tun(4) in -CURRENT: No buffer space available - race condition patch

Marcin Cieslak saper at saper.info
Tue Sep 14 01:10:17 UTC 2010


Output queue of tun(4) gets full after some time when sending lots of data.
I have been observing this on -CURRENT at least since March this year.

Looks like it's a race condition (same in tun(4) and tap(4)), 
the following patch seems to address the issue:

Index: if_tap.c
===================================================================
--- if_tap.c	(revision 212217)
+++ if_tap.c	(working copy)
@@ -881,8 +881,7 @@
 
 			mtx_lock(&tp->tap_mtx);
 			tp->tap_flags |= TAP_RWAIT;
-			mtx_unlock(&tp->tap_mtx);
-			error = tsleep(tp,PCATCH|(PZERO+1),"taprd",0);
+			error = mtx_sleep(tp, &tp->tap_mtx, PDROP|PCATCH|(PZERO+1), "taprd", 0);
 			if (error)
 				return (error);
 		}
Index: if_tun.c
===================================================================
--- if_tun.c	(revision 212217)
+++ if_tun.c	(working copy)
@@ -836,8 +836,7 @@
 			}
 			mtx_lock(&tp->tun_mtx);
 			tp->tun_flags |= TUN_RWAIT;
-			mtx_unlock(&tp->tun_mtx);
-			if ((error = tsleep(tp, PCATCH | (PZERO + 1),
+			if ((error = mtx_sleep(tp, &tp->tun_mtx, PDROP | PCATCH | (PZERO + 1),
 					"tunread", 0)) != 0) {
 				splx(s);
 				return (error);

--Marcin



More information about the freebsd-current mailing list