[PATCH] mm: vmscan: add BUG_ON on illegal return values from scan_objects

[PATCH] mm: vmscan: add BUG_ON on illegal return values from scan_objects

[Oskar Andero]

Add a BUG_ON to catch any illegal value from the shrinkers. This fixes a
potential bug if scan_objects returns a negative other than -1, which
would lead to undefined behaviour.

Cc: Glauber Costa <glommer@openvz.org>
Cc: Dave Chinner <dchinner@redhat.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Hugh Dickins <hughd@google.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Oskar Andero <oskar.andero@sonymobile.com>
---
 mm/vmscan.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/mm/vmscan.c b/mm/vmscan.c
index 6bac41e..fbe6742 100644
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -293,6 +293,7 @@ shrink_slab_one(struct shrinker *shrinker, struct shrink_control *shrinkctl,
 		ret = shrinker->scan_objects(shrinker, shrinkctl);
 		if (ret == -1)
 			break;
+		BUG_ON(ret < -1);
 		freed += ret;
 
 		count_vm_events(SLABS_SCANNED, nr_to_scan);
-- 
1.8.1.5

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

[PATCH] mm: vmscan: add BUG_ON on illegal return values from scan_objects

[Oskar Andero]

Add a BUG_ON to catch any illegal value from the shrinkers. This fixes a
potential bug if scan_objects returns a negative other than -1, which
would lead to undefined behaviour.

Cc: Glauber Costa <glommer@openvz.org>
Cc: Dave Chinner <dchinner@redhat.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Hugh Dickins <hughd@google.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Oskar Andero <oskar.andero@sonymobile.com>
---
 mm/vmscan.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/mm/vmscan.c b/mm/vmscan.c
index 6bac41e..fbe6742 100644
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -293,6 +293,7 @@ shrink_slab_one(struct shrinker *shrinker, struct shrink_control *shrinkctl,
 		ret = shrinker->scan_objects(shrinker, shrinkctl);
 		if (ret == -1)
 			break;
+		BUG_ON(ret < -1);
 		freed += ret;
 
 		count_vm_events(SLABS_SCANNED, nr_to_scan);
-- 
1.8.1.5

Re: [PATCH] mm: vmscan: add BUG_ON on illegal return values from scan_objects

[Greg Kroah-Hartman]

On Mon, May 20, 2013 at 11:14:27AM +0200, Oskar Andero wrote:
> Add a BUG_ON to catch any illegal value from the shrinkers. This fixes a
> potential bug if scan_objects returns a negative other than -1, which
> would lead to undefined behaviour.

So it's better to crash a machine and keep anyone from using it?
Instead of recovering from an error you found?  Not good.

greg k-h

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: [PATCH] mm: vmscan: add BUG_ON on illegal return values from scan_objects

[Greg Kroah-Hartman]

On Mon, May 20, 2013 at 11:14:27AM +0200, Oskar Andero wrote:
> Add a BUG_ON to catch any illegal value from the shrinkers. This fixes a
> potential bug if scan_objects returns a negative other than -1, which
> would lead to undefined behaviour.

So it's better to crash a machine and keep anyone from using it?
Instead of recovering from an error you found?  Not good.

greg k-h

Re: [PATCH] mm: vmscan: add BUG_ON on illegal return values from scan_objects

[David Rientjes]

On Mon, 20 May 2013, Oskar Andero wrote:

> Add a BUG_ON to catch any illegal value from the shrinkers. This fixes a
> potential bug if scan_objects returns a negative other than -1, which
> would lead to undefined behaviour.
> 
> Cc: Glauber Costa <glommer@openvz.org>
> Cc: Dave Chinner <dchinner@redhat.com>
> Cc: Andrew Morton <akpm@linux-foundation.org>
> Cc: Hugh Dickins <hughd@google.com>
> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
> Signed-off-by: Oskar Andero <oskar.andero@sonymobile.com>
> ---
>  mm/vmscan.c | 1 +
>  1 file changed, 1 insertion(+)
> 
> diff --git a/mm/vmscan.c b/mm/vmscan.c
> index 6bac41e..fbe6742 100644
> --- a/mm/vmscan.c
> +++ b/mm/vmscan.c
> @@ -293,6 +293,7 @@ shrink_slab_one(struct shrinker *shrinker, struct shrink_control *shrinkctl,
>  		ret = shrinker->scan_objects(shrinker, shrinkctl);
>  		if (ret == -1)
>  			break;
> +		BUG_ON(ret < -1);
>  		freed += ret;
>  
>  		count_vm_events(SLABS_SCANNED, nr_to_scan);

Nack, this doesn't fix anything.  I can see the intention, and for that it 
might make sense to turn this into VM_BUG_ON() so that anybody debugging 
an issue related to this with CONFIG_DEBUG_VM would get the indication, 
but I don't think we need to enforce the API with BUG_ON().

