[PATCH] RPC: fix accounting bug in the case of a truncated RPC message

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

include/linux/sunrpc/xdr.h

@@ -160,7 +160,7 @@ typedef struct {
 
 typedef size_t (*skb_read_actor_t)(skb_reader_t *desc, void *to, size_t len);
 
-extern int xdr_partial_copy_from_skb(struct xdr_buf *, unsigned int,
+extern ssize_t xdr_partial_copy_from_skb(struct xdr_buf *, unsigned int,
 		skb_reader_t *, skb_read_actor_t);
 
 struct socket;

net/sunrpc/xdr.c

@@ -176,21 +176,23 @@ xdr_inline_pages(struct xdr_buf *xdr, unsigned int offset,
 	xdr->buflen += len;
 }
 
-int
+ssize_t
 xdr_partial_copy_from_skb(struct xdr_buf *xdr, unsigned int base,
 			  skb_reader_t *desc,
 			  skb_read_actor_t copy_actor)
 {
 	struct page	**ppage = xdr->pages;
 	unsigned int	len, pglen = xdr->page_len;
+	ssize_t		copied = 0;
 	int		ret;
 
 	len = xdr->head[0].iov_len;
 	if (base < len) {
 		len -= base;
 		ret = copy_actor(desc, (char *)xdr->head[0].iov_base + base, len);
+		copied += ret;
 		if (ret != len || !desc->count)
-			return 0;
+			goto out;
 		base = 0;
 	} else
 		base -= len;
@@ -214,8 +216,11 @@ xdr_partial_copy_from_skb(struct xdr_buf *xdr, unsigned int base,
 		 * are small by default but can get huge. */
 		if (unlikely(*ppage == NULL)) {
 			*ppage = alloc_page(GFP_ATOMIC);
-			if (unlikely(*ppage == NULL))
-				return -ENOMEM;
+			if (unlikely(*ppage == NULL)) {
+				if (copied == 0)
+					copied = -ENOMEM;
+				goto out;
+			}
 		}
 
 		len = PAGE_CACHE_SIZE;
@@ -233,16 +238,17 @@ xdr_partial_copy_from_skb(struct xdr_buf *xdr, unsigned int base,
 		}
 		flush_dcache_page(*ppage);
 		kunmap_atomic(kaddr, KM_SKB_SUNRPC_DATA);
+		copied += ret;
 		if (ret != len || !desc->count)
-			return 0;
+			goto out;
 		ppage++;
 	} while ((pglen -= len) != 0);
 copy_tail:
 	len = xdr->tail[0].iov_len;
 	if (base < len)
-		copy_actor(desc, (char *)xdr->tail[0].iov_base + base, len - base);
-
-	return 0;
+		copied += copy_actor(desc, (char *)xdr->tail[0].iov_base + base, len - base);
+out:
+	return copied;
 }
 
 

net/sunrpc/xprt.c

@@ -823,10 +823,15 @@ tcp_copy_data(skb_reader_t *desc, void *p, size_t len)
 {
 	if (len > desc->count)
 		len = desc->count;
-	if (skb_copy_bits(desc->skb, desc->offset, p, len))
+	if (skb_copy_bits(desc->skb, desc->offset, p, len)) {
+		dprintk("RPC:      failed to copy %zu bytes from skb. %zu bytes remain\n",
+				len, desc->count);
 		return 0;
+	}
 	desc->offset += len;
 	desc->count -= len;
+	dprintk("RPC:      copied %zu bytes from skb. %zu bytes remain\n",
+			len, desc->count);
 	return len;
 }
 
