ERESTART, it never happens.
2008-03-12 21:02:21 +03:00
if ((rc = rtw_pwrstate(sc, RTW_OFF)) != 0) {
aprint_error_dev(self, "failed to turn off PHY (%d)\n", rc);
return false;
In rtw_detach, flag the rtw as "invalid" (RTW_F_INVALID). Do not try to power-down the RF section of an invalid rtw. The radiotap signal quality field is 16, not 8, bits wide. Convert the byte order.
2004-12-29 22:41:04 +03:00
}
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
Use device_t and its accessors throughout. Use aprint_*_dev(). Improve PMF-ability. Add a 'flags' argument to suspend/resume handlers and callers such as pmf_system_suspend(). Define a flag, PMF_F_SELF, which indicates to PMF that a device is suspending/resuming itself. Add helper routines, pmf_device_suspend_self(dev) and pmf_device_resume_self(dev), that call pmf_device_suspend(dev, PMF_F_SELF) and pmf_device_resume(dev, PMF_F_SELF), respectively. Use PMF_F_SELF to suspend/resume self in ath(4), audio(4), rtw(4), and sip(4). In ath(4) and in rtw(4), replace the icky sc_enable/sc_disable callbacks, provided by the bus front-end, with self-suspension/resumption. Also, clean up the bus front-ends. Make sure that the interrupt handler is disestablished during suspension. Get rid of driver-private flags (e.g., RTW_F_ENABLED, ath_softc->sc_invalid); use device_is_active()/device_has_power() calls, instead. In the network-class suspend handler, call if_stop(, 0) instead of if_stop(, 1), because the latter is superfluous (bus- and driver-suspension hooks will 'disable' the NIC), and it may cause recursion. In the network-class resume handler, prevent infinite recursion through if_init() by getting out early if we are self-suspending (PMF_F_SELF). rtw(4) improvements: Destroy rtw(4) callouts when we detach it. Make rtw at pci detachable. Print some more information with the "rx frame too long" warning. Remove activate() methods: Get rid of rtw_activate() and ath_activate(). The device activate() methods are not good for much these days. Make ath at cardbus resume with crypto functions intact: Introduce a boolean device property, "pmf-powerdown". If pmf-powerdown is present and false, it indicates that a bus back-end should not remove power from a device. Honor this property in cardbus_child_suspend(). Set this property to 'false' in ath_attach(), since removing power from an ath at cardbus seems to lobotomize the WPA crypto engine. XXX Should the pmf-powerdown property propagate toward the root of the device tree? Miscellaneous ath(4) changes: Warn if ath(4) tries to write crypto keys to suspended hardware. Reduce differences between FreeBSD and NetBSD in ath(4) multicast filter setup. Make ath_printrxbuf() print an rx descriptor's status & key index, to help debug crypto errors. Shorten a staircase in ath_ioctl(). Don't check for ieee80211_ioctl() return code ERESTART, it never happens.
2008-03-12 21:02:21 +03:00
rtw_disable_interrupts(&sc->sc_regs);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
Use device_t and its accessors throughout. Use aprint_*_dev(). Improve PMF-ability. Add a 'flags' argument to suspend/resume handlers and callers such as pmf_system_suspend(). Define a flag, PMF_F_SELF, which indicates to PMF that a device is suspending/resuming itself. Add helper routines, pmf_device_suspend_self(dev) and pmf_device_resume_self(dev), that call pmf_device_suspend(dev, PMF_F_SELF) and pmf_device_resume(dev, PMF_F_SELF), respectively. Use PMF_F_SELF to suspend/resume self in ath(4), audio(4), rtw(4), and sip(4). In ath(4) and in rtw(4), replace the icky sc_enable/sc_disable callbacks, provided by the bus front-end, with self-suspension/resumption. Also, clean up the bus front-ends. Make sure that the interrupt handler is disestablished during suspension. Get rid of driver-private flags (e.g., RTW_F_ENABLED, ath_softc->sc_invalid); use device_is_active()/device_has_power() calls, instead. In the network-class suspend handler, call if_stop(, 0) instead of if_stop(, 1), because the latter is superfluous (bus- and driver-suspension hooks will 'disable' the NIC), and it may cause recursion. In the network-class resume handler, prevent infinite recursion through if_init() by getting out early if we are self-suspending (PMF_F_SELF). rtw(4) improvements: Destroy rtw(4) callouts when we detach it. Make rtw at pci detachable. Print some more information with the "rx frame too long" warning. Remove activate() methods: Get rid of rtw_activate() and ath_activate(). The device activate() methods are not good for much these days. Make ath at cardbus resume with crypto functions intact: Introduce a boolean device property, "pmf-powerdown". If pmf-powerdown is present and false, it indicates that a bus back-end should not remove power from a device. Honor this property in cardbus_child_suspend(). Set this property to 'false' in ath_attach(), since removing power from an ath at cardbus seems to lobotomize the WPA crypto engine. XXX Should the pmf-powerdown property propagate toward the root of the device tree? Miscellaneous ath(4) changes: Warn if ath(4) tries to write crypto keys to suspended hardware. Reduce differences between FreeBSD and NetBSD in ath(4) multicast filter setup. Make ath_printrxbuf() print an rx descriptor's status & key index, to help debug crypto errors. Shorten a staircase in ath_ioctl(). Don't check for ieee80211_ioctl() return code ERESTART, it never happens.
2008-03-12 21:02:21 +03:00
return true;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
}
Use device_t and its accessors throughout. Use aprint_*_dev(). Improve PMF-ability. Add a 'flags' argument to suspend/resume handlers and callers such as pmf_system_suspend(). Define a flag, PMF_F_SELF, which indicates to PMF that a device is suspending/resuming itself. Add helper routines, pmf_device_suspend_self(dev) and pmf_device_resume_self(dev), that call pmf_device_suspend(dev, PMF_F_SELF) and pmf_device_resume(dev, PMF_F_SELF), respectively. Use PMF_F_SELF to suspend/resume self in ath(4), audio(4), rtw(4), and sip(4). In ath(4) and in rtw(4), replace the icky sc_enable/sc_disable callbacks, provided by the bus front-end, with self-suspension/resumption. Also, clean up the bus front-ends. Make sure that the interrupt handler is disestablished during suspension. Get rid of driver-private flags (e.g., RTW_F_ENABLED, ath_softc->sc_invalid); use device_is_active()/device_has_power() calls, instead. In the network-class suspend handler, call if_stop(, 0) instead of if_stop(, 1), because the latter is superfluous (bus- and driver-suspension hooks will 'disable' the NIC), and it may cause recursion. In the network-class resume handler, prevent infinite recursion through if_init() by getting out early if we are self-suspending (PMF_F_SELF). rtw(4) improvements: Destroy rtw(4) callouts when we detach it. Make rtw at pci detachable. Print some more information with the "rx frame too long" warning. Remove activate() methods: Get rid of rtw_activate() and ath_activate(). The device activate() methods are not good for much these days. Make ath at cardbus resume with crypto functions intact: Introduce a boolean device property, "pmf-powerdown". If pmf-powerdown is present and false, it indicates that a bus back-end should not remove power from a device. Honor this property in cardbus_child_suspend(). Set this property to 'false' in ath_attach(), since removing power from an ath at cardbus seems to lobotomize the WPA crypto engine. XXX Should the pmf-powerdown property propagate toward the root of the device tree? Miscellaneous ath(4) changes: Warn if ath(4) tries to write crypto keys to suspended hardware. Reduce differences between FreeBSD and NetBSD in ath(4) multicast filter setup. Make ath_printrxbuf() print an rx descriptor's status & key index, to help debug crypto errors. Shorten a staircase in ath_ioctl(). Don't check for ieee80211_ioctl() return code ERESTART, it never happens.
2008-03-12 21:02:21 +03:00
bool
A pointer typedef entails trading too much flexibility to declare const and non-const types, and the kernel uses both const and non-const PMF qualifiers and device suspensors, so change the pmf_qual_t and device_suspensor_t typedefs from "pointers to const" to non-pointer, non-const types.
2010-02-25 01:37:54 +03:00
rtw_resume(device_t self, const pmf_qual_t *qual)
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
{
Use device_t and its accessors throughout. Use aprint_*_dev(). Improve PMF-ability. Add a 'flags' argument to suspend/resume handlers and callers such as pmf_system_suspend(). Define a flag, PMF_F_SELF, which indicates to PMF that a device is suspending/resuming itself. Add helper routines, pmf_device_suspend_self(dev) and pmf_device_resume_self(dev), that call pmf_device_suspend(dev, PMF_F_SELF) and pmf_device_resume(dev, PMF_F_SELF), respectively. Use PMF_F_SELF to suspend/resume self in ath(4), audio(4), rtw(4), and sip(4). In ath(4) and in rtw(4), replace the icky sc_enable/sc_disable callbacks, provided by the bus front-end, with self-suspension/resumption. Also, clean up the bus front-ends. Make sure that the interrupt handler is disestablished during suspension. Get rid of driver-private flags (e.g., RTW_F_ENABLED, ath_softc->sc_invalid); use device_is_active()/device_has_power() calls, instead. In the network-class suspend handler, call if_stop(, 0) instead of if_stop(, 1), because the latter is superfluous (bus- and driver-suspension hooks will 'disable' the NIC), and it may cause recursion. In the network-class resume handler, prevent infinite recursion through if_init() by getting out early if we are self-suspending (PMF_F_SELF). rtw(4) improvements: Destroy rtw(4) callouts when we detach it. Make rtw at pci detachable. Print some more information with the "rx frame too long" warning. Remove activate() methods: Get rid of rtw_activate() and ath_activate(). The device activate() methods are not good for much these days. Make ath at cardbus resume with crypto functions intact: Introduce a boolean device property, "pmf-powerdown". If pmf-powerdown is present and false, it indicates that a bus back-end should not remove power from a device. Honor this property in cardbus_child_suspend(). Set this property to 'false' in ath_attach(), since removing power from an ath at cardbus seems to lobotomize the WPA crypto engine. XXX Should the pmf-powerdown property propagate toward the root of the device tree? Miscellaneous ath(4) changes: Warn if ath(4) tries to write crypto keys to suspended hardware. Reduce differences between FreeBSD and NetBSD in ath(4) multicast filter setup. Make ath_printrxbuf() print an rx descriptor's status & key index, to help debug crypto errors. Shorten a staircase in ath_ioctl(). Don't check for ieee80211_ioctl() return code ERESTART, it never happens.
2008-03-12 21:02:21 +03:00
struct rtw_softc *sc = device_private(self);
/* Power may have been removed, resetting WEP keys.
*/
sc->sc_flags &= ~RTW_F_DK_VALID;
rtw_enable_interrupts(sc);
return true;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
}
static void
rtw_transmit_config(struct rtw_regs *regs)
{
Cosmetic: change u_int{8,16,32,64}_t to C99-style uint{8,16,32,64}_t, which is preferred.
2005-01-02 07:23:03 +03:00
uint32_t tcr;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
tcr = RTW_READ(regs, RTW_TCR);
Realtek sample sources are not consistent about the "Digital PHY" bit in the SROM. It seems as if it is set to 1 when the PHY is *analog*, not *digital*. Fix my sources. In rtw_intr_rx, use units of 500kb/s instead of 100kb/s for rate, to be consistent with net80211's expectations. Polish up some debugging ugly messages. Dump raw 802.11 packets if IFF_DEBUG|IFF_LINK2 and RTW_DEBUG is defined. Polish power-state (on/sleep/off) handling. Especially improve support for RFMD (totally untested) and Maxim. For Philips, take the Digital PHY property into account. Call the net80211 watchdog function from rtw_watchdog, so that we scan again if auth/assoc fails. Be a little more cautious about writing register[RTW_TPPOLL], since other drivers are.... Don't frob the high/low-priority queues right now, since I don't use them. Add rtw_join_bss which programs the card with the BSSID and other properties of a BSS. Use it on state transitions. Factor out rtw_set_nettype. Make rtw_recv_beacon call ieee80211_recv_mgmt instead of dropping beacons on the floor! TBD IBSS merges. Change some rtw_debug=2 printfs to rtw_debug=3 (RTW_DPRINTF3) printfs so the console doesn't get spammed so badly at rtw_debug=2. Change some debugging printfs to RTW_DPRINTFs. E.g., print the "RF programming method" only if debugging is enabled.
2004-12-21 02:05:41 +03:00
tcr |= RTW_TCR_CWMIN;
tcr &= ~RTW_TCR_MXDMA_MASK;
tcr |= RTW_TCR_MXDMA_256;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
tcr |= RTW_TCR_SAT; /* send ACK as fast as possible */
tcr &= ~RTW_TCR_LBK_MASK;
tcr |= RTW_TCR_LBK_NORMAL; /* normal operating mode */
/* set short/long retry limits */
tcr &= ~(RTW_TCR_SRL_MASK|RTW_TCR_LRL_MASK);
Per discussion on tech-kern and tech-userlevel, move the bit-twiddling macros, __BIT, __BITS, SHIFTIN, SHIFTOUT, and __arraycount() from lib/libkern/libkern.h to sys/cdefs.h. Add a __-prefix to SHIFTIN and SHIFTOUT, and add a manual page for the bit-twiddling macros, bits(3). Make the __BIT and __BITS macros "widthless," as best I can, by changing their type to uintmax_t from uint32_t. XXX The manual page lags this change by a bit. Define __PRIxBIT and __PRIxBITS printf(3) format strings.
2006-08-31 23:24:37 +04:00
tcr |= __SHIFTIN(4, RTW_TCR_SRL_MASK) | __SHIFTIN(4, RTW_TCR_LRL_MASK);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
Do not indicate to the NIC that the host has appended the CRC32 to Tx frames.
2004-12-23 08:50:24 +03:00
tcr &= ~RTW_TCR_CRC; /* NIC appends CRC32 */
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
RTW_WRITE(regs, RTW_TCR, tcr);
Add some bus barriers. Remark where they are needed, e.g., read_eeprom. Change some printfs to RTW_DPRINTF2s to quiet rtw(4) when debugging is disabled.
2004-12-20 04:13:45 +03:00
RTW_SYNC(regs, RTW_TCR, RTW_TCR);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
}
Use device_t and its accessors throughout. Use aprint_*_dev(). Improve PMF-ability. Add a 'flags' argument to suspend/resume handlers and callers such as pmf_system_suspend(). Define a flag, PMF_F_SELF, which indicates to PMF that a device is suspending/resuming itself. Add helper routines, pmf_device_suspend_self(dev) and pmf_device_resume_self(dev), that call pmf_device_suspend(dev, PMF_F_SELF) and pmf_device_resume(dev, PMF_F_SELF), respectively. Use PMF_F_SELF to suspend/resume self in ath(4), audio(4), rtw(4), and sip(4). In ath(4) and in rtw(4), replace the icky sc_enable/sc_disable callbacks, provided by the bus front-end, with self-suspension/resumption. Also, clean up the bus front-ends. Make sure that the interrupt handler is disestablished during suspension. Get rid of driver-private flags (e.g., RTW_F_ENABLED, ath_softc->sc_invalid); use device_is_active()/device_has_power() calls, instead. In the network-class suspend handler, call if_stop(, 0) instead of if_stop(, 1), because the latter is superfluous (bus- and driver-suspension hooks will 'disable' the NIC), and it may cause recursion. In the network-class resume handler, prevent infinite recursion through if_init() by getting out early if we are self-suspending (PMF_F_SELF). rtw(4) improvements: Destroy rtw(4) callouts when we detach it. Make rtw at pci detachable. Print some more information with the "rx frame too long" warning. Remove activate() methods: Get rid of rtw_activate() and ath_activate(). The device activate() methods are not good for much these days. Make ath at cardbus resume with crypto functions intact: Introduce a boolean device property, "pmf-powerdown". If pmf-powerdown is present and false, it indicates that a bus back-end should not remove power from a device. Honor this property in cardbus_child_suspend(). Set this property to 'false' in ath_attach(), since removing power from an ath at cardbus seems to lobotomize the WPA crypto engine. XXX Should the pmf-powerdown property propagate toward the root of the device tree? Miscellaneous ath(4) changes: Warn if ath(4) tries to write crypto keys to suspended hardware. Reduce differences between FreeBSD and NetBSD in ath(4) multicast filter setup. Make ath_printrxbuf() print an rx descriptor's status & key index, to help debug crypto errors. Shorten a staircase in ath_ioctl(). Don't check for ieee80211_ioctl() return code ERESTART, it never happens.
2008-03-12 21:02:21 +03:00
static void
rtw_disable_interrupts(struct rtw_regs *regs)
{
RTW_WRITE16(regs, RTW_IMR, 0);
RTW_WBW(regs, RTW_IMR, RTW_ISR);
RTW_WRITE16(regs, RTW_ISR, 0xffff);
RTW_SYNC(regs, RTW_IMR, RTW_ISR);
}
static void
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
rtw_enable_interrupts(struct rtw_softc *sc)
{
struct rtw_regs *regs = &sc->sc_regs;
sc->sc_inten = RTW_INTR_RX|RTW_INTR_TX|RTW_INTR_BEACON|RTW_INTR_ATIMINT;
sc->sc_inten |= RTW_INTR_IOERROR|RTW_INTR_TIMEOUT;
RTW_WRITE16(regs, RTW_IMR, sc->sc_inten);
Add some bus barriers. Remark where they are needed, e.g., read_eeprom. Change some printfs to RTW_DPRINTF2s to quiet rtw(4) when debugging is disabled.
2004-12-20 04:13:45 +03:00
RTW_WBW(regs, RTW_IMR, RTW_ISR);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
RTW_WRITE16(regs, RTW_ISR, 0xffff);
Add some bus barriers. Remark where they are needed, e.g., read_eeprom. Change some printfs to RTW_DPRINTF2s to quiet rtw(4) when debugging is disabled.
2004-12-20 04:13:45 +03:00
RTW_SYNC(regs, RTW_IMR, RTW_ISR);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
/* XXX necessary? */
if (sc->sc_intr_ack != NULL)
(*sc->sc_intr_ack)(regs);
}
Realtek sample sources are not consistent about the "Digital PHY" bit in the SROM. It seems as if it is set to 1 when the PHY is *analog*, not *digital*. Fix my sources. In rtw_intr_rx, use units of 500kb/s instead of 100kb/s for rate, to be consistent with net80211's expectations. Polish up some debugging ugly messages. Dump raw 802.11 packets if IFF_DEBUG|IFF_LINK2 and RTW_DEBUG is defined. Polish power-state (on/sleep/off) handling. Especially improve support for RFMD (totally untested) and Maxim. For Philips, take the Digital PHY property into account. Call the net80211 watchdog function from rtw_watchdog, so that we scan again if auth/assoc fails. Be a little more cautious about writing register[RTW_TPPOLL], since other drivers are.... Don't frob the high/low-priority queues right now, since I don't use them. Add rtw_join_bss which programs the card with the BSSID and other properties of a BSS. Use it on state transitions. Factor out rtw_set_nettype. Make rtw_recv_beacon call ieee80211_recv_mgmt instead of dropping beacons on the floor! TBD IBSS merges. Change some rtw_debug=2 printfs to rtw_debug=3 (RTW_DPRINTF3) printfs so the console doesn't get spammed so badly at rtw_debug=2. Change some debugging printfs to RTW_DPRINTFs. E.g., print the "RF programming method" only if debugging is enabled.
2004-12-21 02:05:41 +03:00
static void
rtw_set_nettype(struct rtw_softc *sc, enum ieee80211_opmode opmode)
{
uint8_t msr;
/* I'm guessing that MSR is protected as CONFIG[0123] are. */
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
rtw_set_access(&sc->sc_regs, RTW_ACCESS_CONFIG);
Realtek sample sources are not consistent about the "Digital PHY" bit in the SROM. It seems as if it is set to 1 when the PHY is *analog*, not *digital*. Fix my sources. In rtw_intr_rx, use units of 500kb/s instead of 100kb/s for rate, to be consistent with net80211's expectations. Polish up some debugging ugly messages. Dump raw 802.11 packets if IFF_DEBUG|IFF_LINK2 and RTW_DEBUG is defined. Polish power-state (on/sleep/off) handling. Especially improve support for RFMD (totally untested) and Maxim. For Philips, take the Digital PHY property into account. Call the net80211 watchdog function from rtw_watchdog, so that we scan again if auth/assoc fails. Be a little more cautious about writing register[RTW_TPPOLL], since other drivers are.... Don't frob the high/low-priority queues right now, since I don't use them. Add rtw_join_bss which programs the card with the BSSID and other properties of a BSS. Use it on state transitions. Factor out rtw_set_nettype. Make rtw_recv_beacon call ieee80211_recv_mgmt instead of dropping beacons on the floor! TBD IBSS merges. Change some rtw_debug=2 printfs to rtw_debug=3 (RTW_DPRINTF3) printfs so the console doesn't get spammed so badly at rtw_debug=2. Change some debugging printfs to RTW_DPRINTFs. E.g., print the "RF programming method" only if debugging is enabled.
2004-12-21 02:05:41 +03:00
msr = RTW_READ8(&sc->sc_regs, RTW_MSR) & ~RTW_MSR_NETYPE_MASK;
switch (opmode) {
case IEEE80211_M_AHDEMO:
case IEEE80211_M_IBSS:
msr |= RTW_MSR_NETYPE_ADHOC_OK;
break;
case IEEE80211_M_HOSTAP:
msr |= RTW_MSR_NETYPE_AP_OK;
break;
case IEEE80211_M_MONITOR:
/* XXX */
msr |= RTW_MSR_NETYPE_NOLINK;
break;
case IEEE80211_M_STA:
msr |= RTW_MSR_NETYPE_INFRA_OK;
break;
}
RTW_WRITE8(&sc->sc_regs, RTW_MSR, msr);
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
rtw_set_access(&sc->sc_regs, RTW_ACCESS_NONE);
Realtek sample sources are not consistent about the "Digital PHY" bit in the SROM. It seems as if it is set to 1 when the PHY is *analog*, not *digital*. Fix my sources. In rtw_intr_rx, use units of 500kb/s instead of 100kb/s for rate, to be consistent with net80211's expectations. Polish up some debugging ugly messages. Dump raw 802.11 packets if IFF_DEBUG|IFF_LINK2 and RTW_DEBUG is defined. Polish power-state (on/sleep/off) handling. Especially improve support for RFMD (totally untested) and Maxim. For Philips, take the Digital PHY property into account. Call the net80211 watchdog function from rtw_watchdog, so that we scan again if auth/assoc fails. Be a little more cautious about writing register[RTW_TPPOLL], since other drivers are.... Don't frob the high/low-priority queues right now, since I don't use them. Add rtw_join_bss which programs the card with the BSSID and other properties of a BSS. Use it on state transitions. Factor out rtw_set_nettype. Make rtw_recv_beacon call ieee80211_recv_mgmt instead of dropping beacons on the floor! TBD IBSS merges. Change some rtw_debug=2 printfs to rtw_debug=3 (RTW_DPRINTF3) printfs so the console doesn't get spammed so badly at rtw_debug=2. Change some debugging printfs to RTW_DPRINTFs. E.g., print the "RF programming method" only if debugging is enabled.
2004-12-21 02:05:41 +03:00
}
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
#define rtw_calchash(addr) \
In sys/dev/ic/rtw.c: Use clue from rtk(4) and re(4) to fix the rtw(4) packet filter. Previously, I was using the wrong CRC32 function to hash multicast addresses; to compensate, I set the multicast filter to all 1s. Now that I hash the addresses correctly, I do not any longer set the filter to all 1s. In rtw_ioctl, avoid gratuitous re-initialization when the interface flags change. If a !IFF_UP -> IFF_UP transition, call rtw_init(); otherwise, only reload the packet filter. In sys/dev/ic/rtwreg.h: Put useful combinations of Receiver Control Register flags in RTW_RCR_PKTFILT_MASK, RTW_RCR_MONITOR, and RTW_RCR_PKTFILT_DEFAULT. (XXX RTW_RCR_MONITOR should be called RTW_RCR_PKTFILT_MONITOR.)
2005-01-03 06:07:12 +03:00
(ether_crc32_be((addr), IEEE80211_ADDR_LEN) >> 26)
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
static void
rtw_pktfilt_load(struct rtw_softc *sc)
{
struct rtw_regs *regs = &sc->sc_regs;
struct ieee80211com *ic = &sc->sc_ic;
Resolve conflicts in importation of 18-May-2005 ath(4) / net80211(9) from FreeBSD. Introduce compatibility shims (sys/dev/ic/ath_netbsd.[ch], sys/net80211/ieee80211_netbsd.[ch]). Update drivers (an, atu, atw, awi, ipw, iwi, rtw, wi) for the new net80211(9) API.
2005-06-22 10:14:51 +04:00
struct ethercom *ec = &sc->sc_ec;
struct ifnet *ifp = &sc->sc_if;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
int hash;
Cosmetic: change u_int{8,16,32,64}_t to C99-style uint{8,16,32,64}_t, which is preferred.
2005-01-02 07:23:03 +03:00
uint32_t hashes[2] = { 0, 0 };
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
struct ether_multi *enm;
struct ether_multistep step;
/* XXX might be necessary to stop Rx/Tx engines while setting filters */
In sys/dev/ic/rtw.c: Use clue from rtk(4) and re(4) to fix the rtw(4) packet filter. Previously, I was using the wrong CRC32 function to hash multicast addresses; to compensate, I set the multicast filter to all 1s. Now that I hash the addresses correctly, I do not any longer set the filter to all 1s. In rtw_ioctl, avoid gratuitous re-initialization when the interface flags change. If a !IFF_UP -> IFF_UP transition, call rtw_init(); otherwise, only reload the packet filter. In sys/dev/ic/rtwreg.h: Put useful combinations of Receiver Control Register flags in RTW_RCR_PKTFILT_MASK, RTW_RCR_MONITOR, and RTW_RCR_PKTFILT_DEFAULT. (XXX RTW_RCR_MONITOR should be called RTW_RCR_PKTFILT_MONITOR.)
2005-01-03 06:07:12 +03:00
sc->sc_rcr &= ~RTW_RCR_PKTFILTER_MASK;
sc->sc_rcr &= ~(RTW_RCR_MXDMA_MASK | RTW_RCR_RXFTH_MASK);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
In sys/dev/ic/rtw.c: Use clue from rtk(4) and re(4) to fix the rtw(4) packet filter. Previously, I was using the wrong CRC32 function to hash multicast addresses; to compensate, I set the multicast filter to all 1s. Now that I hash the addresses correctly, I do not any longer set the filter to all 1s. In rtw_ioctl, avoid gratuitous re-initialization when the interface flags change. If a !IFF_UP -> IFF_UP transition, call rtw_init(); otherwise, only reload the packet filter. In sys/dev/ic/rtwreg.h: Put useful combinations of Receiver Control Register flags in RTW_RCR_PKTFILT_MASK, RTW_RCR_MONITOR, and RTW_RCR_PKTFILT_DEFAULT. (XXX RTW_RCR_MONITOR should be called RTW_RCR_PKTFILT_MONITOR.)
2005-01-03 06:07:12 +03:00
sc->sc_rcr |= RTW_RCR_PKTFILTER_DEFAULT;
/* MAC auto-reset PHY (huh?) */
Realtek sample sources are not consistent about the "Digital PHY" bit in the SROM. It seems as if it is set to 1 when the PHY is *analog*, not *digital*. Fix my sources. In rtw_intr_rx, use units of 500kb/s instead of 100kb/s for rate, to be consistent with net80211's expectations. Polish up some debugging ugly messages. Dump raw 802.11 packets if IFF_DEBUG|IFF_LINK2 and RTW_DEBUG is defined. Polish power-state (on/sleep/off) handling. Especially improve support for RFMD (totally untested) and Maxim. For Philips, take the Digital PHY property into account. Call the net80211 watchdog function from rtw_watchdog, so that we scan again if auth/assoc fails. Be a little more cautious about writing register[RTW_TPPOLL], since other drivers are.... Don't frob the high/low-priority queues right now, since I don't use them. Add rtw_join_bss which programs the card with the BSSID and other properties of a BSS. Use it on state transitions. Factor out rtw_set_nettype. Make rtw_recv_beacon call ieee80211_recv_mgmt instead of dropping beacons on the floor! TBD IBSS merges. Change some rtw_debug=2 printfs to rtw_debug=3 (RTW_DPRINTF3) printfs so the console doesn't get spammed so badly at rtw_debug=2. Change some debugging printfs to RTW_DPRINTFs. E.g., print the "RF programming method" only if debugging is enabled.
2004-12-21 02:05:41 +03:00
sc->sc_rcr |= RTW_RCR_ENMARP;
In sys/dev/ic/rtw.c: Use clue from rtk(4) and re(4) to fix the rtw(4) packet filter. Previously, I was using the wrong CRC32 function to hash multicast addresses; to compensate, I set the multicast filter to all 1s. Now that I hash the addresses correctly, I do not any longer set the filter to all 1s. In rtw_ioctl, avoid gratuitous re-initialization when the interface flags change. If a !IFF_UP -> IFF_UP transition, call rtw_init(); otherwise, only reload the packet filter. In sys/dev/ic/rtwreg.h: Put useful combinations of Receiver Control Register flags in RTW_RCR_PKTFILT_MASK, RTW_RCR_MONITOR, and RTW_RCR_PKTFILT_DEFAULT. (XXX RTW_RCR_MONITOR should be called RTW_RCR_PKTFILT_MONITOR.)
2005-01-03 06:07:12 +03:00
/* DMA whole Rx packets, only. Set Tx DMA burst size to 1024 bytes. */
sc->sc_rcr |= RTW_RCR_MXDMA_1024 | RTW_RCR_RXFTH_WHOLE;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
In sys/dev/ic/rtw.c: Use clue from rtk(4) and re(4) to fix the rtw(4) packet filter. Previously, I was using the wrong CRC32 function to hash multicast addresses; to compensate, I set the multicast filter to all 1s. Now that I hash the addresses correctly, I do not any longer set the filter to all 1s. In rtw_ioctl, avoid gratuitous re-initialization when the interface flags change. If a !IFF_UP -> IFF_UP transition, call rtw_init(); otherwise, only reload the packet filter. In sys/dev/ic/rtwreg.h: Put useful combinations of Receiver Control Register flags in RTW_RCR_PKTFILT_MASK, RTW_RCR_MONITOR, and RTW_RCR_PKTFILT_DEFAULT. (XXX RTW_RCR_MONITOR should be called RTW_RCR_PKTFILT_MONITOR.)
2005-01-03 06:07:12 +03:00
switch (ic->ic_opmode) {
case IEEE80211_M_MONITOR:
sc->sc_rcr |= RTW_RCR_MONITOR;
break;
case IEEE80211_M_AHDEMO:
case IEEE80211_M_IBSS:
/* receive broadcasts in our BSS */
sc->sc_rcr |= RTW_RCR_ADD3;
break;
default:
break;
}
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
ifp->if_flags &= ~IFF_ALLMULTI;
/*
* Program the 64-bit multicast hash filter.
*/
ETHER_FIRST_MULTI(step, ec, enm);
while (enm != NULL) {
/* XXX */
if (memcmp(enm->enm_addrlo, enm->enm_addrhi,
Simplify multicast filter programming, and take care to avoid gratuitous device resets when the kernel adds/deletes a multicast address.
2007-05-29 22:32:07 +04:00
ETHER_ADDR_LEN) != 0) {
ifp->if_flags |= IFF_ALLMULTI;
break;
}
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
hash = rtw_calchash(enm->enm_addrlo);
In sys/dev/ic/rtw.c: Use clue from rtk(4) and re(4) to fix the rtw(4) packet filter. Previously, I was using the wrong CRC32 function to hash multicast addresses; to compensate, I set the multicast filter to all 1s. Now that I hash the addresses correctly, I do not any longer set the filter to all 1s. In rtw_ioctl, avoid gratuitous re-initialization when the interface flags change. If a !IFF_UP -> IFF_UP transition, call rtw_init(); otherwise, only reload the packet filter. In sys/dev/ic/rtwreg.h: Put useful combinations of Receiver Control Register flags in RTW_RCR_PKTFILT_MASK, RTW_RCR_MONITOR, and RTW_RCR_PKTFILT_DEFAULT. (XXX RTW_RCR_MONITOR should be called RTW_RCR_PKTFILT_MONITOR.)
