[PATCH 7/17] arch/arm/common: Add missing spin_unlock_irqrestore

[PATCH 7/17] arch/arm/common: Add missing spin_unlock_irqrestore

[Julia Lawall]

From: Julia Lawall <julia@diku.dk>

Add a spin_unlock_irqrestore missing on the error path.  Although the lock
is destroyed with the rest of the sachip structure in the function
__sa1111_remove, it still seems useful to restore the interrupt state.

The semantic match that finds this problem is as follows:
(http://coccinelle.lip6.fr/)

// <smpl>
@@
expression E1;
@@

* spin_lock_irqsave(E1,...);
  <+... when != E1
  if (...) {
    ... when != E1
*   return ...;
  }
  ...+>
* spin_unlock_irqrestore(E1,...);
// </smpl>

Signed-off-by: Julia Lawall <julia@diku.dk>

---
Perhaps the unlock is now too early?

 arch/arm/common/sa1111.c |    1 +
 1 file changed, 1 insertion(+)

diff --git a/arch/arm/common/sa1111.c b/arch/arm/common/sa1111.c
index a52a27c..59e38ff 100644
--- a/arch/arm/common/sa1111.c
+++ b/arch/arm/common/sa1111.c
@@ -959,6 +959,7 @@ static int sa1111_resume(struct platform_device *dev)
 	 */
 	id = sa1111_readl(sachip->base + SA1111_SKID);
 	if ((id & SKID_ID_MASK) != SKID_SA1111_ID) {
+		spin_unlock_irqrestore(&sachip->lock, flags);
 		__sa1111_remove(sachip);
 		platform_set_drvdata(dev, NULL);
 		kfree(save);

[PATCH 7/17] arch/arm/common: Add missing spin_unlock_irqrestore

[Julia Lawall]

From: Julia Lawall <julia@diku.dk>

Add a spin_unlock_irqrestore missing on the error path.  Although the lock
is destroyed with the rest of the sachip structure in the function
__sa1111_remove, it still seems useful to restore the interrupt state.

The semantic match that finds this problem is as follows:
(http://coccinelle.lip6.fr/)

// <smpl>
@@
expression E1;
@@

* spin_lock_irqsave(E1,...);
  <+... when != E1
  if (...) {
    ... when != E1
*   return ...;
  }
  ...+>
* spin_unlock_irqrestore(E1,...);
// </smpl>

Signed-off-by: Julia Lawall <julia@diku.dk>

---
Perhaps the unlock is now too early?

 arch/arm/common/sa1111.c |    1 +
 1 file changed, 1 insertion(+)

diff --git a/arch/arm/common/sa1111.c b/arch/arm/common/sa1111.c
index a52a27c..59e38ff 100644
--- a/arch/arm/common/sa1111.c
+++ b/arch/arm/common/sa1111.c
@@ -959,6 +959,7 @@ static int sa1111_resume(struct platform_device *dev)
 	 */
 	id = sa1111_readl(sachip->base + SA1111_SKID);
 	if ((id & SKID_ID_MASK) != SKID_SA1111_ID) {
+		spin_unlock_irqrestore(&sachip->lock, flags);
 		__sa1111_remove(sachip);
 		platform_set_drvdata(dev, NULL);
 		kfree(save);

[PATCH 7/17] arch/arm/common: Add missing spin_unlock_irqrestore

[Julia Lawall]

From: Julia Lawall <julia@diku.dk>

Add a spin_unlock_irqrestore missing on the error path.  Although the lock
is destroyed with the rest of the sachip structure in the function
__sa1111_remove, it still seems useful to restore the interrupt state.

The semantic match that finds this problem is as follows:
(http://coccinelle.lip6.fr/)

// <smpl>
@@
expression E1;
@@

* spin_lock_irqsave(E1,...);
  <+... when != E1
  if (...) {
    ... when != E1
*   return ...;
  }
  ...+>
* spin_unlock_irqrestore(E1,...);
// </smpl>

Signed-off-by: Julia Lawall <julia@diku.dk>

---
Perhaps the unlock is now too early?

 arch/arm/common/sa1111.c |    1 +
 1 file changed, 1 insertion(+)

diff --git a/arch/arm/common/sa1111.c b/arch/arm/common/sa1111.c
index a52a27c..59e38ff 100644
--- a/arch/arm/common/sa1111.c
+++ b/arch/arm/common/sa1111.c
@@ -959,6 +959,7 @@ static int sa1111_resume(struct platform_device *dev)
 	 */
 	id = sa1111_readl(sachip->base + SA1111_SKID);
 	if ((id & SKID_ID_MASK) != SKID_SA1111_ID) {
+		spin_unlock_irqrestore(&sachip->lock, flags);
 		__sa1111_remove(sachip);
 		platform_set_drvdata(dev, NULL);
 		kfree(save);

Re: [PATCH 7/17] arch/arm/common: Add missing spin_unlock_irqrestore

[Marek Vasut]

