Re: 2.4.19 BUG in page_alloc.c:91

Salvador Eduardo Tropea (salvador@inti.gov.ar)
Tue, 27 Aug 2002 09:47:42 -0300


I canīt find a final conclusion about this topic.
What I found is that it happends some hours after the kernel killed a
process that was eating the memory.
Example:

Aug 26 13:07:22 ice kernel: Out of Memory: Killed process 18432
(mozilla-bin).
Aug 26 13:07:22 ice kernel: Out of Memory: Killed process 18443
(mozilla-bin).
Aug 26 13:07:22 ice kernel: Out of Memory: Killed process 18444
(mozilla-bin).
Aug 26 13:07:22 ice kernel: Out of Memory: Killed process 18445
(mozilla-bin).
Aug 26 13:07:22 ice kernel: Out of Memory: Killed process 18447
(mozilla-bin).
Aug 26 13:07:22 ice kernel: Out of Memory: Killed process 30410
(mozilla-bin).
....
Aug 27 06:26:00 ice kernel: kernel BUG at page_alloc.c:91!
Aug 27 06:26:00 ice kernel: invalid operand: 0000
Aug 27 06:26:00 ice kernel: CPU: 0
Aug 27 06:26:00 ice kernel: EIP: 0010:[__free_pages_ok+45/624]
Tainted: P
Aug 27 06:26:00 ice kernel: EFLAGS: 00010286
Aug 27 06:26:00 ice kernel: eax: c103a528 ebx: c1105760 ecx:
c1105760 edx: c0228e9c
Aug 27 06:26:00 ice kernel: esi: 00000000 edi: 00000016 ebp:
0000005e esp: c7f93f18
Aug 27 06:26:00 ice kernel: ds: 0018 es: 0018 ss: 0018
Aug 27 06:26:00 ice kernel: Process kswapd (pid: 4, stackpage=c7f93000)
Aug 27 06:26:00 ice kernel: Stack: c22d28c0 c1105760 00000016 0000005e
c013755c c1105760 00000
1d0 00000016
Aug 27 06:26:00 ice kernel: 0000005e c0135a09 c22d28c0 c1105760
c012d482 c012e3fb c012d
4bb 00000020
Aug 27 06:26:00 ice kernel: 000001d0 00000020 00000006 00000006
c7f92000 000003a4 00000
1d0 c0229034
Aug 27 06:26:00 ice kernel: Call Trace: [try_to_free_buffers+140/224]
[try_to_release_page+
73/80] [shrink_cache+482/768] [__free_pages+27/32] [shrink_cache+539/768]
Aug 27 06:26:00 ice kernel: [shrink_caches+86/128]
[try_to_free_pages+60/96] [kswapd_balance
_pgdat+67/144] [kswapd_balance+22/48] [kswapd+157/192] [kernel_thread+40/64]

At 6:25 the cron started the regular locate data base update and other
tasks I guess it made the kernel reduce the size of the cache and fail.
I saw reports it happends with "non-tainted" kernels and with or without
nVidia driver, so perhaps thatīs a real bug in the mechanism used by the
kernel to kill a process that is eating all the memory introduced in 2.4.19.
I can fill a full bug report but perhaps any of you know about the
conclusion of this thread and Iīm just too stupid to find it.
Note that it doesnīt happend if the kernel doesnīt kill mozilla-bin for
days, when kernel kills this process is just a matter of hours before it
and from this point the cache size wonīt be reduced.

SET

-- 
Salvador Eduardo Tropea (SET). (Electronics Engineer)
Visit my home page: http://welcome.to/SetSoft or
http://www.geocities.com/SiliconValley/Vista/6552/
Alternative e-mail: set@computer.org set@ieee.org 
Address: Curapaligue 2124, Caseros, 3 de Febrero
Buenos Aires, (1678), ARGENTINA Phone: +(5411) 4759 0013

- 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/