Re: 2.2.7 Oops on SMP machine in __free_page() while loading executable

Mark Hanson (mbh@netcom.com)
Wed, 16 Jun 1999 14:39:45 -0700 (PDT)


>> This is from our SMP machine which has been experiencing bi-weekly
>> silent freezes with kernels all the way back to 2.0.3x. This morning I
>> actually got an Oops. I don't know if this is related to the silent
>> freezes or not.
>
>> Unable to handle kernel NULL pointer dereference at virtual address 00000004
>> current-> tss.cr3 = 05df6000, %cr3 = 05ef6000
>>>> EIP: c01221b8 <__free_page+ac/ec>
>
>An oops in __free_page usually means hardware memory corruption.
>There's not much here to base a software diagnosis on --- is there any
>particular load pattern which produces the fault?

Hi Stephen,

We have a software package that spawns lots of recursive make processes
and compilers when built. We build it all the time, many multiple
times per day, and usually without any problem, but I've noticed that
the machine usually hangs up when someone is doing a build.

This particular Oops happened while doing a backup of another machine
via "rmt" and running a script that creates different packages of
our product. I'm told that this script creates a lot of temporary
files and thrashes the disk fairly well. The Oops listed the script
as the process that caused the crash.

I'm wondering if these previous "silent freezes" I've been having may
actually be giving Oops's, but I just can't see them because the
screen is blank and the kernel is too hosed to unblank it. What is
it exactly that causes the console to blank after a while. I'm sure
it's something easy, I'm just having a hard time finding it. The
machine is running RH 5.2.

Any suggestions for tracking this sort of thing down are welcome.

Mark

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/