Dne St 26. května 2010 17:56:14 Julia Lawall napsal(a):
> From: Julia Lawall <julia@diku.dk>
> 
> Add a spin_unlock_irqrestore missing on the error path.  Although the lock
> is destroyed with the rest of the sachip structure in the function
> __sa1111_remove, it still seems useful to restore the interrupt state.
> 
> The semantic match that finds this problem is as follows:
> (http://coccinelle.lip6.fr/)
> 
> // <smpl>
> @@
> expression E1;
> @@
> 
> * spin_lock_irqsave(E1,...);
>   <+... when != E1
>   if (...) {
>     ... when != E1
> *   return ...;
>   }
>   ...+>
> * spin_unlock_irqrestore(E1,...);
> // </smpl>
> 
> Signed-off-by: Julia Lawall <julia@diku.dk>
> 
> ---
> Perhaps the unlock is now too early?
> 
>  arch/arm/common/sa1111.c |    1 +
>  1 file changed, 1 insertion(+)
> 
> diff --git a/arch/arm/common/sa1111.c b/arch/arm/common/sa1111.c
> index a52a27c..59e38ff 100644
> --- a/arch/arm/common/sa1111.c
> +++ b/arch/arm/common/sa1111.c
> @@ -959,6 +959,7 @@ static int sa1111_resume(struct platform_device *dev)
>  	 */
>  	id = sa1111_readl(sachip->base + SA1111_SKID);
>  	if ((id & SKID_ID_MASK) != SKID_SA1111_ID) {
> +		spin_unlock_irqrestore(&sachip->lock, flags);
>  		__sa1111_remove(sachip);
>  		platform_set_drvdata(dev, NULL);
>  		kfree(save);

Why are "readl"s protected by spinlock anyway ? Can't we just move the locking 
past the code above ?

I'm no sa1111 expert though, Russell ?

Cheers
> 
> _______________________________________________
> linux-arm-kernel mailing list
> linux-arm-kernel@lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
--
To unsubscribe from this list: send the line "unsubscribe kernel-janitors" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH 7/17] arch/arm/common: Add missing spin_unlock_irqrestore

[Marek Vasut]

Dne St 26. května 2010 17:56:14 Julia Lawall napsal(a):
> From: Julia Lawall <julia@diku.dk>
> 
> Add a spin_unlock_irqrestore missing on the error path.  Although the lock
> is destroyed with the rest of the sachip structure in the function
> __sa1111_remove, it still seems useful to restore the interrupt state.
> 
> The semantic match that finds this problem is as follows:
> (http://coccinelle.lip6.fr/)
> 
> // <smpl>
> @@
> expression E1;
> @@
> 
> * spin_lock_irqsave(E1,...);
>   <+... when != E1
>   if (...) {
>     ... when != E1
> *   return ...;
>   }
>   ...+>
> * spin_unlock_irqrestore(E1,...);
> // </smpl>
> 
> Signed-off-by: Julia Lawall <julia@diku.dk>
> 
> ---
> Perhaps the unlock is now too early?
> 
>  arch/arm/common/sa1111.c |    1 +
>  1 file changed, 1 insertion(+)
> 
> diff --git a/arch/arm/common/sa1111.c b/arch/arm/common/sa1111.c
> index a52a27c..59e38ff 100644
> --- a/arch/arm/common/sa1111.c
> +++ b/arch/arm/common/sa1111.c
> @@ -959,6 +959,7 @@ static int sa1111_resume(struct platform_device *dev)
>  	 */
>  	id = sa1111_readl(sachip->base + SA1111_SKID);
>  	if ((id & SKID_ID_MASK) != SKID_SA1111_ID) {
> +		spin_unlock_irqrestore(&sachip->lock, flags);
>  		__sa1111_remove(sachip);
>  		platform_set_drvdata(dev, NULL);
>  		kfree(save);

Why are "readl"s protected by spinlock anyway ? Can't we just move the locking 
past the code above ?

I'm no sa1111 expert though, Russell ?

Cheers
> 
> _______________________________________________
> linux-arm-kernel mailing list
> linux-arm-kernel@lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

[PATCH 7/17] arch/arm/common: Add missing spin_unlock_irqrestore

[Marek Vasut]

Dne St 26. kv?tna 2010 17:56:14 Julia Lawall napsal(a):
> From: Julia Lawall <julia@diku.dk>
> 
> Add a spin_unlock_irqrestore missing on the error path.  Although the lock
> is destroyed with the rest of the sachip structure in the function
> __sa1111_remove, it still seems useful to restore the interrupt state.
> 
> The semantic match that finds this problem is as follows:
> (http://coccinelle.lip6.fr/)
> 
> // <smpl>
> @@
> expression E1;
> @@
> 
> * spin_lock_irqsave(E1,...);
>   <+... when != E1
>   if (...) {
>     ... when != E1
> *   return ...;
>   }
>   ...+>
> * spin_unlock_irqrestore(E1,...);
> // </smpl>
> 
> Signed-off-by: Julia Lawall <julia@diku.dk>
> 
> ---
> Perhaps the unlock is now too early?
> 
>  arch/arm/common/sa1111.c |    1 +
>  1 file changed, 1 insertion(+)
> 
> diff --git a/arch/arm/common/sa1111.c b/arch/arm/common/sa1111.c
> index a52a27c..59e38ff 100644
> --- a/arch/arm/common/sa1111.c
> +++ b/arch/arm/common/sa1111.c
> @@ -959,6 +959,7 @@ static int sa1111_resume(struct platform_device *dev)
>  	 */
>  	id = sa1111_readl(sachip->base + SA1111_SKID);
>  	if ((id & SKID_ID_MASK) != SKID_SA1111_ID) {
> +		spin_unlock_irqrestore(&sachip->lock, flags);
>  		__sa1111_remove(sachip);
>  		platform_set_drvdata(dev, NULL);
>  		kfree(save);

