Aren/NetBSD
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
129,162 Commits 606 Branches 0 Tags 3.1 GiB
9e8d1d4bd6
Commit Graph

74 Commits

Author SHA1 Message Date
tron
0237ca34a7 Reduce transmit queue len to 16 for the Intel i82547 to avoid FIFO 
overruns. This fixes the horrible send performance reported in
PR kern/23782. Reviewed by Frank van der Linden and Jason R. Thorpe.
 2004-07-15 15:21:57 +00:00
tron
742e1a4221 The Intel i82547EI gigabit ethernet controller isn't really a PCI 
device. So instead trying to find and print out the PCI bus properties
mark it as a CSA device.
 2004-07-13 07:29:37 +00:00
tron
7df6314408 Make this build if WM_DEBUG is defined. 2004-07-12 14:04:36 +00:00
thorpej
47e9c860d9 Add flow control support, based on work by HITOSHI Osada. 2004-05-16 02:34:47 +00:00
ragge
f62c3ab365 Remove the "panic("wm_init: MCLBYTES > 2048 not supported.");" stuff, 
the driver works very well with larger MCLBYTES.
 2004-05-06 10:06:50 +00:00
thorpej
c235249183 Add support for 64-bit PCI DMA addressing. 2004-04-09 17:51:18 +00:00
thorpej
816a869868 s/WTC_DTYP_D/WTX_DTYP_D/ 2004-02-19 05:19:52 +00:00
thorpej
881b8b9966 Fix typo in debug message. From Jeffrey Hsu. 2004-02-19 05:18:47 +00:00
thorpej
fd4acf29d4 Add missing txseg20 counter name. Pointed out by Jeffrey Hsu. 2004-02-19 02:13:23 +00:00
tsutsui
6353b12d1a Fix hardware cksum problem on big endian machines. 
Tested on macppc by HATANO Hiromichi.

Note I guess this bug could be fixed only adding one htole32(),
but I'd rather clean up endianness handling:
- Use htole32() only to accesses against DMA descriptors.
- Don't use uint32_t union member  with htole32()/le32toh()
  to access uint8_t/uint16_t descriptors.
- Add le32toh() in some DPRINTF.
  (XXX: strictly speaking, bus_dmamap_sync() is needed for these DPRINTF)
 2004-01-14 14:29:48 +00:00
thorpej
c0ddd8220b Bump the Tx DMA segment count to 40. Apparently, it's common to have 
DMA segment counts in this ballpark when using jumbo frames (so says
the "ti" driver).
 2003-12-04 06:57:37 +00:00
thorpej
f6c3c3450e Add a missing i82541 ID to the table. 2003-11-22 08:32:12 +00:00
thorpej
32ddb3f486 Fix brain'o in PCIX speed detection. kern/23375 (John R. Shannon). 2003-11-07 06:03:52 +00:00
thorpej
73fea8467b Enable i82541 and i82547 chips - it works well enough to get broader 
testing.  Update the TODO list.
 2003-11-04 19:09:39 +00:00
ichiro
8b115395c7 Fix uninitialized variable warnings 2003-11-03 03:05:25 +00:00
christos
125ccd91b1 Fix uninitialized variable warnings 2003-10-25 18:29:12 +00:00
ragge
c4690aa593 Set snd queue size to max(WM_IFQUEUELEN, IFQ_MAXLEN), instead of hard to 
WM_IFQUEUELEN.
 2003-10-23 12:53:02 +00:00
thorpej
0f781c4abb Add support for the SPI EEPROM found on the newer i82541 and i82547 
chips.  Add the i82541 and i82547 product IDs to the table, #if 0'd
out, for now (there are still more changes to come for these new chips
to work properly).
 2003-10-22 15:50:39 +00:00
thorpej
3e560e2b48 Add a TODO list item. 2003-10-21 16:52:08 +00:00
thorpej
c50f2ac573 - Move the i82540EM (LOM) along with the other i82540 product IDs. 
- Match the i82545GM (1000BASE-T and 1000BASE-X) and i82546GB
  (1000BASE-T and 1000BASE-X) product IDs.
 2003-10-21 16:51:17 +00:00
thorpej
a866babe44 If we're on PCI-X, look up the PCI-X capability. Verify that the 
max memory read byte count in the command register is <= the max designed
memory read byte count in the status register.  If not, reset it to the
status register's value.  This is necessary because some BIOSes configure
the wrong value, and the wrong value confused some versions of the PCI-X
capable chips.
 2003-10-21 16:41:51 +00:00
thorpej
abc7f80ca5 - Map I/O space on chips that support it. While it is not used for 
normal operation, it is required to implement a bug work-around on
  some chips.
- When issuing a reset on the i82544, i82540, i82545, i82546,
  and i82541, use I/O space, if available, to work around a problem
  with the memory write cycle (problem only occurs when issuing
  a reset).
 2003-10-21 05:45:11 +00:00
thorpej
2cadb9ce85 Record characteristics of the bus connection at attach time. 2003-10-21 04:35:01 +00:00
thorpej
e2d086d69d - Do better error checking on EEPROM access. 
- Let the EEPROM wire-protocol handler perform the loop.
 2003-10-20 22:52:19 +00:00
thorpej
12c5273afa Move all the code that handles the MicroWire EEPROM protocol into 
a separate function.
 2003-10-20 16:28:23 +00:00
thorpej
fedb07bc8d Fix a typo in the "failed to acquire eeprom" case. 2003-10-20 15:34:25 +00:00
thorpej
0ebc93fe43 Move the code that reads a series of bits from the EEPROM into a 
separate function.
 2003-10-20 15:33:48 +00:00
thorpej
17aaa92402 Make all of the functions in this file static. 2003-10-20 06:00:26 +00:00
thorpej
10926a6056 Factor out the code that sends bits to the EEPROM into a separate function. 2003-10-20 05:56:17 +00:00
thorpej
f5fc41219d Move the code that handle the EEPROM handshake into separate functions. 2003-10-20 05:40:03 +00:00
thorpej
b0bced8029 Store the number of EEPROM address bits in the softc. 2003-10-17 21:12:48 +00:00
thorpej
2d858d3912 Add internal representation for i82545 rev 3, i82546 rev3, i82541, 
i82541 rev 2, i82547, and i82547 rev 2.
 2003-10-17 20:57:32 +00:00
thorpej
e3ca7c2c73 Make the "align tweak" factor per-interface (since the MTU is per-interface). 2003-10-17 20:41:21 +00:00
tls
ab54d7e94f Add support for jumbo (9K) frames. Add support for larger receive 
descriptor sizes if larger mbuf clusters are in use -- currently
commented out because, for some reason, transmit doesn't work at
all with large mbuf cluster sizes.

