lkml.org 
[lkml]   [2003]   [Aug]   [7]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: 2.4.22-pre lockups (now decoded oops for pre10)


On Thu, 7 Aug 2003, Stephan von Krawczynski wrote:

> On Wed, 6 Aug 2003 15:15:39 -0300 (BRT)
> Marcelo Tosatti <marcelo@conectiva.com.br> wrote:
>
> > Stephan,
> >
> > I'm pretty worried about this problem.
> >
> > Your oopses seem to be the result of some kind of memory corruption. On
> > the other oopses we could see the kernel oopsing on
> > remove_page_from_hash_queue due to corrupted pointers (as Willy pointed
> > out).
> >
> > Can you please try to crash your box again with
> >
> > CONFIG_DEBUG_SLAB=y
> >
> > Again, thanks a lot for your reports.
>
> Ok, I have two things.
> First, another oops. I upgraded the system to rc1 yesterday and it did not
> survive a single day. Here's the decoded oops, the box was "clean" meaning no
> weird modules or the like:
>
>
> ksymoops 2.4.8 on i686 2.4.22-rc1. Options used
> -V (default)
> -k /proc/ksyms (default)
> -l /proc/modules (default)
> -o /lib/modules/2.4.22-rc1/ (default)
> -m /boot/System.map-2.4.22-rc1 (default)
>
> Warning: You did not tell me where to find symbol information. I will
> assume that the log matches the kernel and modules that are running
> right now and I'll use the default options above for symbol resolution.
> If the current kernel and/or modules do not match the log, you can get
> more accurate output by telling me the kernel version and where to find
> map, modules, ksyms etc. ksymoops -h explains the options.
>
> Unable to handle kernel NULL pointer dereference at virtual address 00000004
> c0145060
> *pde = 00000000
> Oops: 0002
> CPU: 1
> EIP: 0010:[<c0145060>] Not tainted
> Using defaults from ksymoops -t elf32-i386 -a i386
> EFLAGS: 00010283
> eax: 00000000 ebx: c822feb4 ecx: c822fe60 edx: e07e7780
> esi: 00000000 edi: e07e7780 ebp: f59bfe3c esp: f59bfe2c
> ds: 0018 es: 0018 ss: 0018
> Process nfsd (pid: 1737, stackpage=f59bf000)
> Stack: f0cce7a0 00000001 f59bfe38 c822fe60 f0cce7f4 eec54ef4 00000000 e07e7760
> f59be000 f59bfea8 c0183ef5 e07e7780 e07e77cc c02ed880 e07e7760 f8c84fc8
> f59bfea8 dfe6c960 00000000 e07e7760 dfe6c960 00000000 f59c6e04 f59bfea8
> Call Trace: [<c0183ef5>] [<f8c84fc8>] [<f8c856f1>] [<f8c8cee4>] [<f8c8e295>]
> [<f8c923f4>] [<f8c80699>] [<f8c65938>] [<f8c923f4>] [<f8c91a38>] [<f8c91a58>]
> [<f8c80411>] [<c010592e>] [<f8c80210>]
> Code: 89 50 04 c7 41 54 00 00 00 00 c7 43 04 00 00 00 00 8b 44 24
>
>
> >>EIP; c0145060 <fsync_buffers_list+50/1b0> <=====
>
> >>ebx; c822feb4 <_end+7e84c94/3852ee40>
> >>ecx; c822fe60 <_end+7e84c40/3852ee40>
> >>edx; e07e7780 <_end+2043c560/3852ee40>
> >>edi; e07e7780 <_end+2043c560/3852ee40>
> >>ebp; f59bfe3c <_end+35614c1c/3852ee40>
> >>esp; f59bfe2c <_end+35614c0c/3852ee40>
>
> Trace; c0183ef5 <reiserfs_sync_file+65/d0>
> Trace; f8c84fc8 <[nfsd]nfsd_sync+78/d0>
> Trace; f8c856f1 <[nfsd]nfsd_commit+a1/b0>
> Trace; f8c8cee4 <[nfsd]nfsd3_proc_commit+94/130>
> Trace; f8c8e295 <[nfsd]nfs3svc_decode_commitargs+35/e0>
> Trace; f8c923f4 <[nfsd]nfsd_procedures3+2f4/320>
> Trace; f8c80699 <[nfsd]nfsd_dispatch+119/21d>
> Trace; f8c65938 <[sunrpc]svc_process+4d8/570>
> Trace; f8c923f4 <[nfsd]nfsd_procedures3+2f4/320>
> Trace; f8c91a38 <[nfsd]nfsd_version3+0/10>
> Trace; f8c91a58 <[nfsd]nfsd_program+0/28>
> Trace; f8c80411 <[nfsd]nfsd+201/370>
> Trace; c010592e <arch_kernel_thread+2e/40>
> Trace; f8c80210 <[nfsd]nfsd+0/370>
>
> Code; c0145060 <fsync_buffers_list+50/1b0>
> 00000000 <_EIP>:
> Code; c0145060 <fsync_buffers_list+50/1b0> <=====
> 0: 89 50 04 mov %edx,0x4(%eax) <=====
> Code; c0145063 <fsync_buffers_list+53/1b0>
> 3: c7 41 54 00 00 00 00 movl $0x0,0x54(%ecx)
> Code; c014506a <fsync_buffers_list+5a/1b0>
> a: c7 43 04 00 00 00 00 movl $0x0,0x4(%ebx)
> Code; c0145071 <fsync_buffers_list+61/1b0>
> 11: 8b 44 24 00 mov 0x0(%esp,1),%eax
>
>
> 1 warning issued. Results may not be reliable.
>
>
> As you can see reiserfs seems involved. Regarding reiserfs and my last postings
> I can assure you that all reiserfs partitions were checked via reiserfsck right
> before installation of rc1 - as Oleg advised - and found:
> "Comparing bitmaps.. vpf-10640: The on-disk and the correct bitmaps differs"
> I was told to use --fix-fixable option which I did and it indeed fixed the
> problem. Trying reiserfsck after that found no errors any more. So I see no
> chance that corrupt data on the media (through former crashes) is responsible
> for this one. Hint: spelling in reiserfsck should be checked ;-)

It might be a problem in reiserfs. You're getting oopses on different
places with different stack traces, which is weird.

I'll take a closer look at this oops now.

> Second, I re-install the box with CONFIG_DEBUG_SLAB="y" right now. Please tell
> me if I should perform special steps (SYSRQ or the like) after the next crash
> happens, or if the decoded oops will be sufficient.

The decoded oops should be sufficient.

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

\
 
 \ /
  Last update: 2005-03-22 13:47    [W:0.100 / U:1.272 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site