Re: 2.6.17-mm4

Manuel Lauss <mano <at> roarinelk.homelinux.net> wrote:
>
> Hello,
> 
> With the attached .config, the kernel reliably panics when someone
> issues a 'sync' (or the kernel decides to write back its buffers):
> 
> reiser4 panicked cowardly: reiser4[sync(8465)]: commit_current_atom (fs/reiser4/txnmgr.c:1062)[zam-597]:
> Kernel panic - not syncing: reiser4[sync(8465)]: commit_current_atom (fs/reiser4/txnmgr.c:1062)[zam-597]:
> 
> (this is all that is printed)
> 
> This happens only with Reiser4 and libata ata_piix driver; it does not
> happen with Reiser4 and 'old' IDE piix driver. Other fs are also not
> affected. I have no idea how to track this, I hope someone else does :)
> 
> Hardware is a pentium-m laptop with ICH4 pata.
> 

Interesting, thanks.

My guess would be that there's a difference in the way in which the two
drivers handle write barriers, and the new driver has confused the reiser4
code.

Are you able to identify any earlier -mm kernel which ran OK with reiser4
and ata_piix?