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

4125 lines
102 KiB
C
Raw Normal View History

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
/* $NetBSD: rtw.c,v 1.47 2005/06/20 02:49:18 atatat 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
/*-
* Copyright (c) 2004, 2005 David Young. All rights reserved.
*
* 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.
* 3. The name of David Young may not be used to endorse or promote
* products derived from this software without specific prior
* written permission.
*
* 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>
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
__KERNEL_RCSID(0, "$NetBSD: rtw.c,v 1.47 2005/06/20 02:49:18 atatat 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 "bpfilter.h"
#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>
#include <machine/endian.h>
#include <machine/bus.h>
#include <machine/intr.h> /* splnet */
#include <uvm/uvm_extern.h>
nuke trailing whitespace
2005-02-27 03:26:58 +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
#include <net/if.h>
#include <net/if_media.h>
#include <net/if_ether.h>
#include <net80211/ieee80211_var.h>
#include <net80211/ieee80211_compat.h>
#include <net80211/ieee80211_radiotap.h>
nuke trailing whitespace
2005-02-27 03:26:58 +03:00
#if NBPFILTER > 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
#include <net/bpf.h>
nuke trailing whitespace
2005-02-27 03:26:58 +03:00
#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
#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>
#define KASSERT2(__cond, __msg) \
do { \
if (!(__cond)) \
panic __msg ; \
} while (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
int rtw_rfprog_fallback = 0;
int rtw_host_rfio = 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
#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;
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 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 */
Add work-in-progress driver rtw(4) for 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 rtw_start(struct ifnet *);
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 *);
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
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 */
/*
* Setup sysctl(3) MIB, hw.rtw.*
*
* TBD condition CTLFLAG_PERMANENT on being an LKM or not
*/
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,
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
MASK_AND_RSHIFT(RTW_CONFIG4_RFTYPE_MASK, RTW_CONFIG4_RFTYPE_MASK));
}
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
{
KASSERT(naccess >= RTW_ACCESS_NONE && naccess <= 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
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);
}
static __inline int
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_chip_reset1(struct rtw_regs *regs, const char *dvname)
Add work-in-progress driver rtw(4) for 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,
("%s: reset in %dus\n", dvname, 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
}
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
printf("%s: reset failed\n", dvname);
Add work-in-progress driver rtw(4) for 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;
}
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 __inline int
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_chip_reset(struct rtw_regs *regs, const char *dvname)
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 |
LSHIFT(7, RTW_TCR_SRL_MASK) | LSHIFT(7, RTW_TCR_LRL_MASK);
RTW_WRITE(regs, RTW_TCR, tcr);
RTW_WBW(regs, RTW_CR, RTW_TCR);
return rtw_chip_reset1(regs, dvname);
}
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_wep_setkeys(struct rtw_softc *sc, struct ieee80211_wepkey *wk, int txkey)
{
uint8_t cfg0, scr;
int i, j, tx_key_len;
struct rtw_regs *regs;
union rtw_keys *rk;
regs = &sc->sc_regs;
rk = &sc->sc_keys;
(void)memset(rk->rk_keys, 0, sizeof(rk->rk_keys));
scr = RTW_READ8(regs, RTW_SCR);
cfg0 = RTW_READ8(regs, RTW_CONFIG0);
scr &= ~(RTW_SCR_KM_MASK | RTW_SCR_TXSECON | RTW_SCR_RXSECON);
cfg0 &= ~(RTW_CONFIG0_WEP104 | RTW_CONFIG0_WEP40);
rtw_set_access(regs, RTW_ACCESS_CONFIG);
if ((sc->sc_ic.ic_flags & IEEE80211_F_PRIVACY) == 0)
goto out;
tx_key_len = wk[txkey].wk_len;
switch (tx_key_len) {
case 5:
scr |= RTW_SCR_TXSECON | RTW_SCR_RXSECON | RTW_SCR_KM_WEP40;
break;
case 13:
scr |= RTW_SCR_TXSECON | RTW_SCR_RXSECON | RTW_SCR_KM_WEP104;
break;
default:
goto out;
}
cfg0 |= RTW_CONFIG0_WEP104 | RTW_CONFIG0_WEP40;
for (i = j = 0; i < IEEE80211_WEP_NKID; i++) {
if (i == txkey)
sc->sc_txkey = j;
else if (wk[i].wk_len != tx_key_len)
continue;
(void)memcpy(rk->rk_keys[j++], wk[i].wk_key, wk[i].wk_len);
}
out:
bus_space_write_region_4(regs->r_bt, regs->r_bh,
RTW_DK0, rk->rk_words,
sizeof(rk->rk_words) / sizeof(rk->rk_words[0]));
bus_space_barrier(regs->r_bt, regs->r_bh, RTW_DK0,
sizeof(rk->rk_words) / sizeof(rk->rk_words[0]),
BUS_SPACE_BARRIER_SYNC);
RTW_WRITE8(regs, RTW_CONFIG0, cfg0);
RTW_WBW(regs, RTW_CONFIG0, RTW_SCR);
RTW_WRITE8(regs, RTW_SCR, scr);
RTW_SYNC(regs, RTW_SCR, RTW_SCR);
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
static __inline int
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_recall_eeprom(struct rtw_regs *regs, const char *dvname)
Add work-in-progress driver rtw(4) for 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);
/* wait 2.5ms for completion */
for (i = 0; i < 25; i++) {
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,
("%s: recall EEPROM in %dus\n", dvname, 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);
}
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
printf("%s: recall EEPROM failed\n", dvname);
Add work-in-progress driver rtw(4) for 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;
}
static __inline int
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
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
if ((rc = rtw_chip_reset(&sc->sc_regs, sc->sc_dev.dv_xname)) != 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;
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
if ((rc = rtw_recall_eeprom(&sc->sc_regs, sc->sc_dev.dv_xname)) != 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
;
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;
}
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;
}
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;
}
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
}
}
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
}
}
static __inline void
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
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_defaults(struct rtw_srom *sr, uint32_t *flags, 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,
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
const char *dvname)
Add work-in-progress driver rtw(4) for 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;
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 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
*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);
printf("%s: SROM version %d.%d", dvname,
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
Sprinkle __UNCONST() and rename variables to quiet uiet -Wcast-qual, -Wshadow complaints.
