Me too! Re: [URGENT ASSISTANCE REQUESTED] production machines dying

John Robinson (
Tue, 25 Nov 1997 18:07:53 GMT

Alan Cox writes ("Re: [URGENT ASSISTANCE REQUESTED] production machines dying"):
> > Kernel panic: skput : over: nnnnnnnn:nnn
> > In swapper task - not syncing
> >
> > I have checked both machines and can find no other error messages as
> > to why it is crashing. I can't recall the exact value of nnnnnnnnn:nnn.
> Those numbers are absolutely critical. Furthermore this is one variant
> of the effects of the "teardrop" fragment attack against Linux/95/NT
> that is fixed in 2.0.32. If you see it with 2.0.32 I -do- want to know

I reported this days ago and had no response. I'm afraid I don't know
how to discover which function it's in apart from grep but tell me how
and I'll do it. Umm, rubbish, I just read more of the thread so now I
do. Ummm.

See also "2.0.32 panic", message-id
<>, excerpted here:

> I'm afraid I didn't get much info out; nothing in the logs, no
> oops. Anyway, the machine froze leaving this on the screen (copied
> down by hand):
> Kernel panic: skput:over: 001a01ac:252
> In swapper task - not syncing
> This has happened twice, both times overnight when the machine has
> been totally idle (it's actually a spare development
> workstation/fileserver) unless someone's prodding it that I don't know
> about (but again nothing in any logs beyond home-time).
> My production server now running 2.0.32 thankfully isn't doing this
> (>40 hours up now).

Update: four times on the dev box, production server now up seven days
without problems. Thanks to Thomas Wouters for his comments, which
unfortunately didn't help. My /boot/ appears to be as old as
my 2.0.30 kernel... Aha! /usr/src/linux/ (suitably
timestamped) says:

0019f8c4 t vortex_start_xmit
0019faac t boomerang_start_xmit
0019fc88 t vortex_interrupt
001a0198 t vortex_rx
001a03ec t boomerang_rx
001a0734 t vortex_close
001a0898 t vortex_get_stats

So it's in vortex_rx ?

Both Thomas Schenk and I have 3c905 cards. Other than that our
hardware configurations are thoroughly different... and my production
server has 3c509 and is running without problems.

This is different from the 2.0.30 3c59x problem, where I discovered
that the network card hangs temporarily from time to time, causing my
DPT SmartRAID IV to go 'peep-peep-peep' and wakeable-up by pressing a
key on the keyboard (that's 2.0.30 with 0.44 of 3c59x), or just wait
for it to clear.

Anyway, /proc/pci under 2.0.32 shows:
PCI devices found:
Bus 0, device 10, function 0:
SCSI storage controller: DPT SmartCache/Raid (rev 2).
Medium devsel. BIST capable. IRQ 12. Master Capable. Latency=160. Min
Gnt=4.Max Lat=8.
I/O at 0x6100.
Bus 0, device 9, function 0:
VGA compatible controller: S3 Inc. Trio32/Trio64 (rev 83).
Medium devsel. IRQ 9.
Non-prefetchable 32 bit memory at 0xf0000000.
Bus 0, device 8, function 0:
Ethernet controller: 3Com 3C905 100bTX (rev 0).
Medium devsel. IRQ 15. Master Capable. Latency=248. Min Gnt=3.Max Lat=
I/O at 0x6000.
Bus 0, device 7, function 0:
ISA bridge: Intel 82371 Triton PIIX (rev 2).
Medium devsel. Fast back-to-back capable. Master Capable. No bursts.
Bus 0, device 0, function 0:
Host bridge: Intel 82437 (rev 1).
Medium devsel. Master Capable. Latency=32.

If you want the kernel config or other of my random wibblings, see my
previous mail.

John. _ _
-- _ | |___| |_ _ _
John Robinson 46 Bank Street, Dumfries DG1 2PA, UK | |_| / . \ ' \| ' \
+44 1387 247249 \___/\___/_||_|_||_|
Her mother calls her beautiful, her daddy says "a whore"... --Marillion