NFSv2/v3: Remove incorrect dprintks from the readdir reply code

The actual size of the directory is unknown to the client, so it is
always requesting the maximum number it can handle. If the server
is replying with fewer entries than was requested, then that will
usually reflect the fact that we've hit the end of the directory.
Flagging it as an error is therefore incorrect.

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>

1 files changed, 2 insertions, 8 deletions

diff --git a/fs/nfs/nfs2xdr.c b/fs/nfs/nfs2xdr.c
index baf759bccd05..db81166182c9 100644
--- a/fs/nfs/nfs2xdr.c
+++ b/fs/nfs/nfs2xdr.c
@@ -978,16 +978,10 @@ static int decode_readdirok(struct xdr_stream *xdr)
         pglen = xdr->buf->page_len;
         hdrlen = (u8 *)xdr->p - (u8 *)xdr->iov->iov_base;
         recvd = xdr->buf->len - hdrlen;
-        if (unlikely(pglen > recvd))
-                goto out_cheating;
-out:
+        if (pglen > recvd)
+                pglen = recvd;
         xdr_read_pages(xdr, pglen);
         return pglen;
-out_cheating:
-        dprintk("NFS: server cheating in readdir result: "
-                "pglen %u > recvd %u\n", pglen, recvd);
-        pglen = recvd;
-        goto out;
 }
 
 static int nfs2_xdr_dec_readdirres(struct rpc_rqst *req,