[03/11] security: commoncap: fix -Wstringop-overread warning

Commit Message

Arnd Bergmann March 22, 2021, 4:02 p.m. UTC

From: Arnd Bergmann <[e-mail]>

gcc-11 introdces a harmless warning for cap_inode_getsecurity:

security/commoncap.c: In function ‘cap_inode_getsecurity’:
security/commoncap.c:440:33: error: ‘memcpy’ reading 16 bytes from a region of size 0 [-Werror=stringop-overread]
  440 |                                 memcpy(&nscap->data, &cap->data, sizeof(__le32) * 2 * VFS_CAP_U32);
      |                                 ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

The problem here is that tmpbuf is initialized to NULL, so gcc assumes
it is not accessible unless it gets set by vfs_getxattr_alloc().  This is
a legitimate warning as far as I can tell, but the code is correct since
it correctly handles the error when that function fails.

Add a separate NULL check to tell gcc about it as well.

Signed-off-by: Arnd Bergmann <[e-mail]>
---
 security/commoncap.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

Comments

Christian Brauner March 22, 2021, 4:31 p.m. UTC | #1

On Mon, Mar 22, 2021 at 05:02:41PM +0100, Arnd Bergmann wrote:
> From: Arnd Bergmann <[e-mail]>
> 
> gcc-11 introdces a harmless warning for cap_inode_getsecurity:
> 
> security/commoncap.c: In function ‘cap_inode_getsecurity’:
> security/commoncap.c:440:33: error: ‘memcpy’ reading 16 bytes from a region of size 0 [-Werror=stringop-overread]
>   440 |                                 memcpy(&nscap->data, &cap->data, sizeof(__le32) * 2 * VFS_CAP_U32);
>       |                                 ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> 
> The problem here is that tmpbuf is initialized to NULL, so gcc assumes
> it is not accessible unless it gets set by vfs_getxattr_alloc().  This is
> a legitimate warning as far as I can tell, but the code is correct since
> it correctly handles the error when that function fails.
> 
> Add a separate NULL check to tell gcc about it as well.
> 
> Signed-off-by: Arnd Bergmann <[e-mail]>
> ---

Seems reasonable,
Acked-by: Christian Brauner <[e-mail]>

James Morris March 24, 2021, 8:50 p.m. UTC | #2

On Mon, 22 Mar 2021, Arnd Bergmann wrote:

> From: Arnd Bergmann <[e-mail]>
> 
> gcc-11 introdces a harmless warning for cap_inode_getsecurity:
> 
> security/commoncap.c: In function ‘cap_inode_getsecurity’:
> security/commoncap.c:440:33: error: ‘memcpy’ reading 16 bytes from a region of size 0 [-Werror=stringop-overread]
>   440 |                                 memcpy(&nscap->data, &cap->data, sizeof(__le32) * 2 * VFS_CAP_U32);
>       |                                 ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> 
> The problem here is that tmpbuf is initialized to NULL, so gcc assumes
> it is not accessible unless it gets set by vfs_getxattr_alloc().  This is
> a legitimate warning as far as I can tell, but the code is correct since
> it correctly handles the error when that function fails.
> 
> Add a separate NULL check to tell gcc about it as well.
> 
> Signed-off-by: Arnd Bergmann <[e-mail]>

Applied to
git://git.kernel.org/pub/scm/linux/kernel/git/jmorris/linux-security.git fixes-v5.12

Patch

diff --git a/security/commoncap.c b/security/commoncap.c
index 28f4d25480df..9a36ed6dd737 100644
--- a/security/commoncap.c
+++ b/security/commoncap.c
@@ -400,7 +400,7 @@  int cap_inode_getsecurity(struct user_namespace *mnt_userns,
 				      &tmpbuf, size, GFP_NOFS);
 	dput(dentry);
 
-	if (ret < 0)
+	if (ret < 0 || !tmpbuf)
 		return ret;
 
 	fs_ns = inode->i_sb->s_user_ns;