NFS client updates for Linux 4.12

Highlights include:
 
 Stable bugfixes:
 - Fix use after free in write error path
 - Use GFP_NOIO for two allocations in writeback
 - Fix a hang in OPEN related to server reboot
 - Check the result of nfs4_pnfs_ds_connect
 - Fix an rcu lock leak
 
 Features:
 - Removal of the unmaintained and unused OSD pNFS layout
 - Cleanup and removal of lots of unnecessary dprintk()s
 - Cleanup and removal of some memory failure paths now that
   GFP_NOFS is guaranteed to never fail.
 - Remove the v3-only data server limitation on pNFS/flexfiles
 
 Bugfixes:
 - RPC/RDMA connection handling bugfixes
 - Copy offload: fixes to ensure the copied data is COMMITed to disk.
 - Readdir: switch back to using the ->iterate VFS interface
 - File locking fixes from Ben Coddington
 - Various use-after-free and deadlock issues in pNFS
 - Write path bugfixes
 -----BEGIN PGP SIGNATURE-----
 
 iQIbBAABAgAGBQJZE0KiAAoJEGcL54qWCgDy/moP93wZ+cGnN5sC+nsirqj4eUiM
 BQKKonweNQIoYRwp5B9jLTxsUMIxRasV5W3BbqEm4PUtBYXfqQ7SfLv7RboKbd4M
 RJB9PS+sjx3Fxf65mhveKziwUFLvQCQ3+we0TpUga6+7SBiGlgPKBfisk7frC0nt
 BbYBuGaWXMPxO0BnR8adNwqiGINPDSzB+8sgjiT8zkZLm4lrew2eV7TDvwVOguD+
 S2vLPGhg1F9wu8aG731MgiSNaeCgsBP6I5D29fTTD7z1DCNMQXOoHcX8k4KwwIDB
 sHRR0tVBsg+1B7WdH4y41GQ03rn3o2DHeJB5cdYGaEu4lx7CecCzt0o0dfAkNizT
 5LxbQxIHPNYMeZmP2T0oD41zQyfjKqrdRSPnXi3dPD98NwaM1Lqv+Kzb/eXzupXp
 vJ7859PQCa3KjQ1IFhwdXTmh53J1c8SzEDpzz7WX0R0saRyxeIJsm30MmdPqKu7Z
 notjsXxrTmjIhC+0vFLey1kejFDh+b0gT6UIwoMdx39VL9AM6DVL7HsrU1kEwCdf
 f8otaLcm0WoUaseF+cMtfRNGEqCMxPywwz7mEKlGiVZgyAM8VfzH+s5j6/u6ncwS
 ASwRclwwPAZN97rzl0exZxuaRwFZd7oFT1zrviPWvv+0SUPuy258J6QpolUSavgi
 Qh7f3QR65K+QX9QbO1g=
 =7Nm2
 -----END PGP SIGNATURE-----

Merge tag 'nfs-for-4.12-1' of git://git.linux-nfs.org/projects/trondmy/linux-nfs

Pull NFS client updates from Trond Myklebust:
 "Highlights include:

  Stable bugfixes:
   - Fix use after free in write error path
   - Use GFP_NOIO for two allocations in writeback
   - Fix a hang in OPEN related to server reboot
   - Check the result of nfs4_pnfs_ds_connect
   - Fix an rcu lock leak

  Features:
   - Removal of the unmaintained and unused OSD pNFS layout
   - Cleanup and removal of lots of unnecessary dprintk()s
   - Cleanup and removal of some memory failure paths now that GFP_NOFS
     is guaranteed to never fail.
   - Remove the v3-only data server limitation on pNFS/flexfiles

  Bugfixes:
   - RPC/RDMA connection handling bugfixes
   - Copy offload: fixes to ensure the copied data is COMMITed to disk.
   - Readdir: switch back to using the ->iterate VFS interface
   - File locking fixes from Ben Coddington
   - Various use-after-free and deadlock issues in pNFS
   - Write path bugfixes"

* tag 'nfs-for-4.12-1' of git://git.linux-nfs.org/projects/trondmy/linux-nfs: (89 commits)
  pNFS/flexfiles: Always attempt to call layoutstats when flexfiles is enabled
  NFSv4.1: Work around a Linux server bug...
  NFS append COMMIT after synchronous COPY
  NFSv4: Fix exclusive create attributes encoding
  NFSv4: Fix an rcu lock leak
  nfs: use kmap/kunmap directly
  NFS: always treat the invocation of nfs_getattr as cache hit when noac is on
  Fix nfs_client refcounting if kmalloc fails in nfs4_proc_exchange_id and nfs4_proc_async_renew
  NFSv4.1: RECLAIM_COMPLETE must handle NFS4ERR_CONN_NOT_BOUND_TO_SESSION
  pNFS: Fix NULL dereference in pnfs_generic_alloc_ds_commits
  pNFS: Fix a typo in pnfs_generic_alloc_ds_commits
  pNFS: Fix a deadlock when coalescing writes and returning the layout
  pNFS: Don't clear the layout return info if there are segments to return
  pNFS: Ensure we commit the layout if it has been invalidated
  pNFS: Don't send COMMITs to the DSes if the server invalidated our layout
  pNFS/flexfiles: Fix up the ff_layout_write_pagelist failure path
  pNFS: Ensure we check layout validity before marking it for return
  NFS4.1 handle interrupted slot reuse from ERR_DELAY
  NFSv4: check return value of xdr_inline_decode
  nfs/filelayout: fix NULL pointer dereference in fl_pnfs_update_layout()
  ...

Documentation/admin-guide/kernel-parameters.txt

@@ -2434,12 +2434,6 @@
 			and gids from such clients.  This is intended to ease
 			migration from NFSv2/v3.
 
-	objlayoutdriver.osd_login_prog=
-			[NFS] [OBJLAYOUT] sets the pathname to the program which
-			is used to automatically discover and login into new
-			osd-targets. Please see:
-			Documentation/filesystems/pnfs.txt for more explanations
-
 	nmi_debug=	[KNL,SH] Specify one or more actions to take
 			when a NMI is triggered.
 			Format: [state][,regs][,debounce][,die]

Documentation/filesystems/nfs/pnfs.txt

@@ -64,46 +64,9 @@ table which are called by the nfs-client pnfs-core to implement the
 different layout types.
 
 Files-layout-driver code is in: fs/nfs/filelayout/.. directory
-Objects-layout-driver code is in: fs/nfs/objlayout/.. directory
 Blocks-layout-driver code is in: fs/nfs/blocklayout/.. directory
 Flexfiles-layout-driver code is in: fs/nfs/flexfilelayout/.. directory
 
-objects-layout setup
---------------------
-
-As part of the full STD implementation the objlayoutdriver.ko needs, at times,
-to automatically login to yet undiscovered iscsi/osd devices. For this the
-driver makes up-calles to a user-mode script called *osd_login*
-
-The path_name of the script to use is by default:
-	/sbin/osd_login.
-This name can be overridden by the Kernel module parameter:
-	objlayoutdriver.osd_login_prog
-
-If Kernel does not find the osd_login_prog path it will zero it out
-and will not attempt farther logins. An admin can then write new value
-to the objlayoutdriver.osd_login_prog Kernel parameter to re-enable it.
-
-The /sbin/osd_login is part of the nfs-utils package, and should usually
-be installed on distributions that support this Kernel version.
-
-The API to the login script is as follows:
-	Usage: $0 -u <URI> -o <OSDNAME> -s <SYSTEMID>
-	Options:
-		-u		target uri e.g. iscsi://<ip>:<port>
-				(always exists)
-				(More protocols can be defined in the future.
-				 The client does not interpret this string it is
-				 passed unchanged as received from the Server)
-		-o		osdname of the requested target OSD
-				(Might be empty)
-				(A string which denotes the OSD name, there is a
-				 limit of 64 chars on this string)
-		-s 		systemid of the requested target OSD
-				(Might be empty)
-				(This string, if not empty is always an hex
-				 representation of the 20 bytes osd_system_id)
-
 blocks-layout setup
 -------------------
 

fs/fuse/file.c

@@ -2177,7 +2177,7 @@ static int fuse_setlk(struct file *file, struct file_lock *fl, int flock)
 	}
 
 	/* Unlock on close is handled by the flush method */
-	if (fl->fl_flags & FL_CLOSE)
+	if ((fl->fl_flags & FL_CLOSE_POSIX) == FL_CLOSE_POSIX)
 		return 0;
 
 	if (pid && pid_nr == 0)

fs/lockd/clntlock.c

@@ -69,6 +69,7 @@ struct nlm_host *nlmclnt_init(const struct nlmclnt_initdata *nlm_init)
 	if (host->h_rpcclnt == NULL && nlm_bind_host(host) == NULL)
 		goto out_nobind;
 
+	host->h_nlmclnt_ops = nlm_init->nlmclnt_ops;
 	return host;
 out_nobind:
 	nlmclnt_release_host(host);

fs/lockd/clntproc.c

@@ -150,17 +150,22 @@ static void nlmclnt_release_lockargs(struct nlm_rqst *req)
  * @host: address of a valid nlm_host context representing the NLM server
  * @cmd: fcntl-style file lock operation to perform
  * @fl: address of arguments for the lock operation
+ * @data: address of data to be sent to callback operations
  *
  */
-int nlmclnt_proc(struct nlm_host *host, int cmd, struct file_lock *fl)
+int nlmclnt_proc(struct nlm_host *host, int cmd, struct file_lock *fl, void *data)
 {
 	struct nlm_rqst		*call;
 	int			status;
+	const struct nlmclnt_operations *nlmclnt_ops = host->h_nlmclnt_ops;
 
 	call = nlm_alloc_call(host);
 	if (call == NULL)
 		return -ENOMEM;
 