2005-01-03 06:07:12 +03:00
hashes[hash >> 5] |= (1 << (hash & 0x1f));
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
ETHER_NEXT_MULTI(step, enm);
}
Simplify multicast filter programming, and take care to avoid gratuitous device resets when the kernel adds/deletes a multicast address.
2007-05-29 22:32:07 +04:00
/* XXX accept all broadcast if scanning */
if ((ifp->if_flags & IFF_BROADCAST) != 0)
sc->sc_rcr |= RTW_RCR_AB; /* accept all broadcast */
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
Simplify multicast filter programming, and take care to avoid gratuitous device resets when the kernel adds/deletes a multicast address.
2007-05-29 22:32:07 +04:00
if (ifp->if_flags & IFF_PROMISC) {
sc->sc_rcr |= RTW_RCR_AB; /* accept all broadcast */
Accommodate the variety of RF front-ends when radiotapping frames: tap only RSSI when there is a Philips RF front-end. Tap both Barker lock quality and RSSI when there is any other RF front-end. Provide radiotap listeners a more complete picture of channel activity: in promiscuous mode, tap frames who do not pass the CRC32 check. Flag packets that were received with a short preamble. Ask the NIC to pass us 802.11 Control frames. Pass frames to radiotap listeners before stripping the FCS. Re-order operations in rtw_intr_rx() in order to accomplish all of this, taking care not to pass a broken packet to net80211! Do not provide a flags field when tapping xmitted frames. Assert sane Rx packet lengths. Really should check and drop instead of KASSERTing. I will revisit this, soon. Update copyright.
2007-11-17 02:35:19 +03:00
sc->sc_rcr |= RTW_RCR_ACRC32; /* accept frames failing CRC */
sc->sc_rcr |= RTW_RCR_AICV; /* accept frames failing ICV */
Simplify multicast filter programming, and take care to avoid gratuitous device resets when the kernel adds/deletes a multicast address.
2007-05-29 22:32:07 +04:00
ifp->if_flags |= IFF_ALLMULTI;
In sys/dev/ic/rtw.c: Use clue from rtk(4) and re(4) to fix the rtw(4) packet filter. Previously, I was using the wrong CRC32 function to hash multicast addresses; to compensate, I set the multicast filter to all 1s. Now that I hash the addresses correctly, I do not any longer set the filter to all 1s. In rtw_ioctl, avoid gratuitous re-initialization when the interface flags change. If a !IFF_UP -> IFF_UP transition, call rtw_init(); otherwise, only reload the packet filter. In sys/dev/ic/rtwreg.h: Put useful combinations of Receiver Control Register flags in RTW_RCR_PKTFILT_MASK, RTW_RCR_MONITOR, and RTW_RCR_PKTFILT_DEFAULT. (XXX RTW_RCR_MONITOR should be called RTW_RCR_PKTFILT_MONITOR.)
2005-01-03 06:07:12 +03:00
}
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
Simplify multicast filter programming, and take care to avoid gratuitous device resets when the kernel adds/deletes a multicast address.
2007-05-29 22:32:07 +04:00
if (ifp->if_flags & IFF_ALLMULTI)
hashes[0] = hashes[1] = 0xffffffff;
if ((hashes[0] | hashes[1]) != 0)
sc->sc_rcr |= RTW_RCR_AM; /* accept multicast */
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
RTW_WRITE(regs, RTW_MAR0, hashes[0]);
RTW_WRITE(regs, RTW_MAR1, hashes[1]);
RTW_WRITE(regs, RTW_RCR, sc->sc_rcr);
RTW_SYNC(regs, RTW_MAR0, RTW_RCR); /* RTW_MAR0 < RTW_MAR1 < RTW_RCR */
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
DPRINTF(sc, RTW_DEBUG_PKTFILT,
("%s: RTW_MAR0 %08x RTW_MAR1 %08x RTW_RCR %08x\n",
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
device_xname(sc->sc_dev), RTW_READ(regs, RTW_MAR0),
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
RTW_READ(regs, RTW_MAR1), RTW_READ(regs, RTW_RCR)));
}
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
static struct mbuf *
rtw_beacon_alloc(struct rtw_softc *sc, struct ieee80211_node *ni)
{
struct ieee80211com *ic = &sc->sc_ic;
struct mbuf *m;
Resolve conflicts in importation of 18-May-2005 ath(4) / net80211(9) from FreeBSD. Introduce compatibility shims (sys/dev/ic/ath_netbsd.[ch], sys/net80211/ieee80211_netbsd.[ch]). Update drivers (an, atu, atw, awi, ipw, iwi, rtw, wi) for the new net80211(9) API.
2005-06-22 10:14:51 +04:00
struct ieee80211_beacon_offsets boff;
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
if ((m = ieee80211_beacon_alloc(ic, ni, &boff)) != NULL) {
RTW_DPRINTF(RTW_DEBUG_BEACON,
("%s: m %p len %u\n", __func__, m, m->m_len));
}
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
return m;
}
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
/* Must be called at splnet. */
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
static int
rtw_init(struct ifnet *ifp)
{
struct rtw_softc *sc = (struct rtw_softc *)ifp->if_softc;
struct ieee80211com *ic = &sc->sc_ic;
struct rtw_regs *regs = &sc->sc_regs;
Use device_t and its accessors throughout. Use aprint_*_dev(). Improve PMF-ability. Add a 'flags' argument to suspend/resume handlers and callers such as pmf_system_suspend(). Define a flag, PMF_F_SELF, which indicates to PMF that a device is suspending/resuming itself. Add helper routines, pmf_device_suspend_self(dev) and pmf_device_resume_self(dev), that call pmf_device_suspend(dev, PMF_F_SELF) and pmf_device_resume(dev, PMF_F_SELF), respectively. Use PMF_F_SELF to suspend/resume self in ath(4), audio(4), rtw(4), and sip(4). In ath(4) and in rtw(4), replace the icky sc_enable/sc_disable callbacks, provided by the bus front-end, with self-suspension/resumption. Also, clean up the bus front-ends. Make sure that the interrupt handler is disestablished during suspension. Get rid of driver-private flags (e.g., RTW_F_ENABLED, ath_softc->sc_invalid); use device_is_active()/device_has_power() calls, instead. In the network-class suspend handler, call if_stop(, 0) instead of if_stop(, 1), because the latter is superfluous (bus- and driver-suspension hooks will 'disable' the NIC), and it may cause recursion. In the network-class resume handler, prevent infinite recursion through if_init() by getting out early if we are self-suspending (PMF_F_SELF). rtw(4) improvements: Destroy rtw(4) callouts when we detach it. Make rtw at pci detachable. Print some more information with the "rx frame too long" warning. Remove activate() methods: Get rid of rtw_activate() and ath_activate(). The device activate() methods are not good for much these days. Make ath at cardbus resume with crypto functions intact: Introduce a boolean device property, "pmf-powerdown". If pmf-powerdown is present and false, it indicates that a bus back-end should not remove power from a device. Honor this property in cardbus_child_suspend(). Set this property to 'false' in ath_attach(), since removing power from an ath at cardbus seems to lobotomize the WPA crypto engine. XXX Should the pmf-powerdown property propagate toward the root of the device tree? Miscellaneous ath(4) changes: Warn if ath(4) tries to write crypto keys to suspended hardware. Reduce differences between FreeBSD and NetBSD in ath(4) multicast filter setup. Make ath_printrxbuf() print an rx descriptor's status & key index, to help debug crypto errors. Shorten a staircase in ath_ioctl(). Don't check for ieee80211_ioctl() return code ERESTART, it never happens.
2008-03-12 21:02:21 +03:00
int rc;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
Use device_t and its accessors throughout. Use aprint_*_dev(). Improve PMF-ability. Add a 'flags' argument to suspend/resume handlers and callers such as pmf_system_suspend(). Define a flag, PMF_F_SELF, which indicates to PMF that a device is suspending/resuming itself. Add helper routines, pmf_device_suspend_self(dev) and pmf_device_resume_self(dev), that call pmf_device_suspend(dev, PMF_F_SELF) and pmf_device_resume(dev, PMF_F_SELF), respectively. Use PMF_F_SELF to suspend/resume self in ath(4), audio(4), rtw(4), and sip(4). In ath(4) and in rtw(4), replace the icky sc_enable/sc_disable callbacks, provided by the bus front-end, with self-suspension/resumption. Also, clean up the bus front-ends. Make sure that the interrupt handler is disestablished during suspension. Get rid of driver-private flags (e.g., RTW_F_ENABLED, ath_softc->sc_invalid); use device_is_active()/device_has_power() calls, instead. In the network-class suspend handler, call if_stop(, 0) instead of if_stop(, 1), because the latter is superfluous (bus- and driver-suspension hooks will 'disable' the NIC), and it may cause recursion. In the network-class resume handler, prevent infinite recursion through if_init() by getting out early if we are self-suspending (PMF_F_SELF). rtw(4) improvements: Destroy rtw(4) callouts when we detach it. Make rtw at pci detachable. Print some more information with the "rx frame too long" warning. Remove activate() methods: Get rid of rtw_activate() and ath_activate(). The device activate() methods are not good for much these days. Make ath at cardbus resume with crypto functions intact: Introduce a boolean device property, "pmf-powerdown". If pmf-powerdown is present and false, it indicates that a bus back-end should not remove power from a device. Honor this property in cardbus_child_suspend(). Set this property to 'false' in ath_attach(), since removing power from an ath at cardbus seems to lobotomize the WPA crypto engine. XXX Should the pmf-powerdown property propagate toward the root of the device tree? Miscellaneous ath(4) changes: Warn if ath(4) tries to write crypto keys to suspended hardware. Reduce differences between FreeBSD and NetBSD in ath(4) multicast filter setup. Make ath_printrxbuf() print an rx descriptor's status & key index, to help debug crypto errors. Shorten a staircase in ath_ioctl(). Don't check for ieee80211_ioctl() return code ERESTART, it never happens.
2008-03-12 21:02:21 +03:00
if (device_is_active(sc->sc_dev)) {
/* Cancel pending I/O and reset. */
rtw_stop(ifp, 0);
In pmf(9), improve the implementation of device self-suspension and make suspension by self, by drvctl(8), and by ACPI system sleep play nice together. Start solidifying some temporary API changes. 1. Extract a new header file, <sys/device_if.h>, from <sys/device.h> and #include it from <sys/pmf.h> instead of <sys/device.h> to break the circular dependency between <sys/device.h> and <sys/pmf.h>. 2. Introduce pmf_qual_t, an aggregate of qualifications on a PMF suspend/resume call. Start to replace instances of PMF_FN_PROTO, PMF_FN_ARGS, et cetera, with a pmf_qual_t. 3. Introduce the notion of a "suspensor," an entity that holds a device in suspension. More than one suspensor may hold a device at once. A device stays suspended as long as at least one suspensor holds it. A device resumes when the last suspensor releases it. Currently, the kernel defines three suspensors, 3a the system-suspensor: for system suspension, initiated by 'sysctl -w machdep.sleep_state=3', by lid closure, by power-button press, et cetera, 3b the drvctl-suspensor: for device suspension by /dev/drvctl ioctl, e.g., drvctl -S sip0. 3c the system self-suspensor: for device drivers that suspend themselves and their children. Several drivers for network interfaces put the network device to sleep while it is not administratively up, that is, after the kernel calls if_stop(, 1). The self-suspensor should not be used directly. See the description of suspensor delegates, below. A suspensor can have one or more "delegates". A suspensor can release devices that its delegates hold suspended. Right now, only the system self-suspensor has delegates. For each device that a self-suspending driver attaches, it creates the device's self-suspensor, a delegate of the system self-suspensor. Suspensors stop a system-wide suspend/resume cycle from waking devices that the operator put to sleep with drvctl before the cycle. They also help self-suspension to work more simply, safely, and in accord with expectations. 4. Add the notion of device activation level, devact_level_t, and a routine for checking the current activation level, device_activation(). Current activation levels are DEVACT_LEVEL_BUS, DEVACT_LEVEL_DRIVER, and DEVACT_LEVEL_CLASS, which respectively indicate that the device's bus is active, that the bus and device are active, and that the bus, device, and the functions of the device's class (network, audio) are active. Suspend/resume calls can be qualified with a devact_level_t. The power-management framework treats a devact_level_t that qualifies a device suspension as the device's current activation level; it only runs hooks to reduce the activation level from the presumed current level to the fully suspended state. The framework treats a devact_level_t qualifying device resumption as the target activation level; it only runs hooks to raise the activation level to the target. 5. Use pmf_qual_t, devact_level_t, and self-suspensors in several drivers. 6. Temporarily add an unused power-management workqueue that I will remove or replace, soon.
2009-09-16 20:34:49 +04:00
} else if (!pmf_device_resume(sc->sc_dev, &sc->sc_qual) ||
!device_is_active(sc->sc_dev))
return 0;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
DPRINTF(sc, RTW_DEBUG_TUNE, ("%s: channel %d freq %d flags 0x%04x\n",
Adapt drivers to the new net80211(9). Most of this is from dyoung@. Thanks!
2005-11-18 19:53:56 +03:00
__func__, ieee80211_chan2ieee(ic, ic->ic_curchan),
ic->ic_curchan->ic_freq, ic->ic_curchan->ic_flags));
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
if ((rc = rtw_pwrstate(sc, RTW_OFF)) != 0)
goto out;
If defined(RTW_DEBUG), provide a sysctl (hw.rtw.rxbufs_limit) for limiting the number of rx buffers an rtw may allocate. Use this sysctl to test the code that copes with buffer exhaustion. Allocate at most RTW_RXQLEN rx buffers, stopping at the sysctl limit. Record in sc_nrxdesc how many were allocated, and put the end-of-ring indication on sc_rxdesc[sc_nrxdesc - 1]. In rtw_init, if no rx buffers could be allocated, log a complaint, clear IFF_RUNNING, and exit with an error. Many changes to accomodate a short rx ring, mainly of the "add a rx-ring length argument" variety. XXX I really should consolidate all of the rx ring variables in one struct and pass that to the rx-ring subroutines. Bug fix: after calling MCLGET, use the (m->m_flags & M_EXT) idiom to check for success, instead of m != NULL. Bug fix: at the top of rtw_start, if IFF_RUNNING is not set, or IFF_OACTIVE is, get out.
2004-12-29 01:21:15 +03:00
if ((rc = rtw_swring_setup(sc)) != 0)
goto out;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
rtw_transmit_config(regs);
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
rtw_set_access(regs, RTW_ACCESS_CONFIG);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
At last, I have rtw w/ Philips RF receiving packets. I added some sysctls to aid debugging: * hw.rtw.debug -- enable debugging * hw.rtw.flush_rfio -- Linux voodoo: possibly makes the MAC "flush" bits down the serial bus to the RF * hw.rtw.host_rfio: force the host to bang bits to the RF, instead of the MAC banging bits * hw.rtw.rfio_delay: after telling the MAC to bang bits to the RF front-end, delay rfio_delay microseconds. * hw.rtw.rfprog_fallback: there is this notion of the "RF programming method." I believe the choice influences the polarity/timing of the serial bus used to program the RF front-end. I know the correct choice for Intersil/RFMD/Philips front-ends, only. For all other front-ends, I "fallback" to rfprog_fallback. Make rtw_txdac_enable take an rtw_softc argument. I will probably revert this change. Add some Linux voodoo to rtw_continuous_tx_enable. I will probably revert this change. Important: add rtw_set_rfprog, which sets the correct RF programming method. This change and the following change are probably responsible for making the Philips RF work. Important: RTW_CONFIG1 is an 8-bit register, treat it that way! Important: RTW_BRSR is 16-bit, RTW_CRCOUNT, RTW_PHYDELAY, and RTW_MSR are 8-bit: treat them that way! Vastly simplify rtw_resume_ticks. Note to self: set the LED state to match the power state. Hedge against the possibility that RTW_MSR is protected as RTW_CONFIG[0123] are, meanwhile reworking that section of rtw_init a little. Add sc_anaparm, which isn't used, yet....
2004-12-13 03:48:02 +03:00
RTW_WRITE8(regs, RTW_MSR, 0x0); /* no link */
Add some bus barriers. Remark where they are needed, e.g., read_eeprom. Change some printfs to RTW_DPRINTF2s to quiet rtw(4) when debugging is disabled.
2004-12-20 04:13:45 +03:00
RTW_WBW(regs, RTW_MSR, RTW_BRSR);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
When checking if a split packet is completed, check the *last* segment, not the first. Otherwise we can overwrite parts of the TX ring that we shouldn't. Kluge the basic rate setting for now. Minor simplification to the dequeueing logic.
2004-12-27 12:40:18 +03:00
/* long PLCP header, 1Mb/2Mb basic rate */
RTW_WRITE16(regs, RTW_BRSR, RTW_BRSR_MBR8180_2MBPS);
Add some bus barriers. Remark where they are needed, e.g., read_eeprom. Change some printfs to RTW_DPRINTF2s to quiet rtw(4) when debugging is disabled.
2004-12-20 04:13:45 +03:00
RTW_SYNC(regs, RTW_BRSR, RTW_BRSR);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
rtw_set_access(regs, RTW_ACCESS_ANAPARM);
rtw_set_access(regs, RTW_ACCESS_NONE);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
/* XXX from reference sources */
RTW_WRITE(regs, RTW_FEMR, 0xffff);
Add some bus barriers. Remark where they are needed, e.g., read_eeprom. Change some printfs to RTW_DPRINTF2s to quiet rtw(4) when debugging is disabled.
2004-12-20 04:13:45 +03:00
RTW_SYNC(regs, RTW_FEMR, RTW_FEMR);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
rtw_set_rfprog(regs, sc->sc_rfchipid, sc->sc_dev);
At last, I have rtw w/ Philips RF receiving packets. I added some sysctls to aid debugging: * hw.rtw.debug -- enable debugging * hw.rtw.flush_rfio -- Linux voodoo: possibly makes the MAC "flush" bits down the serial bus to the RF * hw.rtw.host_rfio: force the host to bang bits to the RF, instead of the MAC banging bits * hw.rtw.rfio_delay: after telling the MAC to bang bits to the RF front-end, delay rfio_delay microseconds. * hw.rtw.rfprog_fallback: there is this notion of the "RF programming method." I believe the choice influences the polarity/timing of the serial bus used to program the RF front-end. I know the correct choice for Intersil/RFMD/Philips front-ends, only. For all other front-ends, I "fallback" to rfprog_fallback. Make rtw_txdac_enable take an rtw_softc argument. I will probably revert this change. Add some Linux voodoo to rtw_continuous_tx_enable. I will probably revert this change. Important: add rtw_set_rfprog, which sets the correct RF programming method. This change and the following change are probably responsible for making the Philips RF work. Important: RTW_CONFIG1 is an 8-bit register, treat it that way! Important: RTW_BRSR is 16-bit, RTW_CRCOUNT, RTW_PHYDELAY, and RTW_MSR are 8-bit: treat them that way! Vastly simplify rtw_resume_ticks. Note to self: set the LED state to match the power state. Hedge against the possibility that RTW_MSR is protected as RTW_CONFIG[0123] are, meanwhile reworking that section of rtw_init a little. Add sc_anaparm, which isn't used, yet....
2004-12-13 03:48:02 +03:00
RTW_WRITE8(regs, RTW_PHYDELAY, sc->sc_phydelay);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
/* from Linux driver */
At last, I have rtw w/ Philips RF receiving packets. I added some sysctls to aid debugging: * hw.rtw.debug -- enable debugging * hw.rtw.flush_rfio -- Linux voodoo: possibly makes the MAC "flush" bits down the serial bus to the RF * hw.rtw.host_rfio: force the host to bang bits to the RF, instead of the MAC banging bits * hw.rtw.rfio_delay: after telling the MAC to bang bits to the RF front-end, delay rfio_delay microseconds. * hw.rtw.rfprog_fallback: there is this notion of the "RF programming method." I believe the choice influences the polarity/timing of the serial bus used to program the RF front-end. I know the correct choice for Intersil/RFMD/Philips front-ends, only. For all other front-ends, I "fallback" to rfprog_fallback. Make rtw_txdac_enable take an rtw_softc argument. I will probably revert this change. Add some Linux voodoo to rtw_continuous_tx_enable. I will probably revert this change. Important: add rtw_set_rfprog, which sets the correct RF programming method. This change and the following change are probably responsible for making the Philips RF work. Important: RTW_CONFIG1 is an 8-bit register, treat it that way! Important: RTW_BRSR is 16-bit, RTW_CRCOUNT, RTW_PHYDELAY, and RTW_MSR are 8-bit: treat them that way! Vastly simplify rtw_resume_ticks. Note to self: set the LED state to match the power state. Hedge against the possibility that RTW_MSR is protected as RTW_CONFIG[0123] are, meanwhile reworking that section of rtw_init a little. Add sc_anaparm, which isn't used, yet....
2004-12-13 03:48:02 +03:00
RTW_WRITE8(regs, RTW_CRCOUNT, RTW_CRCOUNT_MAGIC);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
Add some bus barriers. Remark where they are needed, e.g., read_eeprom. Change some printfs to RTW_DPRINTF2s to quiet rtw(4) when debugging is disabled.
2004-12-20 04:13:45 +03:00
RTW_SYNC(regs, RTW_PHYDELAY, RTW_CRCOUNT);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
rtw_enable_interrupts(sc);
rtw_pktfilt_load(sc);
Miscellaneous changes. Details below. Important changes flagged with []. Using the driver with my Linksys WPC11 ver. 4, it seems to be receiving packets for a change. The WPC11 ver. 4 has a Maxim RF section. My no-name rtw with Philips RF section still does not receive any packets. Keep access-level (analog params > config[0123] registers > none) in sc_access. Add rtw_set_access for changing the access level. Make rtw_continuous_tx_enable and other subroutines use rtw_set_access instead of rtw_config0123_enable and rtw_anaparm_enable. Factor part of the chip-reset code into rtw_chip_reset1. Change the 'struct foo (*bar)[N]'-style arguments to 'struct foo *bar'-style arguments. Consolidate software/hardware Tx/Rx ring setup in rtw_hwring_setup, rtw_swring_setup. Add a new constant, SA2400_OPMODE_DEFAULTS, for the bits that we *always* set in the SA2400 OPMODE register. Factor some code out into rtw_sa2400_calibrate. (Inspired by the Linux driver.) [] When the receiver goes into underrun/overflow state, call a new subroutine, rtw_kick() that stops the Rx/Tx processes, resets the chip, reinitializes the Tx/Rx rings, and restarts Rx/Tx processes. (Inspired by the Linux driver.) [] In rtw_intr_rx, check for too-short packets before calling ieee80211_find_rxnode. I believe this will prevent a repeat of the MCHK exception I saw once on macppc. [] Use seconds-elapased as well as microseconds-elapsed to set the next "due date" for the timeout interrupt. This keeps the driver from programming the timeout to expire too early. [] In rtw_intr, read RTW_ISR at most 10 times, then get out. If the interface is not enabled (RTW_F_ENABLED), then get out. [] In rtw_stop, get out if the interface is not enabled (RTW_F_ENABLED). Block IPL_NET interrupts. Don't read/write any registers if the interface is invalid (RTW_F_INVALID). [] Call rtw_stop in rtw_detach.
2004-12-12 09:37:59 +03:00
rtw_hwring_setup(sc);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
Resolve conflicts in importation of 18-May-2005 ath(4) / net80211(9) from FreeBSD. Introduce compatibility shims (sys/dev/ic/ath_netbsd.[ch], sys/net80211/ieee80211_netbsd.[ch]). Update drivers (an, atu, atw, awi, ipw, iwi, rtw, wi) for the new net80211(9) API.
2005-06-22 10:14:51 +04:00
rtw_wep_setkeys(sc, ic->ic_nw_keys, ic->ic_def_txkey);
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
I have made a small change that suppresses "recovery" from (spurious?) tx timeouts on the beacon queue. The driver's recovery process badly disrupts the MAC's receiver. This needs a closer look, later. Every other fix is by Charles M. Hannum. Charles writes: "With these, the device is quite stable under load, does not experience xmit timeouts, no longer mysteriously freezes, and does not spew a lot of garbage. It does, however, continue to get some "tx fifo underflow"s. global, SYSCTL_SETUP(): a) Removed unused sysctl variables. rtw_io_enable(): b) Change rtw_io_enable to take a softc, and make it call rtw_txring_fixup(). c) Enable the PCI multiple read/write bit. d) Always reset rdb_next to 0 when starting the receive engine. rtw_intr_rx(): e) When processing receive packets, store the next pointer in rdb_next and read it back on the next round, to fix a race condition vs. rtw_tune() during AP discovery and association. f) Remove the special "skip to 0th descriptor" code, because it is no longer needed. rtw_collect_txring(): g) In the "force" case, stop when we catch up to the hardware xmit pointer. rtw_intr_beacon(): h) Resetting tdb_next here looked fishy to me, in light of changes in rtw_collect_txring(). XXX However, this is the one part I haven't tested. rtw_txring_fixup(): i) Add function and line arguments, so we can print out information about the caller. j) In addition to reading the hardware pointer, skip over any additional packets with OWN set -- we may have queued more. rtw_rxring_fixup(): k) Remove this, because it's completely wrong. rtw_intr_ioerror(): l) Don't handle receive overruns here; handle them in the normal receive path. Also changed RTW_INTR_RX and RTW_INTR_IOERROR accordingly. m) Do not fiddle with CR; the transmit engine is still running just fine. n) Do not fiddle with IMR; it was unnecessary. o) Don't touch the receive engine at all; not necessary now that we're leaving CR alone. p) Remove the rtw_do_chip_reset case; it was dangerous and not actually used. rtw_intr(): q) Don't mask off ISR bits when we pass down the ISR value; it's not necessary, and sometimes the other bits useful in debugging output. rtw_seg_too_short(): r) Don't spew here; the caller will do it if necessary. (Stops spewage in an expected case.) rtw_rxdesc_blk: s) Minor change to make the elements have the same order as txdesc_blk. general: t) Added a bunch of debugging output, #ifdef RTW_DIAG, that was useful while investigating the ring management problems. Also moved rtw_txring_fixup() into RTW_DIAG, as I believe the transmit pointer is always kept in sync now, and the function has potential race conditions.
2007-01-06 11:27:53 +03:00
rtw_io_enable(sc, RTW_CR_RE | RTW_CR_TE, 1);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
ifp->if_flags |= IFF_RUNNING;
ic->ic_state = IEEE80211_S_INIT;
RTW_WRITE16(regs, RTW_BSSID16, 0x0);
RTW_WRITE(regs, RTW_BSSID32, 0x0);
rtw_resume_ticks(sc);
Realtek sample sources are not consistent about the "Digital PHY" bit in the SROM. It seems as if it is set to 1 when the PHY is *analog*, not *digital*. Fix my sources. In rtw_intr_rx, use units of 500kb/s instead of 100kb/s for rate, to be consistent with net80211's expectations. Polish up some debugging ugly messages. Dump raw 802.11 packets if IFF_DEBUG|IFF_LINK2 and RTW_DEBUG is defined. Polish power-state (on/sleep/off) handling. Especially improve support for RFMD (totally untested) and Maxim. For Philips, take the Digital PHY property into account. Call the net80211 watchdog function from rtw_watchdog, so that we scan again if auth/assoc fails. Be a little more cautious about writing register[RTW_TPPOLL], since other drivers are.... Don't frob the high/low-priority queues right now, since I don't use them. Add rtw_join_bss which programs the card with the BSSID and other properties of a BSS. Use it on state transitions. Factor out rtw_set_nettype. Make rtw_recv_beacon call ieee80211_recv_mgmt instead of dropping beacons on the floor! TBD IBSS merges. Change some rtw_debug=2 printfs to rtw_debug=3 (RTW_DPRINTF3) printfs so the console doesn't get spammed so badly at rtw_debug=2. Change some debugging printfs to RTW_DPRINTFs. E.g., print the "RF programming method" only if debugging is enabled.
2004-12-21 02:05:41 +03:00
rtw_set_nettype(sc, IEEE80211_M_MONITOR);
At last, I have rtw w/ Philips RF receiving packets. I added some sysctls to aid debugging: * hw.rtw.debug -- enable debugging * hw.rtw.flush_rfio -- Linux voodoo: possibly makes the MAC "flush" bits down the serial bus to the RF * hw.rtw.host_rfio: force the host to bang bits to the RF, instead of the MAC banging bits * hw.rtw.rfio_delay: after telling the MAC to bang bits to the RF front-end, delay rfio_delay microseconds. * hw.rtw.rfprog_fallback: there is this notion of the "RF programming method." I believe the choice influences the polarity/timing of the serial bus used to program the RF front-end. I know the correct choice for Intersil/RFMD/Philips front-ends, only. For all other front-ends, I "fallback" to rfprog_fallback. Make rtw_txdac_enable take an rtw_softc argument. I will probably revert this change. Add some Linux voodoo to rtw_continuous_tx_enable. I will probably revert this change. Important: add rtw_set_rfprog, which sets the correct RF programming method. This change and the following change are probably responsible for making the Philips RF work. Important: RTW_CONFIG1 is an 8-bit register, treat it that way! Important: RTW_BRSR is 16-bit, RTW_CRCOUNT, RTW_PHYDELAY, and RTW_MSR are 8-bit: treat them that way! Vastly simplify rtw_resume_ticks. Note to self: set the LED state to match the power state. Hedge against the possibility that RTW_MSR is protected as RTW_CONFIG[0123] are, meanwhile reworking that section of rtw_init a little. Add sc_anaparm, which isn't used, yet....
2004-12-13 03:48:02 +03:00
if (ic->ic_opmode == IEEE80211_M_MONITOR)
return ieee80211_new_state(ic, IEEE80211_S_RUN, -1);
else
return ieee80211_new_state(ic, IEEE80211_S_SCAN, -1);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
out:
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
aprint_error_dev(sc->sc_dev, "interface not running\n");
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
return rc;
}
Remove leading __ from __(const|inline|signed|volatile) -- it is obsolete.
2005-12-24 23:27:29 +03:00
static inline void
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
rtw_led_init(struct rtw_regs *regs)
{
uint8_t cfg0, cfg1;
rtw_set_access(regs, RTW_ACCESS_CONFIG);
cfg0 = RTW_READ8(regs, RTW_CONFIG0);
cfg0 |= RTW_CONFIG0_LEDGPOEN;
RTW_WRITE8(regs, RTW_CONFIG0, cfg0);
cfg1 = RTW_READ8(regs, RTW_CONFIG1);
RTW_DPRINTF(RTW_DEBUG_LED,
("%s: read %" PRIx8 " from reg[CONFIG1]\n", __func__, cfg1));
cfg1 &= ~RTW_CONFIG1_LEDS_MASK;
cfg1 |= RTW_CONFIG1_LEDS_TX_RX;
RTW_WRITE8(regs, RTW_CONFIG1, cfg1);
rtw_set_access(regs, RTW_ACCESS_NONE);
}
nuke trailing whitespace
2005-02-27 03:26:58 +03:00
/*
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
* IEEE80211_S_INIT: LED1 off
*
* IEEE80211_S_AUTH,
* IEEE80211_S_ASSOC,
* IEEE80211_S_SCAN: LED1 blinks @ 1 Hz, blinks at 5Hz for tx/rx
*
* IEEE80211_S_RUN: LED1 on, blinks @ 5Hz for tx/rx
*/
static void
rtw_led_newstate(struct rtw_softc *sc, enum ieee80211_state nstate)
{
struct rtw_led_state *ls;
ls = &sc->sc_led_state;
switch (nstate) {
case IEEE80211_S_INIT:
rtw_led_init(&sc->sc_regs);
Use device_t and its accessors throughout. Use aprint_*_dev(). Improve PMF-ability. Add a 'flags' argument to suspend/resume handlers and callers such as pmf_system_suspend(). Define a flag, PMF_F_SELF, which indicates to PMF that a device is suspending/resuming itself. Add helper routines, pmf_device_suspend_self(dev) and pmf_device_resume_self(dev), that call pmf_device_suspend(dev, PMF_F_SELF) and pmf_device_resume(dev, PMF_F_SELF), respectively. Use PMF_F_SELF to suspend/resume self in ath(4), audio(4), rtw(4), and sip(4). In ath(4) and in rtw(4), replace the icky sc_enable/sc_disable callbacks, provided by the bus front-end, with self-suspension/resumption. Also, clean up the bus front-ends. Make sure that the interrupt handler is disestablished during suspension. Get rid of driver-private flags (e.g., RTW_F_ENABLED, ath_softc->sc_invalid); use device_is_active()/device_has_power() calls, instead. In the network-class suspend handler, call if_stop(, 0) instead of if_stop(, 1), because the latter is superfluous (bus- and driver-suspension hooks will 'disable' the NIC), and it may cause recursion. In the network-class resume handler, prevent infinite recursion through if_init() by getting out early if we are self-suspending (PMF_F_SELF). rtw(4) improvements: Destroy rtw(4) callouts when we detach it. Make rtw at pci detachable. Print some more information with the "rx frame too long" warning. Remove activate() methods: Get rid of rtw_activate() and ath_activate(). The device activate() methods are not good for much these days. Make ath at cardbus resume with crypto functions intact: Introduce a boolean device property, "pmf-powerdown". If pmf-powerdown is present and false, it indicates that a bus back-end should not remove power from a device. Honor this property in cardbus_child_suspend(). Set this property to 'false' in ath_attach(), since removing power from an ath at cardbus seems to lobotomize the WPA crypto engine. XXX Should the pmf-powerdown property propagate toward the root of the device tree? Miscellaneous ath(4) changes: Warn if ath(4) tries to write crypto keys to suspended hardware. Reduce differences between FreeBSD and NetBSD in ath(4) multicast filter setup. Make ath_printrxbuf() print an rx descriptor's status & key index, to help debug crypto errors. Shorten a staircase in ath_ioctl(). Don't check for ieee80211_ioctl() return code ERESTART, it never happens.
