this one was definitely caused by isdn, some seconds after powering
on the PBX (ascotel). unfortunately I only can see the last oops (and
parts of the one before), because the message just scrolled off my screen!
(I am typing this on a second computer by hand .... *f*ck* ... so I
cannot guarantee this is 100% correct, oh my god so many numbers ...
I'll remove any sti() in printk (if there is any!) and recompile ....)
[<0010976c>] [<0010a6d2>] [<00109470>] [<001092d7>]
Code: 8b 40 08 83 c0 fc 50 8b 44 24 24 83 c0 04 50 8b 74 24 7c 56
Aiee, killing interrupt handler
kfree of non-kmalloced memory: 001b0cc4, next= 00000000, order=0
kfree of non-kmalloced memory: 001b0cb4, next= 00000000, order=0
kfree of non-kmalloced memory: 001b1cc8, next= 00000000, order=0
idle task may not sleep
invalid operand: 0000
CPU: 0
EIP: 0010:[<00000017>]
EFLAGS: 00010202
eax: 00000000 ebx: 01b50f80 ecx: 01680018 edx: 02880534
esi: 019feb70 edi: 0800171c ebp: bffffc24 esp: 01740f3c
ds: 0018 es: 0018 fs: 002b gs: 002b ss: 0018
Process cat (pid: 186, process nr: 10, stackpage=01740000)
Stack: 01740f40 00000018 00000018 00000018 0286153e 01740f8 01b50f80 00000000
(do you really need the stack ????? that's just too much typing)
Call Trace: [<0286153e>] [<0285f7b7>] [<00122bd2>] [<00122c44>] [<0010a6d2>]
Code: f0 4c e1 00 f0 6f ef 00 f0 3c 00 e3 07 45 00 e3 07 6f ef 00
idle task may not sleep
last message repeated several times...
you may wonder why "cat" causes this: for those who know isdn4linux, I'm
doing "cat /dev/isdnctrl | tee ic.out &"
I don't have an EIP for the first oops, and the second EIP=00000017
will cause ksymoops to segfault, so I did the trace manually too...
1)
[<0010976c>] 001096f0 T sys_idle
[<0010a6d2>] 0010a680 T system_call
[<00109470>] 00109470 t init
[<001092d7>] 00109160 T start_kernel
looks like the trace is pointing to some trash data ....
2) the addresses 0286 come from isdn.map.sort
[<0286153e>] 02861484 isdn_close
[<0285f7b7>] 0285f5e4 isdn_read
[<00122bd2>] 00122b70 T close_fp
[<00122c44>] 00122c00 T sys_close
[<0010a6d2>] 0010a680 T system_call
since EIP=00000017, disassembly code doesnt make sense.
/herp