+	if (nlmclnt_ops && nlmclnt_ops->nlmclnt_alloc_call)
+		nlmclnt_ops->nlmclnt_alloc_call(data);
+
 	nlmclnt_locks_init_private(fl, host);
 	if (!fl->fl_u.nfs_fl.owner) {
 		/* lockowner allocation has failed */
@@ -169,6 +174,7 @@ int nlmclnt_proc(struct nlm_host *host, int cmd, struct file_lock *fl)
 	}
 	/* Set up the argument struct */
 	nlmclnt_setlockargs(call, fl);
+	call->a_callback_data = data;
 
 	if (IS_SETLK(cmd) || IS_SETLKW(cmd)) {
 		if (fl->fl_type != F_UNLCK) {
@@ -214,8 +220,12 @@ struct nlm_rqst *nlm_alloc_call(struct nlm_host *host)
 
 void nlmclnt_release_call(struct nlm_rqst *call)
 {
+	const struct nlmclnt_operations *nlmclnt_ops = call->a_host->h_nlmclnt_ops;
+
 	if (!atomic_dec_and_test(&call->a_count))
 		return;
+	if (nlmclnt_ops && nlmclnt_ops->nlmclnt_release_call)
+		nlmclnt_ops->nlmclnt_release_call(call->a_callback_data);
 	nlmclnt_release_host(call->a_host);
 	nlmclnt_release_lockargs(call);
 	kfree(call);
@@ -687,6 +697,19 @@ out:
 	return status;
 }
 
+static void nlmclnt_unlock_prepare(struct rpc_task *task, void *data)
+{
+	struct nlm_rqst	*req = data;
+	const struct nlmclnt_operations *nlmclnt_ops = req->a_host->h_nlmclnt_ops;
+	bool defer_call = false;
+
+	if (nlmclnt_ops && nlmclnt_ops->nlmclnt_unlock_prepare)
+		defer_call = nlmclnt_ops->nlmclnt_unlock_prepare(task, req->a_callback_data);
+
+	if (!defer_call)
+		rpc_call_start(task);
+}
+
 static void nlmclnt_unlock_callback(struct rpc_task *task, void *data)
 {
 	struct nlm_rqst	*req = data;
@@ -720,6 +743,7 @@ die:
 }
 
 static const struct rpc_call_ops nlmclnt_unlock_ops = {
+	.rpc_call_prepare = nlmclnt_unlock_prepare,
 	.rpc_call_done = nlmclnt_unlock_callback,
 	.rpc_release = nlmclnt_rpc_release,
 };

fs/locks.c

@@ -2504,7 +2504,7 @@ locks_remove_flock(struct file *filp, struct file_lock_context *flctx)
 		.fl_owner = filp,
 		.fl_pid = current->tgid,
 		.fl_file = filp,
-		.fl_flags = FL_FLOCK,
+		.fl_flags = FL_FLOCK | FL_CLOSE,
 		.fl_type = F_UNLCK,
 		.fl_end = OFFSET_MAX,
 	};

fs/nfs/Kconfig

@@ -123,11 +123,6 @@ config PNFS_BLOCK
 	depends on NFS_V4_1 && BLK_DEV_DM
 	default NFS_V4
 
-config PNFS_OBJLAYOUT
-	tristate
-	depends on NFS_V4_1 && SCSI_OSD_ULD
-	default NFS_V4
-
 config PNFS_FLEXFILE_LAYOUT
 	tristate
 	depends on NFS_V4_1 && NFS_V3

fs/nfs/Makefile

@@ -31,6 +31,5 @@ nfsv4-$(CONFIG_NFS_V4_1)	+= pnfs.o pnfs_dev.o pnfs_nfs.o
 nfsv4-$(CONFIG_NFS_V4_2)	+= nfs42proc.o
 
 obj-$(CONFIG_PNFS_FILE_LAYOUT) += filelayout/
-obj-$(CONFIG_PNFS_OBJLAYOUT) += objlayout/
 obj-$(CONFIG_PNFS_BLOCK) += blocklayout/
 obj-$(CONFIG_PNFS_FLEXFILE_LAYOUT) += flexfilelayout/

fs/nfs/callback_proc.c

@@ -131,10 +131,11 @@ restart:
 			if (!inode)
 				continue;
 			if (!nfs_sb_active(inode->i_sb)) {
-				rcu_read_lock();
+				rcu_read_unlock();
 				spin_unlock(&clp->cl_lock);
 				iput(inode);
 				spin_lock(&clp->cl_lock);
+				rcu_read_lock();
 				goto restart;
 			}
 			return inode;
@@ -170,10 +171,11 @@ restart:
 			if (!inode)
 				continue;
 			if (!nfs_sb_active(inode->i_sb)) {
-				rcu_read_lock();
+				rcu_read_unlock();
 				spin_unlock(&clp->cl_lock);
 				iput(inode);
 				spin_lock(&clp->cl_lock);
+				rcu_read_lock();
 				goto restart;
 			}
 			return inode;
@@ -317,31 +319,18 @@ static u32 initiate_bulk_draining(struct nfs_client *clp,
 static u32 do_callback_layoutrecall(struct nfs_client *clp,
 				    struct cb_layoutrecallargs *args)
 {
-	u32 res;
-
-	dprintk("%s enter, type=%i\n", __func__, args->cbl_recall_type);
 	if (args->cbl_recall_type == RETURN_FILE)
-		res = initiate_file_draining(clp, args);
+		return initiate_file_draining(clp, args);
-	else
+	return initiate_bulk_draining(clp, args);
-		res = initiate_bulk_draining(clp, args);
-	dprintk("%s returning %i\n", __func__, res);
-	return res;
-
 }
 
 __be32 nfs4_callback_layoutrecall(struct cb_layoutrecallargs *args,
 				  void *dummy, struct cb_process_state *cps)
 {
-	u32 res;
+	u32 res = NFS4ERR_OP_NOT_IN_SESSION;
-
-	dprintk("%s: -->\n", __func__);
 
 	if (cps->clp)
 		res = do_callback_layoutrecall(cps->clp, args);
-	else
-		res = NFS4ERR_OP_NOT_IN_SESSION;
-
-	dprintk("%s: exit with status = %d\n", __func__, res);
 	return cpu_to_be32(res);
 }
 
@@ -364,8 +353,6 @@ __be32 nfs4_callback_devicenotify(struct cb_devicenotifyargs *args,
 	struct nfs_client *clp = cps->clp;
 	struct nfs_server *server = NULL;
 
-	dprintk("%s: -->\n", __func__);
-
 	if (!clp) {
 		res = cpu_to_be32(NFS4ERR_OP_NOT_IN_SESSION);
 		goto out;
@@ -384,8 +371,6 @@ __be32 nfs4_callback_devicenotify(struct cb_devicenotifyargs *args,
 					goto found;
 				}
 			rcu_read_unlock();
-			dprintk("%s: layout type %u not found\n",
-				__func__, dev->cbd_layout_type);
 			continue;
 		}
 
@@ -395,8 +380,6 @@ __be32 nfs4_callback_devicenotify(struct cb_devicenotifyargs *args,
 
 out:
 	kfree(args->devs);
-	dprintk("%s: exit with status = %u\n",
-		__func__, be32_to_cpu(res));
 	return res;
 }
 
@@ -417,16 +400,11 @@ static __be32
 validate_seqid(const struct nfs4_slot_table *tbl, const struct nfs4_slot *slot,
 		const struct cb_sequenceargs * args)
 {
-	dprintk("%s enter. slotid %u seqid %u, slot table seqid: %u\n",
-		__func__, args->csa_slotid, args->csa_sequenceid, slot->seq_nr);
-
 	if (args->csa_slotid > tbl->server_highest_slotid)
 		return htonl(NFS4ERR_BADSLOT);
 
 	/* Replay */
 	if (args->csa_sequenceid == slot->seq_nr) {
-		dprintk("%s seqid %u is a replay\n",
-			__func__, args->csa_sequenceid);
 		if (nfs4_test_locked_slot(tbl, slot->slot_nr))
 			return htonl(NFS4ERR_DELAY);
 		/* Signal process_op to set this error on next op */
@@ -480,15 +458,6 @@ static bool referring_call_exists(struct nfs_client *clp,
 
 		for (j = 0; j < rclist->rcl_nrefcalls; j++) {
 			ref = &rclist->rcl_refcalls[j];
-
-			dprintk("%s: sessionid %x:%x:%x:%x sequenceid %u "
-				"slotid %u\n", __func__,
-				((u32 *)&rclist->rcl_sessionid.data)[0],
-				((u32 *)&rclist->rcl_sessionid.data)[1],
-				((u32 *)&rclist->rcl_sessionid.data)[2],
-				((u32 *)&rclist->rcl_sessionid.data)[3],
-				ref->rc_sequenceid, ref->rc_slotid);
-
 			status = nfs4_slot_wait_on_seqid(tbl, ref->rc_slotid,
 					ref->rc_sequenceid, HZ >> 1) < 0;
 			if (status)
@@ -593,8 +562,6 @@ out:
 		res->csr_status = status;
 
 	trace_nfs4_cb_sequence(args, res, status);
-	dprintk("%s: exit with status = %d res->csr_status %d\n", __func__,
-		ntohl(status), ntohl(res->csr_status));
 	return status;
 }
 

fs/nfs/callback_xdr.c

@@ -171,8 +171,6 @@ static __be32 decode_compound_hdr_arg(struct xdr_stream *xdr, struct cb_compound
 		return htonl(NFS4ERR_MINOR_VERS_MISMATCH);
 	}
 	hdr->nops = ntohl(*p);
-	dprintk("%s: minorversion %d nops %d\n", __func__,
-		hdr->minorversion, hdr->nops);
 	return 0;
 }
 
@@ -192,11 +190,8 @@ static __be32 decode_getattr_args(struct svc_rqst *rqstp, struct xdr_stream *xdr
 
 	status = decode_fh(xdr, &args->fh);
 	if (unlikely(status != 0))
-		goto out;
+		return status;
-	status = decode_bitmap(xdr, args->bitmap);
+	return decode_bitmap(xdr, args->bitmap);
-out:
-	dprintk("%s: exit with status = %d\n", __func__, ntohl(status));
-	return status;
 }
 
 static __be32 decode_recall_args(struct svc_rqst *rqstp, struct xdr_stream *xdr, struct cb_recallargs *args)
@@ -206,17 +201,12 @@ static __be32 decode_recall_args(struct svc_rqst *rqstp, struct xdr_stream *xdr,
 
 	status = decode_delegation_stateid(xdr, &args->stateid);
 	if (unlikely(status != 0))
-		goto out;
+		return status;
 	p = read_buf(xdr, 4);
-	if (unlikely(p == NULL)) {
+	if (unlikely(p == NULL))
-		status = htonl(NFS4ERR_RESOURCE);
+		return htonl(NFS4ERR_RESOURCE);
-		goto out;
-	}
 	args->truncate = ntohl(*p);
-	status = decode_fh(xdr, &args->fh);
+	return decode_fh(xdr, &args->fh);
-out:
-	dprintk("%s: exit with status = %d\n", __func__, ntohl(status));
-	return status;
 }
 