2005-06-10 06:35:34 +04:00
if (srom_version <= 0x0101) {
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
printf(" is not understood, limping along with defaults\n");
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;
}
printf("\n");
for (i = 0; i < IEEE80211_ADDR_LEN; i++)
mac[i] = RTW_SR_GET(sr, RTW_SR_MAC + 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,
("%s: EEPROM MAC %s\n", dvname, 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;
*rcr |= LSHIFT(MASK_AND_RSHIFT(RTW_SR_GET(sr, RTW_SR_RFPARM),
RTW_SR_RFPARM_CS_MASK), RTW_RCR_ENCS1);
*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;
}
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
printf("%s: RF: %s, PA: %s\n", dvname, 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:
*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,
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
const char *dvname)
Add work-in-progress driver rtw(4) for 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) {
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, ("%s: 93c56 SROM\n", dvname));
Add work-in-progress driver rtw(4) for 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 {
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, ("%s: 93c46 SROM\n", dvname));
Add work-in-progress driver rtw(4) for 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) {
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
printf("%s: unable to allocate SROM buffer\n", dvname);
Add work-in-progress driver rtw(4) for 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)) {
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
printf("%s: could not read SROM\n", dvname);
Add work-in-progress driver rtw(4) for 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);
if ((rc = rtw_recall_eeprom(regs, dvname)) != 0)
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,
("\n%s: serial ROM:\n\t", dvname));
Add work-in-progress driver rtw(4) for 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,
const char *dvname)
{
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:
cfg4 |= LSHIFT(rtw_rfprog_fallback, RTW_CONFIG4_RFTYPE_MASK);
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);
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,
("%s: %s RF programming method, %#02x\n", dvname, 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)));
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
static __inline void
rtw_init_channels(enum rtw_locale locale,
struct ieee80211_channel (*chans)[IEEE80211_CHAN_MAX+1],
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
const char *dvname)
Add work-in-progress driver rtw(4) for 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;
}
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
printf("%s: Geographic Location %s\n", dvname, 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
}
static __inline void
rtw_identify_country(struct rtw_regs *regs, enum rtw_locale *locale,
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
const char *dvname)
Add work-in-progress driver rtw(4) for 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:
*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;
}
}
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],
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
const char *dvname)
Add work-in-progress driver rtw(4) for 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);
(*addr)[0] = MASK_AND_RSHIFT(idr0, BITS(0, 7));
(*addr)[1] = MASK_AND_RSHIFT(idr0, BITS(8, 15));
(*addr)[2] = MASK_AND_RSHIFT(idr0, BITS(16, 23));
(*addr)[3] = MASK_AND_RSHIFT(idr0, BITS(24 ,31));
(*addr)[4] = MASK_AND_RSHIFT(idr1, BITS(0, 7));
(*addr)[5] = MASK_AND_RSHIFT(idr1, BITS(8, 15));
if (IEEE80211_ADDR_EQ(addr, empty_macaddr)) {
printf("%s: could not get mac address, attach failed\n",
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
dvname);
Add work-in-progress driver rtw(4) for 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;
}
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
printf("%s: 802.11 address %s\n", dvname, 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;
KASSERT2(idx >= RTW_SR_TXPOWER1 && idx <= RTW_SR_TXPOWER14,
("%s: channel %d out of range", __func__,
idx - RTW_SR_TXPOWER1 + 1));
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;
u_int ndesc[RTW_NTXPRI] =
{RTW_NTXDESCLO, RTW_NTXDESCMD, RTW_NTXDESCHI, RTW_NTXDESCBCN};
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[pri].tdb_nfree = ndesc[pri];
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
}
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
}
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
}
}
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,
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 *ndesc, const char *dvname)
Add work-in-progress driver rtw(4) for 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) {
printf("%s: TEST hit %d-buffer limit\n", dvname, i);
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) {
printf("%s: rtw_rxsoft_alloc failed, %d buffers, "
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 %d\n", dvname, i, rc);
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
}
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,
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
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
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
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
ctl = LSHIFT(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
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)));
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 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
bus_dmamap_sync(rdb->rdb_dmat, rs->rs_dmamap, 0,
rs->rs_dmamap->dm_mapsize, BUS_DMASYNC_PREREAD);
Add work-in-progress driver rtw(4) for 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_rxdesc *rd;
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++) {
rd = &rdb->rdb_desc[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
}
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->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
}
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_io_enable(struct rtw_regs *regs, 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
{
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 */
#if 0
/* enable PCI Read/Write Multiple */
cr |= RTW_CR_MULRW;
#endif
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);
}
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_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;
struct ieee80211_node *ni;
struct ieee80211_frame *wh;
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 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(rdb->rdb_next < rdb->rdb_ndesc);
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
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 (next = rdb->rdb_next; ; next = (next + 1) % rdb->rdb_ndesc) {
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) {
if (nproc > 1)
break;
/* sometimes the NIC skips to the 0th 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
rtw_rxdescs_sync(rdb, 0, 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[0];
if ((rd->rd_stat & htole32(RTW_RXSTAT_OWN)) != 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
break;
The rx descriptor ring bug happens quite often, so printf'ing occurrences is annoying. Only printf about it when rtw_debug contains RTW_DEBUG_BUG.
2004-12-25 10:45:53 +03:00
RTW_DPRINTF(RTW_DEBUG_BUGS,
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: NIC skipped from rxdesc[%u] to rxdesc[0]\n",
sc->sc_dev.dv_xname, next));
next = 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
continue;
}
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
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
#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 = "<";
printf("%s: ", sc->sc_dev.dv_xname);
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) {
printf("%s: DMA error/FIFO overflow %08x, "
"rx descriptor %d\n", sc->sc_dev.dv_xname,
hstat & RTW_RXSTAT_IOERROR, next);
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
sc->sc_if.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;
}
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
len = MASK_AND_RSHIFT(hstat, RTW_RXSTAT_LENGTH_MASK);
if (len < IEEE80211_MIN_LEN) {
sc->sc_ic.ic_stats.is_rx_tooshort++;
goto next;
}
Eliminate use of M_HASFCS.
2005-01-31 05:51:17 +03:00
/* CRC is included with the packet; trim it off. */
len -= IEEE80211_CRC_LEN;
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
hwrate = MASK_AND_RSHIFT(hstat, RTW_RXSTAT_RATE_MASK);
if (hwrate >= sizeof(ratetbl) / sizeof(ratetbl[0])) {
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
printf("%s: unknown rate #%d\n", sc->sc_dev.dv_xname,
MASK_AND_RSHIFT(hstat, RTW_RXSTAT_RATE_MASK));
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
sc->sc_if.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 ((hstat & RTW_RXSTAT_RES) != 0 &&
sc->sc_ic.ic_opmode != IEEE80211_M_MONITOR)
goto next;
/* if bad flags, skip descriptor */
if ((hstat & RTW_RXSTAT_ONESEG) != RTW_RXSTAT_ONESEG) {
printf("%s: too many rx segments\n",
sc->sc_dev.dv_xname);
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:
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
printf("%s: rtw_rxsoft_alloc(, %d) failed, "
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
"dropping packet\n", sc->sc_dev.dv_xname, 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? */
panic("%s: could not load DMA map\n",
sc->sc_dev.dv_xname);
Add work-in-progress driver rtw(4) for 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)
rssi = MASK_AND_RSHIFT(hrssi, RTW_RXRSSI_RSSI);
else {
rssi = MASK_AND_RSHIFT(hrssi, RTW_RXRSSI_IMR_RSSI);
/* TBD find out each front-end's LNA gain in the
* front-end's units
*/
if ((hrssi & RTW_RXRSSI_IMR_LNA) == 0)
rssi |= 0x80;
}
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
sq = MASK_AND_RSHIFT(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
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.