Note that using actual 9K frames is a _lot_ less efficient than
using frames that are a little smaller than 8K so that they fit
in one or two pages (instead of two or three) and avoid the extra
chained descriptor with only 1K of data in it, too.  TCP will
segment things just fine for an 8K total packet size, just as it
does for 1500-byte packets.

On my system, using "mtu 8000" with ifconfig, I get an almost 50%
performance boost for TCP (930Mbit/sec instead of 660Mbit/sec) over
1500-byte frames.  "mtu 9000" yields only 770Mbit/sec.
 2003-09-10 04:02:17 +00:00
thorpej
2f2b9c7cb0 GMII works on the i82543 now. 2003-09-04 19:32:19 +00:00
thorpej
058f44ae26 Add support for the Intel PRO/1000 MT Quad Port server adapter; this 
is just 2 i82546EB chips behind an IBM 133 PCI-X bridge.
 2003-07-29 19:49:50 +00:00
lukem
365cbd9428 add missing __KERNEL_RCSID() 2003-07-14 15:47:00 +00:00
thorpej
b43b1645a2 Use aprint*(). 2003-04-29 01:07:30 +00:00
tron
78a6a01cac Pullup short mbuf in wm_tx_cksum() instead of dropping it which fixes 
PR kern/21190 by myself. Jason R. Thorpe code reviewed the changes.
 2003-04-15 22:52:40 +00:00
thorpej
b2beaa1fbd wm_tx_cksum(): Set the initial offset correctly for packets using 
802.1Q encap.
 2003-04-15 21:12:24 +00:00
kent
d49a3d652c Support for 82540EP and 82540EP (LOM). 
Not tested.
 2003-04-05 13:23:17 +00:00
kent
e7a590c1f1 Support for i82450EP. 2003-04-04 06:45:12 +00:00
thorpej
b6fa4ee6ba Use bus_dmamap_load_mbuf() in wm_add_rxbuf(). 2003-02-04 17:40:31 +00:00
itojun
aa8971e35f m_tag transition mistake (within #if 0) 2003-01-21 05:43:26 +00:00
itojun
40606ab8f2 switch from kame-based m_aux mbuf auxiliary data, to openbsd m_tag 
implementation.  it will simplify porting across *bsd (such as kame/altq),
and make us more synchronized.  from Joel Wilsson
 2003-01-17 08:11:49 +00:00
tsutsui
238efe4697 Replace magic numbers for power management control with PCI_PMCSR* macros. 
XXX Should we use pci_get_powerstate() and pci_set_powerstate() in pci.c?
 2002-12-23 02:58:36 +00:00
kristerw
2ce6431736 wm_copy_small in not used anywhere. Remove. 2002-12-21 16:12:15 +00:00
christos
ab944659c0 change the macro params to longer names 2002-10-23 01:34:58 +00:00
fair
2104199dc3 Change the "dontcare bits" argument to ifmedia_init() to IFM_IMASK 
so that it is possible select PHY instances other than the first
one (instance zero), if there is more than one PHY attached.
 2002-10-21 23:38:10 +00:00
thorpej
b75a007d9f Add trailing ; to CFATTACH_DECL. 2002-10-02 16:51:16 +00:00