2003-08-09 12:39:20

by Jean-Yves LENHOF

[permalink] [raw]
Subject: kernel oops...

Hello all,

I've just add two hangs with a plain vanilla kernel 2.4.21 on a Debian
woody up2date. The machine is an old Pentium 100 with 64 Mb. Softs
running on this machine are :
- Apache+PHP+mod_ssl+...
- Sendmail+spamassassin
- Mailman (This can explain the python)
- proftpd
- squid

I have installed gallery for fotos and while adding many big photos at a
time (the process is quite consuming... the CPU is 100% during this).
The problem seems to not occur when adding photo by photo

The trace on the logs :

Aug 8 20:57:55 www -- MARK --
Aug 8 21:15:01 www kernel: printing eip:
Aug 8 21:15:01 www kernel: c012d765
Aug 8 21:15:01 www kernel: Oops: 0000
Aug 8 21:15:01 www kernel: CPU: 0
Aug 8 21:15:01 www kernel: EIP: 0010:[filemap_nopage+21/524] Not
tainted
Aug 8 21:15:01 www kernel: EFLAGS: 00010282
Aug 8 21:15:01 www kernel: eax: c99cdff0 ebx: c1164260 ecx:
c2d2c260 edx: 08079000
Aug 8 21:15:01 www kernel: esi: 080798c0 edi: 00000000 ebp:
c2d2c260 esp: c3971ea0
Aug 8 21:15:01 www kernel: ds: 0018 es: 0018 ss: 0018
Aug 8 21:15:01 www kernel: Process python (pid: 17362,
stackpage=c3971000)
Aug 8 21:15:01 www kernel: Stack: c1164260 080798c0 00000000 c2d2c260
00000069
c10df214 c170c040 c1171720
Aug 8 21:15:01 www kernel: c0129bd6 c2d2c260 08079000 00000000
c1164260
080798c0 00000000 c2d2c260
Aug 8 21:15:02 www kernel: c0129dc4 c1164260 c2d2c260 080798c0
00000000
c0d811e4 c2d2c260 c1164260
Aug 8 21:15:02 www kernel: Call Trace: [do_no_page+90/492]
[handle_mm_fault+92/192] [do_page_fault+378/1184] [do_page_fault+0/1184]
[open_namei+744/1588]
Aug 8 21:15:02 www kernel: [do_brk+280/536] [sys_brk+188/232]
[error_code+52/64]
Aug 8 21:15:02 www kernel:
Aug 8 21:15:02 www kernel: Code: 8b 40 08 8b 5c 24 24 8b a8 ac 00 00 00
8b 44 24 28 8b 55 20
Aug 8 21:40:11 www syslogd 1.4.1#10: restart.

What is the problem ? And is this problem will be fixed in the newest
2.4.22 ?

Attached the .config of the kernel. If you need more info I can try to
provide them...

OK I've just added a file called results_ksymoops.txt done with the
command "ksymoops /var/log/messages" (I hope this is the good way to do
it)

Thanks in advance, and please CC me as I'm not on the list

Regards,

Have fun and keep up the good work you are doing ;-)

Jean-Yves


Attachments:
config-2.4.21.19062003 (33.92 kB)
results_ksymoops.txt (9.65 kB)
Download all attachments

2003-08-09 20:36:40

by Jean-Yves LENHOF

[permalink] [raw]
Subject: Re: kernel oops...

Le sam 09/08/2003 à 14:39, Jean-Yves LENHOF a écrit :
> Hello all,
>
> I've just add two hangs with a plain vanilla kernel 2.4.21 on a Debian
> woody up2date. The machine is an old Pentium 100 with 64 Mb. Softs
> running on this machine are :

Seems the problem has gone by using the original Debian kernel 2.4.18...
Reading a little bit the different changelog of kernel between 2.4.18
and 2.4.21, it seems that there are some vm change incorporated (from
-aa)...but I'm not an expert at all.

Another thing is that the load is about 2 or 3...

I hope somebody will have an idea off what's going on...

Have fun,

Jean-Yves

PS : Please CC me as I'm not on the list