*/
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.rcvif = &sc->sc_if;
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
wh = mtod(m, struct ieee80211_frame *);
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
/* TBD use _MAR, _BAR, _PAR flags as hints to _find_rxnode? */
ni = ieee80211_find_rxnode(&sc->sc_ic, wh);
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
if ((sc->sc_if.if_flags & (IFF_DEBUG|IFF_LINK2)) ==
(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 NBPFILTER > 0
if (sc->sc_radiobpf != NULL) {
struct ieee80211com *ic = &sc->sc_ic;
struct rtw_rx_radiotap_header *rr = &sc->sc_rxtap;
rr->rr_tsft =
htole64(((uint64_t)htsfth << 32) | htsftl);
if ((hstat & RTW_RXSTAT_SPLCP) != 0)
rr->rr_flags = IEEE80211_RADIOTAP_F_SHORTPRE;
rr->rr_flags = 0;
rr->rr_rate = rate;
rr->rr_chan_freq =
htole16(ic->ic_bss->ni_chan->ic_freq);
rr->rr_chan_flags =
htole16(ic->ic_bss->ni_chan->ic_flags);
rr->rr_antsignal = rssi;
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
rr->rr_barker_lock = htole16(sq);
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
bpf_mtap2(sc->sc_radiobpf, (caddr_t)rr,
sizeof(sc->sc_rxtapu), m);
}
#endif /* NPBFILTER > 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
ieee80211_input(&sc->sc_if, m, ni, rssi, htsftl);
ieee80211_release_node(&sc->sc_ic, ni);
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
}
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->rdb_next = next;
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
KASSERT(rdb->rdb_next < 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
Add work-in-progress driver rtw(4) for 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;
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
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(bus_dma_tag_t dmat, struct ieee80211com *ic,
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);
ieee80211_release_node(ic, ni);
}
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
}
}
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;
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);
Extract the transmission status from a tx descriptor, and convert to host byte-order, once.
2004-12-23 08:44:39 +03:00
rts_retry = MASK_AND_RSHIFT(hstat, RTW_TXSTAT_RTSRETRY_MASK);
data_retry = MASK_AND_RSHIFT(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
sc->sc_if.if_collisions += rts_retry + data_retry;
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 {
sc->sc_if.if_oerrors++;
condstring = "error";
}
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",
sc->sc_dev.dv_xname, 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));
}
/* Collect transmitted packets. */
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_txring(struct rtw_softc *sc, struct rtw_txsoft_blk *tsb,
struct rtw_txdesc_blk *tdb)
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 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) {
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);
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 ((tdb->tdb_desc[ts->ts_last].td_stat &
nuke trailing whitespace
2005-02-27 03:26:58 +03:00
htole32(RTW_TXSTAT_OWN)) != 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
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
if (&sc->sc_txdesc_blk[RTW_TXPRIBCN] == tdb) {
RTW_DPRINTF(RTW_DEBUG_BEACON,
("%s: collected beacon\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
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
sc->sc_if.if_flags &= ~IFF_OACTIVE;
}
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;
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;
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];
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_collect_txring(sc, tsb, tdb);
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
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 ((isr & RTW_INTR_TX) != 0)
rtw_start(&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
}
Add work-in-progress driver rtw(4) for 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;
}
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
{
/* TBD */
return;
}
static void
rtw_intr_atim(struct rtw_softc *sc)
{
/* 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)
{
struct rtw_regs *regs = &sc->sc_regs;
RTW_WRITE(regs, RTW_RDSAR, RTW_RING_BASE(sc, hd_rx));
RTW_WRITE(regs, RTW_TLPDA, RTW_RING_BASE(sc, hd_txlo));
RTW_WRITE(regs, RTW_TNPDA, RTW_RING_BASE(sc, hd_txmd));
RTW_WRITE(regs, RTW_THPDA, RTW_RING_BASE(sc, hd_txhi));
RTW_WRITE(regs, RTW_TBDA, RTW_RING_BASE(sc, hd_bcn));
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_TLPDA, RTW_RDSAR);
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_XMIT_DESC,
("%s: reg[TLPDA] <- %" PRIxPTR "\n", __func__,
(uintptr_t)RTW_RING_BASE(sc, hd_txlo)));
RTW_DPRINTF(RTW_DEBUG_XMIT_DESC,
("%s: reg[TNPDA] <- %" PRIxPTR "\n", __func__,
(uintptr_t)RTW_RING_BASE(sc, hd_txmd)));
RTW_DPRINTF(RTW_DEBUG_XMIT_DESC,
("%s: reg[THPDA] <- %" PRIxPTR "\n", __func__,
(uintptr_t)RTW_RING_BASE(sc, hd_txhi)));
RTW_DPRINTF(RTW_DEBUG_XMIT_DESC,
("%s: reg[TBDA] <- %" PRIxPTR "\n", __func__,
(uintptr_t)RTW_RING_BASE(sc, hd_bcn)));
RTW_DPRINTF(RTW_DEBUG_RECV_DESC,
("%s: reg[RDSAR] <- %" PRIxPTR "\n", __func__,
(uintptr_t)RTW_RING_BASE(sc, hd_rx)));
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,
sc->sc_dev.dv_xname)) != 0 && rdb->rdb_ndesc == 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
printf("%s: could not allocate rx buffers\n",
sc->sc_dev.dv_xname);
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 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
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_reset(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));
tdb->tdb_nfree = tdb->tdb_ndesc;
tdb->tdb_next = 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 void
rtw_txdescs_reset(struct rtw_softc *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
{
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_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
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];
rtw_txsofts_release(sc->sc_dmat, &sc->sc_ic,
&sc->sc_txsoft_blk[pri]);
rtw_txdesc_blk_reset(tdb);
rtw_txdescs_sync(tdb, 0, tdb->tdb_ndesc,
Simplify the reset we do on a transmit timeout. Really, this could be as simple as rtw_stop(, 0), rtw_init().
2004-12-27 22:49:16 +03:00
BUS_DMASYNC_PREWRITE|BUS_DMASYNC_PREREAD);
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_rxdescs_reset(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
rtw_rxdesc_init_all(&sc->sc_rxdesc_blk, &sc->sc_rxsoft[0], 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
}
static void
rtw_intr_ioerror(struct rtw_softc *sc, uint16_t isr)
{
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;
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
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 ((isr & RTW_INTR_TXFOVW) != 0)
printf("%s: tx fifo overflow\n", sc->sc_dev.dv_xname);
if ((isr & (RTW_INTR_RDU|RTW_INTR_RXFOVW)) == 0)
return;
Disable interrupts and stop transmit DMA before disabling the transmitter and receiver.