 #if defined(CONFIG_NFS_V4_1)
@@ -235,10 +225,8 @@ static __be32 decode_layoutrecall_args(struct svc_rqst *rqstp,
 	uint32_t iomode;
 
 	p = read_buf(xdr, 4 * sizeof(uint32_t));
-	if (unlikely(p == NULL)) {
+	if (unlikely(p == NULL))
-		status = htonl(NFS4ERR_BADXDR);
+		return htonl(NFS4ERR_BADXDR);
-		goto out;
-	}
 
 	args->cbl_layout_type = ntohl(*p++);
 	/* Depite the spec's xdr, iomode really belongs in the FILE switch,
@@ -252,37 +240,23 @@ static __be32 decode_layoutrecall_args(struct svc_rqst *rqstp,
 		args->cbl_range.iomode = iomode;
 		status = decode_fh(xdr, &args->cbl_fh);
 		if (unlikely(status != 0))
-			goto out;
+			return status;
 
 		p = read_buf(xdr, 2 * sizeof(uint64_t));
-		if (unlikely(p == NULL)) {
+		if (unlikely(p == NULL))
-			status = htonl(NFS4ERR_BADXDR);
+			return htonl(NFS4ERR_BADXDR);
-			goto out;
-		}
 		p = xdr_decode_hyper(p, &args->cbl_range.offset);
 		p = xdr_decode_hyper(p, &args->cbl_range.length);
-		status = decode_layout_stateid(xdr, &args->cbl_stateid);
+		return decode_layout_stateid(xdr, &args->cbl_stateid);
-		if (unlikely(status != 0))
-			goto out;
 	} else if (args->cbl_recall_type == RETURN_FSID) {
 		p = read_buf(xdr, 2 * sizeof(uint64_t));
-		if (unlikely(p == NULL)) {
+		if (unlikely(p == NULL))
-			status = htonl(NFS4ERR_BADXDR);
+			return htonl(NFS4ERR_BADXDR);
-			goto out;
-		}
 		p = xdr_decode_hyper(p, &args->cbl_fsid.major);
 		p = xdr_decode_hyper(p, &args->cbl_fsid.minor);
-	} else if (args->cbl_recall_type != RETURN_ALL) {
+	} else if (args->cbl_recall_type != RETURN_ALL)
-		status = htonl(NFS4ERR_BADXDR);
+		return htonl(NFS4ERR_BADXDR);
-		goto out;
+	return 0;
-	}
-	dprintk("%s: ltype 0x%x iomode %d changed %d recall_type %d\n",
-		__func__,
-		args->cbl_layout_type, iomode,
-		args->cbl_layoutchanged, args->cbl_recall_type);
-out:
-	dprintk("%s: exit with status = %d\n", __func__, ntohl(status));
-	return status;
 }
 
 static
@@ -437,12 +411,11 @@ static __be32 decode_cb_sequence_args(struct svc_rqst *rqstp,
 
 	status = decode_sessionid(xdr, &args->csa_sessionid);
 	if (status)
-		goto out;
+		return status;
 
-	status = htonl(NFS4ERR_RESOURCE);
 	p = read_buf(xdr, 5 * sizeof(uint32_t));
 	if (unlikely(p == NULL))
-		goto out;
+		return htonl(NFS4ERR_RESOURCE);
 
 	args->csa_addr = svc_addr(rqstp);
 	args->csa_sequenceid = ntohl(*p++);
@@ -456,7 +429,7 @@ static __be32 decode_cb_sequence_args(struct svc_rqst *rqstp,
 						  sizeof(*args->csa_rclists),
 						  GFP_KERNEL);
 		if (unlikely(args->csa_rclists == NULL))
-			goto out;
+			return htonl(NFS4ERR_RESOURCE);
 
 		for (i = 0; i < args->csa_nrclists; i++) {
 			status = decode_rc_list(xdr, &args->csa_rclists[i]);
@@ -466,27 +439,13 @@ static __be32 decode_cb_sequence_args(struct svc_rqst *rqstp,
 			}
 		}
 	}
-	status = 0;
+	return 0;
-
-	dprintk("%s: sessionid %x:%x:%x:%x sequenceid %u slotid %u "
-		"highestslotid %u cachethis %d nrclists %u\n",
-		__func__,
-		((u32 *)&args->csa_sessionid)[0],
-		((u32 *)&args->csa_sessionid)[1],
-		((u32 *)&args->csa_sessionid)[2],
-		((u32 *)&args->csa_sessionid)[3],
-		args->csa_sequenceid, args->csa_slotid,
-		args->csa_highestslotid, args->csa_cachethis,
-		args->csa_nrclists);
-out:
-	dprintk("%s: exit with status = %d\n", __func__, ntohl(status));
-	return status;
 
 out_free:
 	for (i = 0; i < args->csa_nrclists; i++)
 		kfree(args->csa_rclists[i].rcl_refcalls);
 	kfree(args->csa_rclists);
-	goto out;
+	return status;
 }
 
 static __be32 decode_recallany_args(struct svc_rqst *rqstp,
@@ -557,11 +516,8 @@ static __be32 decode_notify_lock_args(struct svc_rqst *rqstp, struct xdr_stream
 
 	status = decode_fh(xdr, &args->cbnl_fh);
 	if (unlikely(status != 0))
-		goto out;
+		return status;
-	status = decode_lockowner(xdr, args);
+	return decode_lockowner(xdr, args);
-out:
-	dprintk("%s: exit with status = %d\n", __func__, ntohl(status));
-	return status;
 }
 
 #endif /* CONFIG_NFS_V4_1 */
@@ -707,7 +663,6 @@ static __be32 encode_getattr_res(struct svc_rqst *rqstp, struct xdr_stream *xdr,
 	status = encode_attr_mtime(xdr, res->bitmap, &res->mtime);
 	*savep = htonl((unsigned int)((char *)xdr->p - (char *)(savep+1)));
 out:
-	dprintk("%s: exit with status = %d\n", __func__, ntohl(status));
 	return status;
 }
 
@@ -734,11 +689,11 @@ static __be32 encode_cb_sequence_res(struct svc_rqst *rqstp,
 	__be32 status = res->csr_status;
 
 	if (unlikely(status != 0))
-		goto out;
+		return status;
 
 	status = encode_sessionid(xdr, &res->csr_sessionid);
 	if (status)
-		goto out;
+		return status;
 
 	p = xdr_reserve_space(xdr, 4 * sizeof(uint32_t));
 	if (unlikely(p == NULL))
@@ -748,9 +703,7 @@ static __be32 encode_cb_sequence_res(struct svc_rqst *rqstp,
 	*p++ = htonl(res->csr_slotid);
 	*p++ = htonl(res->csr_highestslotid);
 	*p++ = htonl(res->csr_target_highestslotid);
-out:
+	return 0;
-	dprintk("%s: exit with status = %d\n", __func__, ntohl(status));
-	return status;
 }
 
 static __be32
@@ -871,14 +824,10 @@ static __be32 process_op(int nop, struct svc_rqst *rqstp,
 	long maxlen;
 	__be32 res;
 
-	dprintk("%s: start\n", __func__);
 	status = decode_op_hdr(xdr_in, &op_nr);
 	if (unlikely(status))
 		return status;
 
-	dprintk("%s: minorversion=%d nop=%d op_nr=%u\n",
-		__func__, cps->minorversion, nop, op_nr);
-
 	switch (cps->minorversion) {
 	case 0:
 		status = preprocess_nfs4_op(op_nr, &op);
@@ -917,7 +866,6 @@ encode_hdr:
 		return res;
 	if (op->encode_res != NULL && status == 0)
 		status = op->encode_res(rqstp, xdr_out, resp);
-	dprintk("%s: done, status = %d\n", __func__, ntohl(status));
 	return status;
 }
 
@@ -937,8 +885,6 @@ static __be32 nfs4_callback_compound(struct svc_rqst *rqstp, void *argp, void *r
 	};
 	unsigned int nops = 0;
 
-	dprintk("%s: start\n", __func__);
-
 	xdr_init_decode(&xdr_in, &rqstp->rq_arg, rqstp->rq_arg.head[0].iov_base);
 
 	p = (__be32*)((char *)rqstp->rq_res.head[0].iov_base + rqstp->rq_res.head[0].iov_len);
@@ -977,7 +923,6 @@ static __be32 nfs4_callback_compound(struct svc_rqst *rqstp, void *argp, void *r
 	*hdr_res.nops = htonl(nops);
 	nfs4_cb_free_slot(&cps);
 	nfs_put_client(cps.clp);
-	dprintk("%s: done, status = %u\n", __func__, ntohl(status));
 	return rpc_success;
 
 out_invalidcred:

fs/nfs/client.c

@@ -218,6 +218,7 @@ static void nfs_cb_idr_remove_locked(struct nfs_client *clp)
 static void pnfs_init_server(struct nfs_server *server)
 {
 	rpc_init_wait_queue(&server->roc_rpcwaitq, "pNFS ROC");
+	rpc_init_wait_queue(&server->uoc_rpcwaitq, "NFS UOC");
 }
 
 #else
@@ -240,8 +241,6 @@ static void pnfs_init_server(struct nfs_server *server)
  */
 void nfs_free_client(struct nfs_client *clp)
 {
-	dprintk("--> nfs_free_client(%u)\n", clp->rpc_ops->version);
-
 	nfs_fscache_release_client_cookie(clp);
 
 	/* -EIO all pending I/O */
@@ -256,8 +255,6 @@ void nfs_free_client(struct nfs_client *clp)
 	kfree(clp->cl_hostname);
 	kfree(clp->cl_acceptor);
 	kfree(clp);
-
-	dprintk("<-- nfs_free_client()\n");
 }
 EXPORT_SYMBOL_GPL(nfs_free_client);
 
@@ -271,7 +268,6 @@ void nfs_put_client(struct nfs_client *clp)
 	if (!clp)
 		return;
 
