[PATCH 4.14] phonet: refcount leak in pep_sock_accep

[PATCH 4.14] phonet: refcount leak in pep_sock_accep

[Aayush Agarwal]

From: Hangyu Hua <hbh25y@gmail.com>

commit bcd0f9335332 ("phonet: refcount leak in pep_sock_accep")
upstream.

sock_hold(sk) is invoked in pep_sock_accept(), but __sock_put(sk) is not
invoked in subsequent failure branches(pep_accept_conn() != 0).

Signed-off-by: Hangyu Hua <hbh25y@gmail.com>
Link: https://lore.kernel.org/r/20211209082839.33985-1-hbh25y@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Aayush Agarwal <aayush.a.agarwal@oracle.com>
---
 net/phonet/pep.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/net/phonet/pep.c b/net/phonet/pep.c
index b0d958cd1823..4c4a8a42ee88 100644
--- a/net/phonet/pep.c
+++ b/net/phonet/pep.c
@@ -881,6 +881,7 @@ static struct sock *pep_sock_accept(struct sock *sk, int flags, int *errp,
 
 	err = pep_accept_conn(newsk, skb);
 	if (err) {
+		__sock_put(sk);
 		sock_put(newsk);
 		newsk = NULL;
 		goto drop;
-- 
2.27.0

Re: [PATCH 4.14] phonet: refcount leak in pep_sock_accep

[Greg KH]

On Fri, Jan 07, 2022 at 02:53:32AM -0800, Aayush Agarwal wrote:
> From: Hangyu Hua <hbh25y@gmail.com>
> 
> commit bcd0f9335332 ("phonet: refcount leak in pep_sock_accep")
> upstream.
> 
> sock_hold(sk) is invoked in pep_sock_accept(), but __sock_put(sk) is not
> invoked in subsequent failure branches(pep_accept_conn() != 0).
> 
> Signed-off-by: Hangyu Hua <hbh25y@gmail.com>
> Link: https://lore.kernel.org/r/20211209082839.33985-1-hbh25y@gmail.com
> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
> Signed-off-by: Aayush Agarwal <aayush.a.agarwal@oracle.com>
> ---
>  net/phonet/pep.c | 1 +
>  1 file changed, 1 insertion(+)

What about releases 5.15.y, 5.10.y, 5.4.y, and 4.19.y?  Is this also
relevant for those trees?

thanks,

greg k-h

Re: [External] : Re: [PATCH 4.14] phonet: refcount leak in pep_sock_accep

[aayush.a.agarwal]

On 07/01/22 4:54 pm, Greg KH wrote:
> On Fri, Jan 07, 2022 at 02:53:32AM -0800, Aayush Agarwal wrote:
>> From: Hangyu Hua <hbh25y@gmail.com>
>>
>> commit bcd0f9335332 ("phonet: refcount leak in pep_sock_accep")
>> upstream.
>>
>> sock_hold(sk) is invoked in pep_sock_accept(), but __sock_put(sk) is not
>> invoked in subsequent failure branches(pep_accept_conn() != 0).
>>
>> Signed-off-by: Hangyu Hua <hbh25y@gmail.com>
>> Link: https://urldefense.com/v3/__https://lore.kernel.org/r/20211209082839.33985-1-hbh25y@gmail.com__;!!ACWV5N9M2RV99hQ!Znc0Oy9gtZZ18UDMwcZiYrfjj4GUibhEq5WJZ44m6azDWCC1hrZpkFh9AmGOqqS94cqz-A$
>> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
>> Signed-off-by: Aayush Agarwal <aayush.a.agarwal@oracle.com>
>> ---
>>   net/phonet/pep.c | 1 +
>>   1 file changed, 1 insertion(+)
> What about releases 5.15.y, 5.10.y, 5.4.y, and 4.19.y?  Is this also
> relevant for those trees?
>
> thanks,
>
> greg k-h

It's relevant for all currently supported stable releases: 4.4.y, 4.9.y, 
4.14.y, 4.19.y, 5.4.y, 5.10.y, 5.15.y . I missed adding the tag "Cc: 
stable@viger.kernel.org #4.4+". Should I send the patch again?

Re: [External] : Re: [PATCH 4.14] phonet: refcount leak in pep_sock_accep

[Greg KH]

On Fri, Jan 07, 2022 at 10:24:53PM +0530, aayush.a.agarwal@oracle.com wrote:
> 
> On 07/01/22 4:54 pm, Greg KH wrote:
> > On Fri, Jan 07, 2022 at 02:53:32AM -0800, Aayush Agarwal wrote:
> > > From: Hangyu Hua <hbh25y@gmail.com>
> > > 
> > > commit bcd0f9335332 ("phonet: refcount leak in pep_sock_accep")
> > > upstream.
> > > 
> > > sock_hold(sk) is invoked in pep_sock_accept(), but __sock_put(sk) is not
> > > invoked in subsequent failure branches(pep_accept_conn() != 0).
> > > 
> > > Signed-off-by: Hangyu Hua <hbh25y@gmail.com>
> > > Link: https://urldefense.com/v3/__https://lore.kernel.org/r/20211209082839.33985-1-hbh25y@gmail.com__;!!ACWV5N9M2RV99hQ!Znc0Oy9gtZZ18UDMwcZiYrfjj4GUibhEq5WJZ44m6azDWCC1hrZpkFh9AmGOqqS94cqz-A$
> > > Signed-off-by: Jakub Kicinski <kuba@kernel.org>
> > > Signed-off-by: Aayush Agarwal <aayush.a.agarwal@oracle.com>
> > > ---
> > >   net/phonet/pep.c | 1 +
> > >   1 file changed, 1 insertion(+)
> > What about releases 5.15.y, 5.10.y, 5.4.y, and 4.19.y?  Is this also
> > relevant for those trees?
> > 
> > thanks,
> > 
> > greg k-h
> 
> It's relevant for all currently supported stable releases: 4.4.y, 4.9.y,
> 4.14.y, 4.19.y, 5.4.y, 5.10.y, 5.15.y . I missed adding the tag "Cc:
> stable@viger.kernel.org #4.4+". Should I send the patch again?

No need, I've queued it up everywhere now, thanks!

greg k-h