linux-kernel - Re: [PATCH] touchscreen: ads7846: please don't touch free'd memory

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]

Date:	Tue, 25 May 2010 12:52:12 -0700
From:	Kevin Hilman <khilman@...prootsystems.com>
To:	Dmitry Torokhov <dmitry.torokhov@...il.com>
Cc:	linux-input@...r.kernel.org, linux-omap@...r.kernel.org,
	Michael Roth <mroth@...sie.de>, Pavel Machek <pavel@....cz>,
	Andrew Morton <akpm@...ux-foundation.org>,
	Mike Frysinger <vapier@...too.org>,
	linux-kernel@...r.kernel.org
Subject: Re: [PATCH] touchscreen: ads7846: please don't touch free'd memory

Dmitry Torokhov <dmitry.torokhov@...il.com> writes:

> On Tue, May 18, 2010 at 04:46:53PM -0700, Kevin Hilman wrote:
>> If the _probe() method fails, the 'ts' struct is freed, yet it is
>> still used as the drvdata passed to suspend/resume/remove methods.
>> Even though the input device does not get registerd, the driver's
>> suspend/resume methods still get called as it's a registered SPI
>> device.  This patch adds sanity checks to these methods to ensure that
>> drvdata is valid before using it.
>> 
>
> This is a load of crap. If probe() fails that means that driver does not
> manage the device and thus ads7846_suspend() and ads7846_resume() should
> not be called _at all_. If SPI core manages to call random methods from
> the drivers that failed to bind to a device that should be fixed in SPI
> core.

Agreed, this is indeed a bug in the SPI driver core.

However, by fixing the SPI core to unregister the driver on failure
(patch below), we prevent the suspend & resume methods from being
called, but the driver's ->remove() method will still be called due to
the driver_unregister().  So at least the remove() method needs fixing
to prevent accessing free'd memory.

Unless there are objections, I'll post the below along with an updated
version of ads7846 patch.

Kevin

diff --git a/drivers/spi/spi.c b/drivers/spi/spi.c
index b3a1f92..42d4d26 100644
--- a/drivers/spi/spi.c
+++ b/drivers/spi/spi.c
@@ -175,6 +175,8 @@ static void spi_drv_shutdown(struct device *dev)
  */
 int spi_register_driver(struct spi_driver *sdrv)
 {
+	int ret;
+
 	sdrv->driver.bus = &spi_bus_type;
 	if (sdrv->probe)
 		sdrv->driver.probe = spi_drv_probe;
@@ -182,7 +184,12 @@ int spi_register_driver(struct spi_driver *sdrv)
 		sdrv->driver.remove = spi_drv_remove;
 	if (sdrv->shutdown)
 		sdrv->driver.shutdown = spi_drv_shutdown;
-	return driver_register(&sdrv->driver);
+
+	ret = driver_register(&sdrv->driver);
+	if (!ret)
+		driver_unregister(&sdrv->driver);
+
+	return ret;
 }
 EXPORT_SYMBOL_GPL(spi_register_driver);
 

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/