2008-03-12 21:02:21 +03:00
aprint_debug_dev(sc->sc_dev, "stopping blink\n");
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
callout_stop(&ls->ls_slow_ch);
callout_stop(&ls->ls_fast_ch);
ls->ls_slowblink = 0;
ls->ls_actblink = 0;
ls->ls_default = 0;
break;
case IEEE80211_S_SCAN:
Use device_t and its accessors throughout. Use aprint_*_dev(). Improve PMF-ability. Add a 'flags' argument to suspend/resume handlers and callers such as pmf_system_suspend(). Define a flag, PMF_F_SELF, which indicates to PMF that a device is suspending/resuming itself. Add helper routines, pmf_device_suspend_self(dev) and pmf_device_resume_self(dev), that call pmf_device_suspend(dev, PMF_F_SELF) and pmf_device_resume(dev, PMF_F_SELF), respectively. Use PMF_F_SELF to suspend/resume self in ath(4), audio(4), rtw(4), and sip(4). In ath(4) and in rtw(4), replace the icky sc_enable/sc_disable callbacks, provided by the bus front-end, with self-suspension/resumption. Also, clean up the bus front-ends. Make sure that the interrupt handler is disestablished during suspension. Get rid of driver-private flags (e.g., RTW_F_ENABLED, ath_softc->sc_invalid); use device_is_active()/device_has_power() calls, instead. In the network-class suspend handler, call if_stop(, 0) instead of if_stop(, 1), because the latter is superfluous (bus- and driver-suspension hooks will 'disable' the NIC), and it may cause recursion. In the network-class resume handler, prevent infinite recursion through if_init() by getting out early if we are self-suspending (PMF_F_SELF). rtw(4) improvements: Destroy rtw(4) callouts when we detach it. Make rtw at pci detachable. Print some more information with the "rx frame too long" warning. Remove activate() methods: Get rid of rtw_activate() and ath_activate(). The device activate() methods are not good for much these days. Make ath at cardbus resume with crypto functions intact: Introduce a boolean device property, "pmf-powerdown". If pmf-powerdown is present and false, it indicates that a bus back-end should not remove power from a device. Honor this property in cardbus_child_suspend(). Set this property to 'false' in ath_attach(), since removing power from an ath at cardbus seems to lobotomize the WPA crypto engine. XXX Should the pmf-powerdown property propagate toward the root of the device tree? Miscellaneous ath(4) changes: Warn if ath(4) tries to write crypto keys to suspended hardware. Reduce differences between FreeBSD and NetBSD in ath(4) multicast filter setup. Make ath_printrxbuf() print an rx descriptor's status & key index, to help debug crypto errors. Shorten a staircase in ath_ioctl(). Don't check for ieee80211_ioctl() return code ERESTART, it never happens.
2008-03-12 21:02:21 +03:00
aprint_debug_dev(sc->sc_dev, "scheduling blink\n");
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
callout_schedule(&ls->ls_slow_ch, RTW_LED_SLOW_TICKS);
callout_schedule(&ls->ls_fast_ch, RTW_LED_FAST_TICKS);
/*FALLTHROUGH*/
case IEEE80211_S_AUTH:
case IEEE80211_S_ASSOC:
ls->ls_default = RTW_LED1;
ls->ls_actblink = RTW_LED1;
ls->ls_slowblink = RTW_LED1;
break;
case IEEE80211_S_RUN:
ls->ls_slowblink = 0;
break;
}
rtw_led_set(ls, &sc->sc_regs, sc->sc_hwverid);
}
static void
rtw_led_set(struct rtw_led_state *ls, struct rtw_regs *regs, int hwverid)
{
uint8_t led_condition;
bus_size_t ofs;
uint8_t mask, newval, val;
led_condition = ls->ls_default;
if (ls->ls_state & RTW_LED_S_SLOW)
led_condition ^= ls->ls_slowblink;
if (ls->ls_state & (RTW_LED_S_RX|RTW_LED_S_TX))
led_condition ^= ls->ls_actblink;
RTW_DPRINTF(RTW_DEBUG_LED,
("%s: LED condition %" PRIx8 "\n", __func__, led_condition));
switch (hwverid) {
default:
case 'F':
ofs = RTW_PSR;
newval = mask = RTW_PSR_LEDGPO0 | RTW_PSR_LEDGPO1;
if (led_condition & RTW_LED0)
newval &= ~RTW_PSR_LEDGPO0;
if (led_condition & RTW_LED1)
newval &= ~RTW_PSR_LEDGPO1;
break;
case 'D':
ofs = RTW_9346CR;
mask = RTW_9346CR_EEM_MASK | RTW_9346CR_EEDI | RTW_9346CR_EECS;
newval = RTW_9346CR_EEM_PROGRAM;
if (led_condition & RTW_LED0)
newval |= RTW_9346CR_EEDI;
if (led_condition & RTW_LED1)
newval |= RTW_9346CR_EECS;
break;
}
val = RTW_READ8(regs, ofs);
RTW_DPRINTF(RTW_DEBUG_LED,
Misc. changes to improve ad hoc mode and to enable hostap mode: Fix the work-around for the NIC bug where it skips to rx descriptor 0. The driver used to skip to rx descriptor 1. Hopefully this stops the out-of-order packet reception that Charles Hannum saw. When debugging is enabled, print rx-descriptor status flags before printing the rx bit rate. Add a debug message for when a beacon tx buffer reclamation. Reset IFF_OACTIVE when we reset the transmitter. Pass the consolidated LED state, a struct rtw_led_state, to rtw_led_attach. Choose the bit-rate for management frames (1Mb/s) at the same place we choose for all other frames. Do not use the NIC's short preamble or RTS options for management frames. Label beacons for the NIC. Following a Linux driver, take care not to zero arbitrary bits in the TPPOLL register. Use the new idiom for IBSS merges: disable transmitter, kick the state machine. Add a second descriptor to the beacon ring. The NIC seems to like this much better.
2005-03-02 08:20:43 +03:00
("%s: read %" PRIx8 " from reg[%#02" PRIxPTR "]\n", __func__, val,
(uintptr_t)ofs));
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
val &= ~mask;
val |= newval;
RTW_WRITE8(regs, ofs, val);
RTW_DPRINTF(RTW_DEBUG_LED,
Misc. changes to improve ad hoc mode and to enable hostap mode: Fix the work-around for the NIC bug where it skips to rx descriptor 0. The driver used to skip to rx descriptor 1. Hopefully this stops the out-of-order packet reception that Charles Hannum saw. When debugging is enabled, print rx-descriptor status flags before printing the rx bit rate. Add a debug message for when a beacon tx buffer reclamation. Reset IFF_OACTIVE when we reset the transmitter. Pass the consolidated LED state, a struct rtw_led_state, to rtw_led_attach. Choose the bit-rate for management frames (1Mb/s) at the same place we choose for all other frames. Do not use the NIC's short preamble or RTS options for management frames. Label beacons for the NIC. Following a Linux driver, take care not to zero arbitrary bits in the TPPOLL register. Use the new idiom for IBSS merges: disable transmitter, kick the state machine. Add a second descriptor to the beacon ring. The NIC seems to like this much better.
2005-03-02 08:20:43 +03:00
("%s: wrote %" PRIx8 " to reg[%#02" PRIxPTR "]\n", __func__, val,
(uintptr_t)ofs));
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
RTW_SYNC(regs, ofs, ofs);
}
static void
rtw_led_fastblink(void *arg)
{
int ostate, s;
struct rtw_softc *sc = (struct rtw_softc *)arg;
struct rtw_led_state *ls = &sc->sc_led_state;
s = splnet();
ostate = ls->ls_state;
ls->ls_state ^= ls->ls_event;
if ((ls->ls_event & RTW_LED_S_TX) == 0)
ls->ls_state &= ~RTW_LED_S_TX;
if ((ls->ls_event & RTW_LED_S_RX) == 0)
ls->ls_state &= ~RTW_LED_S_RX;
ls->ls_event = 0;
if (ostate != ls->ls_state)
rtw_led_set(ls, &sc->sc_regs, sc->sc_hwverid);
splx(s);
Use device_t and its accessors throughout. Use aprint_*_dev(). Improve PMF-ability. Add a 'flags' argument to suspend/resume handlers and callers such as pmf_system_suspend(). Define a flag, PMF_F_SELF, which indicates to PMF that a device is suspending/resuming itself. Add helper routines, pmf_device_suspend_self(dev) and pmf_device_resume_self(dev), that call pmf_device_suspend(dev, PMF_F_SELF) and pmf_device_resume(dev, PMF_F_SELF), respectively. Use PMF_F_SELF to suspend/resume self in ath(4), audio(4), rtw(4), and sip(4). In ath(4) and in rtw(4), replace the icky sc_enable/sc_disable callbacks, provided by the bus front-end, with self-suspension/resumption. Also, clean up the bus front-ends. Make sure that the interrupt handler is disestablished during suspension. Get rid of driver-private flags (e.g., RTW_F_ENABLED, ath_softc->sc_invalid); use device_is_active()/device_has_power() calls, instead. In the network-class suspend handler, call if_stop(, 0) instead of if_stop(, 1), because the latter is superfluous (bus- and driver-suspension hooks will 'disable' the NIC), and it may cause recursion. In the network-class resume handler, prevent infinite recursion through if_init() by getting out early if we are self-suspending (PMF_F_SELF). rtw(4) improvements: Destroy rtw(4) callouts when we detach it. Make rtw at pci detachable. Print some more information with the "rx frame too long" warning. Remove activate() methods: Get rid of rtw_activate() and ath_activate(). The device activate() methods are not good for much these days. Make ath at cardbus resume with crypto functions intact: Introduce a boolean device property, "pmf-powerdown". If pmf-powerdown is present and false, it indicates that a bus back-end should not remove power from a device. Honor this property in cardbus_child_suspend(). Set this property to 'false' in ath_attach(), since removing power from an ath at cardbus seems to lobotomize the WPA crypto engine. XXX Should the pmf-powerdown property propagate toward the root of the device tree? Miscellaneous ath(4) changes: Warn if ath(4) tries to write crypto keys to suspended hardware. Reduce differences between FreeBSD and NetBSD in ath(4) multicast filter setup. Make ath_printrxbuf() print an rx descriptor's status & key index, to help debug crypto errors. Shorten a staircase in ath_ioctl(). Don't check for ieee80211_ioctl() return code ERESTART, it never happens.
2008-03-12 21:02:21 +03:00
aprint_debug_dev(sc->sc_dev, "scheduling fast blink\n");
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
callout_schedule(&ls->ls_fast_ch, RTW_LED_FAST_TICKS);
}
static void
rtw_led_slowblink(void *arg)
{
int s;
struct rtw_softc *sc = (struct rtw_softc *)arg;
struct rtw_led_state *ls = &sc->sc_led_state;
s = splnet();
ls->ls_state ^= RTW_LED_S_SLOW;
rtw_led_set(ls, &sc->sc_regs, sc->sc_hwverid);
splx(s);
Use device_t and its accessors throughout. Use aprint_*_dev(). Improve PMF-ability. Add a 'flags' argument to suspend/resume handlers and callers such as pmf_system_suspend(). Define a flag, PMF_F_SELF, which indicates to PMF that a device is suspending/resuming itself. Add helper routines, pmf_device_suspend_self(dev) and pmf_device_resume_self(dev), that call pmf_device_suspend(dev, PMF_F_SELF) and pmf_device_resume(dev, PMF_F_SELF), respectively. Use PMF_F_SELF to suspend/resume self in ath(4), audio(4), rtw(4), and sip(4). In ath(4) and in rtw(4), replace the icky sc_enable/sc_disable callbacks, provided by the bus front-end, with self-suspension/resumption. Also, clean up the bus front-ends. Make sure that the interrupt handler is disestablished during suspension. Get rid of driver-private flags (e.g., RTW_F_ENABLED, ath_softc->sc_invalid); use device_is_active()/device_has_power() calls, instead. In the network-class suspend handler, call if_stop(, 0) instead of if_stop(, 1), because the latter is superfluous (bus- and driver-suspension hooks will 'disable' the NIC), and it may cause recursion. In the network-class resume handler, prevent infinite recursion through if_init() by getting out early if we are self-suspending (PMF_F_SELF). rtw(4) improvements: Destroy rtw(4) callouts when we detach it. Make rtw at pci detachable. Print some more information with the "rx frame too long" warning. Remove activate() methods: Get rid of rtw_activate() and ath_activate(). The device activate() methods are not good for much these days. Make ath at cardbus resume with crypto functions intact: Introduce a boolean device property, "pmf-powerdown". If pmf-powerdown is present and false, it indicates that a bus back-end should not remove power from a device. Honor this property in cardbus_child_suspend(). Set this property to 'false' in ath_attach(), since removing power from an ath at cardbus seems to lobotomize the WPA crypto engine. XXX Should the pmf-powerdown property propagate toward the root of the device tree? Miscellaneous ath(4) changes: Warn if ath(4) tries to write crypto keys to suspended hardware. Reduce differences between FreeBSD and NetBSD in ath(4) multicast filter setup. Make ath_printrxbuf() print an rx descriptor's status & key index, to help debug crypto errors. Shorten a staircase in ath_ioctl(). Don't check for ieee80211_ioctl() return code ERESTART, it never happens.
2008-03-12 21:02:21 +03:00
aprint_debug_dev(sc->sc_dev, "scheduling slow blink\n");
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
callout_schedule(&ls->ls_slow_ch, RTW_LED_SLOW_TICKS);
}
Use device_t and its accessors throughout. Use aprint_*_dev(). Improve PMF-ability. Add a 'flags' argument to suspend/resume handlers and callers such as pmf_system_suspend(). Define a flag, PMF_F_SELF, which indicates to PMF that a device is suspending/resuming itself. Add helper routines, pmf_device_suspend_self(dev) and pmf_device_resume_self(dev), that call pmf_device_suspend(dev, PMF_F_SELF) and pmf_device_resume(dev, PMF_F_SELF), respectively. Use PMF_F_SELF to suspend/resume self in ath(4), audio(4), rtw(4), and sip(4). In ath(4) and in rtw(4), replace the icky sc_enable/sc_disable callbacks, provided by the bus front-end, with self-suspension/resumption. Also, clean up the bus front-ends. Make sure that the interrupt handler is disestablished during suspension. Get rid of driver-private flags (e.g., RTW_F_ENABLED, ath_softc->sc_invalid); use device_is_active()/device_has_power() calls, instead. In the network-class suspend handler, call if_stop(, 0) instead of if_stop(, 1), because the latter is superfluous (bus- and driver-suspension hooks will 'disable' the NIC), and it may cause recursion. In the network-class resume handler, prevent infinite recursion through if_init() by getting out early if we are self-suspending (PMF_F_SELF). rtw(4) improvements: Destroy rtw(4) callouts when we detach it. Make rtw at pci detachable. Print some more information with the "rx frame too long" warning. Remove activate() methods: Get rid of rtw_activate() and ath_activate(). The device activate() methods are not good for much these days. Make ath at cardbus resume with crypto functions intact: Introduce a boolean device property, "pmf-powerdown". If pmf-powerdown is present and false, it indicates that a bus back-end should not remove power from a device. Honor this property in cardbus_child_suspend(). Set this property to 'false' in ath_attach(), since removing power from an ath at cardbus seems to lobotomize the WPA crypto engine. XXX Should the pmf-powerdown property propagate toward the root of the device tree? Miscellaneous ath(4) changes: Warn if ath(4) tries to write crypto keys to suspended hardware. Reduce differences between FreeBSD and NetBSD in ath(4) multicast filter setup. Make ath_printrxbuf() print an rx descriptor's status & key index, to help debug crypto errors. Shorten a staircase in ath_ioctl(). Don't check for ieee80211_ioctl() return code ERESTART, it never happens.
2008-03-12 21:02:21 +03:00
static void
rtw_led_detach(struct rtw_led_state *ls)
{
callout_destroy(&ls->ls_fast_ch);
callout_destroy(&ls->ls_slow_ch);
}
static void
Misc. changes to improve ad hoc mode and to enable hostap mode: Fix the work-around for the NIC bug where it skips to rx descriptor 0. The driver used to skip to rx descriptor 1. Hopefully this stops the out-of-order packet reception that Charles Hannum saw. When debugging is enabled, print rx-descriptor status flags before printing the rx bit rate. Add a debug message for when a beacon tx buffer reclamation. Reset IFF_OACTIVE when we reset the transmitter. Pass the consolidated LED state, a struct rtw_led_state, to rtw_led_attach. Choose the bit-rate for management frames (1Mb/s) at the same place we choose for all other frames. Do not use the NIC's short preamble or RTS options for management frames. Label beacons for the NIC. Following a Linux driver, take care not to zero arbitrary bits in the TPPOLL register. Use the new idiom for IBSS merges: disable transmitter, kick the state machine. Add a second descriptor to the beacon ring. The NIC seems to like this much better.
2005-03-02 08:20:43 +03:00
rtw_led_attach(struct rtw_led_state *ls, void *arg)
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
{
Merge some of the less invasive changes from the vmlocking branch: - kthread, callout, devsw API changes - select()/poll() improvements - miscellaneous MT safety improvements
2007-07-10 00:51:58 +04:00
callout_init(&ls->ls_fast_ch, 0);
callout_init(&ls->ls_slow_ch, 0);
Misc. changes to improve ad hoc mode and to enable hostap mode: Fix the work-around for the NIC bug where it skips to rx descriptor 0. The driver used to skip to rx descriptor 1. Hopefully this stops the out-of-order packet reception that Charles Hannum saw. When debugging is enabled, print rx-descriptor status flags before printing the rx bit rate. Add a debug message for when a beacon tx buffer reclamation. Reset IFF_OACTIVE when we reset the transmitter. Pass the consolidated LED state, a struct rtw_led_state, to rtw_led_attach. Choose the bit-rate for management frames (1Mb/s) at the same place we choose for all other frames. Do not use the NIC's short preamble or RTS options for management frames. Label beacons for the NIC. Following a Linux driver, take care not to zero arbitrary bits in the TPPOLL register. Use the new idiom for IBSS merges: disable transmitter, kick the state machine. Add a second descriptor to the beacon ring. The NIC seems to like this much better.
2005-03-02 08:20:43 +03:00
callout_setfunc(&ls->ls_fast_ch, rtw_led_fastblink, arg);
callout_setfunc(&ls->ls_slow_ch, rtw_led_slowblink, arg);
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
}
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
static int
Kill caddr_t; there will be some MI fallout, but it will be fixed shortly.
2007-03-04 08:59:00 +03:00
rtw_ioctl(struct ifnet *ifp, u_long cmd, void *data)
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
{
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
int rc = 0, s;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
struct rtw_softc *sc = ifp->if_softc;
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
s = splnet();
Simplify multicast filter programming, and take care to avoid gratuitous device resets when the kernel adds/deletes a multicast address.
2007-05-29 22:32:07 +04:00
if (cmd == SIOCSIFFLAGS) {
*** Summary *** When a link-layer address changes (e.g., ifconfig ex0 link 02:de:ad:be:ef:02 active), send a gratuitous ARP and/or a Neighbor Advertisement to update the network-/link-layer address bindings on our LAN peers. Refuse a change of ethernet address to the address 00:00:00:00:00:00 or to any multicast/broadcast address. (Thanks matt@.) Reorder ifnet ioctl operations so that driver ioctls may inherit the functions of their "class"---ether_ioctl(), fddi_ioctl(), et cetera---and the class ioctls may inherit from the generic ioctl, ifioctl_common(), but both driver- and class-ioctls may override the generic behavior. Make network drivers share more code. Distinguish a "factory" link-layer address from others for the purposes of both protecting that address from deletion and computing EUI64. Return consistent, appropriate error codes from network drivers. Improve readability. KNF. *** Details *** In if_attach(), always initialize the interface ioctl routine, ifnet->if_ioctl, if the driver has not already initialized it. Delete if_ioctl == NULL tests everywhere else, because it cannot happen. In the ioctl routines of network interfaces, inherit common ioctl behaviors by calling either ifioctl_common() or whichever ioctl routine is appropriate for the class of interface---e.g., ether_ioctl() for ethernets. Stop (ab)using SIOCSIFADDR and start to use SIOCINITIFADDR. In the user->kernel interface, SIOCSIFADDR's argument was an ifreq, but on the protocol->ifnet interface, SIOCSIFADDR's argument was an ifaddr. That was confusing, and it would work against me as I make it possible for a network interface to overload most ioctls. On the protocol->ifnet interface, replace SIOCSIFADDR with SIOCINITIFADDR. In ifioctl(), return EPERM if userland tries to invoke SIOCINITIFADDR. In ifioctl(), give the interface the first shot at handling most interface ioctls, and give the protocol the second shot, instead of the other way around. Finally, let compatibility code (COMPAT_OSOCK) take a shot. Pull device initialization out of switch statements under SIOCINITIFADDR. For example, pull ..._init() out of any switch statement that looks like this: switch (...->sa_family) { case ...: ..._init(); ... break; ... default: ..._init(); ... break; } Rewrite many if-else clauses that handle all permutations of IFF_UP and IFF_RUNNING to use a switch statement, switch (x & (IFF_UP|IFF_RUNNING)) { case 0: ... break; case IFF_RUNNING: ... break; case IFF_UP: ... break; case IFF_UP|IFF_RUNNING: ... break; } unifdef lots of code containing #ifdef FreeBSD, #ifdef NetBSD, and #ifdef SIOCSIFMTU, especially in fwip(4) and in ndis(4). In ipw(4), remove an if_set_sadl() call that is out of place. In nfe(4), reuse the jumbo MTU logic in ether_ioctl(). Let ethernets register a callback for setting h/w state such as promiscuous mode and the multicast filter in accord with a change in the if_flags: ether_set_ifflags_cb() registers a callback that returns ENETRESET if the caller should reset the ethernet by calling if_init(), 0 on success, != 0 on failure. Pull common code from ex(4), gem(4), nfe(4), sip(4), tlp(4), vge(4) into ether_ioctl(), and register if_flags callbacks for those drivers. Return ENOTTY instead of EINVAL for inappropriate ioctls. In zyd(4), use ENXIO instead of ENOTTY to indicate that the device is not any longer attached. Add to if_set_sadl() a boolean 'factory' argument that indicates whether a link-layer address was assigned by the factory or some other source. In a comment, recommend using the factory address for generating an EUI64, and update in6_get_hw_ifid() to prefer a factory address to any other link-layer address. Add a routing message, RTM_LLINFO_UPD, that tells protocols to update the binding of network-layer addresses to link-layer addresses. Implement this message in IPv4 and IPv6 by sending a gratuitous ARP or a neighbor advertisement, respectively. Generate RTM_LLINFO_UPD messages on a change of an interface's link-layer address. In ether_ioctl(), do not let SIOCALIFADDR set a link-layer address that is broadcast/multicast or equal to 00:00:00:00:00:00. Make ether_ioctl() call ifioctl_common() to handle ioctls that it does not understand. In gif(4), initialize if_softc and use it, instead of assuming that the gif_softc and ifp overlap. Let ifioctl_common() handle SIOCGIFADDR. Sprinkle rtcache_invariants(), which checks on DIAGNOSTIC kernels that certain invariants on a struct route are satisfied. In agr(4), rewrite agr_ioctl_filter() to be a bit more explicit about the ioctls that we do not allow on an agr(4) member interface. bzero -> memset. Delete unnecessary casts to void *. Use sockaddr_in_init() and sockaddr_in6_init(). Compare pointers with NULL instead of "testing truth". Replace some instances of (type *)0 with NULL. Change some K&R prototypes to ANSI C, and join lines.
2008-11-07 03:20:01 +03:00
if ((rc = ifioctl_common(ifp, cmd, data)) != 0)
;
In pmf(9), improve the implementation of device self-suspension and make suspension by self, by drvctl(8), and by ACPI system sleep play nice together. Start solidifying some temporary API changes. 1. Extract a new header file, <sys/device_if.h>, from <sys/device.h> and #include it from <sys/pmf.h> instead of <sys/device.h> to break the circular dependency between <sys/device.h> and <sys/pmf.h>. 2. Introduce pmf_qual_t, an aggregate of qualifications on a PMF suspend/resume call. Start to replace instances of PMF_FN_PROTO, PMF_FN_ARGS, et cetera, with a pmf_qual_t. 3. Introduce the notion of a "suspensor," an entity that holds a device in suspension. More than one suspensor may hold a device at once. A device stays suspended as long as at least one suspensor holds it. A device resumes when the last suspensor releases it. Currently, the kernel defines three suspensors, 3a the system-suspensor: for system suspension, initiated by 'sysctl -w machdep.sleep_state=3', by lid closure, by power-button press, et cetera, 3b the drvctl-suspensor: for device suspension by /dev/drvctl ioctl, e.g., drvctl -S sip0. 3c the system self-suspensor: for device drivers that suspend themselves and their children. Several drivers for network interfaces put the network device to sleep while it is not administratively up, that is, after the kernel calls if_stop(, 1). The self-suspensor should not be used directly. See the description of suspensor delegates, below. A suspensor can have one or more "delegates". A suspensor can release devices that its delegates hold suspended. Right now, only the system self-suspensor has delegates. For each device that a self-suspending driver attaches, it creates the device's self-suspensor, a delegate of the system self-suspensor. Suspensors stop a system-wide suspend/resume cycle from waking devices that the operator put to sleep with drvctl before the cycle. They also help self-suspension to work more simply, safely, and in accord with expectations. 4. Add the notion of device activation level, devact_level_t, and a routine for checking the current activation level, device_activation(). Current activation levels are DEVACT_LEVEL_BUS, DEVACT_LEVEL_DRIVER, and DEVACT_LEVEL_CLASS, which respectively indicate that the device's bus is active, that the bus and device are active, and that the bus, device, and the functions of the device's class (network, audio) are active. Suspend/resume calls can be qualified with a devact_level_t. The power-management framework treats a devact_level_t that qualifies a device suspension as the device's current activation level; it only runs hooks to reduce the activation level from the presumed current level to the fully suspended state. The framework treats a devact_level_t qualifying device resumption as the target activation level; it only runs hooks to raise the activation level to the target. 5. Use pmf_qual_t, devact_level_t, and self-suspensors in several drivers. 6. Temporarily add an unused power-management workqueue that I will remove or replace, soon.
2009-09-16 20:34:49 +04:00
else switch (ifp->if_flags & (IFF_UP|IFF_RUNNING)) {
case IFF_UP:
rc = rtw_init(ifp);
RTW_PRINT_REGS(&sc->sc_regs, ifp->if_xname, __func__);
break;
case IFF_UP|IFF_RUNNING:
if (device_activation(sc->sc_dev, DEVACT_LEVEL_DRIVER))
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
rtw_pktfilt_load(sc);
sys/dev/ic/rtw.c: Cosmetic: remove two short ladders from rtw_init. Fit the NIC registers print-out onto one line using the RTW_PRINT_REGS macro. sys/dev/ic/rtwvar.h: Add RTW_PRINT_REGS wrapper macro for rtw_print_regs. It is a null op unless defined(RTW_DEBUG).
2005-01-03 06:25:06 +03:00
RTW_PRINT_REGS(&sc->sc_regs, ifp->if_xname, __func__);
In pmf(9), improve the implementation of device self-suspension and make suspension by self, by drvctl(8), and by ACPI system sleep play nice together. Start solidifying some temporary API changes. 1. Extract a new header file, <sys/device_if.h>, from <sys/device.h> and #include it from <sys/pmf.h> instead of <sys/device.h> to break the circular dependency between <sys/device.h> and <sys/pmf.h>. 2. Introduce pmf_qual_t, an aggregate of qualifications on a PMF suspend/resume call. Start to replace instances of PMF_FN_PROTO, PMF_FN_ARGS, et cetera, with a pmf_qual_t. 3. Introduce the notion of a "suspensor," an entity that holds a device in suspension. More than one suspensor may hold a device at once. A device stays suspended as long as at least one suspensor holds it. A device resumes when the last suspensor releases it. Currently, the kernel defines three suspensors, 3a the system-suspensor: for system suspension, initiated by 'sysctl -w machdep.sleep_state=3', by lid closure, by power-button press, et cetera, 3b the drvctl-suspensor: for device suspension by /dev/drvctl ioctl, e.g., drvctl -S sip0. 3c the system self-suspensor: for device drivers that suspend themselves and their children. Several drivers for network interfaces put the network device to sleep while it is not administratively up, that is, after the kernel calls if_stop(, 1). The self-suspensor should not be used directly. See the description of suspensor delegates, below. A suspensor can have one or more "delegates". A suspensor can release devices that its delegates hold suspended. Right now, only the system self-suspensor has delegates. For each device that a self-suspending driver attaches, it creates the device's self-suspensor, a delegate of the system self-suspensor. Suspensors stop a system-wide suspend/resume cycle from waking devices that the operator put to sleep with drvctl before the cycle. They also help self-suspension to work more simply, safely, and in accord with expectations. 4. Add the notion of device activation level, devact_level_t, and a routine for checking the current activation level, device_activation(). Current activation levels are DEVACT_LEVEL_BUS, DEVACT_LEVEL_DRIVER, and DEVACT_LEVEL_CLASS, which respectively indicate that the device's bus is active, that the bus and device are active, and that the bus, device, and the functions of the device's class (network, audio) are active. Suspend/resume calls can be qualified with a devact_level_t. The power-management framework treats a devact_level_t that qualifies a device suspension as the device's current activation level; it only runs hooks to reduce the activation level from the presumed current level to the fully suspended state. The framework treats a devact_level_t qualifying device resumption as the target activation level; it only runs hooks to raise the activation level to the target. 5. Use pmf_qual_t, devact_level_t, and self-suspensors in several drivers. 6. Temporarily add an unused power-management workqueue that I will remove or replace, soon.
2009-09-16 20:34:49 +04:00
break;
case IFF_RUNNING:
sys/dev/ic/rtw.c: Cosmetic: remove two short ladders from rtw_init. Fit the NIC registers print-out onto one line using the RTW_PRINT_REGS macro. sys/dev/ic/rtwvar.h: Add RTW_PRINT_REGS wrapper macro for rtw_print_regs. It is a null op unless defined(RTW_DEBUG).
