e100 breakage located

From: Amin Azez
Date: Wed Nov 22 2006 - 04:47:33 EST


I'm not subscribed to lkm so please Cc me in any responses.

This may be relevant to a post by jesse.brandeburg@xxxxxxxxx who
reported e100 as not working; perhaps if he uses a cross-over cable it
will "work" again.

I notice a patch in 2005 from Micahel O'Donnel to the e100.c driver has
stopped auto-crossover working on some e100 devices we use.

On one system the auto-negotiation was restored by commenting out:
(nic->mac == mac_82551_10) in function e100_phy_init where the MDI/MDI-X
is disabled.

lspci reports:
01:04.0 Ethernet controller: Intel Corp. 82557/8/9 [Ethernet Pro 100]
(rev 10)
01:04.0 Class 0200: 8086:1229 (rev 10)

and on another device
01:05.0 Ethernet controller: Intel Corp. 82559ER (rev 10)
01:01.0 Class 0200: 8086:1209 (rev 10)

So it is true that we are revision 10, but 82557/9 not 82551.

I must confess that having gotten this far, I am lost. Of course I can
"fix" the driver for our hardware but I am not sure how to contrive a
general fix.

Maybe the actual damage is done in
e100_get_defaults(struct nic *nic)
where nic->mac is set to nic->rev_id ?

But it generally seems to be a failure to take into account the actual
hardware type, and only consider the revision.

Any ideas?

Sam
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/