af_unix: return data from multiple SKBs on recv() with MSG_PEEK flag
authorAaron Conole <aconole@bytheb.org>
Sat, 26 Sep 2015 22:50:43 +0000 (18:50 -0400)
committerDavid S. Miller <davem@davemloft.net>
Tue, 29 Sep 2015 20:47:08 +0000 (13:47 -0700)
AF_UNIX sockets now return multiple skbs from recv() when MSG_PEEK flag
is set.

This is referenced in kernel bugzilla #12323 @
https://bugzilla.kernel.org/show_bug.cgi?id=12323

As described both in the BZ and lkml thread @
http://lkml.org/lkml/2008/1/8/444 calling recv() with MSG_PEEK on an
AF_UNIX socket only reads a single skb, where the desired effect is
to return as much skb data has been queued, until hitting the recv
buffer size (whichever comes first).

The modified MSG_PEEK path will now move to the next skb in the tree
and jump to the again: label, rather than following the natural loop
structure. This requires duplicating some of the loop head actions.

This was tested using the python socketpair python code attached to
the bugzilla issue.

Signed-off-by: Aaron Conole <aconole@bytheb.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
net/unix/af_unix.c

index 03ee4d359f6a4922397a1a8a36c015a06aae1cac..ef31b40ad55000a5fd029d7479b546483cc782b3 100644 (file)
@@ -2179,8 +2179,21 @@ unlock:
                        if (UNIXCB(skb).fp)
                                scm.fp = scm_fp_dup(UNIXCB(skb).fp);
 
-                       sk_peek_offset_fwd(sk, chunk);
+                       if (skip) {
+                               sk_peek_offset_fwd(sk, chunk);
+                               skip -= chunk;
+                       }
 
+                       if (UNIXCB(skb).fp)
+                               break;
+
+                       last = skb;
+                       last_len = skb->len;
+                       unix_state_lock(sk);
+                       skb = skb_peek_next(skb, &sk->sk_receive_queue);
+                       if (skb)
+                               goto again;
+                       unix_state_unlock(sk);
                        break;
                }
        } while (size);