2005-01-03 06:25:06 +03:00
RTW_PRINT_REGS(&sc->sc_regs, ifp->if_xname, __func__);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
rtw_stop(ifp, 1);
In pmf(9), improve the implementation of device self-suspension and make suspension by self, by drvctl(8), and by ACPI system sleep play nice together. Start solidifying some temporary API changes. 1. Extract a new header file, <sys/device_if.h>, from <sys/device.h> and #include it from <sys/pmf.h> instead of <sys/device.h> to break the circular dependency between <sys/device.h> and <sys/pmf.h>. 2. Introduce pmf_qual_t, an aggregate of qualifications on a PMF suspend/resume call. Start to replace instances of PMF_FN_PROTO, PMF_FN_ARGS, et cetera, with a pmf_qual_t. 3. Introduce the notion of a "suspensor," an entity that holds a device in suspension. More than one suspensor may hold a device at once. A device stays suspended as long as at least one suspensor holds it. A device resumes when the last suspensor releases it. Currently, the kernel defines three suspensors, 3a the system-suspensor: for system suspension, initiated by 'sysctl -w machdep.sleep_state=3', by lid closure, by power-button press, et cetera, 3b the drvctl-suspensor: for device suspension by /dev/drvctl ioctl, e.g., drvctl -S sip0. 3c the system self-suspensor: for device drivers that suspend themselves and their children. Several drivers for network interfaces put the network device to sleep while it is not administratively up, that is, after the kernel calls if_stop(, 1). The self-suspensor should not be used directly. See the description of suspensor delegates, below. A suspensor can have one or more "delegates". A suspensor can release devices that its delegates hold suspended. Right now, only the system self-suspensor has delegates. For each device that a self-suspending driver attaches, it creates the device's self-suspensor, a delegate of the system self-suspensor. Suspensors stop a system-wide suspend/resume cycle from waking devices that the operator put to sleep with drvctl before the cycle. They also help self-suspension to work more simply, safely, and in accord with expectations. 4. Add the notion of device activation level, devact_level_t, and a routine for checking the current activation level, device_activation(). Current activation levels are DEVACT_LEVEL_BUS, DEVACT_LEVEL_DRIVER, and DEVACT_LEVEL_CLASS, which respectively indicate that the device's bus is active, that the bus and device are active, and that the bus, device, and the functions of the device's class (network, audio) are active. Suspend/resume calls can be qualified with a devact_level_t. The power-management framework treats a devact_level_t that qualifies a device suspension as the device's current activation level; it only runs hooks to reduce the activation level from the presumed current level to the fully suspended state. The framework treats a devact_level_t qualifying device resumption as the target activation level; it only runs hooks to raise the activation level to the target. 5. Use pmf_qual_t, devact_level_t, and self-suspensors in several drivers. 6. Temporarily add an unused power-management workqueue that I will remove or replace, soon.
2009-09-16 20:34:49 +04:00
break;
default:
break;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
}
Simplify multicast filter programming, and take care to avoid gratuitous device resets when the kernel adds/deletes a multicast address.
2007-05-29 22:32:07 +04:00
} else if ((rc = ieee80211_ioctl(&sc->sc_ic, cmd, data)) != ENETRESET)
; /* nothing to do */
else if (cmd == SIOCADDMULTI || cmd == SIOCDELMULTI) {
/* reload packet filter if running */
sys/dev/ic/rtw.c: Cosmetic: remove two short ladders from rtw_init. Fit the NIC registers print-out onto one line using the RTW_PRINT_REGS macro. sys/dev/ic/rtwvar.h: Add RTW_PRINT_REGS wrapper macro for rtw_print_regs. It is a null op unless defined(RTW_DEBUG).
2005-01-03 06:25:06 +03:00
if (ifp->if_flags & IFF_RUNNING)
rtw_pktfilt_load(sc);
rc = 0;
Use device_t and its accessors throughout. Use aprint_*_dev(). Improve PMF-ability. Add a 'flags' argument to suspend/resume handlers and callers such as pmf_system_suspend(). Define a flag, PMF_F_SELF, which indicates to PMF that a device is suspending/resuming itself. Add helper routines, pmf_device_suspend_self(dev) and pmf_device_resume_self(dev), that call pmf_device_suspend(dev, PMF_F_SELF) and pmf_device_resume(dev, PMF_F_SELF), respectively. Use PMF_F_SELF to suspend/resume self in ath(4), audio(4), rtw(4), and sip(4). In ath(4) and in rtw(4), replace the icky sc_enable/sc_disable callbacks, provided by the bus front-end, with self-suspension/resumption. Also, clean up the bus front-ends. Make sure that the interrupt handler is disestablished during suspension. Get rid of driver-private flags (e.g., RTW_F_ENABLED, ath_softc->sc_invalid); use device_is_active()/device_has_power() calls, instead. In the network-class suspend handler, call if_stop(, 0) instead of if_stop(, 1), because the latter is superfluous (bus- and driver-suspension hooks will 'disable' the NIC), and it may cause recursion. In the network-class resume handler, prevent infinite recursion through if_init() by getting out early if we are self-suspending (PMF_F_SELF). rtw(4) improvements: Destroy rtw(4) callouts when we detach it. Make rtw at pci detachable. Print some more information with the "rx frame too long" warning. Remove activate() methods: Get rid of rtw_activate() and ath_activate(). The device activate() methods are not good for much these days. Make ath at cardbus resume with crypto functions intact: Introduce a boolean device property, "pmf-powerdown". If pmf-powerdown is present and false, it indicates that a bus back-end should not remove power from a device. Honor this property in cardbus_child_suspend(). Set this property to 'false' in ath_attach(), since removing power from an ath at cardbus seems to lobotomize the WPA crypto engine. XXX Should the pmf-powerdown property propagate toward the root of the device tree? Miscellaneous ath(4) changes: Warn if ath(4) tries to write crypto keys to suspended hardware. Reduce differences between FreeBSD and NetBSD in ath(4) multicast filter setup. Make ath_printrxbuf() print an rx descriptor's status & key index, to help debug crypto errors. Shorten a staircase in ath_ioctl(). Don't check for ieee80211_ioctl() return code ERESTART, it never happens.
2008-03-12 21:02:21 +03:00
} else if ((ifp->if_flags & IFF_UP) != 0)
Simplify multicast filter programming, and take care to avoid gratuitous device resets when the kernel adds/deletes a multicast address.
2007-05-29 22:32:07 +04:00
rc = rtw_init(ifp);
else
rc = 0;
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
splx(s);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
return rc;
}
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
/* Select a transmit ring with at least one h/w and s/w descriptor free.
* Return 0 on success, -1 on failure.
*/
Remove leading __ from __(const|inline|signed|volatile) -- it is obsolete.
2005-12-24 23:27:29 +03:00
static inline int
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
rtw_txring_choose(struct rtw_softc *sc, struct rtw_txsoft_blk **tsbp,
struct rtw_txdesc_blk **tdbp, int pri)
{
struct rtw_txsoft_blk *tsb;
struct rtw_txdesc_blk *tdb;
KASSERT(pri >= 0 && pri < RTW_NTXPRI);
tsb = &sc->sc_txsoft_blk[pri];
tdb = &sc->sc_txdesc_blk[pri];
if (SIMPLEQ_EMPTY(&tsb->tsb_freeq) || tdb->tdb_nfree == 0) {
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
if (tsb->tsb_tx_timer == 0)
tsb->tsb_tx_timer = 5;
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
*tsbp = NULL;
*tdbp = NULL;
return -1;
}
*tsbp = tsb;
*tdbp = tdb;
return 0;
}
Remove leading __ from __(const|inline|signed|volatile) -- it is obsolete.
2005-12-24 23:27:29 +03:00
static inline struct mbuf *
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
rtw_80211_dequeue(struct rtw_softc *sc, struct ifqueue *ifq, int pri,
struct rtw_txsoft_blk **tsbp, struct rtw_txdesc_blk **tdbp,
struct ieee80211_node **nip, short *if_flagsp)
{
struct mbuf *m;
if (IF_IS_EMPTY(ifq))
return NULL;
if (rtw_txring_choose(sc, tsbp, tdbp, pri) == -1) {
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
DPRINTF(sc, RTW_DEBUG_XMIT_RSRC, ("%s: no ring %d descriptor\n",
__func__, pri));
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
*if_flagsp |= IFF_OACTIVE;
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
sc->sc_if.if_timer = 1;
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
return NULL;
}
IF_DEQUEUE(ifq, m);
*nip = (struct ieee80211_node *)m->m_pkthdr.rcvif;
m->m_pkthdr.rcvif = NULL;
Resolve conflicts in importation of 18-May-2005 ath(4) / net80211(9) from FreeBSD. Introduce compatibility shims (sys/dev/ic/ath_netbsd.[ch], sys/net80211/ieee80211_netbsd.[ch]). Update drivers (an, atu, atw, awi, ipw, iwi, rtw, wi) for the new net80211(9) API.
2005-06-22 10:14:51 +04:00
KASSERT(*nip != NULL);
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
return m;
}
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
/* Point *mp at the next 802.11 frame to transmit. Point *tsbp
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
* at the driver's selection of transmit control block for the packet.
*/
Remove leading __ from __(const|inline|signed|volatile) -- it is obsolete.
2005-12-24 23:27:29 +03:00
static inline int
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
rtw_dequeue(struct ifnet *ifp, struct rtw_txsoft_blk **tsbp,
struct rtw_txdesc_blk **tdbp, struct mbuf **mp,
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
struct ieee80211_node **nip)
{
Resolve conflicts in importation of 18-May-2005 ath(4) / net80211(9) from FreeBSD. Introduce compatibility shims (sys/dev/ic/ath_netbsd.[ch], sys/net80211/ieee80211_netbsd.[ch]). Update drivers (an, atu, atw, awi, ipw, iwi, rtw, wi) for the new net80211(9) API.
2005-06-22 10:14:51 +04:00
int pri;
struct ether_header *eh;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
struct mbuf *m0;
struct rtw_softc *sc;
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
short *if_flagsp;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
In rtw_dequeue(), initialize *mp to NULL before anything else, so that it is either NULL or a valid mbuf. Previously, it was possible for rtw_dequeue() to exit with 0 before initializing *mp. rtw_start() would have used whatever trash was in *mp, which would lead to problems later on.
2007-01-29 10:08:09 +03:00
*mp = NULL;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
sc = (struct rtw_softc *)ifp->if_softc;
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
DPRINTF(sc, RTW_DEBUG_XMIT,
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
("%s: enter %s\n", device_xname(sc->sc_dev), __func__));
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
if_flagsp = &ifp->if_flags;
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
if (sc->sc_ic.ic_state == IEEE80211_S_RUN &&
(*mp = rtw_80211_dequeue(sc, &sc->sc_beaconq, RTW_TXPRIBCN, tsbp,
tdbp, nip, if_flagsp)) != NULL) {
DPRINTF(sc, RTW_DEBUG_XMIT, ("%s: dequeue beacon frame\n",
__func__));
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
return 0;
}
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
if ((*mp = rtw_80211_dequeue(sc, &sc->sc_ic.ic_mgtq, RTW_TXPRIMD, tsbp,
tdbp, nip, if_flagsp)) != NULL) {
DPRINTF(sc, RTW_DEBUG_XMIT, ("%s: dequeue mgt frame\n",
__func__));
return 0;
}
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
if (sc->sc_ic.ic_state != IEEE80211_S_RUN) {
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
DPRINTF(sc, RTW_DEBUG_XMIT, ("%s: not running\n", __func__));
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
return 0;
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
}
Resolve conflicts in importation of 18-May-2005 ath(4) / net80211(9) from FreeBSD. Introduce compatibility shims (sys/dev/ic/ath_netbsd.[ch], sys/net80211/ieee80211_netbsd.[ch]). Update drivers (an, atu, atw, awi, ipw, iwi, rtw, wi) for the new net80211(9) API.
2005-06-22 10:14:51 +04:00
IFQ_POLL(&ifp->if_snd, m0);
if (m0 == NULL) {
DPRINTF(sc, RTW_DEBUG_XMIT, ("%s: no frame ready\n",
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
__func__));
return 0;
}
Resolve conflicts in importation of 18-May-2005 ath(4) / net80211(9) from FreeBSD. Introduce compatibility shims (sys/dev/ic/ath_netbsd.[ch], sys/net80211/ieee80211_netbsd.[ch]). Update drivers (an, atu, atw, awi, ipw, iwi, rtw, wi) for the new net80211(9) API.
2005-06-22 10:14:51 +04:00
pri = ((m0->m_flags & M_PWR_SAV) != 0) ? RTW_TXPRIHI : RTW_TXPRIMD;
if (rtw_txring_choose(sc, tsbp, tdbp, pri) == -1) {
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
DPRINTF(sc, RTW_DEBUG_XMIT_RSRC, ("%s: no ring %d descriptor\n",
__func__, pri));
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
*if_flagsp |= IFF_OACTIVE;
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
sc->sc_if.if_timer = 1;
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
return 0;
}
IFQ_DEQUEUE(&ifp->if_snd, m0);
if (m0 == NULL) {
Resolve conflicts in importation of 18-May-2005 ath(4) / net80211(9) from FreeBSD. Introduce compatibility shims (sys/dev/ic/ath_netbsd.[ch], sys/net80211/ieee80211_netbsd.[ch]). Update drivers (an, atu, atw, awi, ipw, iwi, rtw, wi) for the new net80211(9) API.
2005-06-22 10:14:51 +04:00
DPRINTF(sc, RTW_DEBUG_XMIT, ("%s: no frame ready\n",
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
__func__));
return 0;
}
DPRINTF(sc, RTW_DEBUG_XMIT, ("%s: dequeue data frame\n", __func__));
ifp->if_opackets++;
Push the bpf_ops usage back into bpf.h. Push the common ifp->if_bpf check into the inline functions as well the fourth argument for bpf_attach.
2010-04-05 11:19:28 +04:00
bpf_mtap(ifp, m0);
Resolve conflicts in importation of 18-May-2005 ath(4) / net80211(9) from FreeBSD. Introduce compatibility shims (sys/dev/ic/ath_netbsd.[ch], sys/net80211/ieee80211_netbsd.[ch]). Update drivers (an, atu, atw, awi, ipw, iwi, rtw, wi) for the new net80211(9) API.
2005-06-22 10:14:51 +04:00
eh = mtod(m0, struct ether_header *);
*nip = ieee80211_find_txnode(&sc->sc_ic, eh->ether_dhost);
if (*nip == NULL) {
/* NB: ieee80211_find_txnode does stat+msg */
m_freem(m0);
return -1;
}
if ((m0 = ieee80211_encap(&sc->sc_ic, m0, *nip)) == NULL) {
DPRINTF(sc, RTW_DEBUG_XMIT, ("%s: encap error\n", __func__));
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
ifp->if_oerrors++;
return -1;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
}
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
DPRINTF(sc, RTW_DEBUG_XMIT, ("%s: leave\n", __func__));
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
*mp = m0;
return 0;
}
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
static int
rtw_seg_too_short(bus_dmamap_t dmamap)
{
int i;
for (i = 0; i < dmamap->dm_nsegs; i++) {
I have made a small change that suppresses "recovery" from (spurious?) tx timeouts on the beacon queue. The driver's recovery process badly disrupts the MAC's receiver. This needs a closer look, later. Every other fix is by Charles M. Hannum. Charles writes: "With these, the device is quite stable under load, does not experience xmit timeouts, no longer mysteriously freezes, and does not spew a lot of garbage. It does, however, continue to get some "tx fifo underflow"s. global, SYSCTL_SETUP(): a) Removed unused sysctl variables. rtw_io_enable(): b) Change rtw_io_enable to take a softc, and make it call rtw_txring_fixup(). c) Enable the PCI multiple read/write bit. d) Always reset rdb_next to 0 when starting the receive engine. rtw_intr_rx(): e) When processing receive packets, store the next pointer in rdb_next and read it back on the next round, to fix a race condition vs. rtw_tune() during AP discovery and association. f) Remove the special "skip to 0th descriptor" code, because it is no longer needed. rtw_collect_txring(): g) In the "force" case, stop when we catch up to the hardware xmit pointer. rtw_intr_beacon(): h) Resetting tdb_next here looked fishy to me, in light of changes in rtw_collect_txring(). XXX However, this is the one part I haven't tested. rtw_txring_fixup(): i) Add function and line arguments, so we can print out information about the caller. j) In addition to reading the hardware pointer, skip over any additional packets with OWN set -- we may have queued more. rtw_rxring_fixup(): k) Remove this, because it's completely wrong. rtw_intr_ioerror(): l) Don't handle receive overruns here; handle them in the normal receive path. Also changed RTW_INTR_RX and RTW_INTR_IOERROR accordingly. m) Do not fiddle with CR; the transmit engine is still running just fine. n) Do not fiddle with IMR; it was unnecessary. o) Don't touch the receive engine at all; not necessary now that we're leaving CR alone. p) Remove the rtw_do_chip_reset case; it was dangerous and not actually used. rtw_intr(): q) Don't mask off ISR bits when we pass down the ISR value; it's not necessary, and sometimes the other bits useful in debugging output. rtw_seg_too_short(): r) Don't spew here; the caller will do it if necessary. (Stops spewage in an expected case.) rtw_rxdesc_blk: s) Minor change to make the elements have the same order as txdesc_blk. general: t) Added a bunch of debugging output, #ifdef RTW_DIAG, that was useful while investigating the ring management problems. Also moved rtw_txring_fixup() into RTW_DIAG, as I believe the transmit pointer is always kept in sync now, and the function has potential race conditions.
2007-01-06 11:27:53 +03:00
if (dmamap->dm_segs[i].ds_len < 4)
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
return 1;
}
return 0;
}
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
/* TBD factor with atw_start */
static struct mbuf *
rtw_dmamap_load_txbuf(bus_dma_tag_t dmat, bus_dmamap_t dmam, struct mbuf *chain,
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
u_int ndescfree, device_t dev)
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
{
int first, rc;
struct mbuf *m, *m0;
m0 = chain;
/*
* Load the DMA map. Copy and try (once) again if the packet
* didn't fit in the alloted number of segments.
*/
for (first = 1;
((rc = bus_dmamap_load_mbuf(dmat, dmam, m0,
BUS_DMA_WRITE|BUS_DMA_NOWAIT)) != 0 ||
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
dmam->dm_nsegs > ndescfree || rtw_seg_too_short(dmam)) && first;
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
first = 0) {
I have made a small change that suppresses "recovery" from (spurious?) tx timeouts on the beacon queue. The driver's recovery process badly disrupts the MAC's receiver. This needs a closer look, later. Every other fix is by Charles M. Hannum. Charles writes: "With these, the device is quite stable under load, does not experience xmit timeouts, no longer mysteriously freezes, and does not spew a lot of garbage. It does, however, continue to get some "tx fifo underflow"s. global, SYSCTL_SETUP(): a) Removed unused sysctl variables. rtw_io_enable(): b) Change rtw_io_enable to take a softc, and make it call rtw_txring_fixup(). c) Enable the PCI multiple read/write bit. d) Always reset rdb_next to 0 when starting the receive engine. rtw_intr_rx(): e) When processing receive packets, store the next pointer in rdb_next and read it back on the next round, to fix a race condition vs. rtw_tune() during AP discovery and association. f) Remove the special "skip to 0th descriptor" code, because it is no longer needed. rtw_collect_txring(): g) In the "force" case, stop when we catch up to the hardware xmit pointer. rtw_intr_beacon(): h) Resetting tdb_next here looked fishy to me, in light of changes in rtw_collect_txring(). XXX However, this is the one part I haven't tested. rtw_txring_fixup(): i) Add function and line arguments, so we can print out information about the caller. j) In addition to reading the hardware pointer, skip over any additional packets with OWN set -- we may have queued more. rtw_rxring_fixup(): k) Remove this, because it's completely wrong. rtw_intr_ioerror(): l) Don't handle receive overruns here; handle them in the normal receive path. Also changed RTW_INTR_RX and RTW_INTR_IOERROR accordingly. m) Do not fiddle with CR; the transmit engine is still running just fine. n) Do not fiddle with IMR; it was unnecessary. o) Don't touch the receive engine at all; not necessary now that we're leaving CR alone. p) Remove the rtw_do_chip_reset case; it was dangerous and not actually used. rtw_intr(): q) Don't mask off ISR bits when we pass down the ISR value; it's not necessary, and sometimes the other bits useful in debugging output. rtw_seg_too_short(): r) Don't spew here; the caller will do it if necessary. (Stops spewage in an expected case.) rtw_rxdesc_blk: s) Minor change to make the elements have the same order as txdesc_blk. general: t) Added a bunch of debugging output, #ifdef RTW_DIAG, that was useful while investigating the ring management problems. Also moved rtw_txring_fixup() into RTW_DIAG, as I believe the transmit pointer is always kept in sync now, and the function has potential race conditions.
2007-01-06 11:27:53 +03:00
if (rc == 0) {
#ifdef RTW_DIAGxxx
if (rtw_seg_too_short(dmam)) {
printf("%s: short segment, mbuf lengths:", __func__);
for (m = m0; m; m = m->m_next)
printf(" %d", m->m_len);
printf("\n");
}
#endif
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
bus_dmamap_unload(dmat, dmam);
I have made a small change that suppresses "recovery" from (spurious?) tx timeouts on the beacon queue. The driver's recovery process badly disrupts the MAC's receiver. This needs a closer look, later. Every other fix is by Charles M. Hannum. Charles writes: "With these, the device is quite stable under load, does not experience xmit timeouts, no longer mysteriously freezes, and does not spew a lot of garbage. It does, however, continue to get some "tx fifo underflow"s. global, SYSCTL_SETUP(): a) Removed unused sysctl variables. rtw_io_enable(): b) Change rtw_io_enable to take a softc, and make it call rtw_txring_fixup(). c) Enable the PCI multiple read/write bit. d) Always reset rdb_next to 0 when starting the receive engine. rtw_intr_rx(): e) When processing receive packets, store the next pointer in rdb_next and read it back on the next round, to fix a race condition vs. rtw_tune() during AP discovery and association. f) Remove the special "skip to 0th descriptor" code, because it is no longer needed. rtw_collect_txring(): g) In the "force" case, stop when we catch up to the hardware xmit pointer. rtw_intr_beacon(): h) Resetting tdb_next here looked fishy to me, in light of changes in rtw_collect_txring(). XXX However, this is the one part I haven't tested. rtw_txring_fixup(): i) Add function and line arguments, so we can print out information about the caller. j) In addition to reading the hardware pointer, skip over any additional packets with OWN set -- we may have queued more. rtw_rxring_fixup(): k) Remove this, because it's completely wrong. rtw_intr_ioerror(): l) Don't handle receive overruns here; handle them in the normal receive path. Also changed RTW_INTR_RX and RTW_INTR_IOERROR accordingly. m) Do not fiddle with CR; the transmit engine is still running just fine. n) Do not fiddle with IMR; it was unnecessary. o) Don't touch the receive engine at all; not necessary now that we're leaving CR alone. p) Remove the rtw_do_chip_reset case; it was dangerous and not actually used. rtw_intr(): q) Don't mask off ISR bits when we pass down the ISR value; it's not necessary, and sometimes the other bits useful in debugging output. rtw_seg_too_short(): r) Don't spew here; the caller will do it if necessary. (Stops spewage in an expected case.) rtw_rxdesc_blk: s) Minor change to make the elements have the same order as txdesc_blk. general: t) Added a bunch of debugging output, #ifdef RTW_DIAG, that was useful while investigating the ring management problems. Also moved rtw_txring_fixup() into RTW_DIAG, as I believe the transmit pointer is always kept in sync now, and the function has potential race conditions.
2007-01-06 11:27:53 +03:00
}
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
MGETHDR(m, M_DONTWAIT, MT_DATA);
if (m == NULL) {
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
aprint_error_dev(dev, "unable to allocate Tx mbuf\n");
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
break;
}
if (m0->m_pkthdr.len > MHLEN) {
MCLGET(m, M_DONTWAIT);
if ((m->m_flags & M_EXT) == 0) {
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
aprint_error_dev(dev,
"cannot allocate Tx cluster\n");
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
m_freem(m);
break;
}
}
Kill caddr_t; there will be some MI fallout, but it will be fixed shortly.
2007-03-04 08:59:00 +03:00
m_copydata(m0, 0, m0->m_pkthdr.len, mtod(m, void *));
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
m->m_pkthdr.len = m->m_len = m0->m_pkthdr.len;
m_freem(m0);
m0 = m;
m = NULL;
}
if (rc != 0) {
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
aprint_error_dev(dev, "cannot load Tx buffer, rc = %d\n", rc);
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
m_freem(m0);
return NULL;
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
} else if (rtw_seg_too_short(dmam)) {
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
aprint_error_dev(dev,
"cannot load Tx buffer, segment too short\n");
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
bus_dmamap_unload(dmat, dmam);
m_freem(m0);
return NULL;
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
} else if (dmam->dm_nsegs > ndescfree) {
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
aprint_error_dev(dev, "too many tx segments\n");
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
bus_dmamap_unload(dmat, dmam);
m_freem(m0);
return NULL;
}
return m0;
}
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
#ifdef RTW_DEBUG
Add more comprehensive debug messages as we setup transmit descriptors. Also, dump packets if IFF_DEBUG|IFF_LINK2. Factor out a subroutine for printing transmit-descriptor debug messages.
2004-12-23 08:57:18 +03:00
static void
rtw_print_txdesc(struct rtw_softc *sc, const char *action,
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
struct rtw_txsoft *ts, struct rtw_txdesc_blk *tdb, int desc)
Add more comprehensive debug messages as we setup transmit descriptors. Also, dump packets if IFF_DEBUG|IFF_LINK2. Factor out a subroutine for printing transmit-descriptor debug messages.
2004-12-23 08:57:18 +03:00
{
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
struct rtw_txdesc *td = &tdb->tdb_desc[desc];
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
DPRINTF(sc, RTW_DEBUG_XMIT_DESC, ("%s: %p %s txdesc[%d] next %#08x "
"buf %#08x ctl0 %#08x ctl1 %#08x len %#08x\n",
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
device_xname(sc->sc_dev), ts, action, desc,
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
le32toh(td->td_buf), le32toh(td->td_next),
le32toh(td->td_ctl0), le32toh(td->td_ctl1),
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
le32toh(td->td_len)));
Add more comprehensive debug messages as we setup transmit descriptors. Also, dump packets if IFF_DEBUG|IFF_LINK2. Factor out a subroutine for printing transmit-descriptor debug messages.
2004-12-23 08:57:18 +03:00
}
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
#endif /* RTW_DEBUG */
Add more comprehensive debug messages as we setup transmit descriptors. Also, dump packets if IFF_DEBUG|IFF_LINK2. Factor out a subroutine for printing transmit-descriptor debug messages.
2004-12-23 08:57:18 +03:00
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
static void
rtw_start(struct ifnet *ifp)
{
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
int desc, i, lastdesc, npkt, rate;
Shorten variables, s/txctl/ctl/. This is rtw_start, after all: transmission is implied.
2004-12-23 08:52:27 +03:00
uint32_t proto_ctl0, ctl0, ctl1;
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
bus_dmamap_t dmamap;
struct ieee80211com *ic;
struct ieee80211_duration *d0;
rtw: Try to get hardware WEP to work. It seems to work in the RX direction, but not in the TX direction. The net80211 crypto framework doesn't seem to cope very well with the assymetry (I'm probably missing something), so I will use software WEP for now. net80211: In ieee80211_compute_duration, figure out whether to add the WEP header to the packet overhead by checking the WEP bit in the Frame Control field of the 802.11 header, instead of checking the IEEE80211_F_PRIVACY flag. Also, if the WEP bit is present, assume that the frame described by (wh, len) has already already been WEP encapsulated, and adjust the payload length accordingly. XXX that's a grotty hack that I will have to revisit, later.
2005-06-27 09:49:13 +04:00
struct ieee80211_frame_min *wh;
XXX: GCC uninitialized
2006-05-14 09:42:43 +04:00
struct ieee80211_node *ni = NULL; /* XXX: GCC */
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
struct mbuf *m0;
struct rtw_softc *sc;
XXX: GCC uninitialized
2006-05-14 09:42:43 +04:00
struct rtw_txsoft_blk *tsb = NULL; /* XXX: GCC */
struct rtw_txdesc_blk *tdb = NULL; /* XXX: GCC */
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
struct rtw_txsoft *ts;
struct rtw_txdesc *td;
rtw: Try to get hardware WEP to work. It seems to work in the RX direction, but not in the TX direction. The net80211 crypto framework doesn't seem to cope very well with the assymetry (I'm probably missing something), so I will use software WEP for now. net80211: In ieee80211_compute_duration, figure out whether to add the WEP header to the packet overhead by checking the WEP bit in the Frame Control field of the 802.11 header, instead of checking the IEEE80211_F_PRIVACY flag. Also, if the WEP bit is present, assume that the frame described by (wh, len) has already already been WEP encapsulated, and adjust the payload length accordingly. XXX that's a grotty hack that I will have to revisit, later.
2005-06-27 09:49:13 +04:00
struct ieee80211_key *k;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
sc = (struct rtw_softc *)ifp->if_softc;
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
ic = &sc->sc_ic;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
DPRINTF(sc, RTW_DEBUG_XMIT,
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
("%s: enter %s\n", device_xname(sc->sc_dev), __func__));
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
If defined(RTW_DEBUG), provide a sysctl (hw.rtw.rxbufs_limit) for limiting the number of rx buffers an rtw may allocate. Use this sysctl to test the code that copes with buffer exhaustion. Allocate at most RTW_RXQLEN rx buffers, stopping at the sysctl limit. Record in sc_nrxdesc how many were allocated, and put the end-of-ring indication on sc_rxdesc[sc_nrxdesc - 1]. In rtw_init, if no rx buffers could be allocated, log a complaint, clear IFF_RUNNING, and exit with an error. Many changes to accomodate a short rx ring, mainly of the "add a rx-ring length argument" variety. XXX I really should consolidate all of the rx ring variables in one struct and pass that to the rx-ring subroutines. Bug fix: after calling MCLGET, use the (m->m_flags & M_EXT) idiom to check for success, instead of m != NULL. Bug fix: at the top of rtw_start, if IFF_RUNNING is not set, or IFF_OACTIVE is, get out.
2004-12-29 01:21:15 +03:00
if ((ifp->if_flags & (IFF_RUNNING|IFF_OACTIVE)) != IFF_RUNNING)
goto out;
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
/* XXX do real rate control */
Shorten variables, s/txctl/ctl/. This is rtw_start, after all: transmission is implied.
2004-12-23 08:52:27 +03:00
proto_ctl0 = RTW_TXCTL0_RTSRATE_1MBPS;
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
if ((ic->ic_flags & IEEE80211_F_SHPREAMBLE) != 0)
Shorten variables, s/txctl/ctl/. This is rtw_start, after all: transmission is implied.
2004-12-23 08:52:27 +03:00
proto_ctl0 |= RTW_TXCTL0_SPLCP;
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
for (;;) {
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
if (rtw_dequeue(ifp, &tsb, &tdb, &m0, &ni) == -1)
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
continue;
if (m0 == NULL)
break;
rtw: Try to get hardware WEP to work. It seems to work in the RX direction, but not in the TX direction. The net80211 crypto framework doesn't seem to cope very well with the assymetry (I'm probably missing something), so I will use software WEP for now. net80211: In ieee80211_compute_duration, figure out whether to add the WEP header to the packet overhead by checking the WEP bit in the Frame Control field of the 802.11 header, instead of checking the IEEE80211_F_PRIVACY flag. Also, if the WEP bit is present, assume that the frame described by (wh, len) has already already been WEP encapsulated, and adjust the payload length accordingly. XXX that's a grotty hack that I will have to revisit, later.
2005-06-27 09:49:13 +04:00
wh = mtod(m0, struct ieee80211_frame_min *);
if ((wh->i_fc[1] & IEEE80211_FC1_WEP) != 0 &&
(k = ieee80211_crypto_encap(ic, ni, m0)) == NULL) {
Avoid an unnecessary API difference between NetBSD and FreeBSD: back out my change to ieee80211_crypto_encap that made it free its mbuf argument on error. I had thought it was a bug. It was not. It's the drivers that are broken. Make an(4), atw(4), ipw(4), iwi(4), ral(4), rtw(4), ural(4), and wi(4) free the mbuf when ieee80211_crypto_encap returns NULL. Also, return ath(4) to the way it was---i.e., free the mbuf. Thanks to Sam Leffler to pointing out my mistake.
2005-07-07 03:44:15 +04:00
m_freem(m0);
rtw: Try to get hardware WEP to work. It seems to work in the RX direction, but not in the TX direction. The net80211 crypto framework doesn't seem to cope very well with the assymetry (I'm probably missing something), so I will use software WEP for now. net80211: In ieee80211_compute_duration, figure out whether to add the WEP header to the packet overhead by checking the WEP bit in the Frame Control field of the 802.11 header, instead of checking the IEEE80211_F_PRIVACY flag. Also, if the WEP bit is present, assume that the frame described by (wh, len) has already already been WEP encapsulated, and adjust the payload length accordingly. XXX that's a grotty hack that I will have to revisit, later.
