[patch] move dereference after null check

[patch] move dereference after null check

[Dan Carpenter]

I moved the ops->inherits dereference below the null check.  I moved the 
other assignment as well so that they would be together.

Found by smatch static checker.

regards,
dan carpenter

Signed-off-by: Dan Carpenter <error27@gmail.com>

--- orig/drivers/ata/libata-core.c	2009-11-08 19:40:18.000000000 +0200
+++ devel/drivers/ata/libata-core.c	2009-11-08 19:42:06.000000000 +0200
@@ -5938,13 +5938,14 @@
 {
 	static DEFINE_SPINLOCK(lock);
 	const struct ata_port_operations *cur;
-	void **begin = (void **)ops;
-	void **end = (void **)&ops->inherits;
-	void **pp;
+	void **begin, **end, **pp;
 
 	if (!ops || !ops->inherits)
 		return;
 
+	begin = (void **)ops;
+	end = (void **)&ops->inherits;
+
 	spin_lock(&lock);
 
 	for (cur = ops->inherits; cur; cur = cur->inherits) {

Re: [patch] move dereference after null check

[James Bottomley]

On Tue, 2009-11-10 at 10:55 +0200, Dan Carpenter wrote:
> I moved the ops->inherits dereference below the null check.  I moved the 
> other assignment as well so that they would be together.
> 
> Found by smatch static checker.
> 
> regards,
> dan carpenter
> 
> Signed-off-by: Dan Carpenter <error27@gmail.com>
> 
> --- orig/drivers/ata/libata-core.c	2009-11-08 19:40:18.000000000 +0200
> +++ devel/drivers/ata/libata-core.c	2009-11-08 19:42:06.000000000 +0200
> @@ -5938,13 +5938,14 @@
>  {
>  	static DEFINE_SPINLOCK(lock);
>  	const struct ata_port_operations *cur;
> -	void **begin = (void **)ops;
> -	void **end = (void **)&ops->inherits;

There's no problem here: this isn't a dereference.

James

[patch] move dereference after null check

[Dan Carpenter]

The 'jack' variable was dereference before the NULL check so this patch 
moves the dereference later.

regards,
dan carpenter

Signed-off-by: Dan Carpenter <error27@gmail.com>

--- orig/sound/soc/soc-jack.c	2009-11-14 11:04:04.000000000 +0200
+++ devel/sound/soc/soc-jack.c	2009-11-14 11:05:01.000000000 +0200
@@ -58,7 +58,7 @@
  */
 void snd_soc_jack_report(struct snd_soc_jack *jack, int status, int mask)
 {
-	struct snd_soc_codec *codec = jack->card->codec;
+	struct snd_soc_codec *codec;
 	struct snd_soc_jack_pin *pin;
 	int enable;
 	int oldstatus;
@@ -67,6 +67,7 @@
 		WARN_ON_ONCE(!jack);
 		return;
 	}
+	codec = jack->card->codec;
 
 	mutex_lock(&codec->mutex);

Re: [patch] move dereference after null check

[Liam Girdwood]

On Mon, 2009-11-16 at 11:06 +0200, Dan Carpenter wrote:
> The 'jack' variable was dereference before the NULL check so this patch 
> moves the dereference later.
> 
> regards,
> dan carpenter
> 
> Signed-off-by: Dan Carpenter <error27@gmail.com>
Acked-by: Liam Girdwood <lrg@slimlogic.co.uk>

> 
> --- orig/sound/soc/soc-jack.c	2009-11-14 11:04:04.000000000 +0200
> +++ devel/sound/soc/soc-jack.c	2009-11-14 11:05:01.000000000 +0200
> @@ -58,7 +58,7 @@
>   */
>  void snd_soc_jack_report(struct snd_soc_jack *jack, int status, int mask)
>  {
> -	struct snd_soc_codec *codec = jack->card->codec;
> +	struct snd_soc_codec *codec;
>  	struct snd_soc_jack_pin *pin;
>  	int enable;
>  	int oldstatus;
> @@ -67,6 +67,7 @@
>  		WARN_ON_ONCE(!jack);
>  		return;
>  	}
> +	codec = jack->card->codec;
>  
>  	mutex_lock(&codec->mutex);
>  

Re: [patch] move dereference after null check

[Mark Brown]

On Mon, Nov 16, 2009 at 11:06:33AM +0200, Dan Carpenter wrote:
> The 'jack' variable was dereference before the NULL check so this patch 
> moves the dereference later.
> 
> regards,
> dan carpenter
> 
> Signed-off-by: Dan Carpenter <error27@gmail.com>

This does not apply to the current ASoC code, it appears that someone
else has already provided an equivalent change.  When doing work like
this it's normally best to work against -next to help avoid duplication
such as this.

Re: [patch] move dereference after null check

[Takashi Iwai]

At Mon, 16 Nov 2009 16:11:24 +0000,
Mark Brown wrote:
> 
> On Mon, Nov 16, 2009 at 11:06:33AM +0200, Dan Carpenter wrote:
> > The 'jack' variable was dereference before the NULL check so this patch 
> > moves the dereference later.
> > 
> > regards,
> > dan carpenter
> > 
> > Signed-off-by: Dan Carpenter <error27@gmail.com>
> 
> This does not apply to the current ASoC code, it appears that someone
> else has already provided an equivalent change.  When doing work like
> this it's normally best to work against -next to help avoid duplication
> such as this.

Yep.  FYI, it's the commit below:

  commit 4f066173fe8deb8874f41917e5d26ea2e0c46e3b
  Author: Julia Lawall <julia@diku.dk>
  Date:   Sat Oct 17 08:32:56 2009 +0200
    ASoC: Move dereference after NULL test


thanks,

Takashi

Re: [patch] move dereference after null check

[Dan Carpenter]

Sorry about that.

I wish I could but until I get broadband, or git gets a restartable clone 
I can't use it for kernel development.  The kernel archive is too big to 
download in one shot.

Right now I pay $20 and I can use wifi for a week but it's capped out at 
200MB.  Fortunately, I'm likely to get broadband soon.  I'll pay $156 per 
month and it's capped at 200MB during business hours and "unlimitted" on 
nights and weekends.  Woohoo!

regards,
dan carpenter

On Mon, 16 Nov 2009, Mark Brown wrote:

> On Mon, Nov 16, 2009 at 11:06:33AM +0200, Dan Carpenter wrote:
> > The 'jack' variable was dereference before the NULL check so this patch 
> > moves the dereference later.
> > 
> > regards,
> > dan carpenter
> > 
> > Signed-off-by: Dan Carpenter <error27@gmail.com>
> 
> This does not apply to the current ASoC code, it appears that someone
> else has already provided an equivalent change.  When doing work like
> this it's normally best to work against -next to help avoid duplication
> such as this.
>