2004-12-23 08:54:54 +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_DPRINTF(RTW_DEBUG_BUGS, ("%s: restarting xmit/recv, isr %" PRIx16
"\n", sc->sc_dev.dv_xname, 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
rtw_io_enable(regs, RTW_CR_RE | RTW_CR_TE, 0);
nuke trailing whitespace
2005-02-27 03:26:58 +03:00
/* Collect rx'd packets. Refresh rx buffers. */
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_intr_rx(sc, 0);
nuke trailing whitespace
2005-02-27 03:26:58 +03:00
/* Collect tx'd packets. */
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_intr_tx(sc, 0);
RTW_WRITE16(regs, RTW_IMR, 0);
RTW_SYNC(regs, RTW_IMR, RTW_IMR);
rtw_chip_reset1(regs, sc->sc_dev.dv_xname);
rtw_rxdescs_reset(sc);
rtw_txdescs_reset(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);
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 */
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
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_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_SYNC(regs, RTW_IMR, 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
rtw_io_enable(regs, RTW_CR_RE | RTW_CR_TE, 1);
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
sc->sc_if.if_flags &= ~IFF_OACTIVE;
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
static __inline void
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,
("%s: suspending ticks\n", sc->sc_dev.dv_xname));
Add work-in-progress driver rtw(4) for 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;
}
static __inline void
rtw_resume_ticks(struct rtw_softc *sc)
{
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 tsftrl0, tsftrl1, next_tick;
Add work-in-progress driver rtw(4) for 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);
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
next_tick = tsftrl1 + 1000000;
Add work-in-progress driver rtw(4) for 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(&sc->sc_regs, RTW_TINT, next_tick);
sc->sc_do_tick = 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
RTW_DPRINTF(RTW_DEBUG_TIMEOUT,
("%s: resume ticks delta %#08x now %#08x next %#08x\n",
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
sc->sc_dev.dv_xname, tsftrl1 - tsftrl0, tsftrl1, next_tick));
Add work-in-progress driver rtw(4) for 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)
{
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: timeout\n", sc->sc_dev.dv_xname));
Add work-in-progress driver rtw(4) for 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;
Add work-in-progress driver rtw(4) for 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.
*/
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 ((sc->sc_flags & RTW_F_ENABLED) == 0 ||
(sc->sc_if.if_flags & IFF_RUNNING) == 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
(sc->sc_dev.dv_flags & DVF_ACTIVE) == 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_INTR, ("%s: stray interrupt\n", sc->sc_dev.dv_xname));
Add work-in-progress driver rtw(4) for 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 = "<";
printf("%s: reg[ISR] = %x", sc->sc_dev.dv_xname, isr);
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)
rtw_intr_rx(sc, isr & RTW_INTR_RX);
if ((isr & RTW_INTR_TX) != 0)
rtw_intr_tx(sc, isr & RTW_INTR_TX);
if ((isr & RTW_INTR_BEACON) != 0)
rtw_intr_beacon(sc, isr & RTW_INTR_BEACON);
if ((isr & RTW_INTR_ATIMINT) != 0)
rtw_intr_atim(sc);
if ((isr & RTW_INTR_IOERROR) != 0)
rtw_intr_ioerror(sc, isr & RTW_INTR_IOERROR);
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;
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 ((sc->sc_flags & RTW_F_ENABLED) == 0)
return;
Add work-in-progress driver rtw(4) for 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(sc);
ieee80211_new_state(ic, IEEE80211_S_INIT, -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
if ((sc->sc_flags & RTW_F_INVALID) == 0) {
/* 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);
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_io_enable(&sc->sc_regs, 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
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
if (disable)
Add work-in-progress driver rtw(4) for 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_disable(sc);
/* 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
Add work-in-progress driver rtw(4) for 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,
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
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,
int before_rf, int digphy)
{
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;
u_int chan;
int rc;
int antdiv = sc->sc_flags & RTW_F_ANTDIV,
dflantb = sc->sc_flags & RTW_F_DFLANTB;
KASSERT(ic->ic_bss->ni_chan != NULL);
chan = ieee80211_chan2ieee(ic, ic->ic_bss->ni_chan);
if (chan == IEEE80211_CHAN_ANY)
panic("%s: chan == IEEE80211_CHAN_ANY\n", __func__);
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);
rtw_io_enable(&sc->sc_regs, RTW_CR_RE | RTW_CR_TE, 0);
/* TBD wait for Tx to complete */
KASSERT((sc->sc_flags & RTW_F_ENABLED) != 0);
if ((rc = rtw_phy_init(&sc->sc_regs, sc->sc_rf,
rtw_chan2txpower(&sc->sc_srom, ic, ic->ic_bss->ni_chan),
sc->sc_csthr, ic->ic_bss->ni_chan->ic_freq, antdiv,
dflantb, RTW_ON)) != 0) {
/* XXX condition on powersaving */
printf("%s: phy init failed\n", sc->sc_dev.dv_xname);
}
sc->sc_cur_chan = chan;
rtw_io_enable(&sc->sc_regs, RTW_CR_RE | RTW_CR_TE, 1);
rtw_resume_ticks(sc);
return rc;
}
void
rtw_disable(struct rtw_softc *sc)
{
int rc;
if ((sc->sc_flags & RTW_F_ENABLED) == 0)
return;
/* turn off PHY */
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
if ((sc->sc_flags & RTW_F_INVALID) == 0 &&
(rc = rtw_pwrstate(sc, RTW_OFF)) != 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
printf("%s: failed to turn off PHY (%d)\n",
sc->sc_dev.dv_xname, rc);
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
if (sc->sc_disable != NULL)
(*sc->sc_disable)(sc);
sc->sc_flags &= ~RTW_F_ENABLED;
}
int
rtw_enable(struct rtw_softc *sc)
{
if ((sc->sc_flags & RTW_F_ENABLED) == 0) {
if (sc->sc_enable != NULL && (*sc->sc_enable)(sc) != 0) {
printf("%s: device enable failed\n",
sc->sc_dev.dv_xname);
return (EIO);
}
sc->sc_flags |= RTW_F_ENABLED;
}
return (0);
}
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);
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 |= LSHIFT(4, RTW_TCR_SRL_MASK) | LSHIFT(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
}
static __inline void
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;
struct ethercom *ec = &ic->ic_ec;
struct ifnet *ifp = &sc->sc_ic.ic_if;
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;
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
/* 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
if (ifp->if_flags & IFF_PROMISC) {
sc->sc_rcr |= RTW_RCR_AB; /* accept all broadcast */
allmulti:
ifp->if_flags |= IFF_ALLMULTI;
goto setit;
}
/*
* 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,
ETHER_ADDR_LEN) != 0)
goto allmulti;
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));
sc->sc_rcr |= RTW_RCR_AM;
Add work-in-progress driver rtw(4) for 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);
}
/* all bits set => hash is useless */
if (~(hashes[0] & hashes[1]) == 0)
goto allmulti;
setit:
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
if (ifp->if_flags & IFF_ALLMULTI) {
Add work-in-progress driver rtw(4) for 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_rcr |= RTW_RCR_AM; /* accept all multicast */
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[0] = hashes[1] = 0xffffffff;
}
Add work-in-progress driver rtw(4) for 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",
Add work-in-progress driver rtw(4) for 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_dev.dv_xname, RTW_READ(regs, RTW_MAR0),
RTW_READ(regs, RTW_MAR1), RTW_READ(regs, RTW_RCR)));
return;
}
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 IEEE80211_BEACON_TIMESTAMP_LEN 8
#define IEEE80211_BEACON_BINTVL_LEN 2
#define IEEE80211_BEACON_CAPINFO_LEN 2
#define IEEE80211_TLV_SSID_LEN(__esslen) (2 + (__esslen))
#define IEEE80211_TLV_SUPRATES_LEN(__nrates) (2 + (__nrates))
#define IEEE80211_TLV_XSUPRATES_LEN(__nrates) (2 + (__nrates))
#define IEEE80211_TLV_DSPARMS_LEN 3
#define IEEE80211_TLV_IBSSPARMS 4
#define IEEE80211_TLV_MIN_TIM 6
#define IEEE80211_TLV_ALLRATES_LEN(__nrates) \
(((__nrates) > IEEE80211_RATE_SIZE) ? 4 + (__nrates) : 2 + (__nrates))
/* TBD factor with ieee80211_getmbuf */
static struct mbuf *
rtw_getmbuf(int flags, int type, u_int pktlen)
{
struct mbuf *m;
KASSERT2(pktlen <= MCLBYTES, ("802.11 packet too large: %u", pktlen));
MGETHDR(m, flags, type);
if (m == NULL || pktlen <= MHLEN)
return m;
MCLGET(m, flags);
if ((m->m_flags & M_EXT) != 0)
return m;
m_free(m);
return NULL;
}
/* TBD factor with ath_beacon_alloc */
static struct mbuf *
rtw_beacon_alloc(struct rtw_softc *sc, struct ieee80211_node *ni)
{
struct ieee80211com *ic = &sc->sc_ic;
struct ifnet *ifp = &ic->ic_if;
struct ieee80211_frame *wh;
struct mbuf *m;
int pktlen;
uint8_t *frm;
uint16_t capinfo;
struct ieee80211_rateset *rs;
/*
* NB: the beacon data buffer must be 32-bit aligned;
* we assume the mbuf routines will return us something
* with this alignment (perhaps should assert).