2005-06-27 09:49:13 +04:00
break;
} else
k = NULL;
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
ts = SIMPLEQ_FIRST(&tsb->tsb_freeq);
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
dmamap = ts->ts_dmamap;
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
m0 = rtw_dmamap_load_txbuf(sc->sc_dmat, dmamap, m0,
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
tdb->tdb_nfree, sc->sc_dev);
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
if (m0 == NULL || dmamap->dm_nsegs == 0) {
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
DPRINTF(sc, RTW_DEBUG_XMIT,
("%s: fail dmamap load\n", __func__));
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
goto post_dequeue_err;
}
rtw: Try to get hardware WEP to work. It seems to work in the RX direction, but not in the TX direction. The net80211 crypto framework doesn't seem to cope very well with the assymetry (I'm probably missing something), so I will use software WEP for now. net80211: In ieee80211_compute_duration, figure out whether to add the WEP header to the packet overhead by checking the WEP bit in the Frame Control field of the 802.11 header, instead of checking the IEEE80211_F_PRIVACY flag. Also, if the WEP bit is present, assume that the frame described by (wh, len) has already already been WEP encapsulated, and adjust the payload length accordingly. XXX that's a grotty hack that I will have to revisit, later.
2005-06-27 09:49:13 +04:00
/* Note well: rtw_dmamap_load_txbuf may have created
* a new chain, so we must find the header once
* more.
*/
wh = mtod(m0, struct ieee80211_frame_min *);
Misc. changes to improve ad hoc mode and to enable hostap mode: Fix the work-around for the NIC bug where it skips to rx descriptor 0. The driver used to skip to rx descriptor 1. Hopefully this stops the out-of-order packet reception that Charles Hannum saw. When debugging is enabled, print rx-descriptor status flags before printing the rx bit rate. Add a debug message for when a beacon tx buffer reclamation. Reset IFF_OACTIVE when we reset the transmitter. Pass the consolidated LED state, a struct rtw_led_state, to rtw_led_attach. Choose the bit-rate for management frames (1Mb/s) at the same place we choose for all other frames. Do not use the NIC's short preamble or RTS options for management frames. Label beacons for the NIC. Following a Linux driver, take care not to zero arbitrary bits in the TPPOLL register. Use the new idiom for IBSS merges: disable transmitter, kick the state machine. Add a second descriptor to the beacon ring. The NIC seems to like this much better.
2005-03-02 08:20:43 +03:00
/* XXX do real rate control */
if ((wh->i_fc[0] & IEEE80211_FC0_TYPE_MASK) ==
IEEE80211_FC0_TYPE_MGT)
rate = 2;
else
Revamp ieee80211_get_rate. Now it does not use the rateset in the ic->ic_bss, but it uses the rateset in its new ieee80211_node argument, instead. If the rate is fixed by ic->ic_fixed_rate, but the fixed rate is not in the node's rateset, choose a reasonable default: prefer the lowest basic rate or, if there is no basic rate, prefer the lowest rate, period. Change a printf complaint to a debug message. Adapt drivers to suit new ieee80211_get_rate calling convention. XXX I really need to replace ieee80211_get_rate with a bitrate XXX adaptation algorithm. Soon, soon....
2006-03-28 04:48:10 +04:00
rate = MAX(2, ieee80211_get_rate(ni));
Misc. changes to improve ad hoc mode and to enable hostap mode: Fix the work-around for the NIC bug where it skips to rx descriptor 0. The driver used to skip to rx descriptor 1. Hopefully this stops the out-of-order packet reception that Charles Hannum saw. When debugging is enabled, print rx-descriptor status flags before printing the rx bit rate. Add a debug message for when a beacon tx buffer reclamation. Reset IFF_OACTIVE when we reset the transmitter. Pass the consolidated LED state, a struct rtw_led_state, to rtw_led_attach. Choose the bit-rate for management frames (1Mb/s) at the same place we choose for all other frames. Do not use the NIC's short preamble or RTS options for management frames. Label beacons for the NIC. Following a Linux driver, take care not to zero arbitrary bits in the TPPOLL register. Use the new idiom for IBSS merges: disable transmitter, kick the state machine. Add a second descriptor to the beacon ring. The NIC seems to like this much better.
2005-03-02 08:20:43 +03:00
Add more comprehensive debug messages as we setup transmit descriptors. Also, dump packets if IFF_DEBUG|IFF_LINK2. Factor out a subroutine for printing transmit-descriptor debug messages.
2004-12-23 08:57:18 +03:00
#ifdef RTW_DEBUG
Resolve conflicts in importation of 18-May-2005 ath(4) / net80211(9) from FreeBSD. Introduce compatibility shims (sys/dev/ic/ath_netbsd.[ch], sys/net80211/ieee80211_netbsd.[ch]). Update drivers (an, atu, atw, awi, ipw, iwi, rtw, wi) for the new net80211(9) API.
2005-06-22 10:14:51 +04:00
if ((ifp->if_flags & (IFF_DEBUG|IFF_LINK2)) ==
Add more comprehensive debug messages as we setup transmit descriptors. Also, dump packets if IFF_DEBUG|IFF_LINK2. Factor out a subroutine for printing transmit-descriptor debug messages.
2004-12-23 08:57:18 +03:00
(IFF_DEBUG|IFF_LINK2)) {
ieee80211_dump_pkt(mtod(m0, uint8_t *),
(dmamap->dm_nsegs == 1) ? m0->m_pkthdr.len
: sizeof(wh),
rate, 0);
}
#endif /* RTW_DEBUG */
Shorten variables, s/txctl/ctl/. This is rtw_start, after all: transmission is implied.
2004-12-23 08:52:27 +03:00
ctl0 = proto_ctl0 |
Per discussion on tech-kern and tech-userlevel, move the bit-twiddling macros, __BIT, __BITS, SHIFTIN, SHIFTOUT, and __arraycount() from lib/libkern/libkern.h to sys/cdefs.h. Add a __-prefix to SHIFTIN and SHIFTOUT, and add a manual page for the bit-twiddling macros, bits(3). Make the __BIT and __BITS macros "widthless," as best I can, by changing their type to uintmax_t from uint32_t. XXX The manual page lags this change by a bit. Define __PRIxBIT and __PRIxBITS printf(3) format strings.
2006-08-31 23:24:37 +04:00
__SHIFTIN(m0->m_pkthdr.len, RTW_TXCTL0_TPKTSIZE_MASK);
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
Misc. changes to improve ad hoc mode and to enable hostap mode: Fix the work-around for the NIC bug where it skips to rx descriptor 0. The driver used to skip to rx descriptor 1. Hopefully this stops the out-of-order packet reception that Charles Hannum saw. When debugging is enabled, print rx-descriptor status flags before printing the rx bit rate. Add a debug message for when a beacon tx buffer reclamation. Reset IFF_OACTIVE when we reset the transmitter. Pass the consolidated LED state, a struct rtw_led_state, to rtw_led_attach. Choose the bit-rate for management frames (1Mb/s) at the same place we choose for all other frames. Do not use the NIC's short preamble or RTS options for management frames. Label beacons for the NIC. Following a Linux driver, take care not to zero arbitrary bits in the TPPOLL register. Use the new idiom for IBSS merges: disable transmitter, kick the state machine. Add a second descriptor to the beacon ring. The NIC seems to like this much better.
2005-03-02 08:20:43 +03:00
switch (rate) {
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
default:
case 2:
ctl0 |= RTW_TXCTL0_RATE_1MBPS;
break;
case 4:
ctl0 |= RTW_TXCTL0_RATE_2MBPS;
break;
case 11:
ctl0 |= RTW_TXCTL0_RATE_5MBPS;
break;
case 22:
ctl0 |= RTW_TXCTL0_RATE_11MBPS;
break;
}
Misc. changes to improve ad hoc mode and to enable hostap mode: Fix the work-around for the NIC bug where it skips to rx descriptor 0. The driver used to skip to rx descriptor 1. Hopefully this stops the out-of-order packet reception that Charles Hannum saw. When debugging is enabled, print rx-descriptor status flags before printing the rx bit rate. Add a debug message for when a beacon tx buffer reclamation. Reset IFF_OACTIVE when we reset the transmitter. Pass the consolidated LED state, a struct rtw_led_state, to rtw_led_attach. Choose the bit-rate for management frames (1Mb/s) at the same place we choose for all other frames. Do not use the NIC's short preamble or RTS options for management frames. Label beacons for the NIC. Following a Linux driver, take care not to zero arbitrary bits in the TPPOLL register. Use the new idiom for IBSS merges: disable transmitter, kick the state machine. Add a second descriptor to the beacon ring. The NIC seems to like this much better.
2005-03-02 08:20:43 +03:00
/* XXX >= ? Compare after fragmentation? */
if (m0->m_pkthdr.len > ic->ic_rtsthreshold)
ctl0 |= RTW_TXCTL0_RTSEN;
When rtw_enable is called, power may have been removed and re-applied, so invalidate the WEP key registers to force us to write the keys back to the hardware. Revamp key handling: use the hardware WEP engine for 40-bit and 104-bit keys, only; use software WEP for all other key lengths. Set the hardware key length to the length of the default Tx key, or if there is no default Tx key, use the length of the longest WEP key.
2006-02-19 11:10:06 +03:00
/* XXX Sometimes writes a bogus keyid; h/w doesn't
* seem to care, since we don't activate h/w Tx
* encryption.
*/
Add the WEP capability to our net80211 capability flags, since we do sort of support hardware WEP. Only change the WEP key index in a Tx descriptor from 0 if the type of our transmit key is _WEP, not _TKIP: i.e., only if we're really doing WEP crypto in the hardware. Ignore a watchdog timeout on any Tx ring if we can collect some packets from that ring. Restart both the receiver and the transmitter when a watchdog timeout occurs instead of restarting only the transmitter.
2009-04-02 04:57:20 +04:00
if (k != NULL &&
k->wk_cipher->ic_cipher == IEEE80211_CIPHER_WEP) {
Per discussion on tech-kern and tech-userlevel, move the bit-twiddling macros, __BIT, __BITS, SHIFTIN, SHIFTOUT, and __arraycount() from lib/libkern/libkern.h to sys/cdefs.h. Add a __-prefix to SHIFTIN and SHIFTOUT, and add a manual page for the bit-twiddling macros, bits(3). Make the __BIT and __BITS macros "widthless," as best I can, by changing their type to uintmax_t from uint32_t. XXX The manual page lags this change by a bit. Define __PRIxBIT and __PRIxBITS printf(3) format strings.
2006-08-31 23:24:37 +04:00
ctl0 |= __SHIFTIN(k->wk_keyix, RTW_TXCTL0_KEYID_MASK) &
rtw: Try to get hardware WEP to work. It seems to work in the RX direction, but not in the TX direction. The net80211 crypto framework doesn't seem to cope very well with the assymetry (I'm probably missing something), so I will use software WEP for now. net80211: In ieee80211_compute_duration, figure out whether to add the WEP header to the packet overhead by checking the WEP bit in the Frame Control field of the 802.11 header, instead of checking the IEEE80211_F_PRIVACY flag. Also, if the WEP bit is present, assume that the frame described by (wh, len) has already already been WEP encapsulated, and adjust the payload length accordingly. XXX that's a grotty hack that I will have to revisit, later.
2005-06-27 09:49:13 +04:00
RTW_TXCTL0_KEYID_MASK;
}
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
Misc. changes to improve ad hoc mode and to enable hostap mode: Fix the work-around for the NIC bug where it skips to rx descriptor 0. The driver used to skip to rx descriptor 1. Hopefully this stops the out-of-order packet reception that Charles Hannum saw. When debugging is enabled, print rx-descriptor status flags before printing the rx bit rate. Add a debug message for when a beacon tx buffer reclamation. Reset IFF_OACTIVE when we reset the transmitter. Pass the consolidated LED state, a struct rtw_led_state, to rtw_led_attach. Choose the bit-rate for management frames (1Mb/s) at the same place we choose for all other frames. Do not use the NIC's short preamble or RTS options for management frames. Label beacons for the NIC. Following a Linux driver, take care not to zero arbitrary bits in the TPPOLL register. Use the new idiom for IBSS merges: disable transmitter, kick the state machine. Add a second descriptor to the beacon ring. The NIC seems to like this much better.
2005-03-02 08:20:43 +03:00
if ((wh->i_fc[0] & IEEE80211_FC0_TYPE_MASK) ==
IEEE80211_FC0_TYPE_MGT) {
ctl0 &= ~(RTW_TXCTL0_SPLCP | RTW_TXCTL0_RTSEN);
if ((wh->i_fc[0] & IEEE80211_FC0_SUBTYPE_MASK) ==
IEEE80211_FC0_SUBTYPE_BEACON)
ctl0 |= RTW_TXCTL0_BEACON;
}
In atw(4), use ieee80211_compute_duration() to compute IEEE 802.11 Duration and PLCP Length fields, and delete the abominable atw_frame_setdurs() subroutine. Make rtw(4) use the new ieee80211_compute_duration() calling convention. Add an ieee80211_key argument to ieee80211_compute_duration() and lightly constify arguments. Get the crypto header length from the key argument instead of blithely assuming a WEP header. Add some inline documentation. Account for data padding (IEEE80211_F_DATAPAD).
2005-12-30 00:08:26 +03:00
if (ieee80211_compute_duration(wh, k, m0->m_pkthdr.len,
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
ic->ic_flags, ic->ic_fragthreshold,
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
rate, &ts->ts_d0, &ts->ts_dn, &npkt,
Resolve conflicts in importation of 18-May-2005 ath(4) / net80211(9) from FreeBSD. Introduce compatibility shims (sys/dev/ic/ath_netbsd.[ch], sys/net80211/ieee80211_netbsd.[ch]). Update drivers (an, atu, atw, awi, ipw, iwi, rtw, wi) for the new net80211(9) API.
2005-06-22 10:14:51 +04:00
(ifp->if_flags & (IFF_DEBUG|IFF_LINK2)) ==
Fix a bug in ieee80211_compute_duration: the 802.11 Duration field in an 802.11 unicast data packet is equal to the duration of the SIFS and Acknowledgement. That is, the amount of time reserved *after* the packet has finished transmitting. Change the arguments to ieee80211_compute_duration: pass the entire packet length, not just the payload length. Add a 'debug' argument to ieee80211_compute_duration and its helper subroutine, ieee80211_compute_duration1. If debug != 0, ieee80211_compute_duration printfs its arguments and several local variables. In rtw(4), load the 802.11 Duration field with the result from ieee80211_compute_duration.
2004-12-23 09:08:52 +03:00
(IFF_DEBUG|IFF_LINK2)) == -1) {
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
DPRINTF(sc, RTW_DEBUG_XMIT,
("%s: fail compute duration\n", __func__));
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
goto post_load_err;
}
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
d0 = &ts->ts_d0;
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
Oops, straggler from last commit: set the 802.11 Duration field with the result from ieee80211_compute_duration.
2004-12-23 09:12:43 +03:00
*(uint16_t*)wh->i_dur = htole16(d0->d_data_dur);
Per discussion on tech-kern and tech-userlevel, move the bit-twiddling macros, __BIT, __BITS, SHIFTIN, SHIFTOUT, and __arraycount() from lib/libkern/libkern.h to sys/cdefs.h. Add a __-prefix to SHIFTIN and SHIFTOUT, and add a manual page for the bit-twiddling macros, bits(3). Make the __BIT and __BITS macros "widthless," as best I can, by changing their type to uintmax_t from uint32_t. XXX The manual page lags this change by a bit. Define __PRIxBIT and __PRIxBITS printf(3) format strings.
2006-08-31 23:24:37 +04:00
ctl1 = __SHIFTIN(d0->d_plcp_len, RTW_TXCTL1_LENGTH_MASK) |
__SHIFTIN(d0->d_rts_dur, RTW_TXCTL1_RTSDUR_MASK);
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
Replace d_plcp_svc with d_residue. The latter is the number of whole empty/unused octets to fill out the data time slot. The value is constrained by math to 0 for <= 5.5Mb, 0-1 for 11Mb, and 0-2 for 22Mb. It is used to signal to the MAC that there is residue.
2004-12-27 04:51:49 +03:00
if (d0->d_residue)
Shorten variables, s/txctl/ctl/. This is rtw_start, after all: transmission is implied.
2004-12-23 08:52:27 +03:00
ctl1 |= RTW_TXCTL1_LENGEXT;
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
/* TBD fragmentation */
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
ts->ts_first = tdb->tdb_next;
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
rtw_txdescs_sync(tdb, ts->ts_first, dmamap->dm_nsegs,
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
BUS_DMASYNC_PREWRITE);
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
KASSERT(ts->ts_first < tdb->tdb_ndesc);
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
Push the bpf_ops usage back into bpf.h. Push the common ifp->if_bpf check into the inline functions as well the fourth argument for bpf_attach.
2010-04-05 11:19:28 +04:00
bpf_mtap3(ic->ic_rawbpf, m0);
Add bpf(4) packet taps to rtw(4). Now you can tap DLT_EN10MB (Ethernet II), DLT_IEEE802_11 (802.11), and DLT_IEEE802_11_RADIO (radiotap + 802.11).
2004-12-29 01:30:07 +03:00
if (sc->sc_radiobpf != NULL) {
struct rtw_tx_radiotap_header *rt = &sc->sc_txtap;
rt->rt_rate = rate;
Push the bpf_ops usage back into bpf.h. Push the common ifp->if_bpf check into the inline functions as well the fourth argument for bpf_attach.
2010-04-05 11:19:28 +04:00
bpf_mtap2(sc->sc_radiobpf, rt, sizeof(sc->sc_txtapu),
m0);
Add bpf(4) packet taps to rtw(4). Now you can tap DLT_EN10MB (Ethernet II), DLT_IEEE802_11 (802.11), and DLT_IEEE802_11_RADIO (radiotap + 802.11).
2004-12-29 01:30:07 +03:00
}
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
for (i = 0, lastdesc = desc = ts->ts_first;
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
i < dmamap->dm_nsegs;
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
i++, desc = RTW_NEXT_IDX(tdb, desc)) {
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
if (dmamap->dm_segs[i].ds_len > RTW_TXLEN_LENGTH_MASK) {
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
DPRINTF(sc, RTW_DEBUG_XMIT_DESC,
("%s: seg too long\n", __func__));
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
goto post_load_err;
}
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
td = &tdb->tdb_desc[desc];
td->td_ctl0 = htole32(ctl0);
td->td_ctl1 = htole32(ctl1);
td->td_buf = htole32(dmamap->dm_segs[i].ds_addr);
td->td_len = htole32(dmamap->dm_segs[i].ds_len);
Fix order of operations: rtw_rxdesc_init: before initializing a Rx descriptor's control word, synchronize the Rx buffer rtw_start, be paranoid: set the "NIC owns" bit on a Tx descriptor dead last. Always reinitialize the "next Tx descriptor" pointer, just in case it got clobbered somehow. Let the compiler decide whether to inline rtw_collect_txring or not.
2007-11-16 01:55:50 +03:00
td->td_next = htole32(RTW_NEXT_DESC(tdb, desc));
if (i != 0)
td->td_ctl0 |= htole32(RTW_TXCTL0_OWN);
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
lastdesc = desc;
Add more comprehensive debug messages as we setup transmit descriptors. Also, dump packets if IFF_DEBUG|IFF_LINK2. Factor out a subroutine for printing transmit-descriptor debug messages.
2004-12-23 08:57:18 +03:00
#ifdef RTW_DEBUG
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
rtw_print_txdesc(sc, "load", ts, tdb, desc);
Add more comprehensive debug messages as we setup transmit descriptors. Also, dump packets if IFF_DEBUG|IFF_LINK2. Factor out a subroutine for printing transmit-descriptor debug messages.
2004-12-23 08:57:18 +03:00
#endif /* RTW_DEBUG */
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
}
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
KASSERT(desc < tdb->tdb_ndesc);
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
ts->ts_ni = ni;
Resolve conflicts in importation of 18-May-2005 ath(4) / net80211(9) from FreeBSD. Introduce compatibility shims (sys/dev/ic/ath_netbsd.[ch], sys/net80211/ieee80211_netbsd.[ch]). Update drivers (an, atu, atw, awi, ipw, iwi, rtw, wi) for the new net80211(9) API.
2005-06-22 10:14:51 +04:00
KASSERT(ni != NULL);
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
ts->ts_mbuf = m0;
ts->ts_last = lastdesc;
tdb->tdb_desc[ts->ts_last].td_ctl0 |= htole32(RTW_TXCTL0_LS);
tdb->tdb_desc[ts->ts_first].td_ctl0 |=
Add more comprehensive debug messages as we setup transmit descriptors. Also, dump packets if IFF_DEBUG|IFF_LINK2. Factor out a subroutine for printing transmit-descriptor debug messages.
2004-12-23 08:57:18 +03:00
htole32(RTW_TXCTL0_FS);
#ifdef RTW_DEBUG
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
rtw_print_txdesc(sc, "FS on", ts, tdb, ts->ts_first);
rtw_print_txdesc(sc, "LS on", ts, tdb, ts->ts_last);
Add more comprehensive debug messages as we setup transmit descriptors. Also, dump packets if IFF_DEBUG|IFF_LINK2. Factor out a subroutine for printing transmit-descriptor debug messages.
2004-12-23 08:57:18 +03:00
#endif /* RTW_DEBUG */
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
tdb->tdb_nfree -= dmamap->dm_nsegs;
tdb->tdb_next = desc;
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
rtw_txdescs_sync(tdb, ts->ts_first, dmamap->dm_nsegs,
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
BUS_DMASYNC_PREREAD|BUS_DMASYNC_PREWRITE);
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
tdb->tdb_desc[ts->ts_first].td_ctl0 |=
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
htole32(RTW_TXCTL0_OWN);
Add more comprehensive debug messages as we setup transmit descriptors. Also, dump packets if IFF_DEBUG|IFF_LINK2. Factor out a subroutine for printing transmit-descriptor debug messages.
2004-12-23 08:57:18 +03:00
#ifdef RTW_DEBUG
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
rtw_print_txdesc(sc, "OWN on", ts, tdb, ts->ts_first);
Add more comprehensive debug messages as we setup transmit descriptors. Also, dump packets if IFF_DEBUG|IFF_LINK2. Factor out a subroutine for printing transmit-descriptor debug messages.
2004-12-23 08:57:18 +03:00
#endif /* RTW_DEBUG */
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
rtw_txdescs_sync(tdb, ts->ts_first, 1,
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
BUS_DMASYNC_PREREAD|BUS_DMASYNC_PREWRITE);
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
SIMPLEQ_REMOVE_HEAD(&tsb->tsb_freeq, ts_q);
SIMPLEQ_INSERT_TAIL(&tsb->tsb_dirtyq, ts, ts_q);
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
if (tsb != &sc->sc_txsoft_blk[RTW_TXPRIBCN])
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
sc->sc_led_state.ls_event |= RTW_LED_S_TX;
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
tsb->tsb_tx_timer = 5;
ifp->if_timer = 1;
Extract subroutine rtw_tx_kick() from rtw_start(), and reuse it in rtw_io_enable() to restart the transmitter after resetting it.
2008-03-15 02:59:01 +03:00
rtw_tx_kick(&sc->sc_regs, tsb->tsb_poll);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
}
If defined(RTW_DEBUG), provide a sysctl (hw.rtw.rxbufs_limit) for limiting the number of rx buffers an rtw may allocate. Use this sysctl to test the code that copes with buffer exhaustion. Allocate at most RTW_RXQLEN rx buffers, stopping at the sysctl limit. Record in sc_nrxdesc how many were allocated, and put the end-of-ring indication on sc_rxdesc[sc_nrxdesc - 1]. In rtw_init, if no rx buffers could be allocated, log a complaint, clear IFF_RUNNING, and exit with an error. Many changes to accomodate a short rx ring, mainly of the "add a rx-ring length argument" variety. XXX I really should consolidate all of the rx ring variables in one struct and pass that to the rx-ring subroutines. Bug fix: after calling MCLGET, use the (m->m_flags & M_EXT) idiom to check for success, instead of m != NULL. Bug fix: at the top of rtw_start, if IFF_RUNNING is not set, or IFF_OACTIVE is, get out.
2004-12-29 01:21:15 +03:00
out:
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
DPRINTF(sc, RTW_DEBUG_XMIT, ("%s: leave\n", __func__));
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
return;
post_load_err:
bus_dmamap_unload(sc->sc_dmat, dmamap);
m_freem(m0);
post_dequeue_err:
Resolve conflicts in importation of 18-May-2005 ath(4) / net80211(9) from FreeBSD. Introduce compatibility shims (sys/dev/ic/ath_netbsd.[ch], sys/net80211/ieee80211_netbsd.[ch]). Update drivers (an, atu, atw, awi, ipw, iwi, rtw, wi) for the new net80211(9) API.
2005-06-22 10:14:51 +04:00
ieee80211_free_node(ni);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
return;
}
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
static void
rtw_idle(struct rtw_regs *regs)
{
int active;
Print the TPPoll register when the Tx DMA engine doesn't shut down in the time we allow.
2008-03-12 18:47:49 +03:00
uint8_t tppoll;
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
/* request stop DMA; wait for packets to stop transmitting. */
RTW_WRITE8(regs, RTW_TPPOLL, RTW_TPPOLL_SALL);
RTW_WBR(regs, RTW_TPPOLL, RTW_TPPOLL);
for (active = 0; active < 300 &&
Print the TPPoll register when the Tx DMA engine doesn't shut down in the time we allow.
2008-03-12 18:47:49 +03:00
(tppoll = RTW_READ8(regs, RTW_TPPOLL) & RTW_TPPOLL_ACTIVE) != 0;
active++)
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
DELAY(10);
Print the TPPoll register when the Tx DMA engine doesn't shut down in the time we allow.
2008-03-12 18:47:49 +03:00
printf("%s: transmit DMA idle in %dus, tppoll %02" PRIx8 "\n", __func__,
active * 10, tppoll);
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
}
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
static void
rtw_watchdog(struct ifnet *ifp)
{
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
int pri, tx_timeouts = 0;
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
struct rtw_softc *sc;
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
struct rtw_txsoft_blk *tsb;
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
sc = ifp->if_softc;
ifp->if_timer = 0;
Use device_t and its accessors throughout. Use aprint_*_dev(). Improve PMF-ability. Add a 'flags' argument to suspend/resume handlers and callers such as pmf_system_suspend(). Define a flag, PMF_F_SELF, which indicates to PMF that a device is suspending/resuming itself. Add helper routines, pmf_device_suspend_self(dev) and pmf_device_resume_self(dev), that call pmf_device_suspend(dev, PMF_F_SELF) and pmf_device_resume(dev, PMF_F_SELF), respectively. Use PMF_F_SELF to suspend/resume self in ath(4), audio(4), rtw(4), and sip(4). In ath(4) and in rtw(4), replace the icky sc_enable/sc_disable callbacks, provided by the bus front-end, with self-suspension/resumption. Also, clean up the bus front-ends. Make sure that the interrupt handler is disestablished during suspension. Get rid of driver-private flags (e.g., RTW_F_ENABLED, ath_softc->sc_invalid); use device_is_active()/device_has_power() calls, instead. In the network-class suspend handler, call if_stop(, 0) instead of if_stop(, 1), because the latter is superfluous (bus- and driver-suspension hooks will 'disable' the NIC), and it may cause recursion. In the network-class resume handler, prevent infinite recursion through if_init() by getting out early if we are self-suspending (PMF_F_SELF). rtw(4) improvements: Destroy rtw(4) callouts when we detach it. Make rtw at pci detachable. Print some more information with the "rx frame too long" warning. Remove activate() methods: Get rid of rtw_activate() and ath_activate(). The device activate() methods are not good for much these days. Make ath at cardbus resume with crypto functions intact: Introduce a boolean device property, "pmf-powerdown". If pmf-powerdown is present and false, it indicates that a bus back-end should not remove power from a device. Honor this property in cardbus_child_suspend(). Set this property to 'false' in ath_attach(), since removing power from an ath at cardbus seems to lobotomize the WPA crypto engine. XXX Should the pmf-powerdown property propagate toward the root of the device tree? Miscellaneous ath(4) changes: Warn if ath(4) tries to write crypto keys to suspended hardware. Reduce differences between FreeBSD and NetBSD in ath(4) multicast filter setup. Make ath_printrxbuf() print an rx descriptor's status & key index, to help debug crypto errors. Shorten a staircase in ath_ioctl(). Don't check for ieee80211_ioctl() return code ERESTART, it never happens.
2008-03-12 21:02:21 +03:00
if (!device_is_active(sc->sc_dev))
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
return;
for (pri = 0; pri < RTW_NTXPRI; pri++) {
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
tsb = &sc->sc_txsoft_blk[pri];
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
if (tsb->tsb_tx_timer == 0)
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
continue;
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
else if (--tsb->tsb_tx_timer == 0) {
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
if (SIMPLEQ_EMPTY(&tsb->tsb_dirtyq))
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
continue;
Add the WEP capability to our net80211 capability flags, since we do sort of support hardware WEP. Only change the WEP key index in a Tx descriptor from 0 if the type of our transmit key is _WEP, not _TKIP: i.e., only if we're really doing WEP crypto in the hardware. Ignore a watchdog timeout on any Tx ring if we can collect some packets from that ring. Restart both the receiver and the transmitter when a watchdog timeout occurs instead of restarting only the transmitter.
2009-04-02 04:57:20 +04:00
else if (rtw_collect_txring(sc, tsb,
&sc->sc_txdesc_blk[pri], 0))
continue;
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
printf("%s: transmit timeout, priority %d\n",
ifp->if_xname, pri);
ifp->if_oerrors++;
I have made a small change that suppresses "recovery" from (spurious?) tx timeouts on the beacon queue. The driver's recovery process badly disrupts the MAC's receiver. This needs a closer look, later. Every other fix is by Charles M. Hannum. Charles writes: "With these, the device is quite stable under load, does not experience xmit timeouts, no longer mysteriously freezes, and does not spew a lot of garbage. It does, however, continue to get some "tx fifo underflow"s. global, SYSCTL_SETUP(): a) Removed unused sysctl variables. rtw_io_enable(): b) Change rtw_io_enable to take a softc, and make it call rtw_txring_fixup(). c) Enable the PCI multiple read/write bit. d) Always reset rdb_next to 0 when starting the receive engine. rtw_intr_rx(): e) When processing receive packets, store the next pointer in rdb_next and read it back on the next round, to fix a race condition vs. rtw_tune() during AP discovery and association. f) Remove the special "skip to 0th descriptor" code, because it is no longer needed. rtw_collect_txring(): g) In the "force" case, stop when we catch up to the hardware xmit pointer. rtw_intr_beacon(): h) Resetting tdb_next here looked fishy to me, in light of changes in rtw_collect_txring(). XXX However, this is the one part I haven't tested. rtw_txring_fixup(): i) Add function and line arguments, so we can print out information about the caller. j) In addition to reading the hardware pointer, skip over any additional packets with OWN set -- we may have queued more. rtw_rxring_fixup(): k) Remove this, because it's completely wrong. rtw_intr_ioerror(): l) Don't handle receive overruns here; handle them in the normal receive path. Also changed RTW_INTR_RX and RTW_INTR_IOERROR accordingly. m) Do not fiddle with CR; the transmit engine is still running just fine. n) Do not fiddle with IMR; it was unnecessary. o) Don't touch the receive engine at all; not necessary now that we're leaving CR alone. p) Remove the rtw_do_chip_reset case; it was dangerous and not actually used. rtw_intr(): q) Don't mask off ISR bits when we pass down the ISR value; it's not necessary, and sometimes the other bits useful in debugging output. rtw_seg_too_short(): r) Don't spew here; the caller will do it if necessary. (Stops spewage in an expected case.) rtw_rxdesc_blk: s) Minor change to make the elements have the same order as txdesc_blk. general: t) Added a bunch of debugging output, #ifdef RTW_DIAG, that was useful while investigating the ring management problems. Also moved rtw_txring_fixup() into RTW_DIAG, as I believe the transmit pointer is always kept in sync now, and the function has potential race conditions.
2007-01-06 11:27:53 +03:00
if (pri != RTW_TXPRIBCN)
tx_timeouts++;
Add the transmit section of rtw(4): build up and tear down transmit descriptor rings, move packets from the transmit queues to the transmit rings, handle transmission-completed interrupts. My Linksys WPC11 ver. 4 with Maxim RF, generates interrupts like it is successfully transmitting packets. Clearly there are bugs: my G4 Powerbook locks up hard. I will debug tomorrow.
