Merge tag 'nfs-for-5.13-2' of git://git.linux-nfs.org/projects/trondmy/linux-nfs

Pull NFS client bugfixes from Trond Myklebust:
"Stable fixes:
   - Fix v4.0/v4.1 SEEK_DATA return -ENOTSUPP when set NFS_V4_2 config
   - Fix Oops in xs_tcp_send_request() when transport is disconnected
   - Fix a NULL pointer dereference in pnfs_mark_matching_lsegs_return()

  Bugfixes:
   - Fix instances where signal_pending() should be fatal_signal_pending()
   - fix an incorrect limit in filelayout_decode_layout()
   - Fixes for the SUNRPC backlogged RPC queue
   - Don't corrupt the value of pg_bytes_written in nfs_do_recoalesce()
   - Revert commit 586a0787 ("Clean up rpcrdma_prepare_readch()")"

* tag 'nfs-for-5.13-2' of git://git.linux-nfs.org/projects/trondmy/linux-nfs:
  nfs: Remove trailing semicolon in macros
  xprtrdma: Revert 586a0787
  NFSv4: Fix v4.0/v4.1 SEEK_DATA return -ENOTSUPP when set NFS_V4_2 config
  NFS: Clean up reset of the mirror accounting variables
  NFS: Don't corrupt the value of pg_bytes_written in nfs_do_recoalesce()
  NFS: Fix an Oopsable condition in __nfs_pageio_add_request()
  SUNRPC: More fixes for backlog congestion
  SUNRPC: Fix Oops in xs_tcp_send_request() when transport is disconnected
  NFSv4: Fix a NULL pointer dereference in pnfs_mark_matching_lsegs_return()
  SUNRPC in case of backlog, hand free slots directly to waiting task
  pNFS/NFSv4: Remove redundant initialization of 'rd_size'
  NFS: fix an incorrect limit in filelayout_decode_layout()
  fs/nfs: Use fatal_signal_pending instead of signal_pending

fs/nfs/filelayout/filelayout.c

@@ -718,7 +718,7 @@ filelayout_decode_layout(struct pnfs_layout_hdr *flo,
 		if (unlikely(!p))
 			goto out_err;
 		fl->fh_array[i]->size = be32_to_cpup(p++);
-		if (sizeof(struct nfs_fh) < fl->fh_array[i]->size) {
+		if (fl->fh_array[i]->size > NFS_MAXFHSIZE) {
 			printk(KERN_ERR "NFS: Too big fh %d received %d\n",
 			       i, fl->fh_array[i]->size);
 			goto out_err;

fs/nfs/namespace.c

@@ -362,7 +362,7 @@ static const struct kernel_param_ops param_ops_nfs_timeout = {
 	.set = param_set_nfs_timeout,
 	.get = param_get_nfs_timeout,
 };
-#define param_check_nfs_timeout(name, p) __param_check(name, p, int);
+#define param_check_nfs_timeout(name, p) __param_check(name, p, int)
 
 module_param(nfs_mountpoint_expiry_timeout, nfs_timeout, 0644);
 MODULE_PARM_DESC(nfs_mountpoint_expiry_timeout,

fs/nfs/nfs4file.c

@@ -211,7 +211,7 @@ static loff_t nfs4_file_llseek(struct file *filep, loff_t offset, int whence)
 	case SEEK_HOLE:
 	case SEEK_DATA:
 		ret = nfs42_proc_llseek(filep, offset, whence);
-		if (ret != -ENOTSUPP)
+		if (ret != -EOPNOTSUPP)
 			return ret;
 		fallthrough;
 	default:

fs/nfs/nfs4proc.c

@@ -1706,7 +1706,7 @@ static void nfs_set_open_stateid_locked(struct nfs4_state *state,
 		rcu_read_unlock();
 		trace_nfs4_open_stateid_update_wait(state->inode, stateid, 0);
 
-		if (!signal_pending(current)) {
+		if (!fatal_signal_pending(current)) {
 			if (schedule_timeout(5*HZ) == 0)
 				status = -EAGAIN;
 			else
@@ -3487,7 +3487,7 @@ static bool nfs4_refresh_open_old_stateid(nfs4_stateid *dst,
 		write_sequnlock(&state->seqlock);
 		trace_nfs4_close_stateid_update_wait(state->inode, dst, 0);
 