@@ -865,6 +870,8 @@ tcp_read_fraghdr(struct rpc_xprt *xprt, skb_reader_t *desc)
 static void
 tcp_check_recm(struct rpc_xprt *xprt)
 {
+	dprintk("RPC:      xprt = %p, tcp_copied = %lu, tcp_offset = %u, tcp_reclen = %u, tcp_flags = %lx\n",
+			xprt, xprt->tcp_copied, xprt->tcp_offset, xprt->tcp_reclen, xprt->tcp_flags);
 	if (xprt->tcp_offset == xprt->tcp_reclen) {
 		xprt->tcp_flags |= XPRT_COPY_RECM;
 		xprt->tcp_offset = 0;
@@ -909,7 +916,7 @@ tcp_read_request(struct rpc_xprt *xprt, skb_reader_t *desc)
 	struct rpc_rqst *req;
 	struct xdr_buf *rcvbuf;
 	size_t len;
-	int r;
+	ssize_t r;
 
 	/* Find and lock the request corresponding to this xid */
 	spin_lock(&xprt->sock_lock);
@@ -932,15 +939,17 @@ tcp_read_request(struct rpc_xprt *xprt, skb_reader_t *desc)
 		my_desc.count = len;
 		r = xdr_partial_copy_from_skb(rcvbuf, xprt->tcp_copied,
 					  &my_desc, tcp_copy_data);
-		desc->count -= len;
-		desc->offset += len;
+		desc->count -= r;
+		desc->offset += r;
 	} else
 		r = xdr_partial_copy_from_skb(rcvbuf, xprt->tcp_copied,
 					  desc, tcp_copy_data);
-	xprt->tcp_copied += len;
-	xprt->tcp_offset += len;
 
-	if (r < 0) {
+	if (r > 0) {
+		xprt->tcp_copied += r;
+		xprt->tcp_offset += r;
+	}
+	if (r != len) {
 		/* Error when copying to the receive buffer,
 		 * usually because we weren't able to allocate
 		 * additional buffer pages. All we can do now
@@ -951,9 +960,18 @@ tcp_read_request(struct rpc_xprt *xprt, skb_reader_t *desc)
 		 * be discarded.
 		 */
 		xprt->tcp_flags &= ~XPRT_COPY_DATA;
+		dprintk("RPC:      XID %08x truncated request\n",
+				ntohl(xprt->tcp_xid));
+		dprintk("RPC:      xprt = %p, tcp_copied = %lu, tcp_offset = %u, tcp_reclen = %u\n",
+				xprt, xprt->tcp_copied, xprt->tcp_offset, xprt->tcp_reclen);
 		goto out;
 	}
 
+	dprintk("RPC:      XID %08x read %u bytes\n",
+			ntohl(xprt->tcp_xid), r);
+	dprintk("RPC:      xprt = %p, tcp_copied = %lu, tcp_offset = %u, tcp_reclen = %u\n",
+			xprt, xprt->tcp_copied, xprt->tcp_offset, xprt->tcp_reclen);
+
 	if (xprt->tcp_copied == req->rq_private_buf.buflen)
 		xprt->tcp_flags &= ~XPRT_COPY_DATA;
 	else if (xprt->tcp_offset == xprt->tcp_reclen) {
@@ -961,12 +979,12 @@ tcp_read_request(struct rpc_xprt *xprt, skb_reader_t *desc)
 			xprt->tcp_flags &= ~XPRT_COPY_DATA;
 	}
 
+out:
 	if (!(xprt->tcp_flags & XPRT_COPY_DATA)) {
 		dprintk("RPC: %4d received reply complete\n",
 				req->rq_task->tk_pid);
 		xprt_complete_rqst(xprt, req, xprt->tcp_copied);
 	}
-out:
 	spin_unlock(&xprt->sock_lock);
 	tcp_check_recm(xprt);
 }
@@ -985,6 +1003,7 @@ tcp_read_discard(struct rpc_xprt *xprt, skb_reader_t *desc)
 	desc->count -= len;
 	desc->offset += len;
 	xprt->tcp_offset += len;
+	dprintk("RPC:      discarded %u bytes\n", len);
 	tcp_check_recm(xprt);
 }