2004-12-19 11:19:25 +03:00
} else
ifp->if_timer = 1;
}
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
if (tx_timeouts > 0) {
/* Stop Tx DMA, disable xmtr, flush Tx rings, enable xmtr,
* reset s/w tx-ring pointers, and start transmission.
*
* TBD Stop/restart just the broken rings?
*/
rtw_idle(&sc->sc_regs);
Add the WEP capability to our net80211 capability flags, since we do sort of support hardware WEP. Only change the WEP key index in a Tx descriptor from 0 if the type of our transmit key is _WEP, not _TKIP: i.e., only if we're really doing WEP crypto in the hardware. Ignore a watchdog timeout on any Tx ring if we can collect some packets from that ring. Restart both the receiver and the transmitter when a watchdog timeout occurs instead of restarting only the transmitter.
2009-04-02 04:57:20 +04:00
rtw_io_enable(sc, RTW_CR_RE | RTW_CR_TE, 0);
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
rtw_txdescs_reset(sc);
Add the WEP capability to our net80211 capability flags, since we do sort of support hardware WEP. Only change the WEP key index in a Tx descriptor from 0 if the type of our transmit key is _WEP, not _TKIP: i.e., only if we're really doing WEP crypto in the hardware. Ignore a watchdog timeout on any Tx ring if we can collect some packets from that ring. Restart both the receiver and the transmitter when a watchdog timeout occurs instead of restarting only the transmitter.
2009-04-02 04:57:20 +04:00
rtw_io_enable(sc, RTW_CR_RE | RTW_CR_TE, 1);
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
rtw_start(ifp);
}
Resolve conflicts in importation of 18-May-2005 ath(4) / net80211(9) from FreeBSD. Introduce compatibility shims (sys/dev/ic/ath_netbsd.[ch], sys/net80211/ieee80211_netbsd.[ch]). Update drivers (an, atu, atw, awi, ipw, iwi, rtw, wi) for the new net80211(9) API.
2005-06-22 10:14:51 +04:00
ieee80211_watchdog(&sc->sc_ic);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
return;
}
static void
rtw_next_scan(void *arg)
{
struct ieee80211com *ic = arg;
int s;
/* don't call rtw_start w/o network interrupts blocked */
s = splnet();
if (ic->ic_state == IEEE80211_S_SCAN)
ieee80211_next_scan(ic);
splx(s);
}
Realtek sample sources are not consistent about the "Digital PHY" bit in the SROM. It seems as if it is set to 1 when the PHY is *analog*, not *digital*. Fix my sources. In rtw_intr_rx, use units of 500kb/s instead of 100kb/s for rate, to be consistent with net80211's expectations. Polish up some debugging ugly messages. Dump raw 802.11 packets if IFF_DEBUG|IFF_LINK2 and RTW_DEBUG is defined. Polish power-state (on/sleep/off) handling. Especially improve support for RFMD (totally untested) and Maxim. For Philips, take the Digital PHY property into account. Call the net80211 watchdog function from rtw_watchdog, so that we scan again if auth/assoc fails. Be a little more cautious about writing register[RTW_TPPOLL], since other drivers are.... Don't frob the high/low-priority queues right now, since I don't use them. Add rtw_join_bss which programs the card with the BSSID and other properties of a BSS. Use it on state transitions. Factor out rtw_set_nettype. Make rtw_recv_beacon call ieee80211_recv_mgmt instead of dropping beacons on the floor! TBD IBSS merges. Change some rtw_debug=2 printfs to rtw_debug=3 (RTW_DPRINTF3) printfs so the console doesn't get spammed so badly at rtw_debug=2. Change some debugging printfs to RTW_DPRINTFs. E.g., print the "RF programming method" only if debugging is enabled.
2004-12-21 02:05:41 +03:00
static void
Misc. changes to improve ad hoc mode and to enable hostap mode: Fix the work-around for the NIC bug where it skips to rx descriptor 0. The driver used to skip to rx descriptor 1. Hopefully this stops the out-of-order packet reception that Charles Hannum saw. When debugging is enabled, print rx-descriptor status flags before printing the rx bit rate. Add a debug message for when a beacon tx buffer reclamation. Reset IFF_OACTIVE when we reset the transmitter. Pass the consolidated LED state, a struct rtw_led_state, to rtw_led_attach. Choose the bit-rate for management frames (1Mb/s) at the same place we choose for all other frames. Do not use the NIC's short preamble or RTS options for management frames. Label beacons for the NIC. Following a Linux driver, take care not to zero arbitrary bits in the TPPOLL register. Use the new idiom for IBSS merges: disable transmitter, kick the state machine. Add a second descriptor to the beacon ring. The NIC seems to like this much better.
2005-03-02 08:20:43 +03:00
rtw_join_bss(struct rtw_softc *sc, uint8_t *bssid, uint16_t intval0)
Realtek sample sources are not consistent about the "Digital PHY" bit in the SROM. It seems as if it is set to 1 when the PHY is *analog*, not *digital*. Fix my sources. In rtw_intr_rx, use units of 500kb/s instead of 100kb/s for rate, to be consistent with net80211's expectations. Polish up some debugging ugly messages. Dump raw 802.11 packets if IFF_DEBUG|IFF_LINK2 and RTW_DEBUG is defined. Polish power-state (on/sleep/off) handling. Especially improve support for RFMD (totally untested) and Maxim. For Philips, take the Digital PHY property into account. Call the net80211 watchdog function from rtw_watchdog, so that we scan again if auth/assoc fails. Be a little more cautious about writing register[RTW_TPPOLL], since other drivers are.... Don't frob the high/low-priority queues right now, since I don't use them. Add rtw_join_bss which programs the card with the BSSID and other properties of a BSS. Use it on state transitions. Factor out rtw_set_nettype. Make rtw_recv_beacon call ieee80211_recv_mgmt instead of dropping beacons on the floor! TBD IBSS merges. Change some rtw_debug=2 printfs to rtw_debug=3 (RTW_DPRINTF3) printfs so the console doesn't get spammed so badly at rtw_debug=2. Change some debugging printfs to RTW_DPRINTFs. E.g., print the "RF programming method" only if debugging is enabled.
2004-12-21 02:05:41 +03:00
{
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
uint16_t bcnitv, bintritv, intval;
Realtek sample sources are not consistent about the "Digital PHY" bit in the SROM. It seems as if it is set to 1 when the PHY is *analog*, not *digital*. Fix my sources. In rtw_intr_rx, use units of 500kb/s instead of 100kb/s for rate, to be consistent with net80211's expectations. Polish up some debugging ugly messages. Dump raw 802.11 packets if IFF_DEBUG|IFF_LINK2 and RTW_DEBUG is defined. Polish power-state (on/sleep/off) handling. Especially improve support for RFMD (totally untested) and Maxim. For Philips, take the Digital PHY property into account. Call the net80211 watchdog function from rtw_watchdog, so that we scan again if auth/assoc fails. Be a little more cautious about writing register[RTW_TPPOLL], since other drivers are.... Don't frob the high/low-priority queues right now, since I don't use them. Add rtw_join_bss which programs the card with the BSSID and other properties of a BSS. Use it on state transitions. Factor out rtw_set_nettype. Make rtw_recv_beacon call ieee80211_recv_mgmt instead of dropping beacons on the floor! TBD IBSS merges. Change some rtw_debug=2 printfs to rtw_debug=3 (RTW_DPRINTF3) printfs so the console doesn't get spammed so badly at rtw_debug=2. Change some debugging printfs to RTW_DPRINTFs. E.g., print the "RF programming method" only if debugging is enabled.
2004-12-21 02:05:41 +03:00
int i;
struct rtw_regs *regs = &sc->sc_regs;
for (i = 0; i < IEEE80211_ADDR_LEN; i++)
RTW_WRITE8(regs, RTW_BSSID + i, bssid[i]);
RTW_SYNC(regs, RTW_BSSID16, RTW_BSSID32);
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
rtw_set_access(regs, RTW_ACCESS_CONFIG);
Realtek sample sources are not consistent about the "Digital PHY" bit in the SROM. It seems as if it is set to 1 when the PHY is *analog*, not *digital*. Fix my sources. In rtw_intr_rx, use units of 500kb/s instead of 100kb/s for rate, to be consistent with net80211's expectations. Polish up some debugging ugly messages. Dump raw 802.11 packets if IFF_DEBUG|IFF_LINK2 and RTW_DEBUG is defined. Polish power-state (on/sleep/off) handling. Especially improve support for RFMD (totally untested) and Maxim. For Philips, take the Digital PHY property into account. Call the net80211 watchdog function from rtw_watchdog, so that we scan again if auth/assoc fails. Be a little more cautious about writing register[RTW_TPPOLL], since other drivers are.... Don't frob the high/low-priority queues right now, since I don't use them. Add rtw_join_bss which programs the card with the BSSID and other properties of a BSS. Use it on state transitions. Factor out rtw_set_nettype. Make rtw_recv_beacon call ieee80211_recv_mgmt instead of dropping beacons on the floor! TBD IBSS merges. Change some rtw_debug=2 printfs to rtw_debug=3 (RTW_DPRINTF3) printfs so the console doesn't get spammed so badly at rtw_debug=2. Change some debugging printfs to RTW_DPRINTFs. E.g., print the "RF programming method" only if debugging is enabled.
2004-12-21 02:05:41 +03:00
Per discussion on tech-kern and tech-userlevel, move the bit-twiddling macros, __BIT, __BITS, SHIFTIN, SHIFTOUT, and __arraycount() from lib/libkern/libkern.h to sys/cdefs.h. Add a __-prefix to SHIFTIN and SHIFTOUT, and add a manual page for the bit-twiddling macros, bits(3). Make the __BIT and __BITS macros "widthless," as best I can, by changing their type to uintmax_t from uint32_t. XXX The manual page lags this change by a bit. Define __PRIxBIT and __PRIxBITS printf(3) format strings.
2006-08-31 23:24:37 +04:00
intval = MIN(intval0, __SHIFTOUT_MASK(RTW_BCNITV_BCNITV_MASK));
Realtek sample sources are not consistent about the "Digital PHY" bit in the SROM. It seems as if it is set to 1 when the PHY is *analog*, not *digital*. Fix my sources. In rtw_intr_rx, use units of 500kb/s instead of 100kb/s for rate, to be consistent with net80211's expectations. Polish up some debugging ugly messages. Dump raw 802.11 packets if IFF_DEBUG|IFF_LINK2 and RTW_DEBUG is defined. Polish power-state (on/sleep/off) handling. Especially improve support for RFMD (totally untested) and Maxim. For Philips, take the Digital PHY property into account. Call the net80211 watchdog function from rtw_watchdog, so that we scan again if auth/assoc fails. Be a little more cautious about writing register[RTW_TPPOLL], since other drivers are.... Don't frob the high/low-priority queues right now, since I don't use them. Add rtw_join_bss which programs the card with the BSSID and other properties of a BSS. Use it on state transitions. Factor out rtw_set_nettype. Make rtw_recv_beacon call ieee80211_recv_mgmt instead of dropping beacons on the floor! TBD IBSS merges. Change some rtw_debug=2 printfs to rtw_debug=3 (RTW_DPRINTF3) printfs so the console doesn't get spammed so badly at rtw_debug=2. Change some debugging printfs to RTW_DPRINTFs. E.g., print the "RF programming method" only if debugging is enabled.
2004-12-21 02:05:41 +03:00
bcnitv = RTW_READ16(regs, RTW_BCNITV) & ~RTW_BCNITV_BCNITV_MASK;
Per discussion on tech-kern and tech-userlevel, move the bit-twiddling macros, __BIT, __BITS, SHIFTIN, SHIFTOUT, and __arraycount() from lib/libkern/libkern.h to sys/cdefs.h. Add a __-prefix to SHIFTIN and SHIFTOUT, and add a manual page for the bit-twiddling macros, bits(3). Make the __BIT and __BITS macros "widthless," as best I can, by changing their type to uintmax_t from uint32_t. XXX The manual page lags this change by a bit. Define __PRIxBIT and __PRIxBITS printf(3) format strings.
2006-08-31 23:24:37 +04:00
bcnitv |= __SHIFTIN(intval, RTW_BCNITV_BCNITV_MASK);
Realtek sample sources are not consistent about the "Digital PHY" bit in the SROM. It seems as if it is set to 1 when the PHY is *analog*, not *digital*. Fix my sources. In rtw_intr_rx, use units of 500kb/s instead of 100kb/s for rate, to be consistent with net80211's expectations. Polish up some debugging ugly messages. Dump raw 802.11 packets if IFF_DEBUG|IFF_LINK2 and RTW_DEBUG is defined. Polish power-state (on/sleep/off) handling. Especially improve support for RFMD (totally untested) and Maxim. For Philips, take the Digital PHY property into account. Call the net80211 watchdog function from rtw_watchdog, so that we scan again if auth/assoc fails. Be a little more cautious about writing register[RTW_TPPOLL], since other drivers are.... Don't frob the high/low-priority queues right now, since I don't use them. Add rtw_join_bss which programs the card with the BSSID and other properties of a BSS. Use it on state transitions. Factor out rtw_set_nettype. Make rtw_recv_beacon call ieee80211_recv_mgmt instead of dropping beacons on the floor! TBD IBSS merges. Change some rtw_debug=2 printfs to rtw_debug=3 (RTW_DPRINTF3) printfs so the console doesn't get spammed so badly at rtw_debug=2. Change some debugging printfs to RTW_DPRINTFs. E.g., print the "RF programming method" only if debugging is enabled.
2004-12-21 02:05:41 +03:00
RTW_WRITE16(regs, RTW_BCNITV, bcnitv);
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
/* interrupt host 1ms before the TBTT */
bintritv = RTW_READ16(regs, RTW_BINTRITV) & ~RTW_BINTRITV_BINTRITV;
Per discussion on tech-kern and tech-userlevel, move the bit-twiddling macros, __BIT, __BITS, SHIFTIN, SHIFTOUT, and __arraycount() from lib/libkern/libkern.h to sys/cdefs.h. Add a __-prefix to SHIFTIN and SHIFTOUT, and add a manual page for the bit-twiddling macros, bits(3). Make the __BIT and __BITS macros "widthless," as best I can, by changing their type to uintmax_t from uint32_t. XXX The manual page lags this change by a bit. Define __PRIxBIT and __PRIxBITS printf(3) format strings.
2006-08-31 23:24:37 +04:00
bintritv |= __SHIFTIN(1000, RTW_BINTRITV_BINTRITV);
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
RTW_WRITE16(regs, RTW_BINTRITV, bintritv);
Realtek sample sources are not consistent about the "Digital PHY" bit in the SROM. It seems as if it is set to 1 when the PHY is *analog*, not *digital*. Fix my sources. In rtw_intr_rx, use units of 500kb/s instead of 100kb/s for rate, to be consistent with net80211's expectations. Polish up some debugging ugly messages. Dump raw 802.11 packets if IFF_DEBUG|IFF_LINK2 and RTW_DEBUG is defined. Polish power-state (on/sleep/off) handling. Especially improve support for RFMD (totally untested) and Maxim. For Philips, take the Digital PHY property into account. Call the net80211 watchdog function from rtw_watchdog, so that we scan again if auth/assoc fails. Be a little more cautious about writing register[RTW_TPPOLL], since other drivers are.... Don't frob the high/low-priority queues right now, since I don't use them. Add rtw_join_bss which programs the card with the BSSID and other properties of a BSS. Use it on state transitions. Factor out rtw_set_nettype. Make rtw_recv_beacon call ieee80211_recv_mgmt instead of dropping beacons on the floor! TBD IBSS merges. Change some rtw_debug=2 printfs to rtw_debug=3 (RTW_DPRINTF3) printfs so the console doesn't get spammed so badly at rtw_debug=2. Change some debugging printfs to RTW_DPRINTFs. E.g., print the "RF programming method" only if debugging is enabled.
2004-12-21 02:05:41 +03:00
/* magic from Linux */
Per discussion on tech-kern and tech-userlevel, move the bit-twiddling macros, __BIT, __BITS, SHIFTIN, SHIFTOUT, and __arraycount() from lib/libkern/libkern.h to sys/cdefs.h. Add a __-prefix to SHIFTIN and SHIFTOUT, and add a manual page for the bit-twiddling macros, bits(3). Make the __BIT and __BITS macros "widthless," as best I can, by changing their type to uintmax_t from uint32_t. XXX The manual page lags this change by a bit. Define __PRIxBIT and __PRIxBITS printf(3) format strings.
2006-08-31 23:24:37 +04:00
RTW_WRITE16(regs, RTW_ATIMWND, __SHIFTIN(1, RTW_ATIMWND_ATIMWND));
RTW_WRITE16(regs, RTW_ATIMTRITV, __SHIFTIN(2, RTW_ATIMTRITV_ATIMTRITV));
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
rtw_set_access(regs, RTW_ACCESS_NONE);
Realtek sample sources are not consistent about the "Digital PHY" bit in the SROM. It seems as if it is set to 1 when the PHY is *analog*, not *digital*. Fix my sources. In rtw_intr_rx, use units of 500kb/s instead of 100kb/s for rate, to be consistent with net80211's expectations. Polish up some debugging ugly messages. Dump raw 802.11 packets if IFF_DEBUG|IFF_LINK2 and RTW_DEBUG is defined. Polish power-state (on/sleep/off) handling. Especially improve support for RFMD (totally untested) and Maxim. For Philips, take the Digital PHY property into account. Call the net80211 watchdog function from rtw_watchdog, so that we scan again if auth/assoc fails. Be a little more cautious about writing register[RTW_TPPOLL], since other drivers are.... Don't frob the high/low-priority queues right now, since I don't use them. Add rtw_join_bss which programs the card with the BSSID and other properties of a BSS. Use it on state transitions. Factor out rtw_set_nettype. Make rtw_recv_beacon call ieee80211_recv_mgmt instead of dropping beacons on the floor! TBD IBSS merges. Change some rtw_debug=2 printfs to rtw_debug=3 (RTW_DPRINTF3) printfs so the console doesn't get spammed so badly at rtw_debug=2. Change some debugging printfs to RTW_DPRINTFs. E.g., print the "RF programming method" only if debugging is enabled.
2004-12-21 02:05:41 +03:00
I have made a small change that suppresses "recovery" from (spurious?) tx timeouts on the beacon queue. The driver's recovery process badly disrupts the MAC's receiver. This needs a closer look, later. Every other fix is by Charles M. Hannum. Charles writes: "With these, the device is quite stable under load, does not experience xmit timeouts, no longer mysteriously freezes, and does not spew a lot of garbage. It does, however, continue to get some "tx fifo underflow"s. global, SYSCTL_SETUP(): a) Removed unused sysctl variables. rtw_io_enable(): b) Change rtw_io_enable to take a softc, and make it call rtw_txring_fixup(). c) Enable the PCI multiple read/write bit. d) Always reset rdb_next to 0 when starting the receive engine. rtw_intr_rx(): e) When processing receive packets, store the next pointer in rdb_next and read it back on the next round, to fix a race condition vs. rtw_tune() during AP discovery and association. f) Remove the special "skip to 0th descriptor" code, because it is no longer needed. rtw_collect_txring(): g) In the "force" case, stop when we catch up to the hardware xmit pointer. rtw_intr_beacon(): h) Resetting tdb_next here looked fishy to me, in light of changes in rtw_collect_txring(). XXX However, this is the one part I haven't tested. rtw_txring_fixup(): i) Add function and line arguments, so we can print out information about the caller. j) In addition to reading the hardware pointer, skip over any additional packets with OWN set -- we may have queued more. rtw_rxring_fixup(): k) Remove this, because it's completely wrong. rtw_intr_ioerror(): l) Don't handle receive overruns here; handle them in the normal receive path. Also changed RTW_INTR_RX and RTW_INTR_IOERROR accordingly. m) Do not fiddle with CR; the transmit engine is still running just fine. n) Do not fiddle with IMR; it was unnecessary. o) Don't touch the receive engine at all; not necessary now that we're leaving CR alone. p) Remove the rtw_do_chip_reset case; it was dangerous and not actually used. rtw_intr(): q) Don't mask off ISR bits when we pass down the ISR value; it's not necessary, and sometimes the other bits useful in debugging output. rtw_seg_too_short(): r) Don't spew here; the caller will do it if necessary. (Stops spewage in an expected case.) rtw_rxdesc_blk: s) Minor change to make the elements have the same order as txdesc_blk. general: t) Added a bunch of debugging output, #ifdef RTW_DIAG, that was useful while investigating the ring management problems. Also moved rtw_txring_fixup() into RTW_DIAG, as I believe the transmit pointer is always kept in sync now, and the function has potential race conditions.
2007-01-06 11:27:53 +03:00
rtw_io_enable(sc, RTW_CR_RE | RTW_CR_TE, 1);
Realtek sample sources are not consistent about the "Digital PHY" bit in the SROM. It seems as if it is set to 1 when the PHY is *analog*, not *digital*. Fix my sources. In rtw_intr_rx, use units of 500kb/s instead of 100kb/s for rate, to be consistent with net80211's expectations. Polish up some debugging ugly messages. Dump raw 802.11 packets if IFF_DEBUG|IFF_LINK2 and RTW_DEBUG is defined. Polish power-state (on/sleep/off) handling. Especially improve support for RFMD (totally untested) and Maxim. For Philips, take the Digital PHY property into account. Call the net80211 watchdog function from rtw_watchdog, so that we scan again if auth/assoc fails. Be a little more cautious about writing register[RTW_TPPOLL], since other drivers are.... Don't frob the high/low-priority queues right now, since I don't use them. Add rtw_join_bss which programs the card with the BSSID and other properties of a BSS. Use it on state transitions. Factor out rtw_set_nettype. Make rtw_recv_beacon call ieee80211_recv_mgmt instead of dropping beacons on the floor! TBD IBSS merges. Change some rtw_debug=2 printfs to rtw_debug=3 (RTW_DPRINTF3) printfs so the console doesn't get spammed so badly at rtw_debug=2. Change some debugging printfs to RTW_DPRINTFs. E.g., print the "RF programming method" only if debugging is enabled.
2004-12-21 02:05:41 +03:00
}
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
/* Synchronize the hardware state with the software state. */
static int
rtw_newstate(struct ieee80211com *ic, enum ieee80211_state nstate, int arg)
{
Resolve conflicts in importation of 18-May-2005 ath(4) / net80211(9) from FreeBSD. Introduce compatibility shims (sys/dev/ic/ath_netbsd.[ch], sys/net80211/ieee80211_netbsd.[ch]). Update drivers (an, atu, atw, awi, ipw, iwi, rtw, wi) for the new net80211(9) API.
2005-06-22 10:14:51 +04:00
struct ifnet *ifp = ic->ic_ifp;
struct rtw_softc *sc = (struct rtw_softc *)ifp->if_softc;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
enum ieee80211_state ostate;
int error;
ostate = ic->ic_state;
Use device_t and its accessors throughout. Use aprint_*_dev(). Improve PMF-ability. Add a 'flags' argument to suspend/resume handlers and callers such as pmf_system_suspend(). Define a flag, PMF_F_SELF, which indicates to PMF that a device is suspending/resuming itself. Add helper routines, pmf_device_suspend_self(dev) and pmf_device_resume_self(dev), that call pmf_device_suspend(dev, PMF_F_SELF) and pmf_device_resume(dev, PMF_F_SELF), respectively. Use PMF_F_SELF to suspend/resume self in ath(4), audio(4), rtw(4), and sip(4). In ath(4) and in rtw(4), replace the icky sc_enable/sc_disable callbacks, provided by the bus front-end, with self-suspension/resumption. Also, clean up the bus front-ends. Make sure that the interrupt handler is disestablished during suspension. Get rid of driver-private flags (e.g., RTW_F_ENABLED, ath_softc->sc_invalid); use device_is_active()/device_has_power() calls, instead. In the network-class suspend handler, call if_stop(, 0) instead of if_stop(, 1), because the latter is superfluous (bus- and driver-suspension hooks will 'disable' the NIC), and it may cause recursion. In the network-class resume handler, prevent infinite recursion through if_init() by getting out early if we are self-suspending (PMF_F_SELF). rtw(4) improvements: Destroy rtw(4) callouts when we detach it. Make rtw at pci detachable. Print some more information with the "rx frame too long" warning. Remove activate() methods: Get rid of rtw_activate() and ath_activate(). The device activate() methods are not good for much these days. Make ath at cardbus resume with crypto functions intact: Introduce a boolean device property, "pmf-powerdown". If pmf-powerdown is present and false, it indicates that a bus back-end should not remove power from a device. Honor this property in cardbus_child_suspend(). Set this property to 'false' in ath_attach(), since removing power from an ath at cardbus seems to lobotomize the WPA crypto engine. XXX Should the pmf-powerdown property propagate toward the root of the device tree? Miscellaneous ath(4) changes: Warn if ath(4) tries to write crypto keys to suspended hardware. Reduce differences between FreeBSD and NetBSD in ath(4) multicast filter setup. Make ath_printrxbuf() print an rx descriptor's status & key index, to help debug crypto errors. Shorten a staircase in ath_ioctl(). Don't check for ieee80211_ioctl() return code ERESTART, it never happens.
2008-03-12 21:02:21 +03:00
aprint_debug_dev(sc->sc_dev, "%s: l.%d\n", __func__, __LINE__);
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
rtw_led_newstate(sc, nstate);
Use device_t and its accessors throughout. Use aprint_*_dev(). Improve PMF-ability. Add a 'flags' argument to suspend/resume handlers and callers such as pmf_system_suspend(). Define a flag, PMF_F_SELF, which indicates to PMF that a device is suspending/resuming itself. Add helper routines, pmf_device_suspend_self(dev) and pmf_device_resume_self(dev), that call pmf_device_suspend(dev, PMF_F_SELF) and pmf_device_resume(dev, PMF_F_SELF), respectively. Use PMF_F_SELF to suspend/resume self in ath(4), audio(4), rtw(4), and sip(4). In ath(4) and in rtw(4), replace the icky sc_enable/sc_disable callbacks, provided by the bus front-end, with self-suspension/resumption. Also, clean up the bus front-ends. Make sure that the interrupt handler is disestablished during suspension. Get rid of driver-private flags (e.g., RTW_F_ENABLED, ath_softc->sc_invalid); use device_is_active()/device_has_power() calls, instead. In the network-class suspend handler, call if_stop(, 0) instead of if_stop(, 1), because the latter is superfluous (bus- and driver-suspension hooks will 'disable' the NIC), and it may cause recursion. In the network-class resume handler, prevent infinite recursion through if_init() by getting out early if we are self-suspending (PMF_F_SELF). rtw(4) improvements: Destroy rtw(4) callouts when we detach it. Make rtw at pci detachable. Print some more information with the "rx frame too long" warning. Remove activate() methods: Get rid of rtw_activate() and ath_activate(). The device activate() methods are not good for much these days. Make ath at cardbus resume with crypto functions intact: Introduce a boolean device property, "pmf-powerdown". If pmf-powerdown is present and false, it indicates that a bus back-end should not remove power from a device. Honor this property in cardbus_child_suspend(). Set this property to 'false' in ath_attach(), since removing power from an ath at cardbus seems to lobotomize the WPA crypto engine. XXX Should the pmf-powerdown property propagate toward the root of the device tree? Miscellaneous ath(4) changes: Warn if ath(4) tries to write crypto keys to suspended hardware. Reduce differences between FreeBSD and NetBSD in ath(4) multicast filter setup. Make ath_printrxbuf() print an rx descriptor's status & key index, to help debug crypto errors. Shorten a staircase in ath_ioctl(). Don't check for ieee80211_ioctl() return code ERESTART, it never happens.
2008-03-12 21:02:21 +03:00
aprint_debug_dev(sc->sc_dev, "%s: l.%d\n", __func__, __LINE__);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
if (nstate == IEEE80211_S_INIT) {
callout_stop(&sc->sc_scan_ch);
sc->sc_cur_chan = IEEE80211_CHAN_ANY;
return (*sc->sc_mtbl.mt_newstate)(ic, nstate, arg);
}
if (ostate == IEEE80211_S_INIT && nstate != IEEE80211_S_INIT)
rtw_pwrstate(sc, RTW_ON);
if ((error = rtw_tune(sc)) != 0)
return error;
switch (nstate) {
case IEEE80211_S_INIT:
panic("%s: unexpected state IEEE80211_S_INIT\n", __func__);
break;
case IEEE80211_S_SCAN:
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
if (ostate != IEEE80211_S_SCAN) {
(void)memset(ic->ic_bss->ni_bssid, 0,
IEEE80211_ADDR_LEN);
Misc. changes to improve ad hoc mode and to enable hostap mode: Fix the work-around for the NIC bug where it skips to rx descriptor 0. The driver used to skip to rx descriptor 1. Hopefully this stops the out-of-order packet reception that Charles Hannum saw. When debugging is enabled, print rx-descriptor status flags before printing the rx bit rate. Add a debug message for when a beacon tx buffer reclamation. Reset IFF_OACTIVE when we reset the transmitter. Pass the consolidated LED state, a struct rtw_led_state, to rtw_led_attach. Choose the bit-rate for management frames (1Mb/s) at the same place we choose for all other frames. Do not use the NIC's short preamble or RTS options for management frames. Label beacons for the NIC. Following a Linux driver, take care not to zero arbitrary bits in the TPPOLL register. Use the new idiom for IBSS merges: disable transmitter, kick the state machine. Add a second descriptor to the beacon ring. The NIC seems to like this much better.
2005-03-02 08:20:43 +03:00
rtw_set_nettype(sc, IEEE80211_M_MONITOR);
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
}
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
callout_reset(&sc->sc_scan_ch, rtw_dwelltime * hz / 1000,
rtw_next_scan, ic);
break;
case IEEE80211_S_RUN:
In sys/dev/ic/rtw.c: Use clue from rtk(4) and re(4) to fix the rtw(4) packet filter. Previously, I was using the wrong CRC32 function to hash multicast addresses; to compensate, I set the multicast filter to all 1s. Now that I hash the addresses correctly, I do not any longer set the filter to all 1s. In rtw_ioctl, avoid gratuitous re-initialization when the interface flags change. If a !IFF_UP -> IFF_UP transition, call rtw_init(); otherwise, only reload the packet filter. In sys/dev/ic/rtwreg.h: Put useful combinations of Receiver Control Register flags in RTW_RCR_PKTFILT_MASK, RTW_RCR_MONITOR, and RTW_RCR_PKTFILT_DEFAULT. (XXX RTW_RCR_MONITOR should be called RTW_RCR_PKTFILT_MONITOR.)
2005-01-03 06:07:12 +03:00
switch (ic->ic_opmode) {
case IEEE80211_M_HOSTAP:
case IEEE80211_M_IBSS:
Misc. changes to improve ad hoc mode and to enable hostap mode: Fix the work-around for the NIC bug where it skips to rx descriptor 0. The driver used to skip to rx descriptor 1. Hopefully this stops the out-of-order packet reception that Charles Hannum saw. When debugging is enabled, print rx-descriptor status flags before printing the rx bit rate. Add a debug message for when a beacon tx buffer reclamation. Reset IFF_OACTIVE when we reset the transmitter. Pass the consolidated LED state, a struct rtw_led_state, to rtw_led_attach. Choose the bit-rate for management frames (1Mb/s) at the same place we choose for all other frames. Do not use the NIC's short preamble or RTS options for management frames. Label beacons for the NIC. Following a Linux driver, take care not to zero arbitrary bits in the TPPOLL register. Use the new idiom for IBSS merges: disable transmitter, kick the state machine. Add a second descriptor to the beacon ring. The NIC seems to like this much better.
2005-03-02 08:20:43 +03:00
rtw_set_nettype(sc, IEEE80211_M_MONITOR);
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
/*FALLTHROUGH*/
case IEEE80211_M_AHDEMO:
Misc. changes to improve ad hoc mode and to enable hostap mode: Fix the work-around for the NIC bug where it skips to rx descriptor 0. The driver used to skip to rx descriptor 1. Hopefully this stops the out-of-order packet reception that Charles Hannum saw. When debugging is enabled, print rx-descriptor status flags before printing the rx bit rate. Add a debug message for when a beacon tx buffer reclamation. Reset IFF_OACTIVE when we reset the transmitter. Pass the consolidated LED state, a struct rtw_led_state, to rtw_led_attach. Choose the bit-rate for management frames (1Mb/s) at the same place we choose for all other frames. Do not use the NIC's short preamble or RTS options for management frames. Label beacons for the NIC. Following a Linux driver, take care not to zero arbitrary bits in the TPPOLL register. Use the new idiom for IBSS merges: disable transmitter, kick the state machine. Add a second descriptor to the beacon ring. The NIC seems to like this much better.
