PROBLEM: kernel freezes, possibly via velocity-related

PROBLEM: kernel freezes, possibly via velocity-related

[Arvid Brodin]

[I sent this to lkml a few days ago; sending it to netdev (being the
maillist for the via velocity driver) now after Alan Cox responded
thinking that might be a good idea.]

Hi,

Since I started running deluged (a bittorrent daemon) on my file server,
I've started getting sporadic kernel freezes, a few times a week or so.

Symptoms:
* Keyboard stops responding (SysRq combos does not work),
* No panic/oops message on the console; the screen displays the normal
login prompt with a blinking cursor,
* The network activity LED flashes steadily at 1-2 Hz. The corresponding
LED on my switch blinks too, but the traffic does not seem to get
forwarded to any other port on the switch (no other lights blinking),
* The machine does not respond to pings or ssh logins,
* Nothing in var/log/messages or kern.log that I can relate to the
freeze (I have gotten some "UDP: short packet" and "eth0: excessive work
at interrupt" messages a few hours before the freezes a few times, but
that's all).

I've never gotten a freeze with the deluge daemon stopped (which it has
been for many weeks sometimes), but consistently get freezes, often
within a day or two, when it is started and have active torrents.

In addition to the kernel I currently run (2.6.29.4), the freezes also
occured with gentoo kernels 2.6.23-gentoo-r8 and 2.6.27-gentoo-r10.

# uname -a
Linux sv1 2.6.29.4 #1 Thu May 21 03:26:33 CEST 2009 i686 VIA Esther
processor 1200MHz CentaurHauls GNU/Linux

# lspci -vvv
See http://pastebin.com/f21145883

# cat /proc/cpuinfo
processor	: 0
vendor_id	: CentaurHauls
cpu family	: 6
model		: 10
model name	: VIA Esther processor 1200MHz
stepping	: 9
cpu MHz		: 1199.496
cache size	: 128 KB
fdiv_bug	: no
hlt_bug		: no
f00f_bug	: no
coma_bug	: no
fpu		: yes
fpu_exception	: yes
cpuid level	: 1
wp		: yes
flags		: fpu vme de pse tsc msr pae mce apic sep mtrr pge cmov pat
clflush acpi mmx fxsr sse sse2 tm nx pni est tm2 rng rng_en ace ace_en
ace2 ace2_en phe phe_en pmm pmm_en
bogomips	: 2399.95
clflush size	: 64
power management:


Should I report this to the maintainer of my network driver? I'm not
sure the symptoms are clear enough to directly blame the network driver,
but I do suspect it. Is there some way I can debug a frozen kernel to
zoom in on the problem?

-- 
Arvid