"kernel BUG at page_alloc.c:102!" (kernel 2.4.20)

"kernel BUG at page_alloc.c:102!" (kernel 2.4.20)

[Keith R. John Warno]

Greetings.

On a relatively idle system (used mainly for syslog things) that had
been up for ~42 days, kswapd croaked and the kernel claimed (sorry if
these lines are wrapped):

kernel BUG at page_alloc.c:102!
invalid operand: 0000
CPU:    0
EIP:    0010:[__free_pages_ok+68/656]    Tainted: P 
EFLAGS: 00010286
eax: c11703ec   ebx: c151fabc   ecx: c151fad8   edx: c0254c9c
esi: 00000000   edi: 00000010   ebp: 000001f5   esp: c16f5f14
ds: 0018   es: 0018   ss: 0018
Process kswapd (pid: 5, stackpage=c16f5000)
Stack: d7b64dc0 c151fabc 00000010 000001f5 c0136d2c c151fabc 000001d0 00000010 
       000001f5 c01351b9 d7b64dc0 c151fabc c012cb52 c012db8b c012cb8b 00000020 
       000001d0 00000020 00000006 00000006 c16f4000 000021dd 000001d0 c0254e34 
Call Trace:    [try_to_free_buffers+140/224] [try_to_release_page+73/80] [shrink_cache+498/784] [__free_pages+27/32] [shrink_cache+555/784]
  [shrink_caches+86/128] [try_to_free_pages_zone+60/96] [kswapd_balance_pgdat+65/144] [kswapd_balance+22/48] [kswapd+157/192] [kernel_thread+40/64]

Code: 0f 0b 66 00 53 f6 21 c0 89 d8 2b 05 f0 20 2b c0 69 c0 a3 8b 


The box is essentially a LinuxFromScratch (actually a severely
overhauled SuSE 6.2) box.  Vanilla kernel 2.4.20 (no patches), glibc
2.2.5, gcc 2.95.3.  It's all running on an Abit KT7A mainboard (Athlon
1.4GHz) with 640MB or so of RAM.

I rebooted the box but it hung when trying to turn off swap.  :-)  (BTW,
is there a means by which to "restart" kswapd if need be?)

I've been using 2.4.20 since a few days after its release.  This is the
first time I've seen this "kernel BUG"; I have absolutely no idea what
triggered it, nor how to reproduce it.

If you need more info, just ask!

Regards,
krjw.
-- 
Keith R. John Warno                         [SA, Valaran Corporation]

Re: "kernel BUG at page_alloc.c:102!" (kernel 2.4.20)

[Keith R. John Warno]

* Burton Windle <bwindle@fint.org> [17/03/2003 0847EST]:
> What modules do you have loaded?
> 

Oh yes I forgot to mention those didn't I.  :-/  The kernel is tainted
because of the !@#^%*#@* NVidia closed-source driver (1.0-3123).  I have
ALSA 0.9.0rc6 stuff loaded (yes, old; have yet to update it).
Everything else is stock 2.4.20 kernel stuff. (I am aware there is an
NVidia 1.0-4191 driver set, but  it didn't play nice so I reverted to
the previous release.)

Here is the lsmod output:
Module                  Size  Used by    Tainted: P  
snd-emu10k1            57636   0 
snd-hwdep               3680   0  [snd-emu10k1]
snd-util-mem            1168   0  [snd-emu10k1]
snd-rawmidi            11936   0  [snd-emu10k1]
snd-ac97-codec         25776   0  [snd-emu10k1]
snd-pcm-oss            36964   0 
snd-pcm                54560   0  [snd-emu10k1 snd-pcm-oss]
snd-timer              10400   0  [snd-pcm]
snd-mixer-oss          10752   0  [snd-pcm-oss]
snd                    24008   0  [snd-emu10k1 snd-hwdep snd-util-mem snd-rawmidi snd-ac97-codec snd-pcm-oss snd-pcm snd-timer snd-mixer-oss]
soundcore               3364   6  [snd]
NVdriver             1066528  10 
agpgart                12800   3 
loop                    8496   0  (unused)
autofs4                 8580   1 
nfs                    44932   0  (unused)
nfsd                   44928   0  (unused)
lockd                  37248   0  [nfs nfsd]
sunrpc                 59924   0  [nfs nfsd lockd]
tulip                  37664   1 
mousedev                3904   1 
hid                    13248   0  (unused)
input                   3168   0  [mousedev hid]
usb-uhci               21796   0  (unused)
usbcore                55232   0  [hid usb-uhci]
serial                 44128   0  (unused)

krjw.
-- 
Keith R. John Warno                         [SA, Valaran Corporation]
"When you go into  court you are putting your fate  into the hands of
twelve people who weren't smart enough to get out of jury duty."
                -- Norm Crosby