[Cooker] kernel panic in msec_find

2002-05-06 Thread Dominique LE CAMPION 


Hello,

I have the same problem as reported a month ago on this list:
I repeatedly have kernel oops caused by msec_find (a trace follows)

I use a Mandrake 8.2 (2.4.18-6mdk kernel) on a Celeron+VIA Apollo 133 PC.
I have a NCR 53c810 PCI SCSI adapter, to which
2 hard drives and an Iomega ZIP 100 are connected.

When the kernel oops has occured, only the devices
attached to the SCSI chain are frozen:
- the processes trying to access IDE/ATA devices still work fine.
   (1 hard drive, 1 CDROM, plus 1 CD writer using IDE-SCSI emulation)
- the processes trying to access the ZIP or one of the
   SCSI drives are blocked during a system call (impossible
   to kill them)

I have tried to disable supermount, but it did not seem to
solve the problem. However, removing the ZIP drive from the SCSI chain
seems to solve it... it's not a pleasant solution though !

I did not have this kind of problem with the Mandrake 8.1 kernel or
the standard 2.4.8 kernel.

 From what I can read on the 2.4.18 changelog there has been some work
concerning removable devices:

  pre9:
  [...]
  - Really add devfs fix for removable devices:
its on pre8 changelog but not on pre8 patch(me)

Does anyone know more about this problem ?
Does the 2.4.18-6mdk kernel include this fix ?

Thanks for your help,

Dominique




Unable to handle kernel paging request at virtual address 204f2f8d
c0160783
*pde = 
Oops: 
CPU:0
EIP:0010:[c0160783]Not tainted
Using defaults from ksymoops -t elf32-i386 -a i386
EFLAGS: 00010202
eax: ccdad920   ebx: 204f2f49   ecx:    edx: ccdad920
esi: c6126040   edi: cfe828e0   ebp: c4bb1940   esp: cd58df28
ds: 0018   es: 0018   ss: 0018
Process msec_find (pid: 4955, stackpage=cd58d000)
Stack: c6126040 c0160c16 cfe828e0 c0265a40  c6126040 c61260c0 
c61260ac
c4bb1940 c0141690 c4bb1940 cd58dfa0 c0141b90 c4bb1940 fff7 
000d
bfffece8 c0141d3f c4bb1940 c0141b90 cd58dfa0 cc0c0360 c01338f7 
cc0c0360
Call Trace: [c0160c16] [c0141690] [c0141b90] [c0141d3f] 
[c0141b90]
[c01338f7] [c0106f23]
Code: 66 8b 43 44 25 00 f0 00 00 66 3d 00 60 75 0d f6 43 10 04 74


 EIP; c0160783 devfs_unregister_blkdev+193/18c0   =

 eax; ccdad920 ___strtok+cab1510/12519c50
 ebx; 204f2f49 Before first symbol
 edx; ccdad920 ___strtok+cab1510/12519c50
 esi; c6126040 ___strtok+5e29c30/12519c50
 edi; cfe828e0 ___strtok+fb864d0/12519c50
 ebp; c4bb1940 ___strtok+48b5530/12519c50
 esp; cd58df28 ___strtok+d291b18/12519c50

Trace; c0160c16 devfs_unregister_blkdev+626/18c0
Trace; c0141690 vfs_readdir+60/90
Trace; c0141b90 dcache_readdir+4d0/720
Trace; c0141d3f dcache_readdir+67f/720
Trace; c0141b90 dcache_readdir+4d0/720
Trace; c01338f7 vfs_statfs+c57/1090
Trace; c0106f23 __up_wakeup+1137/25b4

Code;  c0160783 devfs_unregister_blkdev+193/18c0
 _EIP:
Code;  c0160783 devfs_unregister_blkdev+193/18c0   =
0:   66 8b 43 44   mov0x44(%ebx),%ax   =
Code;  c0160787 devfs_unregister_blkdev+197/18c0
4:   25 00 f0 00 00and$0xf000,%eax
Code;  c016078c devfs_unregister_blkdev+19c/18c0
9:   66 3d 00 60   cmp$0x6000,%ax
Code;  c0160790 devfs_unregister_blkdev+1a0/18c0
d:   75 0d jne1c _EIP+0x1c c016079f 
devfs_unregister_blkdev+1af/18c0
Code;  c0160792 devfs_unregister_blkdev+1a2/18c0
f:   f6 43 10 04   testb  $0x4,0x10(%ebx)
Code;  c0160796 devfs_unregister_blkdev+1a6/18c0
   13:   74 00 je 15 _EIP+0x15 c0160798 
devfs_unregister_blkdev+1a8/18c0

Re: [Cooker] kernel panic in msec_find

2002-05-06 Thread Todd Lyons 

Dominique LE CAMPION wrote on Tue, May 07, 2002 at 01:38:03AM +0200 :
 
 I have the same problem as reported a month ago on this list:
 I repeatedly have kernel oops caused by msec_find (a trace follows)
 I use a Mandrake 8.2 (2.4.18-6mdk kernel) on a Celeron+VIA Apollo 133 PC.
 I have a NCR 53c810 PCI SCSI adapter, to which
 2 hard drives and an Iomega ZIP 100 are connected.

It sounds like you are working on a Mandrake 8.2 box.  This is the
Cooker mailing list.  It's great that you're providing detailed
information, but be aware that if you are not running a fully cookerized
system, your bug reports are not 100% useful since many things have
changed in Cooker just since the 8.2 release.

I'm working from memory here, but I'm pretty sure the guys had
determined that the ide-cd module was causing this.  Have you tried a
more recent kernel?  It's upt to -14mdk in Cooker now.

Blue skies...   Todd
-- 
  Todd Lyons -- MandrakeSoft, Inc.   http://www.mandrakesoft.com/
UNIX was not designed to stop you from doing stupid things, because 
  that would also stop you from doing clever things. -- Doug Gwyn
   Cooker Version mandrake-release-8.3-0.2mdk Kernel 2.4.18-13mdk



msg63798/pgp0.pgp
Description: PGP signature

RE: [Cooker] kernel panic in msec_find

2002-05-06 Thread Borsenkow Andrej 


 Dominique LE CAMPION wrote on Tue, May 07, 2002 at 01:38:03AM +0200 :
 
  I have the same problem as reported a month ago on this list:
  I repeatedly have kernel oops caused by msec_find (a trace follows)
  I use a Mandrake 8.2 (2.4.18-6mdk kernel) on a Celeron+VIA Apollo
133 PC.
  I have a NCR 53c810 PCI SCSI adapter, to which
  2 hard drives and an Iomega ZIP 100 are connected.
 
 It sounds like you are working on a Mandrake 8.2 box.  This is the
 Cooker mailing list.  It's great that you're providing detailed
 information, but be aware that if you are not running a fully
cookerized
 system, your bug reports are not 100% useful since many things have
 changed in Cooker just since the 8.2 release.
 
 I'm working from memory here, but I'm pretty sure the guys had
 determined that the ide-cd module was causing this.  Have you tried a
 more recent kernel?  It's upt to -14mdk in Cooker now.
 

Well, even if I run cooker I _am_ interested if 8.2 is going to be fixed
:-)

-andrej