On Mon, 16 Oct 2000 kernel@starlight.noc wrote:
>
> Scenario:
>
> o Linux 2.2.16 (older driver) & Linux 2.4.0-test9 (8139too-fast)
> o Two machines
> o Netcards with RTL8139B chipsets
> o Both hang after seemingly random delays / random amounts of net activity
> o One seems to get a shitload of packet errors if left for a few days and
> used again.
> o Machines have ATI Rage 128 Chipsets for Graphics cards.
>
> Misc:
>
> o If I use w3m all day and don't overdo net activity too much, machine
> will stay stable, if I use netscape and look at high graphic (say high
> bandwidth) sites, *bang* it reboots.
>
> o If I transfer stuff directly across the network (at 100meg, HD) it
> will crash almost instantaneously.
>
> o I have tried all manner of tests to monitor and log net transfer to see
> when the actual hang takes place, but haven't been able to "benchmark"
> it.
>
> o Seems to be rougly after 6.6megs.. found this out by doing,
>
> { while true; sync; done } &
> nc otherbox chargen > bigfile
>
> After reboot, du -h bigfile,
>
> 6.6M bigfile (or something similar)
>
> Heeelp :P
You could try the patch I sent to this list a couple of days ago, and see
if it helps. Let me know if you cannot find it. Search for 8139 in the
subject.
[Btw, you should know that the RTL8139 is a really bad piece of hardware.
I recommend everyone to buy some other network card. (And yes, I'm
upset!).]
/Tobias
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Mon Oct 23 2000 - 21:00:09 EST