When I created UDP socket splicing, I added the goto nextpkt loop

to splice multiple UDP packets in the m_nextpkt list.  Some profiling
with TCP splicing showed that checking so_rcv.sb_mb is wrong.  It
causes several useless runs through the loop.  Better check for
nextrecord which contains the original m_nextpkt value of the mbuf.
OK mikeb@

diff --git a/sys/kern/uipc_socket.c b/sys/kern/uipc_socket.c
index ae9e88c..1a6579c 100644 (file)
--- a/sys/kern/uipc_socket.c
+++ b/sys/kern/uipc_socket.c
@@ -1,4 +1,4 @@
-/*     $OpenBSD: uipc_socket.c,v 1.120 2013/12/10 21:44:50 mikeb Exp $ */
+/*     $OpenBSD: uipc_socket.c,v 1.121 2014/01/11 14:33:48 bluhm Exp $ */
 /*     $NetBSD: uipc_socket.c,v 1.21 1996/02/04 02:17:52 christos Exp $        */
 
 /*
@@ -1405,7 +1405,7 @@ somove(struct socket *so, int wait)
        so->so_splicelen += len;
 
        /* Move several packets if possible. */
-       if (!maxreached && so->so_rcv.sb_mb)
+       if (!maxreached && nextrecord)
                goto nextpkt;
 
  release: