3com 3c905c-txm

Andrew Morton andrewm@uow.edu.au
Fri May 12 11:33:23 2000


Donald Becker wrote:
> 
> [ big snip.  I'll follow up on linux-vortex ]
>
> 4000 PCI cycles is a *very* long time.  Ages.  It should only 0 to 2 PCI
> cycles to queue a packet.

Alas I seem to have lost the ability to reproduce[*].

With four Linux boxes on a 10bT LAN the limit of 2,000 was exceeded a
few times a minute with a particular test case.  With my current setup
of three Linux boxes and one NT, the best I can get is 240 loops, in the
DownStall in boomerang_start_xmit().  3c905B.   Still much higher than
we expect.

More tomorrow....

-- 
-akpm-

[*] The timeout problem, that is.
-------------------------------------------------------------------
To unsubscribe send a message body containing "unsubscribe"
to linux-vortex-request@beowulf.org