[NFS] [PATCH] Stop call_decode() from ignorning RPC header errrors

[Steve Dickson <SteveD@redhat.com>]

[-- Attachment #1: Type: text/plain, Size: 187 bytes --]

This patch stop call_decode() from ignoring errors
that are found while parsing the RPC header. I turns
out the nfs acls routines need these error codes to do
the right thing...

SteveD.

[-- Attachment #2: linux-2.4.21-sunrpc-calldecode.patch --]
[-- Type: text/plain, Size: 703 bytes --]

--- linux-2.4.21/net/sunrpc/clnt.c.diff	2003-08-18 11:11:22.000000000 -0400
+++ linux-2.4.21/net/sunrpc/clnt.c	2003-08-20 12:13:02.000000000 -0400
@@ -784,9 +784,16 @@ call_decode(struct rpc_task *task)
 				__FUNCTION__);
 
 	/* Verify the RPC header */
-	if (!(p = call_verify(task)))
+	if (!(p = call_verify(task))) {
+		/*
+		 * When call_verfiy sets tk_action to NULL (via task_exit)
+		 * a non-retry-able error has occurred (like the server
+		 * not supporting a particular procedure call).
+		 */
+		if (task->tk_action == NULL)
+			return;
 		goto out_retry;
-
+	}
 	/*
 	 * The following is an NFS-specific hack to cater for setuid
 	 * processes whose uid is mapped to nobody on the server.