-		if (signal_pending(current))
+		if (fatal_signal_pending(current))
 			status = -EINTR;
 		else
 			if (schedule_timeout(5*HZ) != 0)

fs/nfs/pagelist.c

@@ -1094,15 +1094,16 @@ nfs_pageio_do_add_request(struct nfs_pageio_descriptor *desc,
 	struct nfs_page *prev = NULL;
 	unsigned int size;
 
-	if (mirror->pg_count != 0) {
-		prev = nfs_list_entry(mirror->pg_list.prev);
-	} else {
+	if (list_empty(&mirror->pg_list)) {
 		if (desc->pg_ops->pg_init)
 			desc->pg_ops->pg_init(desc, req);
 		if (desc->pg_error < 0)
 			return 0;
 		mirror->pg_base = req->wb_pgbase;
-	}
+		mirror->pg_count = 0;
+		mirror->pg_recoalesce = 0;
+	} else
+		prev = nfs_list_entry(mirror->pg_list.prev);
 
 	if (desc->pg_maxretrans && req->wb_nio > desc->pg_maxretrans) {
 		if (NFS_SERVER(desc->pg_inode)->flags & NFS_MOUNT_SOFTERR)
@@ -1127,18 +1128,13 @@ static void nfs_pageio_doio(struct nfs_pageio_descriptor *desc)
 {
 	struct nfs_pgio_mirror *mirror = nfs_pgio_current_mirror(desc);
 
-
 	if (!list_empty(&mirror->pg_list)) {
 		int error = desc->pg_ops->pg_doio(desc);
 		if (error < 0)
 			desc->pg_error = error;
-		else
+		if (list_empty(&mirror->pg_list))
 			mirror->pg_bytes_written += mirror->pg_count;
 	}
-	if (list_empty(&mirror->pg_list)) {
-		mirror->pg_count = 0;
-		mirror->pg_base = 0;
-	}
 }
 
 static void
@@ -1227,10 +1223,6 @@ static int nfs_do_recoalesce(struct nfs_pageio_descriptor *desc)
 
 	do {
 		list_splice_init(&mirror->pg_list, &head);
-		mirror->pg_bytes_written -= mirror->pg_count;
-		mirror->pg_count = 0;
-		mirror->pg_base = 0;
-		mirror->pg_recoalesce = 0;
 
 		while (!list_empty(&head)) {
 			struct nfs_page *req;

fs/nfs/pnfs.c

@@ -1317,6 +1317,11 @@ _pnfs_return_layout(struct inode *ino)
 {
 	struct pnfs_layout_hdr *lo = NULL;
 	struct nfs_inode *nfsi = NFS_I(ino);
+	struct pnfs_layout_range range = {
+		.iomode		= IOMODE_ANY,
+		.offset		= 0,
+		.length		= NFS4_MAX_UINT64,
+	};
 	LIST_HEAD(tmp_list);
 	const struct cred *cred;
 	nfs4_stateid stateid;
@@ -1344,16 +1349,10 @@ _pnfs_return_layout(struct inode *ino)
 	}
 	valid_layout = pnfs_layout_is_valid(lo);
 	pnfs_clear_layoutcommit(ino, &tmp_list);
-	pnfs_mark_matching_lsegs_return(lo, &tmp_list, NULL, 0);
+	pnfs_mark_matching_lsegs_return(lo, &tmp_list, &range, 0);
 
-	if (NFS_SERVER(ino)->pnfs_curr_ld->return_range) {
-		struct pnfs_layout_range range = {
-			.iomode		= IOMODE_ANY,
-			.offset		= 0,
-			.length		= NFS4_MAX_UINT64,
-		};
+	if (NFS_SERVER(ino)->pnfs_curr_ld->return_range)
 		NFS_SERVER(ino)->pnfs_curr_ld->return_range(lo, &range);
-	}
 
