diff mbox

[v3,0/3] Send audit/procinfo/cgroup data in socket-level control message

Message ID 1378308621.7360.110.camel@edumazet-glaptop
State RFC, archived
Delegated to: David Miller
Headers show

Commit Message

Eric Dumazet Sept. 4, 2013, 3:30 p.m. UTC
On Wed, 2013-09-04 at 11:20 -0400, Richard Guy Briggs wrote:
> On Wed, Sep 04, 2013 at 10:58:30AM -0400, Richard Guy Briggs wrote:
> > On Wed, Sep 04, 2013 at 12:42:26AM -0700, Eric W. Biederman wrote:
> > > Jan Kaluza <jkaluza@redhat.com> writes:
> > > > this patchset against net-next (applies also to linux-next) adds 3 new types
> > > > of "Socket"-level control message (SCM_AUDIT, SCM_PROCINFO and SCM_CGROUP).
> > 
> > > By my count you have overflowed cb in struct sk_buff and are stomping on
> > > _skb_refdest.
> > 
> > For patch1/3 I count 56/48, then for patch3 I get 48/48.  Jan, you might
> > do the conversion to a pointer in patch1/3 to avoid bisect breakage.
> 
> Wait, that __aligned(8) is for cb[48], not for the contents.
> 
> For patch1/3 I count 28/48 on 32-bit, 36/48 on 64-bit (or would that be
> 56 by default on 64-bit arches without aligned specified?), then for
> patch3 I get 24/48 on 32 and 40/48 on 64 (or again 48/48 by default?).

Do not count, just add this :



--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Comments

Jan Kaluza Sept. 4, 2013, 3:40 p.m. UTC | #1
On 09/04/2013 05:30 PM, Eric Dumazet wrote:
> On Wed, 2013-09-04 at 11:20 -0400, Richard Guy Briggs wrote:
>> On Wed, Sep 04, 2013 at 10:58:30AM -0400, Richard Guy Briggs wrote:
>>> On Wed, Sep 04, 2013 at 12:42:26AM -0700, Eric W. Biederman wrote:
>>>> Jan Kaluza <jkaluza@redhat.com> writes:
>>>>> this patchset against net-next (applies also to linux-next) adds 3 new types
>>>>> of "Socket"-level control message (SCM_AUDIT, SCM_PROCINFO and SCM_CGROUP).
>>>
>>>> By my count you have overflowed cb in struct sk_buff and are stomping on
>>>> _skb_refdest.
>>>
>>> For patch1/3 I count 56/48, then for patch3 I get 48/48.  Jan, you might
>>> do the conversion to a pointer in patch1/3 to avoid bisect breakage.
>>
>> Wait, that __aligned(8) is for cb[48], not for the contents.
>>
>> For patch1/3 I count 28/48 on 32-bit, 36/48 on 64-bit (or would that be
>> 56 by default on 64-bit arches without aligned specified?), then for
>> patch3 I get 24/48 on 32 and 40/48 on 64 (or again 48/48 by default?).
>
> Do not count, just add this :
>
> diff --git a/net/unix/af_unix.c b/net/unix/af_unix.c
> index 86de99a..5b61320 100644
> --- a/net/unix/af_unix.c
> +++ b/net/unix/af_unix.c
> @@ -1329,6 +1329,7 @@ static void unix_detach_fds(struct scm_cookie *scm, struct sk_buff *skb)
>   {
>   	int i;
>
> +	BUILD_BUG_ON(sizeof(struct unix_skb_parms) > sizeof(skb->cb));
>   	scm->fp = UNIXCB(skb).fp;
>   	UNIXCB(skb).fp = NULL;
>

That's already in af_unix.c:

BUILD_BUG_ON(sizeof(struct unix_skb_parms) > FIELD_SIZEOF(struct 
sk_buff, cb));

This test is passing for me even with only patch1/3 applied when 
building 64bit kernel.

Jan Kaluza

--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
diff mbox

Patch

diff --git a/net/unix/af_unix.c b/net/unix/af_unix.c
index 86de99a..5b61320 100644
--- a/net/unix/af_unix.c
+++ b/net/unix/af_unix.c
@@ -1329,6 +1329,7 @@  static void unix_detach_fds(struct scm_cookie *scm, struct sk_buff *skb)
 {
 	int i;
 
+	BUILD_BUG_ON(sizeof(struct unix_skb_parms) > sizeof(skb->cb));
 	scm->fp = UNIXCB(skb).fp;
 	UNIXCB(skb).fp = NULL;