v0.91 driver released for testing
Donald Becker
becker@cesdis1.gsfc.nasa.gov
Fri Apr 16 23:51:04 1999
OK, after working on the Tulip driver, to the exclusion of most other things
I should have been doing, I'm finally ready to do another release.
ftp://cesdis.gsfc.nasa.gov/pub/linux/drivers/test/tulip.c
I'm assuming that this driver fixes all reported problems, so please
re-report any problems that remain.
v0.91 of 4/14/99
Rewrote the 21143 NWay setup.
This involved extensive changes since it uses the media table in a way
that's incompatible with 21140 .
Fixed some PowerPC issues.
Thanks go to Kazumitsu Hanai <hanai@koto5.kpu-m.ac.jp> for extensive testing.
The driver should now work unmodified with all but the 21040, which
doesn't have hardware mixed-endian support.
Changed the 21040 media selection.
Corrected the media locking, which was ignored during transmit timeouts.
Worked around the ASIX 88140 PCI no-burst-limit bug by modifying CSR0.
Changed tulip_suspend() to not call tulip_close().
This would screw up the module use count if the card was ejected while
suspended.
Restructured the receive mode setup.
Reordered the perfect filter list for better code generation.
Added a multicast hash filter entry for the broadcast address (21140 bug).
Things that I would like a report on:
Large multicast lists with the 21140 or 21140A chips.
Multicast with the ASIX and Comet chips.
Media selection with 21143 cards using symbol mode transceivers.
Things known to be broken:
I've rewritten PNIC symbol mode media selection code, but it breaks with
some cable disconnect-reconnect timing. This will have to go in v0.92.
PNIC2 media selection could use some more work.
Untrackable hangs when resuming CardBus cards.
Donald Becker becker@cesdis.gsfc.nasa.gov
USRA-CESDIS, Center of Excellence in Space Data and Information Sciences.
Code 930.5, Goddard Space Flight Center, Greenbelt, MD. 20771
301-286-0882 http://cesdis.gsfc.nasa.gov/people/becker/whoiam.html