 	/* Don't send a LAYOUTRETURN if list was initially empty */
 	if (!test_bit(NFS_LAYOUT_RETURN_REQUESTED, &lo->plh_flags) ||
@@ -2678,7 +2677,7 @@ EXPORT_SYMBOL_GPL(pnfs_generic_pg_check_range);
 void
 pnfs_generic_pg_init_read(struct nfs_pageio_descriptor *pgio, struct nfs_page *req)
 {
-	u64 rd_size = req->wb_bytes;
+	u64 rd_size;
 
 	pnfs_generic_pg_check_layout(pgio);
 	pnfs_generic_pg_check_range(pgio, req);

fs/nfs/super.c

@@ -1379,7 +1379,7 @@ static const struct kernel_param_ops param_ops_portnr = {
 	.set = param_set_portnr,
 	.get = param_get_uint,
 };
-#define param_check_portnr(name, p) __param_check(name, p, unsigned int);
+#define param_check_portnr(name, p) __param_check(name, p, unsigned int)
 
 module_param_named(callback_tcpport, nfs_callback_set_tcpport, portnr, 0644);
 module_param_named(callback_nr_threads, nfs_callback_nr_threads, ushort, 0644);

include/linux/sunrpc/xprt.h

@@ -368,6 +368,8 @@ struct rpc_xprt *	xprt_alloc(struct net *net, size_t size,
 				unsigned int num_prealloc,
 				unsigned int max_req);
 void			xprt_free(struct rpc_xprt *);
+void			xprt_add_backlog(struct rpc_xprt *xprt, struct rpc_task *task);
+bool			xprt_wake_up_backlog(struct rpc_xprt *xprt, struct rpc_rqst *req);
 
 static inline int
 xprt_enable_swap(struct rpc_xprt *xprt)

net/sunrpc/clnt.c

@@ -1677,13 +1677,6 @@ call_reserveresult(struct rpc_task *task)
 		return;
 	}
 
-	/*
-	 * Even though there was an error, we may have acquired
-	 * a request slot somehow.  Make sure not to leak it.
-	 */
-	if (task->tk_rqstp)
-		xprt_release(task);
-
 	switch (status) {
 	case -ENOMEM:
 		rpc_delay(task, HZ >> 2);

net/sunrpc/xprt.c

@@ -70,6 +70,7 @@
 static void	 xprt_init(struct rpc_xprt *xprt, struct net *net);
 static __be32	xprt_alloc_xid(struct rpc_xprt *xprt);
 static void	 xprt_destroy(struct rpc_xprt *xprt);
+static void	 xprt_request_init(struct rpc_task *task);
 
 static DEFINE_SPINLOCK(xprt_list_lock);
 static LIST_HEAD(xprt_list);
@@ -1606,17 +1607,40 @@ xprt_transmit(struct rpc_task *task)
 	spin_unlock(&xprt->queue_lock);
 }
 
-static void xprt_add_backlog(struct rpc_xprt *xprt, struct rpc_task *task)
+static void xprt_complete_request_init(struct rpc_task *task)
+{
+	if (task->tk_rqstp)
+		xprt_request_init(task);
+}
+
+void xprt_add_backlog(struct rpc_xprt *xprt, struct rpc_task *task)
 {
 	set_bit(XPRT_CONGESTED, &xprt->state);
-	rpc_sleep_on(&xprt->backlog, task, NULL);
+	rpc_sleep_on(&xprt->backlog, task, xprt_complete_request_init);
 }
+EXPORT_SYMBOL_GPL(xprt_add_backlog);
 
