[PATCH] cbus-retu-wdt: Fix bitfield access

[PATCH] cbus-retu-wdt: Fix bitfield access

[Michael Buesch]

An unsigned int pointer must not be casted to an unsigned
long pointer before use. Convert the bitfield to unsigned long
to fix this.
Also use clear_bit() in the release path.

Signed-off-by: Michael Buesch <mb@bu3sch.de>

---

Index: linux-2.6.38-rc6/drivers/cbus/retu-wdt.c
===================================================================
--- linux-2.6.38-rc6.orig/drivers/cbus/retu-wdt.c	2011-03-02 16:46:24.574676092 +0100
+++ linux-2.6.38-rc6/drivers/cbus/retu-wdt.c	2011-03-02 16:48:17.469807413 +0100
@@ -56,7 +56,7 @@ static int counter_param = RETU_WDT_MAX_
 
 struct retu_wdt_dev {
 	struct device		*dev;
-	int			users;
+	unsigned long		users;
 	struct miscdevice	retu_wdt_miscdev;
 	struct delayed_work	ping_work;
 };
@@ -161,7 +161,7 @@ static DEVICE_ATTR(counter, S_IRUGO, ret
 
 static int retu_wdt_open(struct inode *inode, struct file *file)
 {
-	if (test_and_set_bit(1, (unsigned long *)&(retu_wdt->users)))
+	if (test_and_set_bit(0, &retu_wdt->users))
 		return -EBUSY;
 
 	file->private_data = (void *)retu_wdt;
@@ -177,7 +177,7 @@ static int retu_wdt_release(struct inode
 #ifndef CONFIG_WATCHDOG_NOWAYOUT
 	retu_wdt_ping_enable(retu_wdt);
 #endif
-	wdev->users = 0;
+	clear_bit(0, &retu_wdt->users);
 
 	return 0;
 }
@@ -264,7 +264,6 @@ static int __init retu_wdt_probe(struct
 		return -ENOMEM;
 
 	wdev->dev = &pdev->dev;
-	wdev->users = 0;
 
 	ret = device_create_file(&pdev->dev, &dev_attr_period);
 	if (ret) {

-- 
Greetings, Michael.

Re: [PATCH] cbus-retu-wdt: Fix bitfield access

[Felipe Balbi]

On Wed, Mar 02, 2011 at 05:11:53PM +0100, Michael Buesch wrote:
> An unsigned int pointer must not be casted to an unsigned
> long pointer before use. Convert the bitfield to unsigned long
> to fix this.
> Also use clear_bit() in the release path.
> 
> Signed-off-by: Michael Buesch <mb@bu3sch.de>

This looks ok, I'm just wondering if the change on test_and_set_bit()
from bit 1 to bit 0 should be on this patch. I guess not.

-- 
balbi

Re: [PATCH] cbus-retu-wdt: Fix bitfield access

[Michael Büsch]

On Thu, 2011-03-03 at 11:43 +0200, Felipe Balbi wrote: 
> On Wed, Mar 02, 2011 at 05:11:53PM +0100, Michael Buesch wrote:
> > An unsigned int pointer must not be casted to an unsigned
> > long pointer before use. Convert the bitfield to unsigned long
> > to fix this.
> > Also use clear_bit() in the release path.
> > 
> > Signed-off-by: Michael Buesch <mb@bu3sch.de>
> 
> This looks ok, I'm just wondering if the change on test_and_set_bit()
> from bit 1 to bit 0 should be on this patch. I guess not.

Uh, well. If that's worth doing a separate patch, I'll do it.
It doesn't actually really matter if we use bit 0, 1 or ... for this.
I just thought it was weird to use the second bit.

-- 
Greetings Michael.

Re: [PATCH] cbus-retu-wdt: Fix bitfield access

[Felipe Balbi]

On Thu, Mar 03, 2011 at 03:21:29PM +0100, Michael Büsch wrote:
> On Thu, 2011-03-03 at 11:43 +0200, Felipe Balbi wrote: 
> > On Wed, Mar 02, 2011 at 05:11:53PM +0100, Michael Buesch wrote:
> > > An unsigned int pointer must not be casted to an unsigned
> > > long pointer before use. Convert the bitfield to unsigned long
> > > to fix this.
> > > Also use clear_bit() in the release path.
> > > 
> > > Signed-off-by: Michael Buesch <mb@bu3sch.de>
> > 
> > This looks ok, I'm just wondering if the change on test_and_set_bit()
> > from bit 1 to bit 0 should be on this patch. I guess not.
> 
> Uh, well. If that's worth doing a separate patch, I'll do it.
> It doesn't actually really matter if we use bit 0, 1 or ... for this.
> I just thought it was weird to use the second bit.

yeah, I didn't read the whole code before making the comment, if you're
sure no-one is using bit 0, then go for it. Having it separate, would
help on bisecting should things go wrong.

-- 
balbi
--
To unsubscribe from this list: send the line "unsubscribe linux-omap" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH] cbus-retu-wdt: Fix bitfield access

[Michael Büsch]

On Thu, 2011-03-03 at 16:27 +0200, Felipe Balbi wrote: 
> On Thu, Mar 03, 2011 at 03:21:29PM +0100, Michael Büsch wrote:
> > On Thu, 2011-03-03 at 11:43 +0200, Felipe Balbi wrote: 
> > > On Wed, Mar 02, 2011 at 05:11:53PM +0100, Michael Buesch wrote:
> > > > An unsigned int pointer must not be casted to an unsigned
> > > > long pointer before use. Convert the bitfield to unsigned long
> > > > to fix this.
> > > > Also use clear_bit() in the release path.
> > > > 
> > > > Signed-off-by: Michael Buesch <mb@bu3sch.de>
> > > 
> > > This looks ok, I'm just wondering if the change on test_and_set_bit()
> > > from bit 1 to bit 0 should be on this patch. I guess not.
> > 
> > Uh, well. If that's worth doing a separate patch, I'll do it.
> > It doesn't actually really matter if we use bit 0, 1 or ... for this.
> > I just thought it was weird to use the second bit.
> 
> yeah, I didn't read the whole code before making the comment, if you're
> sure no-one is using bit 0, then go for it. Having it separate, would
> help on bisecting should things go wrong.

All users of the bitfield (two) are visible in the patch ;)

-- 
Greetings Michael.

--
To unsubscribe from this list: send the line "unsubscribe linux-omap" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html