2005-03-02 08:20:43 +03:00
case IEEE80211_M_STA:
rtw_join_bss(sc, ic->ic_bss->ni_bssid,
In sys/dev/ic/rtw.c: Use clue from rtk(4) and re(4) to fix the rtw(4) packet filter. Previously, I was using the wrong CRC32 function to hash multicast addresses; to compensate, I set the multicast filter to all 1s. Now that I hash the addresses correctly, I do not any longer set the filter to all 1s. In rtw_ioctl, avoid gratuitous re-initialization when the interface flags change. If a !IFF_UP -> IFF_UP transition, call rtw_init(); otherwise, only reload the packet filter. In sys/dev/ic/rtwreg.h: Put useful combinations of Receiver Control Register flags in RTW_RCR_PKTFILT_MASK, RTW_RCR_MONITOR, and RTW_RCR_PKTFILT_DEFAULT. (XXX RTW_RCR_MONITOR should be called RTW_RCR_PKTFILT_MONITOR.)
2005-01-03 06:07:12 +03:00
ic->ic_bss->ni_intval);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
break;
In sys/dev/ic/rtw.c: Use clue from rtk(4) and re(4) to fix the rtw(4) packet filter. Previously, I was using the wrong CRC32 function to hash multicast addresses; to compensate, I set the multicast filter to all 1s. Now that I hash the addresses correctly, I do not any longer set the filter to all 1s. In rtw_ioctl, avoid gratuitous re-initialization when the interface flags change. If a !IFF_UP -> IFF_UP transition, call rtw_init(); otherwise, only reload the packet filter. In sys/dev/ic/rtwreg.h: Put useful combinations of Receiver Control Register flags in RTW_RCR_PKTFILT_MASK, RTW_RCR_MONITOR, and RTW_RCR_PKTFILT_DEFAULT. (XXX RTW_RCR_MONITOR should be called RTW_RCR_PKTFILT_MONITOR.)
2005-01-03 06:07:12 +03:00
case IEEE80211_M_MONITOR:
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
break;
In sys/dev/ic/rtw.c: Use clue from rtk(4) and re(4) to fix the rtw(4) packet filter. Previously, I was using the wrong CRC32 function to hash multicast addresses; to compensate, I set the multicast filter to all 1s. Now that I hash the addresses correctly, I do not any longer set the filter to all 1s. In rtw_ioctl, avoid gratuitous re-initialization when the interface flags change. If a !IFF_UP -> IFF_UP transition, call rtw_init(); otherwise, only reload the packet filter. In sys/dev/ic/rtwreg.h: Put useful combinations of Receiver Control Register flags in RTW_RCR_PKTFILT_MASK, RTW_RCR_MONITOR, and RTW_RCR_PKTFILT_DEFAULT. (XXX RTW_RCR_MONITOR should be called RTW_RCR_PKTFILT_MONITOR.)
2005-01-03 06:07:12 +03:00
}
Misc. changes to improve ad hoc mode and to enable hostap mode: Fix the work-around for the NIC bug where it skips to rx descriptor 0. The driver used to skip to rx descriptor 1. Hopefully this stops the out-of-order packet reception that Charles Hannum saw. When debugging is enabled, print rx-descriptor status flags before printing the rx bit rate. Add a debug message for when a beacon tx buffer reclamation. Reset IFF_OACTIVE when we reset the transmitter. Pass the consolidated LED state, a struct rtw_led_state, to rtw_led_attach. Choose the bit-rate for management frames (1Mb/s) at the same place we choose for all other frames. Do not use the NIC's short preamble or RTS options for management frames. Label beacons for the NIC. Following a Linux driver, take care not to zero arbitrary bits in the TPPOLL register. Use the new idiom for IBSS merges: disable transmitter, kick the state machine. Add a second descriptor to the beacon ring. The NIC seems to like this much better.
2005-03-02 08:20:43 +03:00
rtw_set_nettype(sc, ic->ic_opmode);
In sys/dev/ic/rtw.c: Use clue from rtk(4) and re(4) to fix the rtw(4) packet filter. Previously, I was using the wrong CRC32 function to hash multicast addresses; to compensate, I set the multicast filter to all 1s. Now that I hash the addresses correctly, I do not any longer set the filter to all 1s. In rtw_ioctl, avoid gratuitous re-initialization when the interface flags change. If a !IFF_UP -> IFF_UP transition, call rtw_init(); otherwise, only reload the packet filter. In sys/dev/ic/rtwreg.h: Put useful combinations of Receiver Control Register flags in RTW_RCR_PKTFILT_MASK, RTW_RCR_MONITOR, and RTW_RCR_PKTFILT_DEFAULT. (XXX RTW_RCR_MONITOR should be called RTW_RCR_PKTFILT_MONITOR.)
2005-01-03 06:07:12 +03:00
break;
Misc. changes to improve ad hoc mode and to enable hostap mode: Fix the work-around for the NIC bug where it skips to rx descriptor 0. The driver used to skip to rx descriptor 1. Hopefully this stops the out-of-order packet reception that Charles Hannum saw. When debugging is enabled, print rx-descriptor status flags before printing the rx bit rate. Add a debug message for when a beacon tx buffer reclamation. Reset IFF_OACTIVE when we reset the transmitter. Pass the consolidated LED state, a struct rtw_led_state, to rtw_led_attach. Choose the bit-rate for management frames (1Mb/s) at the same place we choose for all other frames. Do not use the NIC's short preamble or RTS options for management frames. Label beacons for the NIC. Following a Linux driver, take care not to zero arbitrary bits in the TPPOLL register. Use the new idiom for IBSS merges: disable transmitter, kick the state machine. Add a second descriptor to the beacon ring. The NIC seems to like this much better.
2005-03-02 08:20:43 +03:00
case IEEE80211_S_ASSOC:
In sys/dev/ic/rtw.c: Use clue from rtk(4) and re(4) to fix the rtw(4) packet filter. Previously, I was using the wrong CRC32 function to hash multicast addresses; to compensate, I set the multicast filter to all 1s. Now that I hash the addresses correctly, I do not any longer set the filter to all 1s. In rtw_ioctl, avoid gratuitous re-initialization when the interface flags change. If a !IFF_UP -> IFF_UP transition, call rtw_init(); otherwise, only reload the packet filter. In sys/dev/ic/rtwreg.h: Put useful combinations of Receiver Control Register flags in RTW_RCR_PKTFILT_MASK, RTW_RCR_MONITOR, and RTW_RCR_PKTFILT_DEFAULT. (XXX RTW_RCR_MONITOR should be called RTW_RCR_PKTFILT_MONITOR.)
2005-01-03 06:07:12 +03:00
case IEEE80211_S_AUTH:
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
break;
}
if (nstate != IEEE80211_S_SCAN)
callout_stop(&sc->sc_scan_ch);
return (*sc->sc_mtbl.mt_newstate)(ic, nstate, arg);
}
Add the logic for IBSS merges.
2005-01-04 04:00:30 +03:00
/* Extend a 32-bit TSF timestamp to a 64-bit timestamp. */
static uint64_t
rtw_tsf_extend(struct rtw_regs *regs, uint32_t rstamp)
{
uint32_t tsftl, tsfth;
tsfth = RTW_READ(regs, RTW_TSFTRH);
tsftl = RTW_READ(regs, RTW_TSFTRL);
if (tsftl < rstamp) /* Compensate for rollover. */
tsfth--;
return ((uint64_t)tsfth << 32) | rstamp;
}
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
static void
rtw_recv_mgmt(struct ieee80211com *ic, struct mbuf *m,
Cosmetic: change u_int{8,16,32,64}_t to C99-style uint{8,16,32,64}_t, which is preferred.
2005-01-02 07:23:03 +03:00
struct ieee80211_node *ni, int subtype, int rssi, uint32_t rstamp)
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
{
Resolve conflicts in importation of 18-May-2005 ath(4) / net80211(9) from FreeBSD. Introduce compatibility shims (sys/dev/ic/ath_netbsd.[ch], sys/net80211/ieee80211_netbsd.[ch]). Update drivers (an, atu, atw, awi, ipw, iwi, rtw, wi) for the new net80211(9) API.
2005-06-22 10:14:51 +04:00
struct ifnet *ifp = ic->ic_ifp;
struct rtw_softc *sc = (struct rtw_softc *)ifp->if_softc;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
Add the logic for IBSS merges.
2005-01-04 04:00:30 +03:00
(*sc->sc_mtbl.mt_recv_mgmt)(ic, m, ni, subtype, rssi, rstamp);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
switch (subtype) {
case IEEE80211_FC0_SUBTYPE_PROBE_RESP:
case IEEE80211_FC0_SUBTYPE_BEACON:
Update rtw(4) for the new IBSS merge idiom, where ieee80211_ibss_merge() does not return ENETRESET if the driver should change its BSSID, but it makes a RUN->RUN transition, instead.
2005-12-30 01:23:52 +03:00
if (ic->ic_opmode == IEEE80211_M_IBSS &&
Use device_t and its accessors throughout. Use aprint_*_dev(). Improve PMF-ability. Add a 'flags' argument to suspend/resume handlers and callers such as pmf_system_suspend(). Define a flag, PMF_F_SELF, which indicates to PMF that a device is suspending/resuming itself. Add helper routines, pmf_device_suspend_self(dev) and pmf_device_resume_self(dev), that call pmf_device_suspend(dev, PMF_F_SELF) and pmf_device_resume(dev, PMF_F_SELF), respectively. Use PMF_F_SELF to suspend/resume self in ath(4), audio(4), rtw(4), and sip(4). In ath(4) and in rtw(4), replace the icky sc_enable/sc_disable callbacks, provided by the bus front-end, with self-suspension/resumption. Also, clean up the bus front-ends. Make sure that the interrupt handler is disestablished during suspension. Get rid of driver-private flags (e.g., RTW_F_ENABLED, ath_softc->sc_invalid); use device_is_active()/device_has_power() calls, instead. In the network-class suspend handler, call if_stop(, 0) instead of if_stop(, 1), because the latter is superfluous (bus- and driver-suspension hooks will 'disable' the NIC), and it may cause recursion. In the network-class resume handler, prevent infinite recursion through if_init() by getting out early if we are self-suspending (PMF_F_SELF). rtw(4) improvements: Destroy rtw(4) callouts when we detach it. Make rtw at pci detachable. Print some more information with the "rx frame too long" warning. Remove activate() methods: Get rid of rtw_activate() and ath_activate(). The device activate() methods are not good for much these days. Make ath at cardbus resume with crypto functions intact: Introduce a boolean device property, "pmf-powerdown". If pmf-powerdown is present and false, it indicates that a bus back-end should not remove power from a device. Honor this property in cardbus_child_suspend(). Set this property to 'false' in ath_attach(), since removing power from an ath at cardbus seems to lobotomize the WPA crypto engine. XXX Should the pmf-powerdown property propagate toward the root of the device tree? Miscellaneous ath(4) changes: Warn if ath(4) tries to write crypto keys to suspended hardware. Reduce differences between FreeBSD and NetBSD in ath(4) multicast filter setup. Make ath_printrxbuf() print an rx descriptor's status & key index, to help debug crypto errors. Shorten a staircase in ath_ioctl(). Don't check for ieee80211_ioctl() return code ERESTART, it never happens.
2008-03-12 21:02:21 +03:00
ic->ic_state == IEEE80211_S_RUN &&
device_is_active(sc->sc_dev)) {
Update rtw(4) for the new IBSS merge idiom, where ieee80211_ibss_merge() does not return ENETRESET if the driver should change its BSSID, but it makes a RUN->RUN transition, instead.
2005-12-30 01:23:52 +03:00
uint64_t tsf = rtw_tsf_extend(&sc->sc_regs, rstamp);
if (le64toh(ni->ni_tstamp.tsf) >= tsf)
(void)ieee80211_ibss_merge(ni);
}
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
break;
default:
break;
}
return;
}
static struct ieee80211_node *
Resolve conflicts in importation of 18-May-2005 ath(4) / net80211(9) from FreeBSD. Introduce compatibility shims (sys/dev/ic/ath_netbsd.[ch], sys/net80211/ieee80211_netbsd.[ch]). Update drivers (an, atu, atw, awi, ipw, iwi, rtw, wi) for the new net80211(9) API.
2005-06-22 10:14:51 +04:00
rtw_node_alloc(struct ieee80211_node_table *nt)
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
{
Resolve conflicts in importation of 18-May-2005 ath(4) / net80211(9) from FreeBSD. Introduce compatibility shims (sys/dev/ic/ath_netbsd.[ch], sys/net80211/ieee80211_netbsd.[ch]). Update drivers (an, atu, atw, awi, ipw, iwi, rtw, wi) for the new net80211(9) API.
2005-06-22 10:14:51 +04:00
struct ifnet *ifp = nt->nt_ic->ic_ifp;
struct rtw_softc *sc = (struct rtw_softc *)ifp->if_softc;
struct ieee80211_node *ni = (*sc->sc_mtbl.mt_node_alloc)(nt);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
DPRINTF(sc, RTW_DEBUG_NODE,
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
("%s: alloc node %p\n", device_xname(sc->sc_dev), ni));
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
return ni;
}
static void
Resolve conflicts in importation of 18-May-2005 ath(4) / net80211(9) from FreeBSD. Introduce compatibility shims (sys/dev/ic/ath_netbsd.[ch], sys/net80211/ieee80211_netbsd.[ch]). Update drivers (an, atu, atw, awi, ipw, iwi, rtw, wi) for the new net80211(9) API.
2005-06-22 10:14:51 +04:00
rtw_node_free(struct ieee80211_node *ni)
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
{
Resolve conflicts in importation of 18-May-2005 ath(4) / net80211(9) from FreeBSD. Introduce compatibility shims (sys/dev/ic/ath_netbsd.[ch], sys/net80211/ieee80211_netbsd.[ch]). Update drivers (an, atu, atw, awi, ipw, iwi, rtw, wi) for the new net80211(9) API.
2005-06-22 10:14:51 +04:00
struct ieee80211com *ic = ni->ni_ic;
struct ifnet *ifp = ic->ic_ifp;
struct rtw_softc *sc = (struct rtw_softc *)ifp->if_softc;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
DPRINTF(sc, RTW_DEBUG_NODE,
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
("%s: freeing node %p %s\n", device_xname(sc->sc_dev), ni,
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
ether_sprintf(ni->ni_bssid)));
Resolve conflicts in importation of 18-May-2005 ath(4) / net80211(9) from FreeBSD. Introduce compatibility shims (sys/dev/ic/ath_netbsd.[ch], sys/net80211/ieee80211_netbsd.[ch]). Update drivers (an, atu, atw, awi, ipw, iwi, rtw, wi) for the new net80211(9) API.
2005-06-22 10:14:51 +04:00
(*sc->sc_mtbl.mt_node_free)(ni);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
}
static int
rtw_media_change(struct ifnet *ifp)
{
int error;
error = ieee80211_media_change(ifp);
if (error == ENETRESET) {
if ((ifp->if_flags & (IFF_RUNNING|IFF_UP)) ==
(IFF_RUNNING|IFF_UP))
rtw_init(ifp); /* XXX lose error */
error = 0;
}
return error;
}
static void
rtw_media_status(struct ifnet *ifp, struct ifmediareq *imr)
{
struct rtw_softc *sc = ifp->if_softc;
Use device_t and its accessors throughout. Use aprint_*_dev(). Improve PMF-ability. Add a 'flags' argument to suspend/resume handlers and callers such as pmf_system_suspend(). Define a flag, PMF_F_SELF, which indicates to PMF that a device is suspending/resuming itself. Add helper routines, pmf_device_suspend_self(dev) and pmf_device_resume_self(dev), that call pmf_device_suspend(dev, PMF_F_SELF) and pmf_device_resume(dev, PMF_F_SELF), respectively. Use PMF_F_SELF to suspend/resume self in ath(4), audio(4), rtw(4), and sip(4). In ath(4) and in rtw(4), replace the icky sc_enable/sc_disable callbacks, provided by the bus front-end, with self-suspension/resumption. Also, clean up the bus front-ends. Make sure that the interrupt handler is disestablished during suspension. Get rid of driver-private flags (e.g., RTW_F_ENABLED, ath_softc->sc_invalid); use device_is_active()/device_has_power() calls, instead. In the network-class suspend handler, call if_stop(, 0) instead of if_stop(, 1), because the latter is superfluous (bus- and driver-suspension hooks will 'disable' the NIC), and it may cause recursion. In the network-class resume handler, prevent infinite recursion through if_init() by getting out early if we are self-suspending (PMF_F_SELF). rtw(4) improvements: Destroy rtw(4) callouts when we detach it. Make rtw at pci detachable. Print some more information with the "rx frame too long" warning. Remove activate() methods: Get rid of rtw_activate() and ath_activate(). The device activate() methods are not good for much these days. Make ath at cardbus resume with crypto functions intact: Introduce a boolean device property, "pmf-powerdown". If pmf-powerdown is present and false, it indicates that a bus back-end should not remove power from a device. Honor this property in cardbus_child_suspend(). Set this property to 'false' in ath_attach(), since removing power from an ath at cardbus seems to lobotomize the WPA crypto engine. XXX Should the pmf-powerdown property propagate toward the root of the device tree? Miscellaneous ath(4) changes: Warn if ath(4) tries to write crypto keys to suspended hardware. Reduce differences between FreeBSD and NetBSD in ath(4) multicast filter setup. Make ath_printrxbuf() print an rx descriptor's status & key index, to help debug crypto errors. Shorten a staircase in ath_ioctl(). Don't check for ieee80211_ioctl() return code ERESTART, it never happens.
2008-03-12 21:02:21 +03:00
if (!device_is_active(sc->sc_dev)) {
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
imr->ifm_active = IFM_IEEE80211 | IFM_NONE;
imr->ifm_status = 0;
return;
}
ieee80211_media_status(ifp, imr);
}
Remove leading __ from __(const|inline|signed|volatile) -- it is obsolete.
2005-12-24 23:27:29 +03:00
static inline void
Get rid of the silly char (*dvname)[IFNAMSIZ] pattern. Replace it with const char *dvname. I'm not sure what I was thinking.
2004-12-20 03:28:02 +03:00
rtw_setifprops(struct ifnet *ifp, const char *dvname, void *softc)
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
{
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
(void)strlcpy(ifp->if_xname, dvname, IFNAMSIZ);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
ifp->if_softc = softc;
ifp->if_flags = IFF_SIMPLEX | IFF_BROADCAST | IFF_MULTICAST |
IFF_NOTRAILERS;
ifp->if_ioctl = rtw_ioctl;
ifp->if_start = rtw_start;
ifp->if_watchdog = rtw_watchdog;
ifp->if_init = rtw_init;
ifp->if_stop = rtw_stop;
}
Remove leading __ from __(const|inline|signed|volatile) -- it is obsolete.
2005-12-24 23:27:29 +03:00
static inline void
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
rtw_set80211props(struct ieee80211com *ic)
{
int nrate;
ic->ic_phytype = IEEE80211_T_DS;
ic->ic_opmode = IEEE80211_M_STA;
ic->ic_caps = IEEE80211_C_PMGT | IEEE80211_C_IBSS |
Add the WEP capability to our net80211 capability flags, since we do sort of support hardware WEP. Only change the WEP key index in a Tx descriptor from 0 if the type of our transmit key is _WEP, not _TKIP: i.e., only if we're really doing WEP crypto in the hardware. Ignore a watchdog timeout on any Tx ring if we can collect some packets from that ring. Restart both the receiver and the transmitter when a watchdog timeout occurs instead of restarting only the transmitter.
2009-04-02 04:57:20 +04:00
IEEE80211_C_HOSTAP | IEEE80211_C_MONITOR | IEEE80211_C_WEP;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
nrate = 0;
When we set up the supported rates table, mark 1Mb/s, 2Mb/s as basic rates.
2004-12-23 08:47:42 +03:00
ic->ic_sup_rates[IEEE80211_MODE_11B].rs_rates[nrate++] =
IEEE80211_RATE_BASIC | 2;
ic->ic_sup_rates[IEEE80211_MODE_11B].rs_rates[nrate++] =
IEEE80211_RATE_BASIC | 4;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
ic->ic_sup_rates[IEEE80211_MODE_11B].rs_rates[nrate++] = 11;
ic->ic_sup_rates[IEEE80211_MODE_11B].rs_rates[nrate++] = 22;
ic->ic_sup_rates[IEEE80211_MODE_11B].rs_nrates = nrate;
}
Remove leading __ from __(const|inline|signed|volatile) -- it is obsolete.
2005-12-24 23:27:29 +03:00
static inline void
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
rtw_set80211methods(struct rtw_mtbl *mtbl, struct ieee80211com *ic)
{
mtbl->mt_newstate = ic->ic_newstate;
ic->ic_newstate = rtw_newstate;
mtbl->mt_recv_mgmt = ic->ic_recv_mgmt;
ic->ic_recv_mgmt = rtw_recv_mgmt;
mtbl->mt_node_free = ic->ic_node_free;
ic->ic_node_free = rtw_node_free;
mtbl->mt_node_alloc = ic->ic_node_alloc;
ic->ic_node_alloc = rtw_node_alloc;
rtw: Try to get hardware WEP to work. It seems to work in the RX direction, but not in the TX direction. The net80211 crypto framework doesn't seem to cope very well with the assymetry (I'm probably missing something), so I will use software WEP for now. net80211: In ieee80211_compute_duration, figure out whether to add the WEP header to the packet overhead by checking the WEP bit in the Frame Control field of the 802.11 header, instead of checking the IEEE80211_F_PRIVACY flag. Also, if the WEP bit is present, assume that the frame described by (wh, len) has already already been WEP encapsulated, and adjust the payload length accordingly. XXX that's a grotty hack that I will have to revisit, later.
2005-06-27 09:49:13 +04:00
ic->ic_crypto.cs_key_delete = rtw_key_delete;
ic->ic_crypto.cs_key_set = rtw_key_set;
ic->ic_crypto.cs_key_update_begin = rtw_key_update_begin;
ic->ic_crypto.cs_key_update_end = rtw_key_update_end;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
}
Remove leading __ from __(const|inline|signed|volatile) -- it is obsolete.
2005-12-24 23:27:29 +03:00
static inline void
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
rtw_init_radiotap(struct rtw_softc *sc)
{
Accommodate the variety of RF front-ends when radiotapping frames: tap only RSSI when there is a Philips RF front-end. Tap both Barker lock quality and RSSI when there is any other RF front-end. Provide radiotap listeners a more complete picture of channel activity: in promiscuous mode, tap frames who do not pass the CRC32 check. Flag packets that were received with a short preamble. Ask the NIC to pass us 802.11 Control frames. Pass frames to radiotap listeners before stripping the FCS. Re-order operations in rtw_intr_rx() in order to accomplish all of this, taking care not to pass a broken packet to net80211! Do not provide a flags field when tapping xmitted frames. Assert sane Rx packet lengths. Really should check and drop instead of KASSERTing. I will revisit this, soon. Update copyright.
2007-11-17 02:35:19 +03:00
uint32_t present;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
memset(&sc->sc_rxtapu, 0, sizeof(sc->sc_rxtapu));
Add bpf(4) packet taps to rtw(4). Now you can tap DLT_EN10MB (Ethernet II), DLT_IEEE802_11 (802.11), and DLT_IEEE802_11_RADIO (radiotap + 802.11).
2004-12-29 01:30:07 +03:00
sc->sc_rxtap.rr_ihdr.it_len = htole16(sizeof(sc->sc_rxtapu));
Accommodate the variety of RF front-ends when radiotapping frames: tap only RSSI when there is a Philips RF front-end. Tap both Barker lock quality and RSSI when there is any other RF front-end. Provide radiotap listeners a more complete picture of channel activity: in promiscuous mode, tap frames who do not pass the CRC32 check. Flag packets that were received with a short preamble. Ask the NIC to pass us 802.11 Control frames. Pass frames to radiotap listeners before stripping the FCS. Re-order operations in rtw_intr_rx() in order to accomplish all of this, taking care not to pass a broken packet to net80211! Do not provide a flags field when tapping xmitted frames. Assert sane Rx packet lengths. Really should check and drop instead of KASSERTing. I will revisit this, soon. Update copyright.
2007-11-17 02:35:19 +03:00
if (sc->sc_rfchipid == RTW_RFCHIPID_PHILIPS)
present = htole32(RTW_PHILIPS_RX_RADIOTAP_PRESENT);
else
present = htole32(RTW_RX_RADIOTAP_PRESENT);
sc->sc_rxtap.rr_ihdr.it_present = present;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
memset(&sc->sc_txtapu, 0, sizeof(sc->sc_txtapu));
Add bpf(4) packet taps to rtw(4). Now you can tap DLT_EN10MB (Ethernet II), DLT_IEEE802_11 (802.11), and DLT_IEEE802_11_RADIO (radiotap + 802.11).
2004-12-29 01:30:07 +03:00
sc->sc_txtap.rt_ihdr.it_len = htole16(sizeof(sc->sc_txtapu));
sc->sc_txtap.rt_ihdr.it_present = htole32(RTW_TX_RADIOTAP_PRESENT);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
}
static int
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
rtw_txsoft_blk_setup(struct rtw_txsoft_blk *tsb, u_int qlen)
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
{
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
SIMPLEQ_INIT(&tsb->tsb_dirtyq);
SIMPLEQ_INIT(&tsb->tsb_freeq);
tsb->tsb_ndesc = qlen;
tsb->tsb_desc = malloc(qlen * sizeof(*tsb->tsb_desc), M_DEVBUF,
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
M_NOWAIT);
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
if (tsb->tsb_desc == NULL)
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
return ENOMEM;
return 0;
}
static void
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
rtw_txsoft_blk_cleanup_all(struct rtw_softc *sc)
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
{
int pri;
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
struct rtw_txsoft_blk *tsb;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
for (pri = 0; pri < RTW_NTXPRI; pri++) {
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
tsb = &sc->sc_txsoft_blk[pri];
free(tsb->tsb_desc, M_DEVBUF);
tsb->tsb_desc = NULL;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
}
}
static int
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
rtw_txsoft_blk_setup_all(struct rtw_softc *sc)
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
{
int pri, rc = 0;
int qlen[RTW_NTXPRI] =
{RTW_TXQLENLO, RTW_TXQLENMD, RTW_TXQLENHI, RTW_TXQLENBCN};
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
struct rtw_txsoft_blk *tsbs;
tsbs = sc->sc_txsoft_blk;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
for (pri = 0; pri < RTW_NTXPRI; pri++) {
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
rc = rtw_txsoft_blk_setup(&tsbs[pri], qlen[pri]);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
if (rc != 0)
break;
}
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
tsbs[RTW_TXPRILO].tsb_poll = RTW_TPPOLL_LPQ | RTW_TPPOLL_SLPQ;
tsbs[RTW_TXPRIMD].tsb_poll = RTW_TPPOLL_NPQ | RTW_TPPOLL_SNPQ;
tsbs[RTW_TXPRIHI].tsb_poll = RTW_TPPOLL_HPQ | RTW_TPPOLL_SHPQ;
tsbs[RTW_TXPRIBCN].tsb_poll = RTW_TPPOLL_BQ | RTW_TPPOLL_SBQ;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
return rc;
}
static void
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
rtw_txdesc_blk_setup(struct rtw_txdesc_blk *tdb, struct rtw_txdesc *desc,
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
u_int ndesc, bus_addr_t ofs, bus_addr_t physbase)
{
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
tdb->tdb_ndesc = ndesc;
tdb->tdb_desc = desc;
tdb->tdb_physbase = physbase;
tdb->tdb_ofs = ofs;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
(void)memset(tdb->tdb_desc, 0,
sizeof(tdb->tdb_desc[0]) * tdb->tdb_ndesc);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
rtw_txdesc_blk_init(tdb);
tdb->tdb_next = 0;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
}
static void
rtw_txdesc_blk_setup_all(struct rtw_softc *sc)
{
rtw_txdesc_blk_setup(&sc->sc_txdesc_blk[RTW_TXPRILO],
&sc->sc_descs->hd_txlo[0], RTW_NTXDESCLO,
RTW_RING_OFFSET(hd_txlo), RTW_RING_BASE(sc, hd_txlo));
rtw_txdesc_blk_setup(&sc->sc_txdesc_blk[RTW_TXPRIMD],
&sc->sc_descs->hd_txmd[0], RTW_NTXDESCMD,
RTW_RING_OFFSET(hd_txmd), RTW_RING_BASE(sc, hd_txmd));
rtw_txdesc_blk_setup(&sc->sc_txdesc_blk[RTW_TXPRIHI],
&sc->sc_descs->hd_txhi[0], RTW_NTXDESCHI,
RTW_RING_OFFSET(hd_txhi), RTW_RING_BASE(sc, hd_txhi));
rtw_txdesc_blk_setup(&sc->sc_txdesc_blk[RTW_TXPRIBCN],
&sc->sc_descs->hd_bcn[0], RTW_NTXDESCBCN,
RTW_RING_OFFSET(hd_bcn), RTW_RING_BASE(sc, hd_bcn));
}
static struct rtw_rf *
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
rtw_rf_attach(struct rtw_softc *sc, enum rtw_rfchipid rfchipid, int digphy)
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
{
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
rtw_rf_write_t rf_write;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
struct rtw_rf *rf;
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
switch (rfchipid) {
default:
rf_write = rtw_rf_hostwrite;
break;
case RTW_RFCHIPID_INTERSIL:
case RTW_RFCHIPID_PHILIPS:
case RTW_RFCHIPID_GCT: /* XXX a guess */
case RTW_RFCHIPID_RFMD:
rf_write = (rtw_host_rfio) ? rtw_rf_hostwrite : rtw_rf_macwrite;
break;
}
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
switch (rfchipid) {
Add support for the GCT Semiconductor GRF5101 transceiver/synthesizer. From OpenBSD. GCT will not provide any documentation, so there are many magic numbers in this code.
2005-12-30 01:27:17 +03:00
case RTW_RFCHIPID_GCT:
rf = rtw_grf5101_create(&sc->sc_regs, rf_write, 0);
sc->sc_pwrstate_cb = rtw_maxim_pwrstate;
break;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
case RTW_RFCHIPID_MAXIM:
rf = rtw_max2820_create(&sc->sc_regs, rf_write, 0);
sc->sc_pwrstate_cb = rtw_maxim_pwrstate;
break;
case RTW_RFCHIPID_PHILIPS:
rf = rtw_sa2400_create(&sc->sc_regs, rf_write, digphy);
sc->sc_pwrstate_cb = rtw_philips_pwrstate;
break;
Realtek sample sources are not consistent about the "Digital PHY" bit in the SROM. It seems as if it is set to 1 when the PHY is *analog*, not *digital*. Fix my sources. In rtw_intr_rx, use units of 500kb/s instead of 100kb/s for rate, to be consistent with net80211's expectations. Polish up some debugging ugly messages. Dump raw 802.11 packets if IFF_DEBUG|IFF_LINK2 and RTW_DEBUG is defined. Polish power-state (on/sleep/off) handling. Especially improve support for RFMD (totally untested) and Maxim. For Philips, take the Digital PHY property into account. Call the net80211 watchdog function from rtw_watchdog, so that we scan again if auth/assoc fails. Be a little more cautious about writing register[RTW_TPPOLL], since other drivers are.... Don't frob the high/low-priority queues right now, since I don't use them. Add rtw_join_bss which programs the card with the BSSID and other properties of a BSS. Use it on state transitions. Factor out rtw_set_nettype. Make rtw_recv_beacon call ieee80211_recv_mgmt instead of dropping beacons on the floor! TBD IBSS merges. Change some rtw_debug=2 printfs to rtw_debug=3 (RTW_DPRINTF3) printfs so the console doesn't get spammed so badly at rtw_debug=2. Change some debugging printfs to RTW_DPRINTFs. E.g., print the "RF programming method" only if debugging is enabled.
2004-12-21 02:05:41 +03:00
case RTW_RFCHIPID_RFMD:
/* XXX RFMD has no RF constructor */
sc->sc_pwrstate_cb = rtw_rfmd_pwrstate;
/*FALLTHROUGH*/
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
default:
return NULL;
}
rf->rf_continuous_tx_cb =
(rtw_continuous_tx_cb_t)rtw_continuous_tx_enable;
rf->rf_continuous_tx_arg = (void *)sc;
return rf;
}
/* Revision C and later use a different PHY delay setting than
* revisions A and B.
