[373] in linux-net channel archive
Linux 1.2.8 sudden death - SMC the cause?
daemon@ATHENA.MIT.EDU (Daniel Hollis)
Thu May 25 13:37:34 1995
From: dhollis@hq.jcic.org (Daniel Hollis)
To: linux-net@vger.rutgers.edu
Date: Thu, 25 May 1995 09:18:44 -0700 (PDT)
To: gpg109@rsphy1.anu.edu.au (Paul Gortmaker)
Date: Thu, 25 May 1995 09:16:11 -0700 (PDT)
Cc: linux-network@vger.rutgers.edu
> > Been having problems with 1.2.8 hanging, it locks up hard. No response to
> > pings, no disk activity, and no keyboard response.
> > eth0 is WD80x3 at 0x280 IRQ 15, shared mem at 0xd0000-0xd1fff
> Something is seriously wrong here, as you are using a "16 bit"
> interrupt, buy you only report 8k of packet buffer space, which
> would indicate you are using an 8 bit card (i.e. a wd8003).
> Unless this is an el-cheapo EEPROM-less clone...
This is an SMC Ultra. It's being _auto detected_ by the WD code. The SMC
is a 16 bit card, in a 16 bit slot.
> You can try my most recent patch to 1.2.8 (posted to net channel
> over a week ago) that has more sane locking of the dev xmit functions,
> but that won't help the bogus memory size you report above.
That's being reported by the networking code, not anything _I_ did. If
this is the case, I would have to say the networking code is broken.
Would it also be the cause of crashes, or is it something else?
BTW -- 1.2.1 is relatively stable, 1.2.8 hangs frequently. Same hardware.
-Dan
------------------------------------------------------------------------------
Dan Hollis | Seiyuu Daisuki! |"Ranko-chan, ii oyome-san ni
JCIC System Administrator | Orikasa Ai | nareru wa yo" - Saotome Nodoka
http://www.jcic.org/ | Yokoyama Chisa |
dhollis@jcic.org | ("(^_^)") |
------------------------------------------------------------------------------