-	dprintk("--> nfs_put_client({%d})\n", atomic_read(&clp->cl_count));
 	nn = net_generic(clp->cl_net, nfs_net_id);
 
 	if (atomic_dec_and_lock(&clp->cl_count, &nn->nfs_client_lock)) {
@@ -382,9 +378,6 @@ nfs_found_client(const struct nfs_client_initdata *cl_init,
 	}
 
 	smp_rmb();
-
-	dprintk("<-- %s found nfs_client %p for %s\n",
-		__func__, clp, cl_init->hostname ?: "");
 	return clp;
 }
 
@@ -403,9 +396,6 @@ struct nfs_client *nfs_get_client(const struct nfs_client_initdata *cl_init)
 		return NULL;
 	}
 
-	dprintk("--> nfs_get_client(%s,v%u)\n",
-		cl_init->hostname, rpc_ops->version);
-
 	/* see if the client already exists */
 	do {
 		spin_lock(&nn->nfs_client_lock);
@@ -430,8 +420,6 @@ struct nfs_client *nfs_get_client(const struct nfs_client_initdata *cl_init)
 		new = rpc_ops->alloc_client(cl_init);
 	} while (!IS_ERR(new));
 
-	dprintk("<-- nfs_get_client() Failed to find %s (%ld)\n",
-		cl_init->hostname, PTR_ERR(new));
 	return new;
 }
 EXPORT_SYMBOL_GPL(nfs_get_client);
@@ -558,6 +546,7 @@ static int nfs_start_lockd(struct nfs_server *server)
 		.noresvport	= server->flags & NFS_MOUNT_NORESVPORT ?
 					1 : 0,
 		.net		= clp->cl_net,
+		.nlmclnt_ops 	= clp->cl_nfs_mod->rpc_ops->nlmclnt_ops,
 	};
 
 	if (nlm_init.nfs_version > 3)
@@ -624,27 +613,21 @@ struct nfs_client *nfs_init_client(struct nfs_client *clp,
 {
 	int error;
 
-	if (clp->cl_cons_state == NFS_CS_READY) {
+	/* the client is already initialised */
-		/* the client is already initialised */
+	if (clp->cl_cons_state == NFS_CS_READY)
-		dprintk("<-- nfs_init_client() = 0 [already %p]\n", clp);
 		return clp;
-	}
 
 	/*
 	 * Create a client RPC handle for doing FSSTAT with UNIX auth only
 	 * - RFC 2623, sec 2.3.2
 	 */
 	error = nfs_create_rpc_client(clp, cl_init, RPC_AUTH_UNIX);
-	if (error < 0)
+	nfs_mark_client_ready(clp, error == 0 ? NFS_CS_READY : error);
-		goto error;
+	if (error < 0) {
-	nfs_mark_client_ready(clp, NFS_CS_READY);
+		nfs_put_client(clp);
+		clp = ERR_PTR(error);
+	}
 	return clp;
-
-error:
-	nfs_mark_client_ready(clp, error);
-	nfs_put_client(clp);
-	dprintk("<-- nfs_init_client() = xerror %d\n", error);
-	return ERR_PTR(error);
 }
 EXPORT_SYMBOL_GPL(nfs_init_client);
 
@@ -668,8 +651,6 @@ static int nfs_init_server(struct nfs_server *server,
 	struct nfs_client *clp;
 	int error;
 
-	dprintk("--> nfs_init_server()\n");
-
 	nfs_init_timeout_values(&timeparms, data->nfs_server.protocol,
 			data->timeo, data->retrans);
 	if (data->flags & NFS_MOUNT_NORESVPORT)
@@ -677,10 +658,8 @@ static int nfs_init_server(struct nfs_server *server,
 
 	/* Allocate or find a client reference we can use */
 	clp = nfs_get_client(&cl_init);
-	if (IS_ERR(clp)) {
+	if (IS_ERR(clp))
-		dprintk("<-- nfs_init_server() = error %ld\n", PTR_ERR(clp));
 		return PTR_ERR(clp);
-	}
 
 	server->nfs_client = clp;
 
@@ -725,13 +704,11 @@ static int nfs_init_server(struct nfs_server *server,
 	server->mountd_protocol = data->mount_server.protocol;
 
 	server->namelen  = data->namlen;
-	dprintk("<-- nfs_init_server() = 0 [new %p]\n", clp);
 	return 0;
 
 error:
 	server->nfs_client = NULL;
 	nfs_put_client(clp);
-	dprintk("<-- nfs_init_server() = xerror %d\n", error);
 	return error;
 }
 
@@ -798,12 +775,10 @@ int nfs_probe_fsinfo(struct nfs_server *server, struct nfs_fh *mntfh, struct nfs
 	struct nfs_client *clp = server->nfs_client;
 	int error;
 
-	dprintk("--> nfs_probe_fsinfo()\n");
-
 	if (clp->rpc_ops->set_capabilities != NULL) {
 		error = clp->rpc_ops->set_capabilities(server, mntfh);
 		if (error < 0)
-			goto out_error;
+			return error;
 	}
 
 	fsinfo.fattr = fattr;
@@ -811,7 +786,7 @@ int nfs_probe_fsinfo(struct nfs_server *server, struct nfs_fh *mntfh, struct nfs
 	memset(fsinfo.layouttype, 0, sizeof(fsinfo.layouttype));
 	error = clp->rpc_ops->fsinfo(server, mntfh, &fsinfo);
 	if (error < 0)
-		goto out_error;
+		return error;
 
 	nfs_server_set_fsinfo(server, &fsinfo);
 
@@ -826,12 +801,7 @@ int nfs_probe_fsinfo(struct nfs_server *server, struct nfs_fh *mntfh, struct nfs
 			server->namelen = pathinfo.max_namelen;
 	}
 
-	dprintk("<-- nfs_probe_fsinfo() = 0\n");
 	return 0;
-
-out_error:
-	dprintk("nfs_probe_fsinfo: error = %d\n", -error);
-	return error;
 }
 EXPORT_SYMBOL_GPL(nfs_probe_fsinfo);
 
@@ -927,8 +897,6 @@ EXPORT_SYMBOL_GPL(nfs_alloc_server);
  */
 void nfs_free_server(struct nfs_server *server)
 {
-	dprintk("--> nfs_free_server()\n");
-
 	nfs_server_remove_lists(server);
 
 	if (server->destroy != NULL)
@@ -946,7 +914,6 @@ void nfs_free_server(struct nfs_server *server)
 	nfs_free_iostats(server->io_stats);
 	kfree(server);
 	nfs_release_automount_timer();
-	dprintk("<-- nfs_free_server()\n");
 }
 EXPORT_SYMBOL_GPL(nfs_free_server);
 
@@ -1026,10 +993,6 @@ struct nfs_server *nfs_clone_server(struct nfs_server *source,
 	struct nfs_fattr *fattr_fsinfo;
 	int error;
 
-	dprintk("--> nfs_clone_server(,%llx:%llx,)\n",
-		(unsigned long long) fattr->fsid.major,
-		(unsigned long long) fattr->fsid.minor);
-
 	server = nfs_alloc_server();
 	if (!server)
 		return ERR_PTR(-ENOMEM);
@@ -1061,10 +1024,6 @@ struct nfs_server *nfs_clone_server(struct nfs_server *source,
 	if (server->namelen == 0 || server->namelen > NFS4_MAXNAMLEN)
 		server->namelen = NFS4_MAXNAMLEN;
 
-	dprintk("Cloned FSID: %llx:%llx\n",
-		(unsigned long long) server->fsid.major,
-		(unsigned long long) server->fsid.minor);
-
 	error = nfs_start_lockd(server);
 	if (error < 0)
 		goto out_free_server;
@@ -1073,13 +1032,11 @@ struct nfs_server *nfs_clone_server(struct nfs_server *source,
 	server->mount_time = jiffies;
 
 	nfs_free_fattr(fattr_fsinfo);
-	dprintk("<-- nfs_clone_server() = %p\n", server);
 	return server;
 
 out_free_server:
 	nfs_free_fattr(fattr_fsinfo);
 	nfs_free_server(server);
-	dprintk("<-- nfs_clone_server() = error %d\n", error);
 	return ERR_PTR(error);
 }
 EXPORT_SYMBOL_GPL(nfs_clone_server);

fs/nfs/dir.c

@@ -57,7 +57,7 @@ static void nfs_readdir_clear_array(struct page*);
 const struct file_operations nfs_dir_operations = {
 	.llseek		= nfs_llseek_dir,
 	.read		= generic_read_dir,
-	.iterate_shared	= nfs_readdir,
+	.iterate	= nfs_readdir,
 	.open		= nfs_opendir,
 	.release	= nfs_closedir,
 	.fsync		= nfs_fsync_dir,
@@ -145,7 +145,6 @@ struct nfs_cache_array_entry {
 };
 
 struct nfs_cache_array {
-	atomic_t refcount;
 	int size;
 	int eof_index;
 	u64 last_cookie;
@@ -170,27 +169,6 @@ typedef struct {
 	unsigned int	eof:1;
 } nfs_readdir_descriptor_t;
 
-/*
- * The caller is responsible for calling nfs_readdir_release_array(page)
- */
-static
-struct nfs_cache_array *nfs_readdir_get_array(struct page *page)
-{
-	void *ptr;
-	if (page == NULL)
-		return ERR_PTR(-EIO);
-	ptr = kmap(page);
-	if (ptr == NULL)
-		return ERR_PTR(-ENOMEM);
-	return ptr;
-}
-
-static
-void nfs_readdir_release_array(struct page *page)
-{
-	kunmap(page);
-}
-
 /*
  * we are freeing strings created by nfs_add_to_readdir_array()
  */
@@ -201,20 +179,11 @@ void nfs_readdir_clear_array(struct page *page)
 	int i;
 
 	array = kmap_atomic(page);
-	if (atomic_dec_and_test(&array->refcount))
+	for (i = 0; i < array->size; i++)
-		for (i = 0; i < array->size; i++)
+		kfree(array->array[i].string.name);
-			kfree(array->array[i].string.name);
 	kunmap_atomic(array);
 }
 
