Re: linux-next: Tree for May 22

[Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>]

Hi,

On Friday 23 May 2008, Rafael J. Wysocki wrote:
> On Thursday, 22 of May 2008, Stephen Rothwell wrote:
> > Hi all,
> > 
> > Changes since next-20080522:
> > 
> > The three patches reverted from the driver-core tree yesterday have been
> > removed from there.
> > 
> > The net tree (actually the wireless tree) had a trivial conflict with the
> > wireless-current tree.
> > 
> > One of the sound tree build patches has been merged.
> > 
> > The semaphore-removal tree still needed three patches reverted because similar
> > patches were merged into upstream and the scsi-rc-fixes trees.
> > 
> > The bkl-removal tree had a conflict with the arm tree because the arm
> > tree removed a file that was changed.
> > 
> > I have applied the following temporary patch for known build problems:
> > 
> > 	"Fix various 8390 builds" - the net tree broke builds on various
> > architectures - hopefully this patch will go into the net tree shortly.
> > 	"build failure with netfilter on __nf_ct_helper_find()" - breakage
> > due to the moving of stuff from list.h to rculist.h
> 
> This one and the previous linux-next can't resume from suspend to RAM on my
> HP nx6325.  The problem is reliably reproducible 100% of the time.
> 
> The patch that causes it to happen, as identified by bisection, is:
> 
> commit 3dcefa9419f3b8a49921d2218c93a3ddba8e0855
> Author: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>
> Date:   Thu May 22 10:40:19 2008 +1000
> 
>     ide: use __generic_unplug_device() in ide_do_drive_cmd()
> 
>     Cc: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
>     Cc: Borislav Petkov <petkovbb@gmail.com>
>     Cc: Jens Axboe <jens.axboe@oracle.com>
>     Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>

Thanks for narrowing it down.

> Reverting it along with the next IDE patches (up to and including
> commit fd42e6df0fd163411f5c655b4bde4fb6a8910edc
> "ide: remove action argument in ide_do_drive_cmd") makes the problem go away.

Please check if the patch below (against IDE tree but should also apply
to the current linux-next) fixes the issue (for blk_pm_resume_request()
the queue is stopped so we need to call ->request_fn explicitly).

---
 block/blk-exec.c     |    3 +++
 drivers/ide/ide-io.c |    3 +++
 2 files changed, 6 insertions(+)

Index: b/block/blk-exec.c
===================================================================
--- a/block/blk-exec.c
+++ b/block/blk-exec.c
@@ -57,6 +57,9 @@ void blk_execute_rq_nowait(struct reques
 	spin_lock_irq(q->queue_lock);
 	__elv_add_request(q, rq, where, 1);
 	__generic_unplug_device(q);
+	/* the queue is stopped so it won't be plugged+unplugged */
+	if (blk_pm_resume_request(rq))
+		q->request_fn(q);
 	spin_unlock_irq(q->queue_lock);
 }
 EXPORT_SYMBOL_GPL(blk_execute_rq_nowait);
Index: b/drivers/ide/ide-io.c
===================================================================
--- a/drivers/ide/ide-io.c
+++ b/drivers/ide/ide-io.c
@@ -1540,6 +1540,9 @@ void ide_do_drive_cmd(ide_drive_t *drive
 	hwgroup->rq = NULL;
 	__elv_add_request(drive->queue, rq, ELEVATOR_INSERT_FRONT, 1);
 	__generic_unplug_device(drive->queue);
+	/* the queue is stopped so it won't be plugged+unplugged */
+	if (blk_pm_resume_request(rq))
+		do_ide_request(drive->queue);
 	spin_unlock_irqrestore(&ide_lock, flags);
 }