2.4.21pre7(ac1) - BUG at filemap.c:81

2.4.21pre7(ac1) - BUG at filemap.c:81

[Tom Winkler]

Hi,

With kernel 2.4.21pre7 as well as kernel 2.4.21pre7-ac1 I get a message
about a "kernel BUG at filemap.c:81!" (see below). The message comes
"from time to time" (but the system seems to keep running without
problems). The bug can be reproduced by doing a hdparm -tT. The T part
(disk reads) is not completed but the message below is displayed.

The Kernel is also patched with the latest acpi patch as well as with
the i810 (lite) framebuffer driver.
A full dmesg along with an lspci can be found at
http://www.wnk.at/tmp/dmesg.txt and
http://www.wnk.at/tmp/lspci.txt


kernel message when running "hdparm -tT"; system keeps running:

kernel BUG at filemap.c:81!
invalid operand: 0000
CPU:    0
EIP:    0010:[<c012d03e>]    Not tainted
EFLAGS: 00010206
eax: 0a610000   ebx: 00000000   ecx: c11a5fe0   edx: cfe75c5c
esi: c11a5fe0   edi: cfe75c5c   ebp: cd98a820   esp: cd66feb4
ds: 0018   es: 0018   ss: 0018
Process hdparm (pid: 248, stackpage=cd66f000)
Stack: 00000000 c012da32 cd6767e4 00003643 cfe75c5c c012dad6 c11a5fe0 cd6767e4
       00003643 cfe75c5c c11a5fe0 00000004 0000363f cd98a820 004a8530 c012e119
       0000001f 00000020 00003621 cd98a820 c11a663c cd6767e4 00003621 c012e50c
Call Trace:    [<c012da32>] [<c012dad6>] [<c012e119>] [<c012e50c>] 
[<c012e8d0>]
  [<c012ea18>] [<c012e8d0>] [<c013be03>] [<c010740f>]

Code: 0f 0b 51 00 3d 55 2e c0 ff 05 64 a2 32 c0 5b c3 89 f6 83 ec


If you need any further information about the setup please let me know. If 
there are any patches you'd like me to test just drop me a line.
I'm not subscribed to lkm so please cc me on your replies.

Bye,
-- 
Tom Winkler
e-mail: tom@qwws.net

Re: 2.4.21pre7(ac1) - BUG at filemap.c:81

[Tom Winkler]

Hello,

This is a followup message to my previous report with more detailed
information this time.


summary:

Kernel reports BUG in filemap.c


full description:

The bug message shows up "from time to time" but can also be
reproduced on my system (see below).
After such a message gets displayed the system keeps running
without noticeable problems.

kernel version:

Linux version 2.4.21-pre7-ac1 (tom@thunder) (gcc version 3.2.3 20030331 
(Debian prerelease))


additional patches applied to the kernel:

acpi-20030328 [ url: 
http://sourceforge.net/project/showfiles.php?group_id=36832 ]
i810fb (0.0.35-lite) [ url: http://i810fb.sourceforge.net/ ]


output of Oops:

ksymoops 2.4.8 on i686 2.4.21-pre7-ac1.  Options used
     -V (default)
     -k /proc/ksyms (default)
     -l /proc/modules (default)
     -o /lib/modules/2.4.21-pre7-ac1/ (default)
     -m /boot/System.map-2.4.21-pre7-ac1 (specified)

kernel BUG at filemap.c:81!
invalid operand: 0000
CPU:    0
EIP:    0010:[<c012d03e>]    Not tainted
Using defaults from ksymoops -t elf32-i386 -a i386
EFLAGS: 00010206
eax: 0a610000   ebx: 00000000   ecx: c11a5fe0   edx: cfe75c5c
esi: c11a5fe0   edi: cfe75c5c   ebp: cd98a820   esp: cd66feb4
ds: 0018   es: 0018   ss: 0018
Process hdparm (pid: 248, stackpage=cd66f000)
Stack: 00000000 c012da32 cd6767e4 00003643 cfe75c5c c012dad6 c11a5fe0 cd6767e4
       00003643 cfe75c5c c11a5fe0 00000004 0000363f cd98a820 004a8530 c012e119
       0000001f 00000020 00003621 cd98a820 c11a663c cd6767e4 00003621 c012e50c
Call Trace:    [<c012da32>] [<c012dad6>] [<c012e119>] [<c012e50c>] 
[<c012e8d0>]
  [<c012ea18>] [<c012e8d0>] [<c013be03>] [<c010740f>]
Code: 0f 0b 51 00 3d 55 2e c0 ff 05 64 a2 32 c0 5b c3 89 f6 83 ec


