[patch] afs: potential null dereference

[patch] afs: potential null dereference

[Dan Carpenter]

It seems clear from the surrounding code that xpermits is allowed to be 
NULL here.

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

diff --git a/fs/afs/security.c b/fs/afs/security.c
index 3ef5043..bb4ed14 100644
--- a/fs/afs/security.c
+++ b/fs/afs/security.c
@@ -189,8 +189,9 @@ void afs_cache_permit(struct afs_vnode *vnode, struct key *key, long acl_order)
 	if (!permits)
 		goto out_unlock;
 
-	memcpy(permits->permits, xpermits->permits,
-	       count * sizeof(struct afs_permit));
+	if (xpermits)
+		memcpy(permits->permits, xpermits->permits,
+			count * sizeof(struct afs_permit));
 
 	_debug("key %x access %x",
 	       key_serial(key), vnode->status.caller_access);

Re: [patch] afs: potential null dereference

[David Howells]

Dan Carpenter <error27@gmail.com> wrote:

> It seems clear from the surrounding code that xpermits is allowed to be 
> NULL here.

Interesting.  The memcpy() won't oops due to this because if it is given a
NULL pointer, it will also be given a zero count.  I wonder if this means the
if-statement your patch adds is actually unnecessary...

David

Re: [patch] afs: potential null dereference

[Dan Carpenter]

On Mon, Mar 22, 2010 at 12:05:20PM +0000, David Howells wrote:
> Dan Carpenter <error27@gmail.com> wrote:
> 
> > It seems clear from the surrounding code that xpermits is allowed to be 
> > NULL here.
> 
> Interesting.  The memcpy() won't oops due to this because if it is given a
> NULL pointer, it will also be given a zero count.  I wonder if this means the
> if-statement your patch adds is actually unnecessary...
> 

I was concerned about the dereference here:

+       if (xpermits)
+               memcpy(permits->permits, xpermits->permits,
                                         ^^^^^^^^^^^^^^^^^
+                       count * sizeof(struct afs_permit));

This code has been there for three years, so yeah, you would think if it
were a problem someone would have complained.  My theory was "xpermits"
was almost always non-null.

regards,
dan carpenter

> David

Re: [patch] afs: potential null dereference

[David Howells]

Dan Carpenter <error27@gmail.com> wrote:

> I was concerned about the dereference here:
> 
> +       if (xpermits)
> +               memcpy(permits->permits, xpermits->permits,
>                                          ^^^^^^^^^^^^^^^^^
> +                       count * sizeof(struct afs_permit));

That's a good point - in which case your patch should definitely go in.

David

[PATCH] AFS: Potential null dereference

[David Howells]

From: Dan Carpenter <error27@gmail.com>

It seems clear from the surrounding code that xpermits is allowed to be NULL
here.

Signed-off-by: Dan Carpenter <error27@gmail.com>
Signed-off-by: David Howells <dhowells@redhat.com>
---

 fs/afs/security.c |    5 +++--
 1 files changed, 3 insertions(+), 2 deletions(-)

diff --git a/fs/afs/security.c b/fs/afs/security.c
index 3ef5043..bb4ed14 100644
--- a/fs/afs/security.c
+++ b/fs/afs/security.c
@@ -189,8 +189,9 @@ void afs_cache_permit(struct afs_vnode *vnode, struct key *key, long acl_order)
 	if (!permits)
 		goto out_unlock;
 
-	memcpy(permits->permits, xpermits->permits,
-	       count * sizeof(struct afs_permit));
+	if (xpermits)
+		memcpy(permits->permits, xpermits->permits,
+			count * sizeof(struct afs_permit));
 
 	_debug("key %x access %x",
 	       key_serial(key), vnode->status.caller_access);