*/
rs = &ni->ni_rates;
pktlen = sizeof(struct ieee80211_frame)
+ IEEE80211_BEACON_TIMESTAMP_LEN
+ IEEE80211_BEACON_BINTVL_LEN
+ IEEE80211_BEACON_CAPINFO_LEN
+ IEEE80211_TLV_SSID_LEN(ni->ni_esslen)
+ IEEE80211_TLV_ALLRATES_LEN(rs->rs_nrates)
+ IEEE80211_TLV_DSPARMS_LEN
+ MAX(IEEE80211_TLV_IBSSPARMS, IEEE80211_TLV_MIN_TIM);
m = rtw_getmbuf(M_DONTWAIT, MT_DATA, pktlen);
if (m == NULL) {
RTW_DPRINTF(RTW_DEBUG_BEACON,
("%s: cannot get mbuf/cluster; size %u\n",
__func__, pktlen));
#if 0
sc->sc_stats.ast_be_nombuf++;
#endif
return NULL;
}
wh = mtod(m, struct ieee80211_frame *);
wh->i_fc[0] = IEEE80211_FC0_VERSION_0 | IEEE80211_FC0_TYPE_MGT |
IEEE80211_FC0_SUBTYPE_BEACON;
wh->i_fc[1] = IEEE80211_FC1_DIR_NODS;
*(u_int16_t *)wh->i_dur = 0;
memcpy(wh->i_addr1, ifp->if_broadcastaddr, IEEE80211_ADDR_LEN);
memcpy(wh->i_addr2, ic->ic_myaddr, IEEE80211_ADDR_LEN);
memcpy(wh->i_addr3, ni->ni_bssid, IEEE80211_ADDR_LEN);
*(u_int16_t *)wh->i_seq = 0;
/*
* beacon frame format
* [8] time stamp
* [2] beacon interval
* [2] cabability information
* [tlv] ssid
* [tlv] supported rates
* [tlv] parameter set (IBSS)
* [tlv] extended supported rates
*/
frm = (u_int8_t *)&wh[1];
/* timestamp is set by hardware */
memset(frm, 0, IEEE80211_BEACON_TIMESTAMP_LEN);
frm += IEEE80211_BEACON_TIMESTAMP_LEN;
*(u_int16_t *)frm = htole16(ni->ni_intval);
frm += IEEE80211_BEACON_BINTVL_LEN;
if (ic->ic_opmode == IEEE80211_M_IBSS)
capinfo = IEEE80211_CAPINFO_IBSS;
else
capinfo = IEEE80211_CAPINFO_ESS;
if (ic->ic_flags & IEEE80211_F_PRIVACY)
capinfo |= IEEE80211_CAPINFO_PRIVACY;
if ((ic->ic_flags & IEEE80211_F_SHPREAMBLE) &&
IEEE80211_IS_CHAN_2GHZ(ni->ni_chan))
capinfo |= IEEE80211_CAPINFO_SHORT_PREAMBLE;
if (ic->ic_flags & IEEE80211_F_SHSLOT)
capinfo |= IEEE80211_CAPINFO_SHORT_SLOTTIME;
*(u_int16_t *)frm = htole16(capinfo);
frm += IEEE80211_BEACON_CAPINFO_LEN;
*frm++ = IEEE80211_ELEMID_SSID;
*frm++ = ni->ni_esslen;
memcpy(frm, ni->ni_essid, ni->ni_esslen);
frm += ni->ni_esslen;
frm = ieee80211_add_rates(frm, rs);
*frm++ = IEEE80211_ELEMID_DSPARMS;
*frm++ = 1;
*frm++ = ieee80211_chan2ieee(ic, ni->ni_chan);
if (ic->ic_opmode == IEEE80211_M_IBSS) {
*frm++ = IEEE80211_ELEMID_IBSSPARMS;
*frm++ = 2;
*frm++ = 0; *frm++ = 0; /* TODO: ATIM window */
} else {
/* TODO: TIM */
*frm++ = IEEE80211_ELEMID_TIM;
*frm++ = 4; /* length */
nuke trailing whitespace
2005-02-27 03:26:58 +03:00
*frm++ = 0; /* DTIM count */
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
*frm++ = 1; /* DTIM period */
*frm++ = 0; /* bitmap control */
*frm++ = 0; /* Partial Virtual Bitmap (variable length) */
}
frm = ieee80211_add_xrates(frm, rs);
m->m_pkthdr.len = m->m_len = frm - mtod(m, u_int8_t *);
m->m_pkthdr.rcvif = (void *)ni;
KASSERT2(m->m_pkthdr.len <= pktlen,
("beacon bigger than expected, len %u calculated %u",
m->m_pkthdr.len, pktlen));
RTW_DPRINTF(RTW_DEBUG_BEACON,
("%s: m %p len %u\n", __func__, m, m->m_len));
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;
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
int rc = 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
if ((rc = rtw_enable(sc)) != 0)
goto out;
/* Cancel pending I/O and reset. */
rtw_stop(ifp, 0);
ic->ic_bss->ni_chan = ic->ic_ibss_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
DPRINTF(sc, RTW_DEBUG_TUNE, ("%s: channel %d freq %d flags 0x%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
__func__, ieee80211_chan2ieee(ic, ic->ic_bss->ni_chan),
ic->ic_bss->ni_chan->ic_freq, ic->ic_bss->ni_chan->ic_flags));
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
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_set_rfprog(regs, sc->sc_rfchipid, sc->sc_dev.dv_xname);
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
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_wep_setkeys(sc, ic->ic_nw_keys, ic->ic_wep_txkey);
Add work-in-progress driver rtw(4) for 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_io_enable(regs, RTW_CR_RE|RTW_CR_TE, 1);
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:
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: interface not running\n", sc->sc_dev.dv_xname);
Add work-in-progress driver rtw(4) for 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
static __inline void
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);
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:
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);
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);
callout_schedule(&ls->ls_slow_ch, RTW_LED_SLOW_TICKS);
}
static __inline 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
{
callout_init(&ls->ls_fast_ch);
callout_init(&ls->ls_slow_ch);
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
rtw_ioctl(struct ifnet *ifp, u_long cmd, caddr_t data)
{
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;
struct ifreq *ifr = (struct ifreq *)data;
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();
Add work-in-progress driver rtw(4) for 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 (cmd) {
case SIOCSIFFLAGS:
if ((ifp->if_flags & IFF_UP) != 0) {
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
if ((sc->sc_flags & RTW_F_ENABLED) != 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_pktfilt_load(sc);
} else
rc = rtw_init(ifp);
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
} else if ((sc->sc_flags & RTW_F_ENABLED) != 0) {
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);
}
break;
case SIOCADDMULTI:
case SIOCDELMULTI:
if (cmd == SIOCADDMULTI)
rc = ether_addmulti(ifr, &sc->sc_ic.ic_ec);
else
rc = ether_delmulti(ifr, &sc->sc_ic.ic_ec);
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 (rc != ENETRESET)
break;
if (ifp->if_flags & IFF_RUNNING)
rtw_pktfilt_load(sc);
rc = 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
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 SIOCS80211NWKEY:
if ((rc = ieee80211_ioctl(ifp, cmd, data)) != ENETRESET)
break;
rc = 0;
if ((ifp->if_flags & IFF_RUNNING) == 0)
break;
rtw_wep_setkeys(sc, sc->sc_ic.ic_nw_keys,
sc->sc_ic.ic_wep_txkey);
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
default:
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 ((rc = ieee80211_ioctl(ifp, cmd, data)) != ENETRESET)
break;
if ((sc->sc_flags & RTW_F_ENABLED) != 0)
rc = rtw_init(ifp);
else
rc = 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
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
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.
