Boomerang driver: can't recover from cabling error

Boris Tobotras (tobotras@jet.msk.su)
Thu, 07 Aug 1997 16:12:23 +0400


Hi,

I have this repeatable behavior of Ethernet driver when it has an
error:

Aug 7 15:41:51 goliath kernel: eth0: transmit timed out, tx_status 88 status e000.
Aug 7 15:41:51 goliath kernel: eth0: Transmitter encountered 16 collisions -- network network cable problem?
Aug 7 15:41:51 goliath kernel: Flags; bus-master 0, full 0; dirty 0 current 0.
Aug 7 15:41:51 goliath kernel: Down list 00000000 vs. 07011a10.
Aug 7 15:41:51 goliath kernel: 0: 07011a10 length 00000000 status 00000000
Aug 7 15:41:51 goliath kernel: 1: 07011a20 length 00000000 status 00000000
Aug 7 15:41:51 goliath kernel: 2: 07011a30 length 00000000 status 00000000
Aug 7 15:41:51 goliath kernel: 3: 07011a40 length 00000000 status 00000000
Aug 7 15:41:51 goliath kernel: 4: 07011a50 length 00000000 status 00000000
Aug 7 15:41:51 goliath kernel: 5: 07011a60 length 00000000 status 00000000
Aug 7 15:41:51 goliath kernel: 6: 07011a70 length 00000000 status 00000000
Aug 7 15:41:51 goliath kernel: 7: 07011a80 length 00000000 status 00000000
Aug 7 15:41:51 goliath kernel: 8: 07011a90 length 00000000 status 00000000
Aug 7 15:41:51 goliath kernel: 9: 07011aa0 length 00000000 status 00000000
Aug 7 15:41:51 goliath kernel: 10: 07011ab0 length 00000000 status 00000000
Aug 7 15:41:51 goliath kernel: 11: 07011ac0 length 00000000 status 00000000
Aug 7 15:41:51 goliath kernel: 12: 07011ad0 length 00000000 status 00000000
Aug 7 15:41:51 goliath kernel: 13: 07011ae0 length 00000000 status 00000000
Aug 7 15:41:51 goliath kernel: 14: 07011af0 length 00000000 status 00000000
Aug 7 15:41:51 goliath kernel: 15: 07011b00 length 00000000 status 00000000

This repeats on each attempt to send packets. That is OK, but bad
thing is it does not recover when cabling gets fixed. Removing/loading
3c59x module fixes that, but this is very sad solution, since it breaks
existing users' sessions. Can this be fixed somehow?

My setup: 2.0.30 kernel, this is what module prints on load:

3c900.c:v0.40 4/16/97 becker@cesdis.gsfc.nasa.gov
loading device 'eth0'...
eth0: 3Com 3c905 Boomerang 100baseTx at 0xef00, 00:60:97:5b:3f:ac, IRQ 9
8K word-wide RAM 3:5 Rx:Tx split, autoselect/MII interface.
Rx Pacing bug exists, disabling bus-master receives.

Card runs on 10Mbit, if this makes any difference.

-- 
	Best regards, -- Boris.