Fix lockup for loopback NFS mounts. The pipelined I/O limitations could be

hit on the client side and prevent the server side from retiring writes.
Pipeline operations turned off for all READs (no big loss since reads are
usually synchronous) and for NFS writes, and left on for the default bwrite().
(MFC expected prior to 4.3 freeze)

Testing by: mjacob, dillon

1 files changed, 8 insertions, 0 deletions

diff --git a/sys/nfs/nfs_vnops.c b/sys/nfs/nfs_vnops.c
index 37b977ca53d5..d889cf433db8 100644
--- a/sys/nfs/nfs_vnops.c
+++ b/sys/nfs/nfs_vnops.c
@@ -2869,6 +2869,9 @@ again:
 			 * NOTE: we are not clearing B_DONE here, so we have
 			 * to do it later on in this routine if we intend to 
 			 * initiate I/O on the bp.
+			 *
+			 * Note: to avoid loopback deadlocks, we do not
+			 * assign b_runningbufspace.
 			 */
 			if (wcred == NULL)
 				wcred = bp->b_wcred;
@@ -3142,7 +3145,12 @@ nfs_writebp(bp, force, procp)
 	curproc->p_stats->p_ru.ru_oublock++;
 	splx(s);
 
+	/*
+	 * Note: to avoid loopback deadlocks, we do not
+	 * assign b_runningbufspace.
+	 */
 	vfs_busy_pages(bp, 1);
+
 	if (force)
 		bp->b_flags |= B_WRITEINPROG;
 	BUF_KERNPROC(bp);