linux-kernel - [PATCH v2 2/2] iio: pressure: mpl115: do not rely on structure field ordering

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <1485981657-14586-3-git-send-email-peda@axentia.se>
Date:   Wed, 1 Feb 2017 21:40:57 +0100
From:   Peter Rosin <peda@...ntia.se>
To:     <linux-kernel@...r.kernel.org>
CC:     Peter Rosin <peda@...ntia.se>, Jonathan Cameron <jic23@...nel.org>,
        Hartmut Knaack <knaack.h@....de>,
        Lars-Peter Clausen <lars@...afoo.de>,
        Peter Meerwald-Stadler <pmeerw@...erw.net>,
        Sanchayan Maity <maitysanchayan@...il.com>,
        Gregor Boirie <gregor.boirie@...rot.com>,
        Alison Schofield <amsfield22@...il.com>,
        Ken Lin <ken.lin@...antech.com>, <linux-iio@...r.kernel.org>
Subject: [PATCH v2 2/2] iio: pressure: mpl115: do not rely on structure field ordering

Fixes a regression triggered by a change in the layout of
struct iio_chan_spec, but the real bug is in the driver which assumed
a specific structure layout in the first place. Hint: the three bits were
not OR:ed together as implied by the indentation prior to this patch,
there was a comma between the first two, which accidentally moved the
..._SCALE and ..._OFFSET bits to the next structure field. That field
was .info_mask_shared_by_type before the _available attributes was added
by commit 51239600074b ("iio:core: add a callback to allow drivers to
provide _available attributes") and .info_mask_separate_available
afterwards, and the regression happened.

info_mask_shared_by_type is actually a better choice than the originally
intended info_mask_separate for the ..._SCALE and ..._OFFSET bits since
a constant is returned from mpl115_read_raw for the scale/offset. Using
info_mask_shared_by_type also preserves the behavior from before the
regression and is therefore less likely to cause other interesting side
effects.

The above mentioned regression causes unintended sysfs attibutes to
show up that are not backed by code, in turn causing a NULL pointer
defererence to happen on access.

Fixes: 3017d90e8931 ("iio: Add Freescale MPL115A2 pressure / temperature sensor driver")
Fixes: 51239600074b ("iio:core: add a callback to allow drivers to provide _available attributes")
Signed-off-by: Peter Rosin <peda@...ntia.se>
---
 drivers/iio/pressure/mpl115.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/iio/pressure/mpl115.c b/drivers/iio/pressure/mpl115.c
index 73f2f0c46e62..8f2bce213248 100644
--- a/drivers/iio/pressure/mpl115.c
+++ b/drivers/iio/pressure/mpl115.c
@@ -137,6 +137,7 @@ static const struct iio_chan_spec mpl115_channels[] = {
 	{
 		.type = IIO_TEMP,
 		.info_mask_separate = BIT(IIO_CHAN_INFO_RAW),
+		.info_mask_shared_by_type =
 			BIT(IIO_CHAN_INFO_OFFSET) | BIT(IIO_CHAN_INFO_SCALE),
 	},
 };
-- 
2.1.4