[PATCH] spi: core: Fix Oops in spi_pump_messages error path

[PATCH] spi: core: Fix Oops in spi_pump_messages error path

[Maxime Ripard]

When the generic implementation of the transfer_one_message callback was called
by the spi_pump_messages function, if that transfer was to fail, the
spi_finalize_current_message was called twice, once in
spi_transfer_one_message, and one in spi_pump_messages.

This was causing a null pointer dereference in the second call, because the
first one set the ->cur_msg field to NULL.

Since the SPI framework expect the transfer_one_message callback to call
spi_finalize_current_message, we can remove it from spi_pump_messages, together
with any dereference of the ->cur_msg pointer.

Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
Cc: stable@vger.kernel.org
---
 drivers/spi/spi.c | 2 --
 1 file changed, 2 deletions(-)

diff --git a/drivers/spi/spi.c b/drivers/spi/spi.c
index 23756b0..39f12be 100644
--- a/drivers/spi/spi.c
+++ b/drivers/spi/spi.c
@@ -756,8 +756,6 @@ static void spi_pump_messages(struct kthread_work *work)
 	if (ret) {
 		dev_err(&master->dev,
 			"failed to transfer one message from queue: %d\n", ret);
-		master->cur_msg->status = ret;
-		spi_finalize_current_message(master);
 		return;
 	}
 }
-- 
1.8.4.2

Re: [PATCH] spi: core: Fix Oops in spi_pump_messages error path

[Geert Uytterhoeven]

On Mon, Feb 17, 2014 at 6:20 PM, Maxime Ripard
<maxime.ripard@free-electrons.com> wrote:
> When the generic implementation of the transfer_one_message callback was called
> by the spi_pump_messages function, if that transfer was to fail, the
> spi_finalize_current_message was called twice, once in
> spi_transfer_one_message, and one in spi_pump_messages.
>
> This was causing a null pointer dereference in the second call, because the
> first one set the ->cur_msg field to NULL.
>
> Since the SPI framework expect the transfer_one_message callback to call
> spi_finalize_current_message, we can remove it from spi_pump_messages, together
> with any dereference of the ->cur_msg pointer.
>
> Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
> Cc: stable@vger.kernel.org

Already fixed in v3.14-rc3 in 1f802f8249a0da536877842c43c7204064c4de8b
("spi: Fix crash with double message finalisation on error handling").

There's no need to inform stable, as the problem was introduced in v3.14-rc1.

Gr{oetje,eeting}s,

                        Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

Re: [PATCH] spi: core: Fix Oops in spi_pump_messages error path

[Maxime Ripard]

[-- Attachment #1: Type: text/plain, Size: 1401 bytes --]

Hi Geert,

On Mon, Feb 17, 2014 at 07:02:09PM +0100, Geert Uytterhoeven wrote:
> On Mon, Feb 17, 2014 at 6:20 PM, Maxime Ripard
> <maxime.ripard-wi1+55ScJUtKEb57/3fJTNBPR1lH4CV8@public.gmane.org> wrote:
> > When the generic implementation of the transfer_one_message callback was called
> > by the spi_pump_messages function, if that transfer was to fail, the
> > spi_finalize_current_message was called twice, once in
> > spi_transfer_one_message, and one in spi_pump_messages.
> >
> > This was causing a null pointer dereference in the second call, because the
> > first one set the ->cur_msg field to NULL.
> >
> > Since the SPI framework expect the transfer_one_message callback to call
> > spi_finalize_current_message, we can remove it from spi_pump_messages, together
> > with any dereference of the ->cur_msg pointer.
> >
> > Signed-off-by: Maxime Ripard <maxime.ripard-wi1+55ScJUtKEb57/3fJTNBPR1lH4CV8@public.gmane.org>
> > Cc: stable-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
> 
> Already fixed in v3.14-rc3 in 1f802f8249a0da536877842c43c7204064c4de8b
> ("spi: Fix crash with double message finalisation on error handling").
> 
> There's no need to inform stable, as the problem was introduced in v3.14-rc1.

Oops, totally missed that. Thanks!

Maxime

-- 
Maxime Ripard, Free Electrons
Embedded Linux, Kernel and Android engineering
http://free-electrons.com

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

Re: [PATCH] spi: core: Fix Oops in spi_pump_messages error path

[Maxime Ripard]

[-- Attachment #1: Type: text/plain, Size: 1315 bytes --]

Hi Geert,

On Mon, Feb 17, 2014 at 07:02:09PM +0100, Geert Uytterhoeven wrote:
> On Mon, Feb 17, 2014 at 6:20 PM, Maxime Ripard
> <maxime.ripard@free-electrons.com> wrote:
> > When the generic implementation of the transfer_one_message callback was called
> > by the spi_pump_messages function, if that transfer was to fail, the
> > spi_finalize_current_message was called twice, once in
> > spi_transfer_one_message, and one in spi_pump_messages.
> >
> > This was causing a null pointer dereference in the second call, because the
> > first one set the ->cur_msg field to NULL.
> >
> > Since the SPI framework expect the transfer_one_message callback to call
> > spi_finalize_current_message, we can remove it from spi_pump_messages, together
> > with any dereference of the ->cur_msg pointer.
> >
> > Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
> > Cc: stable@vger.kernel.org
> 
> Already fixed in v3.14-rc3 in 1f802f8249a0da536877842c43c7204064c4de8b
> ("spi: Fix crash with double message finalisation on error handling").
> 
> There's no need to inform stable, as the problem was introduced in v3.14-rc1.

Oops, totally missed that. Thanks!

Maxime

-- 
Maxime Ripard, Free Electrons
Embedded Linux, Kernel and Android engineering
http://free-electrons.com

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