Re: Bug: slab corruption in 2.5.67-mm1

[Szonyi Calin]

Andrew Morton said:
> J Sloan <[e-mail]> wrote:
>>
>> This may be of interest -
>>
>> kernel: 2.5.67-mm1
>>
>> Linux distro: Red Hat 8.0 + updates
>>
>> Hardware:
>> Celeron 1.2 Ghz on Intel Motherboard
>> 512 MB RAM, 2x e100 ethernet
>
> whoa.  Uniprocessor.
>
>> ---- snip ----
>> Freeing unused kernel memory: 312k freed
>> EXT3 FS 2.4-0.9.16, 02 Dec 2001 on ide0(3,3), internal journal
>> Adding 514072k swap on /dev/hda2.  Priority:42 extents:1
>> kjournald starting.  Commit interval 5 seconds
>> EXT3 FS 2.4-0.9.16, 02 Dec 2001 on ide0(3,1), internal journal
>> EXT3-fs: mounted filesystem with ordered data mode.
>> Slab corruption: start=dfa2f320, expend=dfa2f97f, problemat=dfa2f328
>> Data: ********6A
>
> Yes, this means that someone ran put_task_struct() against an
> already-freed task_struct.  There's some deubg code in -mm which is
> supposed to trap this, but it obviously didn't trigger for some reason.
>
> Until someone finds a way to reproduce this we're a bit stuck.  A code
> audit may find it.

FYI
Same thing happend on my machine *once*, in the same place
(i.e mounting root fs -- JFS) but i forgot the full report on my home
machine :-(.
Kernel 2.5.67 mm1 compiled by gcc 3.2.2 and booted with "debug"
on kernel command line. Couldn't reproduce it :-(
Machine: Duron 700 MHz, 256 M RAM, chipset Via KT133, ATI Radeon VE QY
graphics card, CS4239 sound card, no ethernet.

Distribution Slackware 8.1 + updates.

Calin
-- 
# fortune
fortune: write error on /dev/null --- please empty the bit bucket

-----------------------------------------
This email was sent using SquirrelMail.
   "Webmail for nuts!"
http://squirrelmail.org/