swapoff i...

[Jarosław Kamper]

divide error: 0000 [#1]
Modules linked in: snd_emu10k1 snd_rawmidi snd_seq_device snd_ac97_codec 
snd_util_mem snd_hwdep radeon drm button nfs 8139too mii md5 ipv6 ext2 
mbcache nfsd lockd sunrpc uhci_hcd usbcore via_agp agpgart loop ide_cd 
cdrom psmouse snd_pcm_oss snd_pcm snd_timer snd_page_alloc snd_mixer_oss 
snd soundcore ide_disk xfs exportfs via82cxxx ide_core
CPU:    0
EIP:    0060:[<c0175009>]    Not tainted VLI
EFLAGS: 00010246   (2.6.11.6-4)
EIP is at refill_inactive_zone+0x1e9/0x570
eax: 00000000   ebx: 00000000   ecx: 00000023   edx: 00000000
esi: 51eb851f   edi: c031b2c4   ebp: c15c7f00   esp: c15c7e2c
ds: 007b   es: 007b   ss: 0068
Process kswapd0 (pid: 130, threadinfo=c15c6000 task=c159eaa0)
Stack: 000147b3 c15c7e50 00000046 c031b394 00000046 0000001f 00000000 
00000000
       00000020 00000020 00000000 c15c7f50 c15c7e5c c15c7e5c c15c7e64 
c15c7e64
       c13f1bd8 c13dd218 ced31060 ced31060 dfc64be0 00000005 0001f7bc 
00000576
Call Trace:
 [<c013a34a>] show_stack+0x7a/0x90
 [<c013a4cd>] show_registers+0x14d/0x1b0
 [<c013a692>] die+0xc2/0x140
 [<c013a874>] do_divide_error+0xa4/0xb0
 [<c0139f2f>] error_code+0x4f/0x60
 [<c017542b>] shrink_zone+0x9b/0xe0
 [<c0175870>] balance_pgdat+0x240/0x310
 [<c0175a21>] kswapd+0xe1/0x100
 [<c01372e5>] kernel_thread_helper+0x5/0x10
Code: 85 2c ff ff ff c1 ea 05 89 d1 31 d2 f7 f1 8b 55 94 b9 64 00 00 00 29 
c1 39 d3 0f 46 d3 89 d0 f7 e6 8b 45 98 c1 ea 05 89 d3 31 d2 <f7> f3 39 c8 
0f 4ec8 89 0d a0 bb 3d c0 8b 85 40 ff ff ff c1 e8

To chyba nie do końca normalne? Robiłem to na serwerkach z 64mb 
(czyszczenie buforów ;) ) i nigdy nie miałem takiej sytuacji (rzadko coś 
zdechło z braku ramu), ale to desktop z 512mb, gdzie 300mb jest cached.
Pomysły?

-- 
Jarosław Kamper <jack w jack.eu.org> http://jack.eu.org/