-static bool grab_page(struct page *page)
-{
-	struct nfs_cache_array *array = kmap_atomic(page);
-	bool res = atomic_inc_not_zero(&array->refcount);
-	kunmap_atomic(array);
-	return res;
-}
-
 /*
  * the caller is responsible for freeing qstr.name
  * when called by nfs_readdir_add_to_array, the strings will be freed in
@@ -239,13 +208,10 @@ int nfs_readdir_make_qstr(struct qstr *string, const char *name, unsigned int le
 static
 int nfs_readdir_add_to_array(struct nfs_entry *entry, struct page *page)
 {
-	struct nfs_cache_array *array = nfs_readdir_get_array(page);
+	struct nfs_cache_array *array = kmap(page);
 	struct nfs_cache_array_entry *cache_entry;
 	int ret;
 
-	if (IS_ERR(array))
-		return PTR_ERR(array);
-
 	cache_entry = &array->array[array->size];
 
 	/* Check that this entry lies within the page bounds */
@@ -264,7 +230,7 @@ int nfs_readdir_add_to_array(struct nfs_entry *entry, struct page *page)
 	if (entry->eof != 0)
 		array->eof_index = array->size;
 out:
-	nfs_readdir_release_array(page);
+	kunmap(page);
 	return ret;
 }
 
@@ -353,11 +319,7 @@ int nfs_readdir_search_array(nfs_readdir_descriptor_t *desc)
 	struct nfs_cache_array *array;
 	int status;
 
-	array = nfs_readdir_get_array(desc->page);
+	array = kmap(desc->page);
-	if (IS_ERR(array)) {
-		status = PTR_ERR(array);
-		goto out;
-	}
 
 	if (*desc->dir_cookie == 0)
 		status = nfs_readdir_search_for_pos(array, desc);
@@ -369,8 +331,7 @@ int nfs_readdir_search_array(nfs_readdir_descriptor_t *desc)
 		desc->current_index += array->size;
 		desc->page_index++;
 	}
-	nfs_readdir_release_array(desc->page);
+	kunmap(desc->page);
-out:
 	return status;
 }
 
@@ -606,13 +567,10 @@ int nfs_readdir_page_filler(nfs_readdir_descriptor_t *desc, struct nfs_entry *en
 
 out_nopages:
 	if (count == 0 || (status == -EBADCOOKIE && entry->eof != 0)) {
-		array = nfs_readdir_get_array(page);
+		array = kmap(page);
-		if (!IS_ERR(array)) {
+		array->eof_index = array->size;
-			array->eof_index = array->size;
+		status = 0;
-			status = 0;
+		kunmap(page);
-			nfs_readdir_release_array(page);
-		} else
-			status = PTR_ERR(array);
 	}
 
 	put_page(scratch);
@@ -674,13 +632,8 @@ int nfs_readdir_xdr_to_array(nfs_readdir_descriptor_t *desc, struct page *page,
 		goto out;
 	}
 
-	array = nfs_readdir_get_array(page);
+	array = kmap(page);
-	if (IS_ERR(array)) {
-		status = PTR_ERR(array);
-		goto out_label_free;
-	}
 	memset(array, 0, sizeof(struct nfs_cache_array));
-	atomic_set(&array->refcount, 1);
 	array->eof_index = -1;
 
 	status = nfs_readdir_alloc_pages(pages, array_size);
@@ -703,8 +656,7 @@ int nfs_readdir_xdr_to_array(nfs_readdir_descriptor_t *desc, struct page *page,
 
 	nfs_readdir_free_pages(pages, array_size);
 out_release_array:
-	nfs_readdir_release_array(page);
+	kunmap(page);
-out_label_free:
 	nfs4_label_free(entry.label);
 out:
 	nfs_free_fattr(entry.fattr);
@@ -743,7 +695,8 @@ int nfs_readdir_filler(nfs_readdir_descriptor_t *desc, struct page* page)
 static
 void cache_page_release(nfs_readdir_descriptor_t *desc)
 {
-	nfs_readdir_clear_array(desc->page);
+	if (!desc->page->mapping)
+		nfs_readdir_clear_array(desc->page);
 	put_page(desc->page);
 	desc->page = NULL;
 }
@@ -751,16 +704,8 @@ void cache_page_release(nfs_readdir_descriptor_t *desc)
 static
 struct page *get_cache_page(nfs_readdir_descriptor_t *desc)
 {
-	struct page *page;
+	return read_cache_page(desc->file->f_mapping,
-
-	for (;;) {
-		page = read_cache_page(desc->file->f_mapping,
 			desc->page_index, (filler_t *)nfs_readdir_filler, desc);
-		if (IS_ERR(page) || grab_page(page))
-			break;
-		put_page(page);
-	}
-	return page;
 }
 
 /*
@@ -809,12 +754,7 @@ int nfs_do_filldir(nfs_readdir_descriptor_t *desc)
 	struct nfs_cache_array *array = NULL;
 	struct nfs_open_dir_context *ctx = file->private_data;
 
-	array = nfs_readdir_get_array(desc->page);
+	array = kmap(desc->page);
-	if (IS_ERR(array)) {
-		res = PTR_ERR(array);
-		goto out;
-	}
-
 	for (i = desc->cache_entry_index; i < array->size; i++) {
 		struct nfs_cache_array_entry *ent;
 
@@ -835,8 +775,7 @@ int nfs_do_filldir(nfs_readdir_descriptor_t *desc)
 	if (array->eof_index >= 0)
 		desc->eof = 1;
 
-	nfs_readdir_release_array(desc->page);
+	kunmap(desc->page);
-out:
 	cache_page_release(desc);
 	dfprintk(DIRCACHE, "NFS: nfs_do_filldir() filling ended @ cookie %Lu; returning = %d\n",
 			(unsigned long long)*desc->dir_cookie, res);
@@ -966,11 +905,13 @@ out:
 
 static loff_t nfs_llseek_dir(struct file *filp, loff_t offset, int whence)
 {
+	struct inode *inode = file_inode(filp);
 	struct nfs_open_dir_context *dir_ctx = filp->private_data;
 
 	dfprintk(FILE, "NFS: llseek dir(%pD2, %lld, %d)\n",
 			filp, offset, whence);
 
+	inode_lock(inode);
 	switch (whence) {
 		case 1:
 			offset += filp->f_pos;
@@ -978,13 +919,16 @@ static loff_t nfs_llseek_dir(struct file *filp, loff_t offset, int whence)
 			if (offset >= 0)
 				break;
 		default:
-			return -EINVAL;
+			offset = -EINVAL;
+			goto out;
 	}
 	if (offset != filp->f_pos) {
 		filp->f_pos = offset;
 		dir_ctx->dir_cookie = 0;
 		dir_ctx->duped = 0;
 	}
+out:
+	inode_unlock(inode);
 	return offset;
 }
 

fs/nfs/direct.c

@@ -392,16 +392,6 @@ static void nfs_direct_complete(struct nfs_direct_req *dreq)
 	nfs_direct_req_release(dreq);
 }
 
-static void nfs_direct_readpage_release(struct nfs_page *req)
-{
-	dprintk("NFS: direct read done (%s/%llu %d@%lld)\n",
-		req->wb_context->dentry->d_sb->s_id,
-		(unsigned long long)NFS_FILEID(d_inode(req->wb_context->dentry)),
-		req->wb_bytes,
-		(long long)req_offset(req));
-	nfs_release_request(req);
-}
-
 static void nfs_direct_read_completion(struct nfs_pgio_header *hdr)
 {
 	unsigned long bytes = 0;
@@ -426,7 +416,7 @@ static void nfs_direct_read_completion(struct nfs_pgio_header *hdr)
 			set_page_dirty(page);
 		bytes += req->wb_bytes;
 		nfs_list_remove_request(req);
-		nfs_direct_readpage_release(req);
+		nfs_release_request(req);
 	}
 out_put:
 	if (put_dreq(dreq))
@@ -700,16 +690,9 @@ static void nfs_direct_commit_complete(struct nfs_commit_data *data)
 	int status = data->task.tk_status;
 
 	nfs_init_cinfo_from_dreq(&cinfo, dreq);
-	if (status < 0) {
+	if (status < 0 || nfs_direct_cmp_commit_data_verf(dreq, data))
-		dprintk("NFS: %5u commit failed with error %d.\n",
-			data->task.tk_pid, status);
 		dreq->flags = NFS_ODIRECT_RESCHED_WRITES;
-	} else if (nfs_direct_cmp_commit_data_verf(dreq, data)) {
-		dprintk("NFS: %5u commit verify failed\n", data->task.tk_pid);
-		dreq->flags = NFS_ODIRECT_RESCHED_WRITES;
-	}
 
-	dprintk("NFS: %5u commit returned %d\n", data->task.tk_pid, status);
 	while (!list_empty(&data->pages)) {
 		req = nfs_list_entry(data->pages.next);
 		nfs_list_remove_request(req);

fs/nfs/file.c

@@ -482,7 +482,7 @@ static int nfs_launder_page(struct page *page)
 		inode->i_ino, (long long)page_offset(page));
 
 	nfs_fscache_wait_on_page_write(nfsi, page);
-	return nfs_wb_launder_page(inode, page);
+	return nfs_wb_page(inode, page);
 }
 
 static int nfs_swap_activate(struct swap_info_struct *sis, struct file *file,
@@ -697,14 +697,14 @@ do_unlk(struct file *filp, int cmd, struct file_lock *fl, int is_local)
 	if (!IS_ERR(l_ctx)) {
 		status = nfs_iocounter_wait(l_ctx);
 		nfs_put_lock_context(l_ctx);
-		if (status < 0)
+		/*  NOTE: special case
+		 * 	If we're signalled while cleaning up locks on process exit, we
+		 * 	still need to complete the unlock.
+		 */
+		if (status < 0 && !(fl->fl_flags & FL_CLOSE))
 			return status;
 	}
 
-	/* NOTE: special case
-	 * 	If we're signalled while cleaning up locks on process exit, we
-	 * 	still need to complete the unlock.
-	 */
 	/*
 	 * Use local locking if mounted with "-onolock" or with appropriate
 	 * "-olocal_lock="
@@ -820,9 +820,23 @@ int nfs_flock(struct file *filp, int cmd, struct file_lock *fl)
 	if (NFS_SERVER(inode)->flags & NFS_MOUNT_LOCAL_FLOCK)
 		is_local = 1;
 
