Re: Please add 98fb1ffd8154890d7051750e61ff5548c3ee2ab2

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Mon, Sep 28, 2015 at 04:12:55PM -0500, Kevin Cernekee wrote:
> On Mon, Sep 28, 2015 at 4:03 PM, Richard Weinberger <richard@xxxxxx> wrote:
> > CC'in patch authors.
> >
> > Am 28.09.2015 um 15:29 schrieb Luis Henriques:
> >> On Tue, Sep 15, 2015 at 04:48:07PM +0200, Richard Weinberger wrote:
> >>> Hi!
> >>>
> >>> Can you please add this commit to stable?
> >>> commit 98fb1ffd8154890d7051750e61ff5548c3ee2ab2
> >>> Author: Kevin Cernekee <cernekee@xxxxxxxxxxxx>
> >>> Date:   Wed Apr 22 09:30:53 2015 -0300
> >>>
> >>>     UBI: block: Add missing cache flushes
> >>>
> >>> The fix affects all kernels >= v3.15.
> >>>
> >>
> >> I was trying to apply this commit to 3.16, but it requires a backport.
> >> Could you please confirm the following backport is correct?  It can
> >> probably be applied to other stable kernels, but I haven't check.
> >>
> >> Cheers,
> >> --
> >> Luís
> >>
> >> From e781e504616d2652e9fc11f6536dd8be221298f9 Mon Sep 17 00:00:00 2001
> >> From: Kevin Cernekee <cernekee@xxxxxxxxxxxx>
> >> Date: Wed, 22 Apr 2015 09:30:53 -0300
> >> Subject: [PATCH] UBI: block: Add missing cache flushes
> >>
> >> commit 98fb1ffd8154890d7051750e61ff5548c3ee2ab2 upstream.
> >>
> >> Block drivers are responsible for calling flush_dcache_page() on each
> >> BIO request. This operation keeps the I$ coherent with the D$ on
> >> architectures that don't have hardware coherency support. Without this
> >> flush, random crashes are seen when executing user programs from an ext4
> >> filesystem backed by a ubiblock device.
> >>
> >> This patch is based on the change implemented in commit 2d4dc890b5c8
> >> ("block: add helpers to run flush_dcache_page() against a bio and a
> >> request's pages").
> >>
> >> Fixes: 9d54c8a33eec ("UBI: R/O block driver on top of UBI volumes")
> >> Signed-off-by: Kevin Cernekee <cernekee@xxxxxxxxxxxx>
> >> Signed-off-by: Ezequiel Garcia <ezequiel.garcia@xxxxxxxxxx>
> >> Signed-off-by: Richard Weinberger <richard@xxxxxx>
> >> [ luis: backported to 3.16: adjusted context ]
> >> Signed-off-by: Luis Henriques <luis.henriques@xxxxxxxxxxxxx>
> >> ---
> >>  drivers/mtd/ubi/block.c | 1 +
> >>  1 file changed, 1 insertion(+)
> >>
> >> diff --git a/drivers/mtd/ubi/block.c b/drivers/mtd/ubi/block.c
> >> index 8457df7ec5af..d3fbda7eae98 100644
> >> --- a/drivers/mtd/ubi/block.c
> >> +++ b/drivers/mtd/ubi/block.c
> >> @@ -275,6 +275,7 @@ static void ubiblock_do_work(struct work_struct *work)
> >>               spin_unlock_irq(rq->queue_lock);
> >>               res = do_ubiblock_request(dev, req);
> >>               spin_lock_irq(rq->queue_lock);
> >> +             rq_flush_dcache_pages(req);
> >>
> >>               /*
> >>                * If we're done with this request,
> >>
> 
> In our project's 3.14 tree I added the flush in a slightly different
> place, but it looks like your patch would work too:
> 
> https://chromium-review.googlesource.com/#/c/251693/2/drivers/mtd/ubi/block.c

Thank you, Kevin.  I'll just use your 3.14 backport as it doesn't seem
likely that this operation actually requires the rq->queue_lock (and it's
already tested!).

Just for completeness, I'm attaching the backport below.

Cheers,
--
Luís

>From 41e93187e983724944d9c70ddd822cada09004c8 Mon Sep 17 00:00:00 2001
From: Kevin Cernekee <cernekee@xxxxxxxxxxxx>
Date: Wed, 22 Apr 2015 09:30:53 -0300
Subject: [PATCH] UBI: block: Add missing cache flushes

commit 98fb1ffd8154890d7051750e61ff5548c3ee2ab2 upstream.

Block drivers are responsible for calling flush_dcache_page() on each
BIO request. This operation keeps the I$ coherent with the D$ on
architectures that don't have hardware coherency support. Without this
flush, random crashes are seen when executing user programs from an ext4
filesystem backed by a ubiblock device.

This patch is based on the change implemented in commit 2d4dc890b5c8
("block: add helpers to run flush_dcache_page() against a bio and a
request's pages").

Fixes: 9d54c8a33eec ("UBI: R/O block driver on top of UBI volumes")
Signed-off-by: Kevin Cernekee <cernekee@xxxxxxxxxxxx>
Signed-off-by: Ezequiel Garcia <ezequiel.garcia@xxxxxxxxxx>
Signed-off-by: Richard Weinberger <richard@xxxxxx>
[ luis: backported to 3.16: used Kevin's backport to 3.14 ]
Signed-off-by: Luis Henriques <luis.henriques@xxxxxxxxxxxxx>
---
 drivers/mtd/ubi/block.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/mtd/ubi/block.c b/drivers/mtd/ubi/block.c
index 8457df7ec5af..8b4a182d7cf9 100644
--- a/drivers/mtd/ubi/block.c
+++ b/drivers/mtd/ubi/block.c
@@ -255,6 +255,7 @@ static int do_ubiblock_request(struct ubiblock *dev, struct request *req)
 	mutex_lock(&dev->dev_mutex);
 	ret = ubiblock_read(dev, bio_data(req->bio), sec, len);
 	mutex_unlock(&dev->dev_mutex);
+	rq_flush_dcache_pages(req);
 
 	return ret;
 }
--
To unsubscribe from this list: send the line "unsubscribe stable" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html



[Index of Archives]     [Linux Kernel]     [Kernel Development Newbies]     [Linux USB Devel]     [Video for Linux]     [Linux Audio Users]     [Yosemite Hiking]     [Linux Kernel]     [Linux SCSI]