Re: 2.6.11-rc3-mm1 bad scheduling while atomic + lockup

[Jens Axboe]

On Tue, Feb 08 2005, Andrew Morton wrote:
> Jens Axboe <axboe@xxxxxxx> wrote:
> >
> >  > 	        ->axboe!
> > 
> >  :-)
> > 
> >  The thing wants a rewrite. Ideally the serializing point would be a
> >  special request. The patch is still better than nothing right now, it's
> >  really easy to hang the device with hdparm in -linus since it's
> >  impossible to guess when it is safe to issue tuning actions from user
> >  space.
> 
> I'm not sure which is worse, really.  I've never hung an interface with
> hdparm, nor seen any reports of it.  Making I/O errors deadly rather hurts.

I've gotten several reports of it, try to tune the drive settings with
any kind of drive activity and it will barf.

>  Will it happen on all I/O errors, or was this a special case?

Not sure, at least the crc stuff will trigger it. If you want you can
deactivate the patch for now, I'll get it fixed properly.

-- 
Jens Axboe

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/