*/
static __inline int
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) {
*tsbp = NULL;
*tdbp = NULL;
return -1;
}
*tsbp = tsb;
*tdbp = tdb;
return 0;
}
static __inline struct mbuf *
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) {
*if_flagsp |= IFF_OACTIVE;
return NULL;
}
IF_DEQUEUE(ifq, m);
*nip = (struct ieee80211_node *)m->m_pkthdr.rcvif;
m->m_pkthdr.rcvif = NULL;
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.
*/
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)
{
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
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,
("%s: enter %s\n", sc->sc_dev.dv_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
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
}
if ((*mp = rtw_80211_dequeue(sc, &sc->sc_ic.ic_pwrsaveq, RTW_TXPRIHI,
tsbp, tdbp, nip, if_flagsp)) != NULL) {
DPRINTF(sc, RTW_DEBUG_XMIT, ("%s: dequeue pwrsave frame\n",
__func__));
return 0;
}
if (rtw_txring_choose(sc, tsbp, tdbp, RTW_TXPRIMD) == -1) {
DPRINTF(sc, RTW_DEBUG_XMIT, ("%s: no descriptor\n", __func__));
*if_flagsp |= IFF_OACTIVE;
return 0;
}
*mp = NULL;
IFQ_DEQUEUE(&ifp->if_snd, m0);
if (m0 == NULL) {
DPRINTF(sc, RTW_DEBUG_XMIT, ("%s: no frame/ring ready\n",
__func__));
return 0;
}
DPRINTF(sc, RTW_DEBUG_XMIT, ("%s: dequeue data frame\n", __func__));
ifp->if_opackets++;
Add work-in-progress driver rtw(4) for 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 NBPFILTER > 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 (ifp->if_bpf)
bpf_mtap(ifp->if_bpf, m0);
Add work-in-progress driver rtw(4) for 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
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 ((m0 = ieee80211_encap(ifp, m0, nip)) == NULL) {
DPRINTF(sc, RTW_DEBUG_XMIT,
("%s: encap error\n", __func__));
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++) {
if (dmamap->dm_segs[i].ds_len < 4) {
printf("%s: segment too short\n", __func__);
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,
u_int ndescfree, short *ifflagsp, const char *dvname)
{
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) {
if (rc == 0)
bus_dmamap_unload(dmat, dmam);
MGETHDR(m, M_DONTWAIT, MT_DATA);
if (m == NULL) {
printf("%s: unable to allocate Tx mbuf\n",
dvname);
break;
}
if (m0->m_pkthdr.len > MHLEN) {
MCLGET(m, M_DONTWAIT);
if ((m->m_flags & M_EXT) == 0) {
printf("%s: cannot allocate Tx cluster\n",
dvname);
m_freem(m);
break;
}
}
m_copydata(m0, 0, m0->m_pkthdr.len, mtod(m, caddr_t));
m->m_pkthdr.len = m->m_len = m0->m_pkthdr.len;
m_freem(m0);
m0 = m;
m = NULL;
}
if (rc != 0) {
printf("%s: cannot load Tx buffer, rc = %d\n", dvname, rc);
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)) {
printf("%s: cannot load Tx buffer, segment too short\n",
dvname);
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) {
*ifflagsp |= IFF_OACTIVE;
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];
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: %p %s txdesc[%d] ctl0 %#08x "
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
"ctl1 %#08x buf %#08x len %#08x\n",
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_dev.dv_xname, ts, action, desc,
le32toh(td->td_ctl0),
le32toh(td->td_ctl1), le32toh(td->td_buf),
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)
{
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
uint8_t tppoll;
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;
struct ieee80211_frame *wh;
struct ieee80211_node *ni;
struct mbuf *m0;
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;
struct rtw_txdesc_blk *tdb;
struct rtw_txsoft *ts;
struct rtw_txdesc *td;
Add work-in-progress driver rtw(4) for 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,
("%s: enter %s\n", sc->sc_dev.dv_xname, __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;
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,
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, &ifp->if_flags, sc->sc_dev.dv_xname);
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;
}
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
wh = mtod(m0, struct ieee80211_frame *);
/* XXX do real rate control */
if ((wh->i_fc[0] & IEEE80211_FC0_TYPE_MASK) ==
IEEE80211_FC0_TYPE_MGT)
rate = 2;
else
rate = MAX(2, ieee80211_get_rate(ic));
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
if ((sc->sc_if.if_flags & (IFF_DEBUG|IFF_LINK2)) ==
(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 |
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
LSHIFT(m0->m_pkthdr.len, RTW_TXCTL0_TPKTSIZE_MASK);
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;
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 ((wh->i_fc[1] & IEEE80211_FC1_WEP) != 0)
ctl0 |= LSHIFT(sc->sc_txkey, RTW_TXCTL0_KEYID_MASK);
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;
}
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
if (ieee80211_compute_duration(wh, 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,
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
(sc->sc_if.if_flags & (IFF_DEBUG|IFF_LINK2)) ==
(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);
Shorten variables, s/txctl/ctl/. This is rtw_start, after all: transmission is implied.
