Re: pm8001: locking in mpi_sata_completion() is broken

[Dan Carpenter <dan.carpenter@oracle.com>]

[-- Attachment #1: Type: text/plain, Size: 945 bytes --]

On Thu, Dec 15, 2011 at 09:26:46AM +0800, Jack Wang wrote:
> >   2018                          mb();/*in order to force CPU ordering*/
> >   2019                          spin_unlock_irqrestore(&pm8001_ha->lock,
> > flags);
> > 
> > ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> > Calling irqrestore before we've done an irqsave.  What are we trying to
> > restore?
> [Jack Wang] 
> We hold spin_lock_irqsave when we process interrupt, so here we need restore
> it to do task done clean.


You're not restoring a previous saved state, flags is set to zero
here.  You're calling:
				spin_unlock_irqrestore(&pm8001_ha->lock, 0);

The whole function is badly broken.  mpi_sata_event() has similar
problems.

With spin_lock_irqsave() and spin_unlock_irqrestore() you first have
to save the state to "flags" before you can restore it.  Doing it
the other way round doesn't make any sort sense.

regards,
dan carpenter

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 836 bytes --]