-static void xprt_wake_up_backlog(struct rpc_xprt *xprt)
+static bool __xprt_set_rq(struct rpc_task *task, void *data)
 {
-	if (rpc_wake_up_next(&xprt->backlog) == NULL)
+	struct rpc_rqst *req = data;
+
+	if (task->tk_rqstp == NULL) {
+		memset(req, 0, sizeof(*req));	/* mark unused */
+		task->tk_rqstp = req;
+		return true;
+	}
+	return false;
+}
+
+bool xprt_wake_up_backlog(struct rpc_xprt *xprt, struct rpc_rqst *req)
+{
+	if (rpc_wake_up_first(&xprt->backlog, __xprt_set_rq, req) == NULL) {
 		clear_bit(XPRT_CONGESTED, &xprt->state);
+		return false;
+	}
+	return true;
 }
+EXPORT_SYMBOL_GPL(xprt_wake_up_backlog);
 
 static bool xprt_throttle_congested(struct rpc_xprt *xprt, struct rpc_task *task)
 {
@@ -1626,7 +1650,7 @@ static bool xprt_throttle_congested(struct rpc_xprt *xprt, struct rpc_task *task
 		goto out;
 	spin_lock(&xprt->reserve_lock);
 	if (test_bit(XPRT_CONGESTED, &xprt->state)) {
-		rpc_sleep_on(&xprt->backlog, task, NULL);
+		xprt_add_backlog(xprt, task);
 		ret = true;
 	}
 	spin_unlock(&xprt->reserve_lock);
@@ -1703,11 +1727,11 @@ EXPORT_SYMBOL_GPL(xprt_alloc_slot);
 void xprt_free_slot(struct rpc_xprt *xprt, struct rpc_rqst *req)
 {
 	spin_lock(&xprt->reserve_lock);
-	if (!xprt_dynamic_free_slot(xprt, req)) {
+	if (!xprt_wake_up_backlog(xprt, req) &&
+	    !xprt_dynamic_free_slot(xprt, req)) {
 		memset(req, 0, sizeof(*req));	/* mark unused */
 		list_add(&req->rq_list, &xprt->free);
 	}
-	xprt_wake_up_backlog(xprt);
 	spin_unlock(&xprt->reserve_lock);
 }
 EXPORT_SYMBOL_GPL(xprt_free_slot);
@@ -1894,10 +1918,10 @@ void xprt_release(struct rpc_task *task)
 	xdr_free_bvec(&req->rq_snd_buf);
 	if (req->rq_cred != NULL)
 		put_rpccred(req->rq_cred);
-	task->tk_rqstp = NULL;
 	if (req->rq_release_snd_buf)
 		req->rq_release_snd_buf(req);
 
+	task->tk_rqstp = NULL;
 	if (likely(!bc_prealloc(req)))
 		xprt->ops->free_slot(xprt, req);
 	else

net/sunrpc/xprtrdma/rpc_rdma.c

@@ -628,8 +628,9 @@ static bool rpcrdma_prepare_pagelist(struct rpcrdma_req *req,
 	return false;
 }
 
-/* The tail iovec might not reside in the same page as the
- * head iovec.
+/* The tail iovec may include an XDR pad for the page list,
+ * as well as additional content, and may not reside in the
+ * same page as the head iovec.
  */
 static bool rpcrdma_prepare_tail_iov(struct rpcrdma_req *req,
 				     struct xdr_buf *xdr,
@@ -747,19 +748,27 @@ static bool rpcrdma_prepare_readch(struct rpcrdma_xprt *r_xprt,
 				   struct rpcrdma_req *req,
 				   struct xdr_buf *xdr)
 {
-	struct kvec *tail = &xdr->tail[0];
-
 	if (!rpcrdma_prepare_head_iov(r_xprt, req, xdr->head[0].iov_len))
 		return false;
 
-	/* If there is a Read chunk, the page list is handled
+	/* If there is a Read chunk, the page list is being handled
 	 * via explicit RDMA, and thus is skipped here.
 	 */
 
-	if (tail->iov_len) {
-		if (!rpcrdma_prepare_tail_iov(req, xdr,
-					      offset_in_page(tail->iov_base),
-					      tail->iov_len))
+	/* Do not include the tail if it is only an XDR pad */
+	if (xdr->tail[0].iov_len > 3) {
+		unsigned int page_base, len;
+
+		/* If the content in the page list is an odd length,
+		 * xdr_write_pages() adds a pad at the beginning of
+		 * the tail iovec. Force the tail's non-pad content to
+		 * land at the next XDR position in the Send message.
+		 */
+		page_base = offset_in_page(xdr->tail[0].iov_base);
+		len = xdr->tail[0].iov_len;
+		page_base += len & 3;
+		len -= len & 3;
+		if (!rpcrdma_prepare_tail_iov(req, xdr, page_base, len))
 			return false;
 		kref_get(&req->rl_kref);
 	}