-	/* We're simulating flock() locks using posix locks on the server */
+	/*
-	if (fl->fl_type == F_UNLCK)
+	 * VFS doesn't require the open mode to match a flock() lock's type.
+	 * NFS, however, may simulate flock() locking with posix locking which
+	 * requires the open mode to match the lock type.
+	 */
+	switch (fl->fl_type) {
+	case F_UNLCK:
 		return do_unlk(filp, cmd, fl, is_local);
+	case F_RDLCK:
+		if (!(filp->f_mode & FMODE_READ))
+			return -EBADF;
+		break;
+	case F_WRLCK:
+		if (!(filp->f_mode & FMODE_WRITE))
+			return -EBADF;
+	}
+
 	return do_setlk(filp, cmd, fl, is_local);
 }
 EXPORT_SYMBOL_GPL(nfs_flock);

fs/nfs/filelayout/filelayout.c

@@ -921,11 +921,11 @@ fl_pnfs_update_layout(struct inode *ino,
 	fl = FILELAYOUT_LSEG(lseg);
 
 	status = filelayout_check_deviceid(lo, fl, gfp_flags);
-	if (status)
+	if (status) {
-		lseg = ERR_PTR(status);
-out:
-	if (IS_ERR(lseg))
 		pnfs_put_lseg(lseg);
+		lseg = ERR_PTR(status);
+	}
+out:
 	return lseg;
 }
 
@@ -933,6 +933,7 @@ static void
 filelayout_pg_init_read(struct nfs_pageio_descriptor *pgio,
 			struct nfs_page *req)
 {
+	pnfs_generic_pg_check_layout(pgio);
 	if (!pgio->pg_lseg) {
 		pgio->pg_lseg = fl_pnfs_update_layout(pgio->pg_inode,
 						      req->wb_context,
@@ -959,6 +960,7 @@ filelayout_pg_init_write(struct nfs_pageio_descriptor *pgio,
 	struct nfs_commit_info cinfo;
 	int status;
 
+	pnfs_generic_pg_check_layout(pgio);
 	if (!pgio->pg_lseg) {
 		pgio->pg_lseg = fl_pnfs_update_layout(pgio->pg_inode,
 						      req->wb_context,

fs/nfs/flexfilelayout/flexfilelayout.c

@@ -846,6 +846,7 @@ ff_layout_pg_init_read(struct nfs_pageio_descriptor *pgio,
 	int ds_idx;
 
 retry:
+	pnfs_generic_pg_check_layout(pgio);
 	/* Use full layout for now */
 	if (!pgio->pg_lseg)
 		ff_layout_pg_get_read(pgio, req, false);
@@ -894,6 +895,7 @@ ff_layout_pg_init_write(struct nfs_pageio_descriptor *pgio,
 	int status;
 
 retry:
+	pnfs_generic_pg_check_layout(pgio);
 	if (!pgio->pg_lseg) {
 		pgio->pg_lseg = pnfs_update_layout(pgio->pg_inode,
 						   req->wb_context,
@@ -1800,16 +1802,16 @@ ff_layout_write_pagelist(struct nfs_pgio_header *hdr, int sync)
 
 	ds = nfs4_ff_layout_prepare_ds(lseg, idx, true);
 	if (!ds)
-		return PNFS_NOT_ATTEMPTED;
+		goto out_failed;
 
 	ds_clnt = nfs4_ff_find_or_create_ds_client(lseg, idx, ds->ds_clp,
 						   hdr->inode);
 	if (IS_ERR(ds_clnt))
-		return PNFS_NOT_ATTEMPTED;
+		goto out_failed;
 
 	ds_cred = ff_layout_get_ds_cred(lseg, idx, hdr->cred);
 	if (!ds_cred)
-		return PNFS_NOT_ATTEMPTED;
+		goto out_failed;
 
 	vers = nfs4_ff_layout_ds_version(lseg, idx);
 
@@ -1839,6 +1841,11 @@ ff_layout_write_pagelist(struct nfs_pgio_header *hdr, int sync)
 			  sync, RPC_TASK_SOFTCONN);
 	put_rpccred(ds_cred);
 	return PNFS_ATTEMPTED;
+
+out_failed:
+	if (ff_layout_avoid_mds_available_ds(lseg))
+		return PNFS_TRY_AGAIN;
+	return PNFS_NOT_ATTEMPTED;
 }
 
 static u32 calc_ds_index_from_commit(struct pnfs_layout_segment *lseg, u32 i)
@@ -2354,10 +2361,21 @@ ff_layout_prepare_layoutstats(struct nfs42_layoutstat_args *args)
 	return 0;
 }
 
+static int
+ff_layout_set_layoutdriver(struct nfs_server *server,
+		const struct nfs_fh *dummy)
+{
+#if IS_ENABLED(CONFIG_NFS_V4_2)
+	server->caps |= NFS_CAP_LAYOUTSTATS;
+#endif
+	return 0;
+}
+
 static struct pnfs_layoutdriver_type flexfilelayout_type = {
 	.id			= LAYOUT_FLEX_FILES,
 	.name			= "LAYOUT_FLEX_FILES",
 	.owner			= THIS_MODULE,
+	.set_layoutdriver	= ff_layout_set_layoutdriver,
 	.alloc_layout_hdr	= ff_layout_alloc_layout_hdr,
 	.free_layout_hdr	= ff_layout_free_layout_hdr,
 	.alloc_lseg		= ff_layout_alloc_lseg,

fs/nfs/flexfilelayout/flexfilelayoutdev.c

@@ -119,7 +119,13 @@ nfs4_ff_alloc_deviceid_node(struct nfs_server *server, struct pnfs_device *pdev,
 		if (ds_versions[i].wsize > NFS_MAX_FILE_IO_SIZE)
 			ds_versions[i].wsize = NFS_MAX_FILE_IO_SIZE;
 
-		if (ds_versions[i].version != 3 || ds_versions[i].minor_version != 0) {
+		/*
+		 * check for valid major/minor combination.
+		 * currently we support dataserver which talk:
+		 *   v3, v4.0, v4.1, v4.2
+		 */
+		if (!((ds_versions[i].version == 3 && ds_versions[i].minor_version == 0) ||
+			(ds_versions[i].version == 4 && ds_versions[i].minor_version < 3))) {
 			dprintk("%s: [%d] unsupported ds version %d-%d\n", __func__,
 				i, ds_versions[i].version,
 				ds_versions[i].minor_version);
@@ -415,7 +421,7 @@ nfs4_ff_layout_prepare_ds(struct pnfs_layout_segment *lseg, u32 ds_idx,
 			     mirror->mirror_ds->ds_versions[0].minor_version);
 
 	/* connect success, check rsize/wsize limit */
-	if (ds->ds_clp) {
+	if (!status) {
 		max_payload =
 			nfs_block_size(rpc_max_payload(ds->ds_clp->cl_rpcclient),
 				       NULL);

fs/nfs/inode.c

@@ -734,7 +734,10 @@ int nfs_getattr(const struct path *path, struct kstat *stat,
 	if (need_atime || nfs_need_revalidate_inode(inode)) {
 		struct nfs_server *server = NFS_SERVER(inode);
 
-		nfs_readdirplus_parent_cache_miss(path->dentry);
+		if (!(server->flags & NFS_MOUNT_NOAC))
+			nfs_readdirplus_parent_cache_miss(path->dentry);
+		else
+			nfs_readdirplus_parent_cache_hit(path->dentry);
 		err = __nfs_revalidate_inode(server, inode);
 	} else
 		nfs_readdirplus_parent_cache_hit(path->dentry);

fs/nfs/internal.h

@@ -495,7 +495,6 @@ void nfs_mark_request_commit(struct nfs_page *req,
 			     u32 ds_commit_idx);
 int nfs_write_need_commit(struct nfs_pgio_header *);
 void nfs_writeback_update_inode(struct nfs_pgio_header *hdr);
-int nfs_commit_file(struct file *file, struct nfs_write_verifier *verf);
 int nfs_generic_commit_list(struct inode *inode, struct list_head *head,
 			    int how, struct nfs_commit_info *cinfo);
 void nfs_retry_commit(struct list_head *page_list,
@@ -756,9 +755,13 @@ static inline bool nfs_error_is_fatal(int err)
 {
 	switch (err) {
 	case -ERESTARTSYS:
+	case -EACCES:
+	case -EDQUOT:
+	case -EFBIG:
 	case -EIO:
 	case -ENOSPC:
 	case -EROFS:
+	case -ESTALE:
 	case -E2BIG:
 		return true;
 	default:

fs/nfs/namespace.c

@@ -143,11 +143,8 @@ struct vfsmount *nfs_d_automount(struct path *path)
 	struct nfs_fh *fh = NULL;
 	struct nfs_fattr *fattr = NULL;
 
-	dprintk("--> nfs_d_automount()\n");
-
-	mnt = ERR_PTR(-ESTALE);
 	if (IS_ROOT(path->dentry))
-		goto out_nofree;
+		return ERR_PTR(-ESTALE);
 
 	mnt = ERR_PTR(-ENOMEM);
 	fh = nfs_alloc_fhandle();
@@ -155,13 +152,10 @@ struct vfsmount *nfs_d_automount(struct path *path)
 	if (fh == NULL || fattr == NULL)
 		goto out;
 
-	dprintk("%s: enter\n", __func__);
-
 	mnt = server->nfs_client->rpc_ops->submount(server, path->dentry, fh, fattr);
 	if (IS_ERR(mnt))
 		goto out;
 
-	dprintk("%s: done, success\n", __func__);
 	mntget(mnt); /* prevent immediate expiration */
 	mnt_set_expiry(mnt, &nfs_automount_list);
 	schedule_delayed_work(&nfs_automount_task, nfs_mountpoint_expiry_timeout);
@@ -169,11 +163,6 @@ struct vfsmount *nfs_d_automount(struct path *path)
 out:
 	nfs_free_fattr(fattr);
 	nfs_free_fhandle(fh);
-out_nofree:
-	if (IS_ERR(mnt))
-		dprintk("<-- %s(): error %ld\n", __func__, PTR_ERR(mnt));
-	else
-		dprintk("<-- %s() = %p\n", __func__, mnt);
 	return mnt;
 }
 
