Ooops, you're right, ide_kill_rq() sets hwif->rq (and rq in turn) to NULL.
Please disregard this patch.
Vasily
On Mon, Jan 10, 2011 at 03:25:47PM +0300, Sergei Shtylyov wrote:
> Hello.
>
> On 10-01-2011 2:51, Vasily Tarasov wrote:
>
> >From: Vasily Tarasov<sectorsize512@xxxxxxxxx>
>
> >No need to check for the rq pointer to be NULL before calling
> >blk_fetch_request(). First we call it, then we test the pointer that it
> >returns.
>
> >Signed-off-by: Vasily Tarasov<sectorsize512@xxxxxxxxx>
>
> I'm not feeling sure about this patch...
>
> >--- linux-2.6.37/drivers/ide/ide-io.c.orig 2011-01-04 19:50:19.000000000 -0500
> >+++ linux-2.6.37/drivers/ide/ide-io.c 2011-01-09 18:15:44.000000000 -0500
> >@@ -438,7 +438,7 @@ void do_ide_request(struct request_queue
> > ide_drive_t *drive = q->queuedata;
> > ide_hwif_t *hwif = drive->hwif;
> > struct ide_host *host = hwif->host;
> >- struct request *rq = NULL;
> >+ struct request *rq;
> > ide_startstop_t startstop;
> >
> > spin_unlock_irq(q->queue_lock);
> >@@ -485,16 +485,13 @@ repeat:
> >
> > spin_unlock_irq(&hwif->lock);
> > spin_lock_irq(q->queue_lock);
> >+ rq = blk_fetch_request(drive->queue);
> >+ spin_unlock_irq(q->queue_lock);
> >+ spin_lock_irq(&hwif->lock);
> > /*
> > * we know that the queue isn't empty, but this can happen
> > * if the q->prep_rq_fn() decides to kill a request
> > */
> >- if (!rq)
> >- rq = blk_fetch_request(drive->queue);
>
> Did you consider the scenario where the 'goto' is taken at the
> end of this *if* branch? 'rq' is non-NULL then...
>
> >-
> >- spin_unlock_irq(q->queue_lock);
> >- spin_lock_irq(&hwif->lock);
> >-
> > if (!rq) {
> > ide_unlock_port(hwif);
> > goto out;
>
> WBR, Serge
--
To unsubscribe from this list: send the line "unsubscribe linux-ide" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