net/sunrpc/xprtrdma/transport.c

@@ -520,9 +520,8 @@ xprt_rdma_alloc_slot(struct rpc_xprt *xprt, struct rpc_task *task)
 	return;
 
 out_sleep:
-	set_bit(XPRT_CONGESTED, &xprt->state);
-	rpc_sleep_on(&xprt->backlog, task, NULL);
 	task->tk_status = -EAGAIN;
+	xprt_add_backlog(xprt, task);
 }
 
 /**
@@ -537,10 +536,11 @@ xprt_rdma_free_slot(struct rpc_xprt *xprt, struct rpc_rqst *rqst)
 	struct rpcrdma_xprt *r_xprt =
 		container_of(xprt, struct rpcrdma_xprt, rx_xprt);
 
+	rpcrdma_reply_put(&r_xprt->rx_buf, rpcr_to_rdmar(rqst));
+	if (!xprt_wake_up_backlog(xprt, rqst)) {
 		memset(rqst, 0, sizeof(*rqst));
 		rpcrdma_buffer_put(&r_xprt->rx_buf, rpcr_to_rdmar(rqst));
-	if (unlikely(!rpc_wake_up_next(&xprt->backlog)))
-		clear_bit(XPRT_CONGESTED, &xprt->state);
+	}
 }
 
 static bool rpcrdma_check_regbuf(struct rpcrdma_xprt *r_xprt,

net/sunrpc/xprtrdma/verbs.c

@@ -1200,6 +1200,20 @@ rpcrdma_mr_get(struct rpcrdma_xprt *r_xprt)
 	return mr;
 }
 
+/**
+ * rpcrdma_reply_put - Put reply buffers back into pool
+ * @buffers: buffer pool
+ * @req: object to return
+ *
+ */
+void rpcrdma_reply_put(struct rpcrdma_buffer *buffers, struct rpcrdma_req *req)
+{
+	if (req->rl_reply) {
+		rpcrdma_rep_put(buffers, req->rl_reply);
+		req->rl_reply = NULL;
+	}
+}
+
 /**
  * rpcrdma_buffer_get - Get a request buffer
  * @buffers: Buffer pool from which to obtain a buffer
@@ -1228,9 +1242,7 @@ rpcrdma_buffer_get(struct rpcrdma_buffer *buffers)
  */
 void rpcrdma_buffer_put(struct rpcrdma_buffer *buffers, struct rpcrdma_req *req)
 {
-	if (req->rl_reply)
-		rpcrdma_rep_put(buffers, req->rl_reply);
-	req->rl_reply = NULL;
+	rpcrdma_reply_put(buffers, req);
 
 	spin_lock(&buffers->rb_lock);
 	list_add(&req->rl_list, &buffers->rb_send_bufs);

net/sunrpc/xprtrdma/xprt_rdma.h

@@ -479,6 +479,7 @@ struct rpcrdma_req *rpcrdma_buffer_get(struct rpcrdma_buffer *);
 void rpcrdma_buffer_put(struct rpcrdma_buffer *buffers,
 			struct rpcrdma_req *req);
 void rpcrdma_rep_put(struct rpcrdma_buffer *buf, struct rpcrdma_rep *rep);
+void rpcrdma_reply_put(struct rpcrdma_buffer *buffers, struct rpcrdma_req *req);
 
 bool rpcrdma_regbuf_realloc(struct rpcrdma_regbuf *rb, size_t size,
 			    gfp_t flags);

net/sunrpc/xprtsock.c

@@ -1010,6 +1010,8 @@ static int xs_tcp_send_request(struct rpc_rqst *req)
 			kernel_sock_shutdown(transport->sock, SHUT_RDWR);
 		return -ENOTCONN;
 	}
+	if (!transport->inet)
+		return -ENOTCONN;
 
 	xs_pktdump("packet data:",
 				req->rq_svec->iov_base,