@@ -248,27 +237,20 @@ struct vfsmount *nfs_do_submount(struct dentry *dentry, struct nfs_fh *fh,
 		.fattr = fattr,
 		.authflavor = authflavor,
 	};
-	struct vfsmount *mnt = ERR_PTR(-ENOMEM);
+	struct vfsmount *mnt;
 	char *page = (char *) __get_free_page(GFP_USER);
 	char *devname;
 
-	dprintk("--> nfs_do_submount()\n");
-
-	dprintk("%s: submounting on %pd2\n", __func__,
-			dentry);
 	if (page == NULL)
-		goto out;
+		return ERR_PTR(-ENOMEM);
-	devname = nfs_devname(dentry, page, PAGE_SIZE);
-	mnt = (struct vfsmount *)devname;
-	if (IS_ERR(devname))
-		goto free_page;
-	mnt = nfs_do_clone_mount(NFS_SB(dentry->d_sb), devname, &mountdata);
-free_page:
-	free_page((unsigned long)page);
-out:
-	dprintk("%s: done\n", __func__);
 
-	dprintk("<-- nfs_do_submount() = %p\n", mnt);
+	devname = nfs_devname(dentry, page, PAGE_SIZE);
+	if (IS_ERR(devname))
+		mnt = (struct vfsmount *)devname;
+	else
+		mnt = nfs_do_clone_mount(NFS_SB(dentry->d_sb), devname, &mountdata);
+
+	free_page((unsigned long)page);
 	return mnt;
 }
 EXPORT_SYMBOL_GPL(nfs_do_submount);

fs/nfs/nfs3proc.c

@@ -865,12 +865,63 @@ static void nfs3_proc_commit_setup(struct nfs_commit_data *data, struct rpc_mess
 	msg->rpc_proc = &nfs3_procedures[NFS3PROC_COMMIT];
 }
 
+static void nfs3_nlm_alloc_call(void *data)
+{
+	struct nfs_lock_context *l_ctx = data;
+	if (l_ctx && test_bit(NFS_CONTEXT_UNLOCK, &l_ctx->open_context->flags)) {
+		get_nfs_open_context(l_ctx->open_context);
+		nfs_get_lock_context(l_ctx->open_context);
+	}
+}
+
+static bool nfs3_nlm_unlock_prepare(struct rpc_task *task, void *data)
+{
+	struct nfs_lock_context *l_ctx = data;
+	if (l_ctx && test_bit(NFS_CONTEXT_UNLOCK, &l_ctx->open_context->flags))
+		return nfs_async_iocounter_wait(task, l_ctx);
+	return false;
+
+}
+
+static void nfs3_nlm_release_call(void *data)
+{
+	struct nfs_lock_context *l_ctx = data;
+	struct nfs_open_context *ctx;
+	if (l_ctx && test_bit(NFS_CONTEXT_UNLOCK, &l_ctx->open_context->flags)) {
+		ctx = l_ctx->open_context;
+		nfs_put_lock_context(l_ctx);
+		put_nfs_open_context(ctx);
+	}
+}
+
+const struct nlmclnt_operations nlmclnt_fl_close_lock_ops = {
+	.nlmclnt_alloc_call = nfs3_nlm_alloc_call,
+	.nlmclnt_unlock_prepare = nfs3_nlm_unlock_prepare,
+	.nlmclnt_release_call = nfs3_nlm_release_call,
+};
+
 static int
 nfs3_proc_lock(struct file *filp, int cmd, struct file_lock *fl)
 {
 	struct inode *inode = file_inode(filp);
+	struct nfs_lock_context *l_ctx = NULL;
+	struct nfs_open_context *ctx = nfs_file_open_context(filp);
+	int status;
 
-	return nlmclnt_proc(NFS_SERVER(inode)->nlm_host, cmd, fl);
+	if (fl->fl_flags & FL_CLOSE) {
+		l_ctx = nfs_get_lock_context(ctx);
+		if (IS_ERR(l_ctx))
+			l_ctx = NULL;
+		else
+			set_bit(NFS_CONTEXT_UNLOCK, &ctx->flags);
+	}
+
+	status = nlmclnt_proc(NFS_SERVER(inode)->nlm_host, cmd, fl, l_ctx);
+
+	if (l_ctx)
+		nfs_put_lock_context(l_ctx);
+
+	return status;
 }
 
 static int nfs3_have_delegation(struct inode *inode, fmode_t flags)
@@ -921,6 +972,7 @@ const struct nfs_rpc_ops nfs_v3_clientops = {
 	.dir_inode_ops	= &nfs3_dir_inode_operations,
 	.file_inode_ops	= &nfs3_file_inode_operations,
 	.file_ops	= &nfs_file_operations,
+	.nlmclnt_ops	= &nlmclnt_fl_close_lock_ops,
 	.getroot	= nfs3_proc_get_root,
 	.submount	= nfs_submount,
 	.try_mount	= nfs_try_mount,

fs/nfs/nfs42proc.c

@@ -167,23 +167,29 @@ static ssize_t _nfs42_proc_copy(struct file *src,
 	if (status)
 		return status;
 
+	res->commit_res.verf = kzalloc(sizeof(struct nfs_writeverf), GFP_NOFS);
+	if (!res->commit_res.verf)
+		return -ENOMEM;
 	status = nfs4_call_sync(server->client, server, &msg,
 				&args->seq_args, &res->seq_res, 0);
 	if (status == -ENOTSUPP)
 		server->caps &= ~NFS_CAP_COPY;
 	if (status)
-		return status;
+		goto out;
 
-	if (res->write_res.verifier.committed != NFS_FILE_SYNC) {
+	if (!nfs_write_verifier_cmp(&res->write_res.verifier.verifier,
-		status = nfs_commit_file(dst, &res->write_res.verifier.verifier);
+				    &res->commit_res.verf->verifier)) {
-		if (status)
+		status = -EAGAIN;
-			return status;
+		goto out;
 	}
 
 	truncate_pagecache_range(dst_inode, pos_dst,
 				 pos_dst + res->write_res.count);
 
-	return res->write_res.count;
+	status = res->write_res.count;
+out:
+	kfree(res->commit_res.verf);
+	return status;
 }
 
 ssize_t nfs42_proc_copy(struct file *src, loff_t pos_src,
@@ -240,6 +246,9 @@ ssize_t nfs42_proc_copy(struct file *src, loff_t pos_src,
 		if (err == -ENOTSUPP) {
 			err = -EOPNOTSUPP;
 			break;
+		} if (err == -EAGAIN) {
+			dst_exception.retry = 1;
+			continue;
 		}
 
 		err2 = nfs4_handle_exception(server, err, &src_exception);
@@ -379,6 +388,7 @@ nfs42_layoutstat_done(struct rpc_task *task, void *calldata)
 			pnfs_mark_layout_stateid_invalid(lo, &head);
 			spin_unlock(&inode->i_lock);
 			pnfs_free_lseg_list(&head);
+			nfs_commit_inode(inode, 0);
 		} else
 			spin_unlock(&inode->i_lock);
 		break;
@@ -400,8 +410,6 @@ nfs42_layoutstat_done(struct rpc_task *task, void *calldata)
 	case -EOPNOTSUPP:
 		NFS_SERVER(inode)->caps &= ~NFS_CAP_LAYOUTSTATS;
 	}
-
-	dprintk("%s server returns %d\n", __func__, task->tk_status);
 }
 
 static void

fs/nfs/nfs42xdr.c

@@ -66,12 +66,14 @@
 					 encode_putfh_maxsz + \
 					 encode_savefh_maxsz + \
 					 encode_putfh_maxsz + \
-					 encode_copy_maxsz)
+					 encode_copy_maxsz + \
+					 encode_commit_maxsz)
 #define NFS4_dec_copy_sz		(compound_decode_hdr_maxsz + \
 					 decode_putfh_maxsz + \
 					 decode_savefh_maxsz + \
 					 decode_putfh_maxsz + \
-					 decode_copy_maxsz)
+					 decode_copy_maxsz + \
+					 decode_commit_maxsz)
 #define NFS4_enc_deallocate_sz		(compound_encode_hdr_maxsz + \
 					 encode_putfh_maxsz + \
 					 encode_deallocate_maxsz + \
@@ -222,6 +224,18 @@ static void nfs4_xdr_enc_allocate(struct rpc_rqst *req,
 	encode_nops(&hdr);
 }
 
+static void encode_copy_commit(struct xdr_stream *xdr,
+			  struct nfs42_copy_args *args,
+			  struct compound_hdr *hdr)
+{
+	__be32 *p;
+
+	encode_op_hdr(xdr, OP_COMMIT, decode_commit_maxsz, hdr);
+	p = reserve_space(xdr, 12);
+	p = xdr_encode_hyper(p, args->dst_pos);
+	*p = cpu_to_be32(args->count);
+}
+
 /*
  * Encode COPY request
  */
@@ -239,6 +253,7 @@ static void nfs4_xdr_enc_copy(struct rpc_rqst *req,
 	encode_savefh(xdr, &hdr);
 	encode_putfh(xdr, args->dst_fh, &hdr);
 	encode_copy(xdr, args, &hdr);
+	encode_copy_commit(xdr, args, &hdr);
 	encode_nops(&hdr);
 }
 
@@ -481,6 +496,9 @@ static int nfs4_xdr_dec_copy(struct rpc_rqst *rqstp,
 	if (status)
 		goto out;
 	status = decode_copy(xdr, res);
+	if (status)
+		goto out;
+	status = decode_commit(xdr, &res->commit_res);
 out:
 	return status;
 }

fs/nfs/nfs4client.c

@@ -359,11 +359,9 @@ struct nfs_client *nfs4_init_client(struct nfs_client *clp,
 	struct nfs_client *old;
 	int error;
 
-	if (clp->cl_cons_state == NFS_CS_READY) {
+	if (clp->cl_cons_state == NFS_CS_READY)
 		/* the client is initialised already */
-		dprintk("<-- nfs4_init_client() = 0 [already %p]\n", clp);
 		return clp;
-	}
 
 	/* Check NFS protocol revision and initialize RPC op vector */
 	clp->rpc_ops = &nfs_v4_clientops;
@@ -421,7 +419,6 @@ struct nfs_client *nfs4_init_client(struct nfs_client *clp,
 error:
 	nfs_mark_client_ready(clp, error);
 	nfs_put_client(clp);
-	dprintk("<-- nfs4_init_client() = xerror %d\n", error);
 	return ERR_PTR(error);
 }
 
