If a mount -u is done to either NFS client that switches it

from TCP to UDP and the rsize/wsize/readdirsize is greater
than NFS_MAXDGRAMDATA, it is possible for a thread doing an
I/O RPC to get stuck repeatedly doing retries. This happens
because the RPC will use a resize/wsize/readdirsize that won't
work for UDP and, as such, it will keep failing indefinitely.
This patch returns an error for this case, to avoid the problem.
A discussion on freebsd-fs@ seemed to indicate that returning
an error was preferable to silently ignoring the "udp"/"mntudp"
option.
This problem was discovered while investigating a problem reported
by pjd@ via email.

MFC after:	2 weeks

1 files changed, 17 insertions, 0 deletions

diff --git a/sys/nfsclient/nfs_vfsops.c b/sys/nfsclient/nfs_vfsops.c
index fafc2537ba6f..e43c048f517b 100644
--- a/sys/nfsclient/nfs_vfsops.c
+++ b/sys/nfsclient/nfs_vfsops.c
@@ -1106,6 +1106,23 @@ nfs_mount(struct mount *mp)
 			error = EIO;
 			goto out;
 		}
+
+		/*
+		 * Cannot switch to UDP if current rsize/wsize/readdirsize is
+		 * too large, since there may be an I/O RPC in progress that
+		 * will get retried after the switch to the UDP socket. These
+		 * retries will fail over and over and over again.
+		 */
+		if (args.sotype == SOCK_DGRAM &&
+		    (nmp->nm_rsize > NFS_MAXDGRAMDATA ||
+		     nmp->nm_wsize > NFS_MAXDGRAMDATA ||
+		     nmp->nm_readdirsize > NFS_MAXDGRAMDATA)) {
+			vfs_mount_error(mp,
+			    "old rsize/wsize/readdirsize greater than UDP max");
+			error = EINVAL;
+			goto out;
+		}
+
 		/*
 		 * When doing an update, we can't change from or to
 		 * v3, switch lockd strategies or change cookie translation