Why are "readl"s protected by spinlock anyway ? Can't we just move the locking 
past the code above ?

I'm no sa1111 expert though, Russell ?

Cheers
> 
> _______________________________________________
> linux-arm-kernel mailing list
> linux-arm-kernel at lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

Re: [PATCH 7/17] arch/arm/common: Add missing

[Russell King - ARM Linux]

On Wed, May 26, 2010 at 07:07:06PM +0200, Marek Vasut wrote:
> Why are "readl"s protected by spinlock anyway ? Can't we just move the locking 
> past the code above ?

Good question - and there seems to be a deadlock waiting to happen -
sa1111_wake() re-takes the same lock.

I think we should kill all the spinlock in sa1111_resume().

Re: [PATCH 7/17] arch/arm/common: Add missing spin_unlock_irqrestore

[Russell King - ARM Linux]

On Wed, May 26, 2010 at 07:07:06PM +0200, Marek Vasut wrote:
> Why are "readl"s protected by spinlock anyway ? Can't we just move the locking 
> past the code above ?

Good question - and there seems to be a deadlock waiting to happen -
sa1111_wake() re-takes the same lock.

I think we should kill all the spinlock in sa1111_resume().

[PATCH 7/17] arch/arm/common: Add missing spin_unlock_irqrestore

[Russell King - ARM Linux]

On Wed, May 26, 2010 at 07:07:06PM +0200, Marek Vasut wrote:
> Why are "readl"s protected by spinlock anyway ? Can't we just move the locking 
> past the code above ?

Good question - and there seems to be a deadlock waiting to happen -
sa1111_wake() re-takes the same lock.

I think we should kill all the spinlock in sa1111_resume().

Re: [PATCH 7/17] arch/arm/common: Add missing spin_unlock_irqrestore

[Marek Vasut]

Dne St 26. května 2010 20:43:58 Russell King - ARM Linux napsal(a):
> On Wed, May 26, 2010 at 07:07:06PM +0200, Marek Vasut wrote:
> > Why are "readl"s protected by spinlock anyway ? Can't we just move the
> > locking past the code above ?
> 
> Good question - and there seems to be a deadlock waiting to happen -
> sa1111_wake() re-takes the same lock.
> 
> I think we should kill all the spinlock in sa1111_resume().

Russell, Julia, check the patch I posted. Comments are welcome.

btw. Russell, killing the spinlock in resume might be overkill. If there was any 
reason for that being there, it was for the write ops so I'd rather keep those 
locked. Anyone with actual hardware available should look into that though.
--
To unsubscribe from this list: send the line "unsubscribe kernel-janitors" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH 7/17] arch/arm/common: Add missing spin_unlock_irqrestore

[Marek Vasut]

Dne St 26. května 2010 20:43:58 Russell King - ARM Linux napsal(a):
> On Wed, May 26, 2010 at 07:07:06PM +0200, Marek Vasut wrote:
> > Why are "readl"s protected by spinlock anyway ? Can't we just move the
> > locking past the code above ?
> 
> Good question - and there seems to be a deadlock waiting to happen -
> sa1111_wake() re-takes the same lock.
> 
> I think we should kill all the spinlock in sa1111_resume().

Russell, Julia, check the patch I posted. Comments are welcome.

btw. Russell, killing the spinlock in resume might be overkill. If there was any 
reason for that being there, it was for the write ops so I'd rather keep those 
locked. Anyone with actual hardware available should look into that though.

[PATCH 7/17] arch/arm/common: Add missing spin_unlock_irqrestore

[Marek Vasut]

Dne St 26. kv?tna 2010 20:43:58 Russell King - ARM Linux napsal(a):
> On Wed, May 26, 2010 at 07:07:06PM +0200, Marek Vasut wrote:
> > Why are "readl"s protected by spinlock anyway ? Can't we just move the
> > locking past the code above ?
> 
> Good question - and there seems to be a deadlock waiting to happen -
> sa1111_wake() re-takes the same lock.
> 
> I think we should kill all the spinlock in sa1111_resume().

Russell, Julia, check the patch I posted. Comments are welcome.

btw. Russell, killing the spinlock in resume might be overkill. If there was any 
reason for that being there, it was for the write ops so I'd rather keep those 
locked. Anyone with actual hardware available should look into that though.