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: [PATCH] mm: vmscan: add BUG_ON on illegal return values from scan_objects

[David Rientjes]

On Mon, 20 May 2013, Oskar Andero wrote:

> Add a BUG_ON to catch any illegal value from the shrinkers. This fixes a
> potential bug if scan_objects returns a negative other than -1, which
> would lead to undefined behaviour.
> 
> Cc: Glauber Costa <glommer@openvz.org>
> Cc: Dave Chinner <dchinner@redhat.com>
> Cc: Andrew Morton <akpm@linux-foundation.org>
> Cc: Hugh Dickins <hughd@google.com>
> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
> Signed-off-by: Oskar Andero <oskar.andero@sonymobile.com>
> ---
>  mm/vmscan.c | 1 +
>  1 file changed, 1 insertion(+)
> 
> diff --git a/mm/vmscan.c b/mm/vmscan.c
> index 6bac41e..fbe6742 100644
> --- a/mm/vmscan.c
> +++ b/mm/vmscan.c
> @@ -293,6 +293,7 @@ shrink_slab_one(struct shrinker *shrinker, struct shrink_control *shrinkctl,
>  		ret = shrinker->scan_objects(shrinker, shrinkctl);
>  		if (ret == -1)
>  			break;
> +		BUG_ON(ret < -1);
>  		freed += ret;
>  
>  		count_vm_events(SLABS_SCANNED, nr_to_scan);

Nack, this doesn't fix anything.  I can see the intention, and for that it 
might make sense to turn this into VM_BUG_ON() so that anybody debugging 
an issue related to this with CONFIG_DEBUG_VM would get the indication, 
but I don't think we need to enforce the API with BUG_ON().

Re: [PATCH] mm: vmscan: add BUG_ON on illegal return values from scan_objects

[Dave Chinner]

On Mon, May 20, 2013 at 11:14:27AM +0200, Oskar Andero wrote:
> Add a BUG_ON to catch any illegal value from the shrinkers. This fixes a
> potential bug if scan_objects returns a negative other than -1, which
> would lead to undefined behaviour.
> 
> Cc: Glauber Costa <glommer@openvz.org>
> Cc: Dave Chinner <dchinner@redhat.com>
> Cc: Andrew Morton <akpm@linux-foundation.org>
> Cc: Hugh Dickins <hughd@google.com>
> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
> Signed-off-by: Oskar Andero <oskar.andero@sonymobile.com>
> ---
>  mm/vmscan.c | 1 +
>  1 file changed, 1 insertion(+)
> 
> diff --git a/mm/vmscan.c b/mm/vmscan.c
> index 6bac41e..fbe6742 100644
> --- a/mm/vmscan.c
> +++ b/mm/vmscan.c
> @@ -293,6 +293,7 @@ shrink_slab_one(struct shrinker *shrinker, struct shrink_control *shrinkctl,
>  		ret = shrinker->scan_objects(shrinker, shrinkctl);
>  		if (ret == -1)
>  			break;
> +		BUG_ON(ret < -1);
>  		freed += ret;
>  
>  		count_vm_events(SLABS_SCANNED, nr_to_scan);

NACK. we've got to fix the damn shrinkers first.

If you want this sort of guard added to the patchset Glauber and I
are working on that does this, then discuss it in the context of
that patch set.

Even if you do, you'll get the same answer: we need to first all the
busted shrinkers before we even consider being nasty about enforcing
the API constraints to prevent furture breakage.

If you want to do something useful, look at all the comments about
broken shrinkers in Glauber's patch set and work with the owners of
the code to understand what they really need and get them fixed.

-Dave.
-- 
Dave Chinner
dchinner@redhat.com

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: [PATCH] mm: vmscan: add BUG_ON on illegal return values from scan_objects

[Dave Chinner]

On Mon, May 20, 2013 at 11:14:27AM +0200, Oskar Andero wrote:
> Add a BUG_ON to catch any illegal value from the shrinkers. This fixes a
> potential bug if scan_objects returns a negative other than -1, which
> would lead to undefined behaviour.
> 
> Cc: Glauber Costa <glommer@openvz.org>
> Cc: Dave Chinner <dchinner@redhat.com>
> Cc: Andrew Morton <akpm@linux-foundation.org>
> Cc: Hugh Dickins <hughd@google.com>
> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
> Signed-off-by: Oskar Andero <oskar.andero@sonymobile.com>
> ---
>  mm/vmscan.c | 1 +
>  1 file changed, 1 insertion(+)
> 
> diff --git a/mm/vmscan.c b/mm/vmscan.c
> index 6bac41e..fbe6742 100644
> --- a/mm/vmscan.c
> +++ b/mm/vmscan.c
> @@ -293,6 +293,7 @@ shrink_slab_one(struct shrinker *shrinker, struct shrink_control *shrinkctl,
>  		ret = shrinker->scan_objects(shrinker, shrinkctl);
>  		if (ret == -1)
>  			break;
> +		BUG_ON(ret < -1);
>  		freed += ret;
>  
>  		count_vm_events(SLABS_SCANNED, nr_to_scan);

NACK. we've got to fix the damn shrinkers first.

If you want this sort of guard added to the patchset Glauber and I
are working on that does this, then discuss it in the context of
that patch set.

Even if you do, you'll get the same answer: we need to first all the
busted shrinkers before we even consider being nasty about enforcing
the API constraints to prevent furture breakage.

If you want to do something useful, look at all the comments about
broken shrinkers in Glauber's patch set and work with the owners of
the code to understand what they really need and get them fixed.

-Dave.
-- 
Dave Chinner
dchinner@redhat.com