2004-12-23 08:52:27 +03:00
ctl1 = LSHIFT(d0->d_plcp_len, RTW_TXCTL1_LENGTH_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
LSHIFT(d0->d_rts_dur, RTW_TXCTL1_RTSDUR_MASK);
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
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 NBPFILTER > 0
if (ic->ic_rawbpf != NULL)
bpf_mtap((caddr_t)ic->ic_rawbpf, m0);
if (sc->sc_radiobpf != NULL) {
struct rtw_tx_radiotap_header *rt = &sc->sc_txtap;
rt->rt_flags = 0;
rt->rt_rate = rate;
rt->rt_chan_freq =
htole16(ic->ic_bss->ni_chan->ic_freq);
rt->rt_chan_flags =
htole16(ic->ic_bss->ni_chan->ic_flags);
bpf_mtap2(sc->sc_radiobpf, (caddr_t)rt,
sizeof(sc->sc_txtapu), m0);
}
#endif /* NPBFILTER > 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
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);
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 (i != 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
td->td_ctl0 |= htole32(RTW_TXCTL0_OWN);
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);
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;
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
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 (tsb != &sc->sc_txsoft_blk[RTW_TXPRIBCN]) {
sc->sc_led_state.ls_event |= RTW_LED_S_TX;
tsb->tsb_tx_timer = 5;
ifp->if_timer = 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
tppoll = RTW_READ8(&sc->sc_regs, RTW_TPPOLL);
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
tppoll &= ~RTW_TPPOLL_SALL;
tppoll |= tsb->tsb_poll & RTW_TPPOLL_ALL;
RTW_WRITE8(&sc->sc_regs, RTW_TPPOLL, tppoll);
Fix bug in last commit: use &sc->sc_regs, regs is not defined.
2004-12-20 04:28:24 +03:00
RTW_SYNC(&sc->sc_regs, RTW_TPPOLL, RTW_TPPOLL);
Add work-in-progress driver rtw(4) for 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:
ieee80211_release_node(&sc->sc_ic, 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;
}
static void
rtw_watchdog(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 pri;
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;
if ((sc->sc_flags & RTW_F_ENABLED) == 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
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;
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) {
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;
printf("%s: transmit timeout, priority %d\n",
ifp->if_xname, pri);
ifp->if_oerrors++;
Simplify the reset we do on a transmit timeout. Really, this could be as simple as rtw_stop(, 0), rtw_init().
2004-12-27 22:49:16 +03:00
/* Stop Tx DMA, disable transmitter, clear
* Tx rings, and restart.
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
*
* TBD Stop/restart just the broken ring?
Simplify the reset we do on a transmit timeout. Really, this could be as simple as rtw_stop(, 0), rtw_init().
2004-12-27 22:49:16 +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_WRITE8(&sc->sc_regs, RTW_TPPOLL, RTW_TPPOLL_SALL);
Simplify the reset we do on a transmit timeout. Really, this could be as simple as rtw_stop(, 0), rtw_init().
2004-12-27 22:49:16 +03:00
RTW_SYNC(&sc->sc_regs, RTW_TPPOLL, RTW_TPPOLL);
rtw_io_enable(&sc->sc_regs, RTW_CR_TE, 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_txdescs_reset(sc);
Simplify the reset we do on a transmit timeout. Really, this could be as simple as rtw_stop(, 0), rtw_init().
2004-12-27 22:49:16 +03:00
rtw_io_enable(&sc->sc_regs, RTW_CR_TE, 1);
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
sc->sc_if.if_flags &= ~IFF_OACTIVE;
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
rtw_start(ifp);
} else
ifp->if_timer = 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
ieee80211_watchdog(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
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
{
uint16_t bcnitv, intval;
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
intval = MIN(intval0, PRESHIFT(RTW_BCNITV_BCNITV_MASK));
bcnitv = RTW_READ16(regs, RTW_BCNITV) & ~RTW_BCNITV_BCNITV_MASK;
bcnitv |= LSHIFT(intval, RTW_BCNITV_BCNITV_MASK);
RTW_WRITE16(regs, RTW_BCNITV, bcnitv);
/* magic from Linux */
RTW_WRITE16(regs, RTW_ATIMWND, LSHIFT(1, RTW_ATIMWND_ATIMWND));
RTW_WRITE16(regs, RTW_ATIMTRITV, LSHIFT(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
/* TBD WEP */
RTW_WRITE8(regs, RTW_SCR, 0);
rtw_io_enable(regs, 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
/* Synchronize the hardware state with the software state. */
static int
rtw_newstate(struct ieee80211com *ic, enum ieee80211_state nstate, int arg)
{
struct ifnet *ifp = &ic->ic_if;
struct rtw_softc *sc = ifp->if_softc;
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 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
enum ieee80211_state ostate;
int error;
ostate = ic->ic_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
rtw_led_newstate(sc, nstate);
Add work-in-progress driver rtw(4) for 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
m = rtw_beacon_alloc(sc, ic->ic_bss);
if (m == NULL) {
printf("%s: could not allocate beacon\n",
sc->sc_dev.dv_xname);
} else
IF_ENQUEUE(&sc->sc_beaconq, m);
/*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);
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
/* Start beacon transmission. */
Add work-in-progress driver rtw(4) for 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_RUN &&
(ic->ic_opmode == IEEE80211_M_HOSTAP ||
ic->ic_opmode == 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_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
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
Add the logic for IBSS merges.
2005-01-04 04:00:30 +03:00
rtw_ibss_merge(struct rtw_softc *sc, struct ieee80211_node *ni, 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
{
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
uint8_t tppoll;
Add the logic for IBSS merges.
2005-01-04 04:00:30 +03:00
struct ieee80211com *ic = &sc->sc_ic;
if (le64toh(ni->ni_tsf) >= rtw_tsf_extend(&sc->sc_regs, rstamp) &&
ieee80211_ibss_merge(ic, ni) == ENETRESET) {
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
/* Stop beacon queue. Kick state machine to synchronize
* with the new IBSS.
*/
tppoll = RTW_READ8(&sc->sc_regs, RTW_TPPOLL);
tppoll |= RTW_TPPOLL_SBQ;
RTW_WRITE8(&sc->sc_regs, RTW_TPPOLL, tppoll);
(void)ieee80211_new_state(&sc->sc_ic, IEEE80211_S_RUN, -1);
Add the logic for IBSS merges.
2005-01-04 04:00:30 +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;
}
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
{
struct rtw_softc *sc = (struct rtw_softc*)ic->ic_softc;
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:
Add the logic for IBSS merges.