>>EIP; c012d03e <do_readahead+7e/a0>   <=====

>>ecx; c11a5fe0 <_end+de47fc/145b287c>
>>edx; cfe75c5c <_end+fab4478/145b287c>
>>esi; c11a5fe0 <_end+de47fc/145b287c>
>>edi; cfe75c5c <_end+fab4478/145b287c>
>>ebp; cd98a820 <_end+d5c903c/145b287c>
>>esp; cd66feb4 <_end+d2ae6d0/145b287c>

Trace; c012da32 <madvise_fixup_middle+a2/190>
Trace; c012dad6 <madvise_fixup_middle+146/190>
Trace; c012e119 <sys_mincore+79/140>
Trace; c012e50c <generic_file_write+27c/760>
Trace; c012e8d0 <generic_file_write+640/760>
Trace; c012ea18 <remove_suid+28/70>
Trace; c012e8d0 <generic_file_write+640/760>
Trace; c013be03 <write_some_buffers+13/e0>
Trace; c010740f <simd_coprocessor_error+3/c>

Code;  c012d03e <do_readahead+7e/a0>
00000000 <_EIP>:
Code;  c012d03e <do_readahead+7e/a0>   <=====
   0:   0f 0b                     ud2a      <=====
Code;  c012d040 <do_readahead+80/a0>
   2:   51                        push   %ecx
Code;  c012d041 <do_readahead+81/a0>
   3:   00 3d 55 2e c0 ff         add    %bh,0xffc02e55
Code;  c012d047 <do_readahead+87/a0>
   9:   05 64 a2 32 c0            add    $0xc032a264,%eax
Code;  c012d04c <do_readahead+8c/a0>
   e:   5b                        pop    %ebx
Code;  c012d04d <do_readahead+8d/a0>
   f:   c3                        ret
Code;  c012d04e <do_readahead+8e/a0>
  10:   89 f6                     mov    %esi,%esi
Code;  c012d050 <do_readahead+90/a0>
  12:   83 ec 00                  sub    $0x0,%esp


problem can be triggered by:

hdparm -tT


software:

Debian Sid (unstable)

ver_linux script reports:

Linux thunder 2.4.21-pre7-ac1 #2 Die Apr 15 23:02:39 CEST 2003 i686 unknown 
unknown GNU/Linux

Gnu C                  3.2.3
Gnu make               3.80
util-linux             2.11y
mount                  2.11y
modutils               2.4.21
e2fsprogs              1.32
pcmcia-cs              3.2.2
Linux C Library        2.3.1
Dynamic linker (ldd)   2.3.1
Procps                 3.1.6
Net-tools              1.60
Console-tools          0.2.3
Sh-utils               4.5.9
Modules Loaded         nls_iso8859-15 nls_cp437 sonypi rtc



processor information:

processor       : 0
vendor_id       : GenuineIntel
cpu family      : 6
model           : 8
model name      : Pentium III (Coppermine)
stepping        : 10
cpu MHz         : 695.567
cache size      : 256 KB
fdiv_bug        : no
hlt_bug         : no
f00f_bug        : no
coma_bug        : no
fpu             : yes
fpu_exception   : yes
cpuid level     : 2
wp              : yes
flags           : fpu vme de pse tsc msr pae mce cx8 sep mtrr pge mca cmov pat 
pse36 mmx fxsr sse
bogomips        : 1389.36



lspci output:

http://www.wnk.at/tmp/lspci.txt


dmesg output:

http://www.wnk.at/tmp/dmesg.txt


bye,
-- 
Tom Winkler
e-mail: tom@qwws.net