FC10, crash tool, kexec-tools, kdump, and uncompressed vmlinux

[THX-1138]

My Thinkpad T400 running FC10 keeps crashing, usually within moments of being booted. In an effort to find out why, I spent a couple of days figuring out how to get a system dump. (Hint: kexec-tools.)

I finally have a system vmcore file -- but I can't look at it with the 'crash' tool because the latter wants an uncompressed vmlinux (not vmlinuz) to read, presumably unstripped. And I haven't been able to find out where to get one.

Where, oh where has my little vmlinux gone? :-)

Thanks for any help..

[THX-1138]

Maybe I'm using the wrong tools after all? I did a lot of searching to find out how to get a crash dump when the kernel panics -- I was, in fact, amazed at the scarcity of info -- and the path I outlined is the best I could locate.

Is there a better and/or more canonical way to capture info about a panic so I can see what's going on and try to resolve/report it?

Most of the time the crash occurs when X is running, so I don't get to see the panic output on the console. The few times I have seen it, I wasn't able to transcribe it and it overflowed the screen anyway. If I could reproduce it at runlevel 3, I'd do so and copy it by hand, but I have no idea what sets it off and I might be there for a long time waiting.

So, what's the best way to capture this? The T400 doesn't have a serial port, so setting up a serial console to another system to capture the trace doesn't appear to be an option.

Thanks for any help!

[THX-1138]

I got a partial traceback on the F3 console while X was running. It wasn't a complete system crash, but it might be indicative of something. (As soon as I switched to X, there was a complete system crash.)

The top of the message said something about CPU#1 being stuck for 61s (or something like that).

What I was able to capture (i.e., transcribe manually):

[<c0414a2b>] ? lapic_next_event+0x15/0x1c
[<c04464e6>] ? clockevents_program_event+0xd4/0xe3
[<c0447399>] ? tick_dev_program_event+0x28/0x95
[<c0447450>] ? tick_program_event+0x22/0x29
[<c04420ea>] ? hrtimer_interrupt+0x135/0x155
[<c0406f63>] ? do_softirq+0xbe/0xdb
[<c0432431>] ? irq_exit+0x5f/0x83
[<c052047c>] ? trace_hardirqs_on_thunk+0xc/0x10
[<c06acf92>] ? do_page_fault+0x0/0x710
[<c06ab86a>] error_code+0x72/0x78
[<c0520205>] ? __get_user_4+0x11/0x17
[<c044920d>] ? exit_robust_list+0x27/0xf9
[<c06ab717>] ? _spin_lock_irq+0x1c/0x20
[<c0453ac7>] ? acct_collect+0x156/0x166
[<c0430b17>] do_exit+0x197/0x736
[<c0431126>] do_group_exit+0x70/0x97
[<c04396d2>] get_signal_to_deliver+0x2af/0x2d6
[<c0403f10>] do_notify_resume+0x71/0x679
[<c044464f>] ? getnstimeofday+0x54/0xe1
[<c0441f82>] ? ktime_get_ts+0x4a/0x4e
[<c0498d8a>] ? path_put+0x15/0x18
[<c0461c84>] ? audit_syscall_exit+0xb2/0xc7
[<c0404d75>] work_notifysig+0x13/0x1a
=======================

Perhaps there's a bad spot in the swapfile or something. I'm going to boot up a CD and run a badblocks scan. That'll take a few hours..

In the meantime, any ideas?

Thanks!

[THX-1138]

Well, no bad blocks were found on any of the partitions (non-destructive read/write test, one pass). I went through the BIOS settings again, and turned off something about 'multi-CPU' (it's a 'Intel Core 2 Duo processor P7370 (2.0 GHz)' unit). ISTR having to disable hyperthreading to avoid crashes on an earlier laptop..