2005-01-04 04:00:30 +03:00
if (ic->ic_opmode != IEEE80211_M_IBSS ||
ic->ic_state != IEEE80211_S_RUN)
return;
rtw_ibss_merge(sc, ni, 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
break;
default:
break;
}
return;
}
static struct ieee80211_node *
rtw_node_alloc(struct ieee80211com *ic)
{
struct rtw_softc *sc = (struct rtw_softc *)ic->ic_if.if_softc;
struct ieee80211_node *ni = (*sc->sc_mtbl.mt_node_alloc)(ic);
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,
("%s: alloc node %p\n", sc->sc_dev.dv_xname, 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
rtw_node_free(struct ieee80211com *ic, struct ieee80211_node *ni)
{
struct rtw_softc *sc = (struct rtw_softc *)ic->ic_if.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_NODE,
("%s: freeing node %p %s\n", sc->sc_dev.dv_xname, 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)));
(*sc->sc_mtbl.mt_node_free)(ic, ni);
}
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;
if ((sc->sc_flags & RTW_F_ENABLED) == 0) {
imr->ifm_active = IFM_IEEE80211 | IFM_NONE;
imr->ifm_status = 0;
return;
}
ieee80211_media_status(ifp, imr);
}
void
rtw_power(int why, void *arg)
{
struct rtw_softc *sc = arg;
struct ifnet *ifp = &sc->sc_ic.ic_if;
int s;
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_PWR,
("%s: rtw_power(%d,)\n", sc->sc_dev.dv_xname, why));
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
s = splnet();
switch (why) {
case PWR_STANDBY:
/* XXX do nothing. */
break;
case PWR_SUSPEND:
rtw_stop(ifp, 0);
if (sc->sc_power != NULL)
(*sc->sc_power)(sc, why);
break;
case PWR_RESUME:
if (ifp->if_flags & IFF_UP) {
if (sc->sc_power != NULL)
(*sc->sc_power)(sc, why);
rtw_init(ifp);
}
break;
case PWR_SOFTSUSPEND:
case PWR_SOFTSTANDBY:
case PWR_SOFTRESUME:
break;
}
splx(s);
}
/* rtw_shutdown: make sure the interface is stopped at reboot time. */
void
rtw_shutdown(void *arg)
{
struct rtw_softc *sc = arg;
rtw_stop(&sc->sc_ic.ic_if, 1);
}
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
{
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
(void)memcpy(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;
}
static __inline void
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 |
IEEE80211_C_HOSTAP | IEEE80211_C_MONITOR | IEEE80211_C_WEP;
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;
}
static __inline void
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;
}
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_establish_hooks(struct rtw_hooks *hooks, const char *dvname,
Add work-in-progress driver rtw(4) for 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 *arg)
{
/*
* Make sure the interface is shutdown during reboot.
*/
hooks->rh_shutdown = shutdownhook_establish(rtw_shutdown, arg);
if (hooks->rh_shutdown == NULL)
printf("%s: WARNING: unable to establish shutdown hook\n",
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
dvname);
Add work-in-progress driver rtw(4) for 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 a suspend hook to make sure we come back up after a
* resume.
*/
hooks->rh_power = powerhook_establish(rtw_power, arg);
if (hooks->rh_power == NULL)
printf("%s: WARNING: unable to establish power hook\n",
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
dvname);
Add work-in-progress driver rtw(4) for 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 __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_disestablish_hooks(struct rtw_hooks *hooks, const char *dvname,
Add work-in-progress driver rtw(4) for 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 *arg)
{
if (hooks->rh_shutdown != NULL)
shutdownhook_disestablish(hooks->rh_shutdown);
if (hooks->rh_power != NULL)
powerhook_disestablish(hooks->rh_power);
}
static __inline void
rtw_init_radiotap(struct rtw_softc *sc)
{
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));
sc->sc_rxtap.rr_ihdr.it_present = htole32(RTW_RX_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
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
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_reset(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
}
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) {
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)
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 phydelay = LSHIFT(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)
{
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
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;
}
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
printf("%s: hardware version %c\n", sc->sc_dev.dv_xname,
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) {
printf("%s: could not allocate hw descriptors, error %d\n",
sc->sc_dev.dv_xname, rc);
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),
(caddr_t*)&sc->sc_descs, BUS_DMA_COHERENT);
if (rc != 0) {
printf("%s: could not map hw descriptors, error %d\n",
sc->sc_dev.dv_xname, rc);
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) {
printf("%s: could not create DMA map for hw descriptors, "
"error %d\n", sc->sc_dev.dv_xname, rc);
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) {
printf("%s: could not load DMA map for hw descriptors, "
"error %d\n", sc->sc_dev.dv_xname, rc);
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) {
Add work-in-progress driver rtw(4) for Realtek RTL8180 MAC/baseband. Add register definitions for Maxim MAX2820 and Philips SA2400 radio front-ends.
2004-09-26 06:29:15 +04:00
printf("%s: could not load DMA map for "
"hw tx descriptors, error %d\n",
sc->sc_dev.dv_xname, rc);
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) {
printf("%s: could not load DMA map for hw rx descriptors, "
"error %d\n", sc->sc_dev.dv_xname, rc);
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,
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
sc->sc_dev.dv_xname) != 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,
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
sc->sc_dev.dv_xname) != 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
printf("%s: attach failed, malformed serial ROM\n",
sc->sc_dev.dv_xname);
goto err;
}
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
printf("%s: %s PHY\n", sc->sc_dev.dv_xname,
((sc->sc_flags & RTW_F_DIGPHY) != 0) ? "digital" : "analog");
printf("%s: CS threshold %u\n", sc->sc_dev.dv_xname, 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) {
printf("%s: attach failed, could not attach RF\n",
sc->sc_dev.dv_xname);
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,
("%s: PHY delay %d\n", sc->sc_dev.dv_xname, 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)
rtw_identify_country(&sc->sc_regs, &sc->sc_locale,
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
sc->sc_dev.dv_xname);
Add work-in-progress driver rtw(4) for 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(sc->sc_locale, &sc->sc_ic.ic_channels,
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
sc->sc_dev.dv_xname);
Add work-in-progress driver rtw(4) for 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,
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
sc->sc_dev.dv_xname) != 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);
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(&sc->sc_if, sc->sc_dev.dv_xname, (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
IFQ_SET_READY(&sc->sc_if.if_snd);
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.
*/
if_attach(&sc->sc_if);
ieee80211_ifattach(&sc->sc_if);
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 */
ieee80211_media_init(&sc->sc_if, rtw_media_change, rtw_media_status);
callout_init(&sc->sc_scan_ch);
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);
Add work-in-progress driver rtw(4) for 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 NBPFILTER > 0
bpfattach2(&sc->sc_if, DLT_IEEE802_11_RADIO,
sizeof(struct ieee80211_frame) + 64, &sc->sc_radiobpf);
#endif
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_establish_hooks(&sc->sc_hooks, sc->sc_dev.dv_xname, (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
NEXT_ATTACH_STATE(sc, FINISHED);
return;
err:
rtw_detach(sc);
return;
}
int
rtw_detach(struct rtw_softc *sc)
{
int pri;
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
sc->sc_flags |= RTW_F_INVALID;
Add work-in-progress driver rtw(4) for 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:
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_stop(&sc->sc_if, 1);
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_disestablish_hooks(&sc->sc_hooks, sc->sc_dev.dv_xname,
Add work-in-progress driver rtw(4) for 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*)sc);
callout_stop(&sc->sc_scan_ch);
ieee80211_ifdetach(&sc->sc_if);
if_detach(&sc->sc_if);
break;
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:
bus_dmamem_unmap(sc->sc_dmat, (caddr_t)sc->sc_descs,
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;
}
return 0;
}
int
rtw_activate(struct device *self, enum devact act)
{
struct rtw_softc *sc = (struct rtw_softc *)self;
int rc = 0, s;
s = splnet();
switch (act) {
case DVACT_ACTIVATE:
rc = EOPNOTSUPP;
break;
case DVACT_DEACTIVATE:
if_deactivate(&sc->sc_ic.ic_if);
break;
}
splx(s);
return rc;
}