*/
Cosmetic: change u_int{8,16,32,64}_t to C99-style uint{8,16,32,64}_t, which is preferred.
2005-01-02 07:23:03 +03:00
static uint8_t
Sprinkle __UNCONST() and rename variables to quiet uiet -Wcast-qual, -Wshadow complaints.
2005-06-10 06:35:34 +04:00
rtw_check_phydelay(struct rtw_regs *regs, uint32_t old_rcr)
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
{
#define REVAB (RTW_RCR_MXDMA_UNLIMITED | RTW_RCR_AICV)
#define REVC (REVAB | RTW_RCR_RXFTH_WHOLE)
Per discussion on tech-kern and tech-userlevel, move the bit-twiddling macros, __BIT, __BITS, SHIFTIN, SHIFTOUT, and __arraycount() from lib/libkern/libkern.h to sys/cdefs.h. Add a __-prefix to SHIFTIN and SHIFTOUT, and add a manual page for the bit-twiddling macros, bits(3). Make the __BIT and __BITS macros "widthless," as best I can, by changing their type to uintmax_t from uint32_t. XXX The manual page lags this change by a bit. Define __PRIxBIT and __PRIxBITS printf(3) format strings.
2006-08-31 23:24:37 +04:00
uint8_t phydelay = __SHIFTIN(0x6, RTW_PHYDELAY_PHYDELAY);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
RTW_WRITE(regs, RTW_RCR, REVAB);
Add some bus barriers. Remark where they are needed, e.g., read_eeprom. Change some printfs to RTW_DPRINTF2s to quiet rtw(4) when debugging is disabled.
2004-12-20 04:13:45 +03:00
RTW_WBW(regs, RTW_RCR, RTW_RCR);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
RTW_WRITE(regs, RTW_RCR, REVC);
RTW_WBR(regs, RTW_RCR, RTW_RCR);
if ((RTW_READ(regs, RTW_RCR) & REVC) == REVC)
phydelay |= RTW_PHYDELAY_REVC_MAGIC;
Sprinkle __UNCONST() and rename variables to quiet uiet -Wcast-qual, -Wshadow complaints.
2005-06-10 06:35:34 +04:00
RTW_WRITE(regs, RTW_RCR, old_rcr); /* restore RCR */
Add some bus barriers. Remark where they are needed, e.g., read_eeprom. Change some printfs to RTW_DPRINTF2s to quiet rtw(4) when debugging is disabled.
2004-12-20 04:13:45 +03:00
RTW_SYNC(regs, RTW_RCR, RTW_RCR);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
return phydelay;
#undef REVC
}
void
rtw_attach(struct rtw_softc *sc)
{
Resolve conflicts in importation of 18-May-2005 ath(4) / net80211(9) from FreeBSD. Introduce compatibility shims (sys/dev/ic/ath_netbsd.[ch], sys/net80211/ieee80211_netbsd.[ch]). Update drivers (an, atu, atw, awi, ipw, iwi, rtw, wi) for the new net80211(9) API.
2005-06-22 10:14:51 +04:00
struct ifnet *ifp = &sc->sc_if;
Add undefined variable 'ic' to rtw_attach.
2005-07-11 10:01:42 +04:00
struct ieee80211com *ic = &sc->sc_ic;
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
struct rtw_txsoft_blk *tsb;
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
int pri, rc;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
In pmf(9), improve the implementation of device self-suspension and make suspension by self, by drvctl(8), and by ACPI system sleep play nice together. Start solidifying some temporary API changes. 1. Extract a new header file, <sys/device_if.h>, from <sys/device.h> and #include it from <sys/pmf.h> instead of <sys/device.h> to break the circular dependency between <sys/device.h> and <sys/pmf.h>. 2. Introduce pmf_qual_t, an aggregate of qualifications on a PMF suspend/resume call. Start to replace instances of PMF_FN_PROTO, PMF_FN_ARGS, et cetera, with a pmf_qual_t. 3. Introduce the notion of a "suspensor," an entity that holds a device in suspension. More than one suspensor may hold a device at once. A device stays suspended as long as at least one suspensor holds it. A device resumes when the last suspensor releases it. Currently, the kernel defines three suspensors, 3a the system-suspensor: for system suspension, initiated by 'sysctl -w machdep.sleep_state=3', by lid closure, by power-button press, et cetera, 3b the drvctl-suspensor: for device suspension by /dev/drvctl ioctl, e.g., drvctl -S sip0. 3c the system self-suspensor: for device drivers that suspend themselves and their children. Several drivers for network interfaces put the network device to sleep while it is not administratively up, that is, after the kernel calls if_stop(, 1). The self-suspensor should not be used directly. See the description of suspensor delegates, below. A suspensor can have one or more "delegates". A suspensor can release devices that its delegates hold suspended. Right now, only the system self-suspensor has delegates. For each device that a self-suspending driver attaches, it creates the device's self-suspensor, a delegate of the system self-suspensor. Suspensors stop a system-wide suspend/resume cycle from waking devices that the operator put to sleep with drvctl before the cycle. They also help self-suspension to work more simply, safely, and in accord with expectations. 4. Add the notion of device activation level, devact_level_t, and a routine for checking the current activation level, device_activation(). Current activation levels are DEVACT_LEVEL_BUS, DEVACT_LEVEL_DRIVER, and DEVACT_LEVEL_CLASS, which respectively indicate that the device's bus is active, that the bus and device are active, and that the bus, device, and the functions of the device's class (network, audio) are active. Suspend/resume calls can be qualified with a devact_level_t. The power-management framework treats a devact_level_t that qualifies a device suspension as the device's current activation level; it only runs hooks to reduce the activation level from the presumed current level to the fully suspended state. The framework treats a devact_level_t qualifying device resumption as the target activation level; it only runs hooks to raise the activation level to the target. 5. Use pmf_qual_t, devact_level_t, and self-suspensors in several drivers. 6. Temporarily add an unused power-management workqueue that I will remove or replace, soon.
2009-09-16 20:34:49 +04:00
pmf_self_suspensor_init(sc->sc_dev, &sc->sc_suspensor, &sc->sc_qual);
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
rtw_cipher_wep = ieee80211_cipher_wep;
rtw_cipher_wep.ic_decap = rtw_wep_decap;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
NEXT_ATTACH_STATE(sc, DETACHED);
switch (RTW_READ(&sc->sc_regs, RTW_TCR) & RTW_TCR_HWVERID_MASK) {
case RTW_TCR_HWVERID_F:
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
sc->sc_hwverid = 'F';
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
break;
case RTW_TCR_HWVERID_D:
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
sc->sc_hwverid = 'D';
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
break;
default:
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
sc->sc_hwverid = '?';
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
break;
}
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
aprint_verbose_dev(sc->sc_dev, "hardware version %c\n",
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
sc->sc_hwverid);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
rc = bus_dmamem_alloc(sc->sc_dmat, sizeof(struct rtw_descs),
RTW_DESC_ALIGNMENT, 0, &sc->sc_desc_segs, 1, &sc->sc_desc_nsegs,
0);
if (rc != 0) {
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
aprint_error_dev(sc->sc_dev,
"could not allocate hw descriptors, error %d\n", rc);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
goto err;
}
NEXT_ATTACH_STATE(sc, FINISH_DESC_ALLOC);
rc = bus_dmamem_map(sc->sc_dmat, &sc->sc_desc_segs,
sc->sc_desc_nsegs, sizeof(struct rtw_descs),
Kill caddr_t; there will be some MI fallout, but it will be fixed shortly.
2007-03-04 08:59:00 +03:00
(void **)&sc->sc_descs, BUS_DMA_COHERENT);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
if (rc != 0) {
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
aprint_error_dev(sc->sc_dev,
"could not map hw descriptors, error %d\n", rc);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
goto err;
}
NEXT_ATTACH_STATE(sc, FINISH_DESC_MAP);
rc = bus_dmamap_create(sc->sc_dmat, sizeof(struct rtw_descs), 1,
sizeof(struct rtw_descs), 0, 0, &sc->sc_desc_dmamap);
if (rc != 0) {
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
aprint_error_dev(sc->sc_dev,
"could not create DMA map for hw descriptors, error %d\n",
rc);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
goto err;
}
NEXT_ATTACH_STATE(sc, FINISH_DESCMAP_CREATE);
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
sc->sc_rxdesc_blk.rdb_dmat = sc->sc_dmat;
sc->sc_rxdesc_blk.rdb_dmamap = sc->sc_desc_dmamap;
Consolidate variables related to the rx ring in sc_rxdesc_blk, which is a struct rtw_rxdesc_blk. Put a copy of the DMA tag and the DMA map into the rx- and tx-ring blocks so that I don't have to pass them to subroutines all of the time.
2004-12-29 03:21:37 +03:00
for (pri = 0; pri < RTW_NTXPRI; pri++) {
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
sc->sc_txdesc_blk[pri].tdb_dmat = sc->sc_dmat;
sc->sc_txdesc_blk[pri].tdb_dmamap = sc->sc_desc_dmamap;
Consolidate variables related to the rx ring in sc_rxdesc_blk, which is a struct rtw_rxdesc_blk. Put a copy of the DMA tag and the DMA map into the rx- and tx-ring blocks so that I don't have to pass them to subroutines all of the time.
2004-12-29 03:21:37 +03:00
}
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
rc = bus_dmamap_load(sc->sc_dmat, sc->sc_desc_dmamap, sc->sc_descs,
sizeof(struct rtw_descs), NULL, 0);
if (rc != 0) {
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
aprint_error_dev(sc->sc_dev,
"could not load DMA map for hw descriptors, error %d\n",
rc);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
goto err;
}
NEXT_ATTACH_STATE(sc, FINISH_DESCMAP_LOAD);
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
if (rtw_txsoft_blk_setup_all(sc) != 0)
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
goto err;
NEXT_ATTACH_STATE(sc, FINISH_TXCTLBLK_SETUP);
rtw_txdesc_blk_setup_all(sc);
NEXT_ATTACH_STATE(sc, FINISH_TXDESCBLK_SETUP);
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
sc->sc_rxdesc_blk.rdb_desc = &sc->sc_descs->hd_rx[0];
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
for (pri = 0; pri < RTW_NTXPRI; pri++) {
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
tsb = &sc->sc_txsoft_blk[pri];
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
if ((rc = rtw_txdesc_dmamaps_create(sc->sc_dmat,
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
&tsb->tsb_desc[0], tsb->tsb_ndesc)) != 0) {
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
aprint_error_dev(sc->sc_dev,
"could not load DMA map for hw tx descriptors, "
"error %d\n", rc);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
goto err;
}
}
NEXT_ATTACH_STATE(sc, FINISH_TXMAPS_CREATE);
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
if ((rc = rtw_rxdesc_dmamaps_create(sc->sc_dmat, &sc->sc_rxsoft[0],
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
RTW_RXQLEN)) != 0) {
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
aprint_error_dev(sc->sc_dev,
"could not load DMA map for hw rx descriptors, error %d\n",
rc);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
goto err;
}
NEXT_ATTACH_STATE(sc, FINISH_RXMAPS_CREATE);
/* Reset the chip to a known state. */
if (rtw_reset(sc) != 0)
goto err;
NEXT_ATTACH_STATE(sc, FINISH_RESET);
sc->sc_rcr = RTW_READ(&sc->sc_regs, RTW_RCR);
if ((sc->sc_rcr & RTW_RCR_9356SEL) != 0)
sc->sc_flags |= RTW_F_9356SROM;
if (rtw_srom_read(&sc->sc_regs, sc->sc_flags, &sc->sc_srom,
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
sc->sc_dev) != 0)
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
goto err;
NEXT_ATTACH_STATE(sc, FINISH_READ_SROM);
if (rtw_srom_parse(&sc->sc_srom, &sc->sc_flags, &sc->sc_csthr,
&sc->sc_rfchipid, &sc->sc_rcr, &sc->sc_locale,
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
sc->sc_dev) != 0) {
aprint_error_dev(sc->sc_dev,
"attach failed, malformed serial ROM\n");
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
goto err;
}
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
aprint_verbose_dev(sc->sc_dev, "%s PHY\n",
Realtek sample sources are not consistent about the "Digital PHY" bit in the SROM. It seems as if it is set to 1 when the PHY is *analog*, not *digital*. Fix my sources. In rtw_intr_rx, use units of 500kb/s instead of 100kb/s for rate, to be consistent with net80211's expectations. Polish up some debugging ugly messages. Dump raw 802.11 packets if IFF_DEBUG|IFF_LINK2 and RTW_DEBUG is defined. Polish power-state (on/sleep/off) handling. Especially improve support for RFMD (totally untested) and Maxim. For Philips, take the Digital PHY property into account. Call the net80211 watchdog function from rtw_watchdog, so that we scan again if auth/assoc fails. Be a little more cautious about writing register[RTW_TPPOLL], since other drivers are.... Don't frob the high/low-priority queues right now, since I don't use them. Add rtw_join_bss which programs the card with the BSSID and other properties of a BSS. Use it on state transitions. Factor out rtw_set_nettype. Make rtw_recv_beacon call ieee80211_recv_mgmt instead of dropping beacons on the floor! TBD IBSS merges. Change some rtw_debug=2 printfs to rtw_debug=3 (RTW_DPRINTF3) printfs so the console doesn't get spammed so badly at rtw_debug=2. Change some debugging printfs to RTW_DPRINTFs. E.g., print the "RF programming method" only if debugging is enabled.
2004-12-21 02:05:41 +03:00
((sc->sc_flags & RTW_F_DIGPHY) != 0) ? "digital" : "analog");
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
aprint_verbose_dev(sc->sc_dev, "carrier-sense threshold %u\n",
sc->sc_csthr);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
NEXT_ATTACH_STATE(sc, FINISH_PARSE_SROM);
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
sc->sc_rf = rtw_rf_attach(sc, sc->sc_rfchipid,
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
sc->sc_flags & RTW_F_DIGPHY);
if (sc->sc_rf == NULL) {
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
aprint_verbose_dev(sc->sc_dev,
"attach failed, could not attach RF\n");
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
goto err;
}
NEXT_ATTACH_STATE(sc, FINISH_RF_ATTACH);
sc->sc_phydelay = rtw_check_phydelay(&sc->sc_regs, sc->sc_rcr);
Change rtw_debug from a debug level to a debug mask. Add a lot of debug flags. From Linux: handle an RTL8180 bug. Sometimes the NIC skips from the middle of the ring to the 0th rx descriptor. Now the driver resynchronizes. Handle a receive descriptor underrun or Rx FIFO overflow condition in the way that the Linux driver does. This kind of seems like overkill, but whatever. Protect rtw_ioctl with splnet(). Do not load a tx descriptor with a buffer shorter than 4 bytes. Handle a transmit timeout less disruptively.
2004-12-25 09:58:37 +03:00
RTW_DPRINTF(RTW_DEBUG_ATTACH,
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
("%s: PHY delay %d\n", device_xname(sc->sc_dev), sc->sc_phydelay));
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
if (sc->sc_locale == RTW_LOCALE_UNKNOWN)
Miscellaneous Realtek RTL8180L driver improvements: 1 Added new sysctl controls for debugging. 2 Improve detection & support for hardware WEP. 3 Revamp handling of transmit descriptor rings. 4 Reliably IFF_OACTIVE when transmit descriptors are available, to stop the transmit section of the driver from freezing up. 5 Fix beacon transmission in adhoc and hostap modes. XXX There is a wart in hostap mode, where beacons are transmitted at 1/2 the correct rate. Load beacon descriptors when the RTW_INTR_BINT interrupt arrives; schedule RTW_INTR_BINT 1ms ahead of the target beacon time. 6 Recover more gracefully from tx/rx errors: avoid transmitter/receiver/chip resets. Try to re-synchronize software state with hardware state---e.g., load next descriptor pointer from hardware. 7 Activate the transmit watchdog timer for beacons as well as other packets. 8 Introduce rtw_idle() that waits for transmit DMA to finish; call it before resetting the transmitter.
2005-11-23 07:12:33 +03:00
rtw_identify_country(&sc->sc_regs, &sc->sc_locale);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
rtw_init_channels(sc->sc_locale, &sc->sc_ic.ic_channels, sc->sc_dev);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
if (rtw_identify_sta(&sc->sc_regs, &sc->sc_ic.ic_myaddr,
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
sc->sc_dev) != 0)
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
goto err;
NEXT_ATTACH_STATE(sc, FINISH_ID_STA);
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
rtw_setifprops(ifp, device_xname(sc->sc_dev), (void*)sc);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
Make arg of IFQ_SET_READY be &ifp->if_snd, to match all other drivers and be compilable with altq enabled. Reviewed by dyoung@
2005-09-22 20:15:20 +04:00
IFQ_SET_READY(&ifp->if_snd);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
Resolve conflicts in importation of 18-May-2005 ath(4) / net80211(9) from FreeBSD. Introduce compatibility shims (sys/dev/ic/ath_netbsd.[ch], sys/net80211/ieee80211_netbsd.[ch]). Update drivers (an, atu, atw, awi, ipw, iwi, rtw, wi) for the new net80211(9) API.
2005-06-22 10:14:51 +04:00
sc->sc_ic.ic_ifp = ifp;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
rtw_set80211props(&sc->sc_ic);
Misc. changes to improve ad hoc mode and to enable hostap mode: Fix the work-around for the NIC bug where it skips to rx descriptor 0. The driver used to skip to rx descriptor 1. Hopefully this stops the out-of-order packet reception that Charles Hannum saw. When debugging is enabled, print rx-descriptor status flags before printing the rx bit rate. Add a debug message for when a beacon tx buffer reclamation. Reset IFF_OACTIVE when we reset the transmitter. Pass the consolidated LED state, a struct rtw_led_state, to rtw_led_attach. Choose the bit-rate for management frames (1Mb/s) at the same place we choose for all other frames. Do not use the NIC's short preamble or RTS options for management frames. Label beacons for the NIC. Following a Linux driver, take care not to zero arbitrary bits in the TPPOLL register. Use the new idiom for IBSS merges: disable transmitter, kick the state machine. Add a second descriptor to the beacon ring. The NIC seems to like this much better.
2005-03-02 08:20:43 +03:00
rtw_led_attach(&sc->sc_led_state, (void *)sc);
Move the register access mode into struct rtw_regs. Change rtw_set_access, rtw_set_access1 to match. Add a subroutine for setting WEP keys. WEP isn't quite finished, because I have to add the WEP header to Tx packets. Implement the SIOCS80211NWKEY ioctl for setting WEP keys. Program the LEDs based on operating state and packet activity. * On a Revision F RTL8180, blink LED1 at 1Hz to indicate scan/authenticate/associate states. In the run state, turn LED1 on. In every state, blink LED1 at 5Hz to indicate non-beacon tx/rx activity. I would like to use two LEDs, but in all my Rev. F instances, LED0 is not wired to an LED; instead, the first LED is wired to indicate that the card's power is on. * On a Revision D RTL8180, program the LEDs so that LED0 indicates Tx, and LED1 indicates Rx. The Rx LED will blink annoyingly if there are beacons in the air, but at least the Tx LED is useful. * Store the hardware revision in the softc to support my futile attempt at programming LEDs for both Rev. D and Rev. F parts; I never did get Rev. D LEDs to work right. * Add a debug flag RTW_DEBUG_LED for the LED transitions. Add RTW_TPPOLL_ALL, RTW_TPPOLL_SALL to start and stop, respectively, all of the transmit rings. In ad hoc mode, allocate a beacon and load it into the beacon ring. Start the ring. In one trial, the card re-transmitted the beacon ring's contents several times before stopping. More programming and testing for ad hoc mode is necessary. I'm not setting the beacon flag in the transmit descriptor. Revamp the transmit section to make better use of all the transmit rings: beacon queue, high-, low-, and medium-priority rings. Put beacon frames on the beacon ring. All other management frames, and data frames, go on the medium-priority ring. Power-save data frames go on the high-priority ring. (Note that powersaving is not implemented!) This is a work in progress. Send all 802.11 Management frames at 1Mbps. After we put a packet on a transmit ring, tickle the right bit in the TPPOLL to tell RTL8180. Stop all rings on error and in rtw_stop. Use the RF chip type, not the RTL8180 revision, to choose between host- and MAC-controlled RF serial I/O. Now the Netgear MA521 works. Remove bogus definition of bit RTW_TPPOLL_FSWINT.
2005-01-16 14:50:43 +03:00
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
/*
* Call MI attach routines.
*/
Resolve conflicts in importation of 18-May-2005 ath(4) / net80211(9) from FreeBSD. Introduce compatibility shims (sys/dev/ic/ath_netbsd.[ch], sys/net80211/ieee80211_netbsd.[ch]). Update drivers (an, atu, atw, awi, ipw, iwi, rtw, wi) for the new net80211(9) API.
2005-06-22 10:14:51 +04:00
if_attach(ifp);
ieee80211_ifattach(&sc->sc_ic);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
rtw_set80211methods(&sc->sc_mtbl, &sc->sc_ic);
/* possibly we should fill in our own sc_send_prresp, since
* the RTL8180 is probably sending probe responses in ad hoc
* mode.
*/
/* complete initialization */
Resolve conflicts in importation of 18-May-2005 ath(4) / net80211(9) from FreeBSD. Introduce compatibility shims (sys/dev/ic/ath_netbsd.[ch], sys/net80211/ieee80211_netbsd.[ch]). Update drivers (an, atu, atw, awi, ipw, iwi, rtw, wi) for the new net80211(9) API.
2005-06-22 10:14:51 +04:00
ieee80211_media_init(&sc->sc_ic, rtw_media_change, rtw_media_status);
Merge some of the less invasive changes from the vmlocking branch: - kthread, callout, devsw API changes - select()/poll() improvements - miscellaneous MT safety improvements
2007-07-10 00:51:58 +04:00
callout_init(&sc->sc_scan_ch, 0);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
Add bpf(4) packet taps to rtw(4). Now you can tap DLT_EN10MB (Ethernet II), DLT_IEEE802_11 (802.11), and DLT_IEEE802_11_RADIO (radiotap + 802.11).
2004-12-29 01:30:07 +03:00
rtw_init_radiotap(sc);
Push the bpf_ops usage back into bpf.h. Push the common ifp->if_bpf check into the inline functions as well the fourth argument for bpf_attach.
2010-04-05 11:19:28 +04:00
bpf_attach2(ifp, DLT_IEEE802_11_RADIO,
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
sizeof(struct ieee80211_frame) + 64, &sc->sc_radiobpf);
NEXT_ATTACH_STATE(sc, FINISHED);
Historically, an(4), ath(4), atw(4), rtw(4), and wi(4) have printed out their modes and rates at boot. Revert to the historical behavior.
2005-07-07 03:58:14 +04:00
ieee80211_announce(ic);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
return;
err:
rtw_detach(sc);
return;
}
int
rtw_detach(struct rtw_softc *sc)
{
Resolve conflicts in importation of 18-May-2005 ath(4) / net80211(9) from FreeBSD. Introduce compatibility shims (sys/dev/ic/ath_netbsd.[ch], sys/net80211/ieee80211_netbsd.[ch]). Update drivers (an, atu, atw, awi, ipw, iwi, rtw, wi) for the new net80211(9) API.
2005-06-22 10:14:51 +04:00
struct ifnet *ifp = &sc->sc_if;
Synchronize device detachment with network interrupts. Use the pmf network class instead of powerhooks.
2007-12-16 03:04:07 +03:00
int pri, s;
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
Synchronize device detachment with network interrupts. Use the pmf network class instead of powerhooks.
2007-12-16 03:04:07 +03:00
s = splnet();
In rtw_detach, flag the rtw as "invalid" (RTW_F_INVALID). Do not try to power-down the RF section of an invalid rtw. The radiotap signal quality field is 16, not 8, bits wide. Convert the byte order.
2004-12-29 22:41:04 +03:00
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
switch (sc->sc_attach_state) {
case FINISHED:
Resolve conflicts in importation of 18-May-2005 ath(4) / net80211(9) from FreeBSD. Introduce compatibility shims (sys/dev/ic/ath_netbsd.[ch], sys/net80211/ieee80211_netbsd.[ch]). Update drivers (an, atu, atw, awi, ipw, iwi, rtw, wi) for the new net80211(9) API.
2005-06-22 10:14:51 +04:00
rtw_stop(ifp, 1);
Miscellaneous changes. Details below. Important changes flagged with []. Using the driver with my Linksys WPC11 ver. 4, it seems to be receiving packets for a change. The WPC11 ver. 4 has a Maxim RF section. My no-name rtw with Philips RF section still does not receive any packets. Keep access-level (analog params > config[0123] registers > none) in sc_access. Add rtw_set_access for changing the access level. Make rtw_continuous_tx_enable and other subroutines use rtw_set_access instead of rtw_config0123_enable and rtw_anaparm_enable. Factor part of the chip-reset code into rtw_chip_reset1. Change the 'struct foo (*bar)[N]'-style arguments to 'struct foo *bar'-style arguments. Consolidate software/hardware Tx/Rx ring setup in rtw_hwring_setup, rtw_swring_setup. Add a new constant, SA2400_OPMODE_DEFAULTS, for the bits that we *always* set in the SA2400 OPMODE register. Factor some code out into rtw_sa2400_calibrate. (Inspired by the Linux driver.) [] When the receiver goes into underrun/overflow state, call a new subroutine, rtw_kick() that stops the Rx/Tx processes, resets the chip, reinitializes the Tx/Rx rings, and restarts Rx/Tx processes. (Inspired by the Linux driver.) [] In rtw_intr_rx, check for too-short packets before calling ieee80211_find_rxnode. I believe this will prevent a repeat of the MCHK exception I saw once on macppc. [] Use seconds-elapased as well as microseconds-elapsed to set the next "due date" for the timeout interrupt. This keeps the driver from programming the timeout to expire too early. [] In rtw_intr, read RTW_ISR at most 10 times, then get out. If the interface is not enabled (RTW_F_ENABLED), then get out. [] In rtw_stop, get out if the interface is not enabled (RTW_F_ENABLED). Block IPL_NET interrupts. Don't read/write any registers if the interface is invalid (RTW_F_INVALID). [] Call rtw_stop in rtw_detach.
2004-12-12 09:37:59 +03:00
Do not embed a struct device in rtw_softc any longer. Register both pci and cardbus attachments with CFATTACH_DECL_NEW(). Access the softc through the device_t using device_private(). While I'm here, change a couple of KASSERT()s about the Rx buffer length to a warning.
2007-12-21 21:22:43 +03:00
pmf_device_deregister(sc->sc_dev);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
callout_stop(&sc->sc_scan_ch);
Resolve conflicts in importation of 18-May-2005 ath(4) / net80211(9) from FreeBSD. Introduce compatibility shims (sys/dev/ic/ath_netbsd.[ch], sys/net80211/ieee80211_netbsd.[ch]). Update drivers (an, atu, atw, awi, ipw, iwi, rtw, wi) for the new net80211(9) API.
2005-06-22 10:14:51 +04:00
ieee80211_ifdetach(&sc->sc_ic);
if_detach(ifp);
Use device_t and its accessors throughout. Use aprint_*_dev(). Improve PMF-ability. Add a 'flags' argument to suspend/resume handlers and callers such as pmf_system_suspend(). Define a flag, PMF_F_SELF, which indicates to PMF that a device is suspending/resuming itself. Add helper routines, pmf_device_suspend_self(dev) and pmf_device_resume_self(dev), that call pmf_device_suspend(dev, PMF_F_SELF) and pmf_device_resume(dev, PMF_F_SELF), respectively. Use PMF_F_SELF to suspend/resume self in ath(4), audio(4), rtw(4), and sip(4). In ath(4) and in rtw(4), replace the icky sc_enable/sc_disable callbacks, provided by the bus front-end, with self-suspension/resumption. Also, clean up the bus front-ends. Make sure that the interrupt handler is disestablished during suspension. Get rid of driver-private flags (e.g., RTW_F_ENABLED, ath_softc->sc_invalid); use device_is_active()/device_has_power() calls, instead. In the network-class suspend handler, call if_stop(, 0) instead of if_stop(, 1), because the latter is superfluous (bus- and driver-suspension hooks will 'disable' the NIC), and it may cause recursion. In the network-class resume handler, prevent infinite recursion through if_init() by getting out early if we are self-suspending (PMF_F_SELF). rtw(4) improvements: Destroy rtw(4) callouts when we detach it. Make rtw at pci detachable. Print some more information with the "rx frame too long" warning. Remove activate() methods: Get rid of rtw_activate() and ath_activate(). The device activate() methods are not good for much these days. Make ath at cardbus resume with crypto functions intact: Introduce a boolean device property, "pmf-powerdown". If pmf-powerdown is present and false, it indicates that a bus back-end should not remove power from a device. Honor this property in cardbus_child_suspend(). Set this property to 'false' in ath_attach(), since removing power from an ath at cardbus seems to lobotomize the WPA crypto engine. XXX Should the pmf-powerdown property propagate toward the root of the device tree? Miscellaneous ath(4) changes: Warn if ath(4) tries to write crypto keys to suspended hardware. Reduce differences between FreeBSD and NetBSD in ath(4) multicast filter setup. Make ath_printrxbuf() print an rx descriptor's status & key index, to help debug crypto errors. Shorten a staircase in ath_ioctl(). Don't check for ieee80211_ioctl() return code ERESTART, it never happens.
2008-03-12 21:02:21 +03:00
rtw_led_detach(&sc->sc_led_state);
Fix a bug I introduced to rtw_chan2txpower() the other night. Fix a bug in device detachment: change a break statement to a /*FALLTHROUGH*/ in rtw_detach().
2007-12-17 00:39:33 +03:00
/*FALLTHROUGH*/
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
case FINISH_ID_STA:
case FINISH_RF_ATTACH:
rtw_rf_destroy(sc->sc_rf);
sc->sc_rf = NULL;
/*FALLTHROUGH*/
case FINISH_PARSE_SROM:
case FINISH_READ_SROM:
rtw_srom_free(&sc->sc_srom);
/*FALLTHROUGH*/
case FINISH_RESET:
case FINISH_RXMAPS_CREATE:
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
rtw_rxdesc_dmamaps_destroy(sc->sc_dmat, &sc->sc_rxsoft[0],
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
RTW_RXQLEN);
/*FALLTHROUGH*/
case FINISH_TXMAPS_CREATE:
for (pri = 0; pri < RTW_NTXPRI; pri++) {
rtw_txdesc_dmamaps_destroy(sc->sc_dmat,
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
sc->sc_txsoft_blk[pri].tsb_desc,
sc->sc_txsoft_blk[pri].tsb_ndesc);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
}
/*FALLTHROUGH*/
case FINISH_TXDESCBLK_SETUP:
case FINISH_TXCTLBLK_SETUP:
I like the tlp/atw-style names for software descriptors, txsoft, better than txctl. Change from rtw_txctl/rtw_rxctl to rtw_txsoft/rtw_rxsoft. Change the descriptor blocks' names to match: rtw_txctl_blk becomes rtw_txsoft_blk. Change the member-name prefixes for both software and hardware descriptors.
2004-12-29 04:06:52 +03:00
rtw_txsoft_blk_cleanup_all(sc);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
/*FALLTHROUGH*/
case FINISH_DESCMAP_LOAD:
bus_dmamap_unload(sc->sc_dmat, sc->sc_desc_dmamap);
/*FALLTHROUGH*/
case FINISH_DESCMAP_CREATE:
bus_dmamap_destroy(sc->sc_dmat, sc->sc_desc_dmamap);
/*FALLTHROUGH*/
case FINISH_DESC_MAP:
Kill caddr_t; there will be some MI fallout, but it will be fixed shortly.
2007-03-04 08:59:00 +03:00
bus_dmamem_unmap(sc->sc_dmat, (void *)sc->sc_descs,
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
sizeof(struct rtw_descs));
/*FALLTHROUGH*/
case FINISH_DESC_ALLOC:
bus_dmamem_free(sc->sc_dmat, &sc->sc_desc_segs,
sc->sc_desc_nsegs);
/*FALLTHROUGH*/
case DETACHED:
NEXT_ATTACH_STATE(sc, DETACHED);
break;
}
Synchronize device detachment with network interrupts. Use the pmf network class instead of powerhooks.
2007-12-16 03:04:07 +03:00
splx(s);
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
return 0;
}
Reference in New Issue Copy Permalink