Stop setting a kick for TCP's RQ
authorBarret Rhoden <brho@cs.berkeley.edu>
Fri, 25 Mar 2016 21:42:22 +0000 (17:42 -0400)
committerBarret Rhoden <brho@cs.berkeley.edu>
Thu, 31 Mar 2016 20:53:42 +0000 (16:53 -0400)
We never call it, since the RQ is written to by qpassnolim(), which doesn't
call kick internally.  I verified (out-of-tree) with a fake kick method and
netperf that the kick wasn't called, at least in normal operation.

It caused a problem when I tried to clean up qio.

Signed-off-by: Barret Rhoden <brho@cs.berkeley.edu>
kern/src/net/tcp.c

index cd6affd..14c5b23 100644 (file)
@@ -638,7 +638,7 @@ void tcpacktimer(void *v)
 
 static void tcpcreate(struct conv *c)
 {
-       c->rq = qopen(QMAX, Qcoalesce, tcpacktimer, c);
+       c->rq = qopen(QMAX, Qcoalesce, 0, 0);
        c->wq = qopen(8 * QMAX, Qkick, tcpkick, c);
 }