Skip to contentSkip to conversation

parport_pc Oops with 2.6.0-test3

DISCUSSION Christian Kujau 2 participants 6 messages Aug 18, 2003 – Aug 25, 2003

lore
Share & tools
lore
#0 lore raw

hi,

as the subject might presume, the parport_pc modules oopses when it's unloaded. so on every shutdown i see an Oops on the console, which get logged to syslog then:

http://christian.go4more.de/parport/parport_pc-init6.txt

yes, the kernel is tainted with the nvidia module, i try to reprodduce later on without nvidia.

upon manual "rmmod parport_pc", this exits with a seg-fault, the log shows:

http://christian.go4more.de/parport/parport_pc-rmmod.txt

This is with vanilla 2.6.0-test3, i386, Debian unstable (glibc 2.3.2-2, kernel compiled with gcc (GCC) 3.3.2 20030812 (Debian prerelease).

more info available, of course.

Thanks,
Christian.

PS: i have 2.6.0-test3 (and previous versions) running now for days/weeks und X11, normal use, sometimes under load, some quirks, but *no* crashes! Thank you!

Signature · 3 lines
BOFH excuse #333:

A plumber is needed, the network drain is clogged
#1 lore raw

i was able to reproduce with a non tainted kernel:

http://christian.go4more.de/parport/parport_oops.txt

on this machine i am only using PLIP sometimes (to connect a very old laptop). so, i don't need parport_pc very often. but as it's used for printing too, this could be an issue....

Thanks,
Christian.

Signature · 3 lines
BOFH excuse #39:

terrorist activities
author
#3 lore raw
Quoted 9 lines · Andrew Morton
Quoted 3 lines · Christian Kujau

as the subject might presume, the parport_pc modules oopses when it's unloaded.

Can you retest on 2.6.0-test4?

sorry, for replying lately: yes, i'll do so in the evening.

Christian.

#4 lore raw
In reply to Andrew Morton
Quoted 1 line · Andrew Morton

Can you retest on 2.6.0-test4?

i have, but with no (?) changes:

prinz:~$ modprobe parport_pc
parport0: PC-style at 0x378 (0x778), irq 7, using FIFO [PCSPP,TRISTATE,COMPAT,ECP]
prinz:~$ rmmod parport_pc
Unable to handle kernel paging request at virtual address e1a4f164
 printing eip:
 e1a4757a
 *pde = 1fe2a067
 *pte = 00000000
 Oops: 0000 [#1]
 CPU:    0
 EIP:    0060:[<e1a4757a>]    Tainted: P
 EFLAGS: 00010282
 EIP is at cleanup_module+0xa/0x50 [parport_pc]
 eax: deab8980   ebx: e1a4b2c0   ecx: 00000000   edx: e1a4b2c0
 esi: 00000880   edi: 00000000   ebp: c0320318   esp: dae47f50
 ds: 007b   es: 007b   ss: 0068
 Process rmmod (pid: 618, threadinfo=dae46000 task=dcda4640)
 Stack: dae47f70 e1a4b2c0 00000880 c01346a8 e1a4b2c0 bffffce0 \
        0000003b 00000000
        70726170 5f74726f c0006370 dfd69940 db6be180 dfd60280 \
        40014000 40015000
        40015000 dfd60280 dfd69940 dfd69960 00000000 dae46000 \
        00148634 dfd69940
 Call Trace:
 [<c01346a8>] sys_delete_module+0x138/0x1b0
 [<c010929b>] syscall_call+0x7/0xb
Code: 8b 15 64 f1 a4 e1 89 c3 85 d2 74 29 85 db 74 15 8d b6 00 00

Segmentation fault
prinz:~$

(oh, yes, it's tainted again. the nvidia.ko module. but it oopsed with a non-tainted 2.6.0-test3 too: http://christian.go4more.de/parport/parport_oops.txt )

thank you for your time,
Christian.

Signature · 3 lines
BOFH excuse #161:

monitor VLF leakage
#5 lore raw
Quoted 5 lines · Christian Kujau
Quoted 1 line · Andrew Morton

Can you retest on 2.6.0-test4?

i have, but with no (?) changes:

Please send your .config.