Advice on hangs in 2.5.73-bk6?

Advice on hangs in 2.5.73-bk6?

[Matthew Hughes]

I'm looking for any advice on how to handle machine hangs in current
2.5 series kernels.

My impression from the periodicals is that the 2.5 series is to the
point where general beta testing is appropriate. I'm interested in the
performance improvements for multimedia apps, so I've been trying out
the new kernels since around 2.5.67 - while the performance is great, 
each version that I've tried hangs consistently within a day or so. 

The box is a uniprocessor AMD Athlon XP1900+ machine that runs 2.4.20
solidly. The distribution is RH 8.0. Earlier versions had occasional
OOPSes in the logs, though not necessarily related to the hangs (see
http://bugme.osdl.org/show_bug.cgi?id=580 for details). 2.5.73-bk6 still
hangs after 4-36 hours of operation, but there are generally no kernel 
messages within 15 or 20 minutes of the hang, and no OOPS messages. 

The hangs are pretty hard - no reply to pings, Ctrl-Alt-Del does
nothing, Magic Sysrq does nothing (though I don't know the status of
this feature in 2.5). 

I'm looking for any advice on steps to take to debug this issue - I know
the problem description is probably too vague to be useful A pointer to
a web site with debugging steps would be great, or if anyone would like
more info or to outline some steps to take it would be much appreciated.
I'm not a kernel developer, but I am a somewhat experienced embedded 
developer, so I could probably manage to get a serial debugger or
something running if that's the appropriate action at this point.

I'm not a subscriber to lkml, so please CC me on any replies.
							Thanks!
							Matt Hughes
						mhughes@mhughes.dhs.org

2.6.0-test1 hangs silently in under an hour

[Matthew Hughes]

I've got a desktop machine which is hanging regularly with 
recent 2.5 and now 2.6.0-test1 releases. It seems very solid 
running 2.4.20 ...

The machine functions as a normal desktop (email, X, web browser,
etc) and as an NFS server for music served to a machine attached to 
the stereo (so light but consistent disk and network activity).

The symptom is simply a totally hung system - no response to keyboard, 
mouse, network pings, ctrl-alt-del, or magic sysrq. There is no output
to /var/log/messages that survives a reboot. I've compiled the kernel
with all the kernel debugging options set to true with no difference in
behavior:
CONFIG_DEBUG_KERNEL=y
CONFIG_DEBUG_STACKOVERFLOW=y
CONFIG_DEBUG_SLAB=y
CONFIG_DEBUG_IOVIRT=y
CONFIG_MAGIC_SYSRQ=y
CONFIG_DEBUG_SPINLOCK=y
CONFIG_DEBUG_PAGEALLOC=y
CONFIG_DEBUG_SPINLOCK_SLEEP=y
CONFIG_FRAME_POINTER=y
CONFIG_X86_EXTRA_IRQS=y
CONFIG_X86_FIND_SMP_CONFIG=y
CONFIG_X86_MPPARSE=y

Some general info - 
Distribution: Red Hat 8.0
Hardware Environment: Athlon XP1900
Software Environment: gcc 3.2
The motherboard is an Asus A7V266-E. 

I'd be happy to run test patches, change config options, etc if that
would help ... also, if there's anything else I should be doing to 
track this down before reporting, please let me know or point me 
to a document.

I'm not a subscriber to lkml, so please CC me on any replies.
						Thanks!
						Matt Hughes
					mhughes@mhughes.dhs.org