@@ -469,6 +466,50 @@ static bool nfs4_same_verifier(nfs4_verifier *v1, nfs4_verifier *v2)
 	return memcmp(v1->data, v2->data, sizeof(v1->data)) == 0;
 }
 
+static int nfs4_match_client(struct nfs_client  *pos,  struct nfs_client *new,
+			     struct nfs_client **prev, struct nfs_net *nn)
+{
+	int status;
+
+	if (pos->rpc_ops != new->rpc_ops)
+		return 1;
+
+	if (pos->cl_minorversion != new->cl_minorversion)
+		return 1;
+
+	/* If "pos" isn't marked ready, we can't trust the
+	 * remaining fields in "pos", especially the client
+	 * ID and serverowner fields.  Wait for CREATE_SESSION
+	 * to finish. */
+	if (pos->cl_cons_state > NFS_CS_READY) {
+		atomic_inc(&pos->cl_count);
+		spin_unlock(&nn->nfs_client_lock);
+
+		nfs_put_client(*prev);
+		*prev = pos;
+
+		status = nfs_wait_client_init_complete(pos);
+		spin_lock(&nn->nfs_client_lock);
+
+		if (status < 0)
+			return status;
+	}
+
+	if (pos->cl_cons_state != NFS_CS_READY)
+		return 1;
+
+	if (pos->cl_clientid != new->cl_clientid)
+		return 1;
+
+	/* NFSv4.1 always uses the uniform string, however someone
+	 * might switch the uniquifier string on us.
+	 */
+	if (!nfs4_match_client_owner_id(pos, new))
+		return 1;
+
+	return 0;
+}
+
 /**
  * nfs40_walk_client_list - Find server that recognizes a client ID
  *
@@ -497,34 +538,10 @@ int nfs40_walk_client_list(struct nfs_client *new,
 	spin_lock(&nn->nfs_client_lock);
 	list_for_each_entry(pos, &nn->nfs_client_list, cl_share_link) {
 
-		if (pos->rpc_ops != new->rpc_ops)
+		status = nfs4_match_client(pos, new, &prev, nn);
-			continue;
+		if (status < 0)
-
+			goto out_unlock;
-		if (pos->cl_minorversion != new->cl_minorversion)
+		if (status != 0)
-			continue;
-
-		/* If "pos" isn't marked ready, we can't trust the
-		 * remaining fields in "pos" */
-		if (pos->cl_cons_state > NFS_CS_READY) {
-			atomic_inc(&pos->cl_count);
-			spin_unlock(&nn->nfs_client_lock);
-
-			nfs_put_client(prev);
-			prev = pos;
-
-			status = nfs_wait_client_init_complete(pos);
-			if (status < 0)
-				goto out;
-			status = -NFS4ERR_STALE_CLIENTID;
-			spin_lock(&nn->nfs_client_lock);
-		}
-		if (pos->cl_cons_state != NFS_CS_READY)
-			continue;
-
-		if (pos->cl_clientid != new->cl_clientid)
-			continue;
-
-		if (!nfs4_match_client_owner_id(pos, new))
 			continue;
 		/*
 		 * We just sent a new SETCLIENTID, which should have
@@ -557,8 +574,6 @@ int nfs40_walk_client_list(struct nfs_client *new,
 
 			prev = NULL;
 			*result = pos;
-			dprintk("NFS: <-- %s using nfs_client = %p ({%d})\n",
-				__func__, pos, atomic_read(&pos->cl_count));
 			goto out;
 		case -ERESTARTSYS:
 		case -ETIMEDOUT:
@@ -567,36 +582,22 @@ int nfs40_walk_client_list(struct nfs_client *new,
 			 */
 			nfs4_schedule_path_down_recovery(pos);
 		default:
+			spin_lock(&nn->nfs_client_lock);
 			goto out;
 		}
 
 		spin_lock(&nn->nfs_client_lock);
 	}
+out_unlock:
 	spin_unlock(&nn->nfs_client_lock);
 
 	/* No match found. The server lost our clientid */
 out:
 	nfs_put_client(prev);
-	dprintk("NFS: <-- %s status = %d\n", __func__, status);
 	return status;
 }
 
 #ifdef CONFIG_NFS_V4_1
-/*
- * Returns true if the client IDs match
- */
-static bool nfs4_match_clientids(u64 a, u64 b)
-{
-	if (a != b) {
-		dprintk("NFS: --> %s client ID %llx does not match %llx\n",
-			__func__, a, b);
-		return false;
-	}
-	dprintk("NFS: --> %s client ID %llx matches %llx\n",
-		__func__, a, b);
-	return true;
-}
-
 /*
  * Returns true if the server major ids match
  */
@@ -605,36 +606,8 @@ nfs4_check_serverowner_major_id(struct nfs41_server_owner *o1,
 				struct nfs41_server_owner *o2)
 {
 	if (o1->major_id_sz != o2->major_id_sz)
-		goto out_major_mismatch;
+		return false;
-	if (memcmp(o1->major_id, o2->major_id, o1->major_id_sz) != 0)
+	return memcmp(o1->major_id, o2->major_id, o1->major_id_sz) == 0;
-		goto out_major_mismatch;
-
-	dprintk("NFS: --> %s server owner major IDs match\n", __func__);
-	return true;
-
-out_major_mismatch:
-	dprintk("NFS: --> %s server owner major IDs do not match\n",
-		__func__);
-	return false;
-}
-
-/*
- * Returns true if server minor ids match
- */
-static bool
-nfs4_check_serverowner_minor_id(struct nfs41_server_owner *o1,
-				struct nfs41_server_owner *o2)
-{
-	/* Check eir_server_owner so_minor_id */
-	if (o1->minor_id != o2->minor_id)
-		goto out_minor_mismatch;
-
-	dprintk("NFS: --> %s server owner minor IDs match\n", __func__);
-	return true;
-
-out_minor_mismatch:
-	dprintk("NFS: --> %s server owner minor IDs do not match\n", __func__);
-	return false;
 }
 
 /*
@@ -645,18 +618,9 @@ nfs4_check_server_scope(struct nfs41_server_scope *s1,
 			struct nfs41_server_scope *s2)
 {
 	if (s1->server_scope_sz != s2->server_scope_sz)
-		goto out_scope_mismatch;
+		return false;
-	if (memcmp(s1->server_scope, s2->server_scope,
+	return memcmp(s1->server_scope, s2->server_scope,
-		   s1->server_scope_sz) != 0)
+					s1->server_scope_sz) == 0;
-		goto out_scope_mismatch;
-
-	dprintk("NFS: --> %s server scopes match\n", __func__);
-	return true;
-
-out_scope_mismatch:
-	dprintk("NFS: --> %s server scopes do not match\n",
-		__func__);
-	return false;
 }
 
 /**
@@ -680,7 +644,7 @@ int nfs4_detect_session_trunking(struct nfs_client *clp,
 				 struct rpc_xprt *xprt)
 {
 	/* Check eir_clientid */
-	if (!nfs4_match_clientids(clp->cl_clientid, res->clientid))
+	if (clp->cl_clientid != res->clientid)
 		goto out_err;
 
 	/* Check eir_server_owner so_major_id */
@@ -689,8 +653,7 @@ int nfs4_detect_session_trunking(struct nfs_client *clp,
 		goto out_err;
 
 	/* Check eir_server_owner so_minor_id */
-	if (!nfs4_check_serverowner_minor_id(clp->cl_serverowner,
+	if (clp->cl_serverowner->minor_id != res->server_owner->minor_id)
-					     res->server_owner))
 		goto out_err;
 
 	/* Check eir_server_scope */
@@ -739,33 +702,10 @@ int nfs41_walk_client_list(struct nfs_client *new,
 		if (pos == new)
 			goto found;
 
-		if (pos->rpc_ops != new->rpc_ops)
+		status = nfs4_match_client(pos, new, &prev, nn);
-			continue;
+		if (status < 0)
-
+			goto out;
-		if (pos->cl_minorversion != new->cl_minorversion)
+		if (status != 0)
-			continue;
-
-		/* If "pos" isn't marked ready, we can't trust the
-		 * remaining fields in "pos", especially the client
-		 * ID and serverowner fields.  Wait for CREATE_SESSION
-		 * to finish. */
-		if (pos->cl_cons_state > NFS_CS_READY) {
-			atomic_inc(&pos->cl_count);
-			spin_unlock(&nn->nfs_client_lock);
-
-			nfs_put_client(prev);
-			prev = pos;
-
-			status = nfs_wait_client_init_complete(pos);
-			spin_lock(&nn->nfs_client_lock);
-			if (status < 0)
-				break;
-			status = -NFS4ERR_STALE_CLIENTID;
-		}
-		if (pos->cl_cons_state != NFS_CS_READY)
-			continue;
-
-		if (!nfs4_match_clientids(pos->cl_clientid, new->cl_clientid))
 			continue;
 
 		/*
@@ -777,23 +717,15 @@ int nfs41_walk_client_list(struct nfs_client *new,
 						     new->cl_serverowner))
 			continue;
 
-		/* Unlike NFSv4.0, we know that NFSv4.1 always uses the
-		 * uniform string, however someone might switch the
-		 * uniquifier string on us.
-		 */
-		if (!nfs4_match_client_owner_id(pos, new))
-			continue;
 found:
 		atomic_inc(&pos->cl_count);
 		*result = pos;
 		status = 0;
-		dprintk("NFS: <-- %s using nfs_client = %p ({%d})\n",
-			__func__, pos, atomic_read(&pos->cl_count));
 		break;
 	}
 
+out:
 	spin_unlock(&nn->nfs_client_lock);
-	dprintk("NFS: <-- %s status = %d\n", __func__, status);
 	nfs_put_client(prev);
 	return status;
 }
@@ -916,9 +848,6 @@ static int nfs4_set_client(struct nfs_server *server,
 		.timeparms = timeparms,
 	};
 	struct nfs_client *clp;
-	int error;
-
-	dprintk("--> nfs4_set_client()\n");
 
 	if (server->flags & NFS_MOUNT_NORESVPORT)
 		set_bit(NFS_CS_NORESVPORT, &cl_init.init_flags);
@@ -927,15 +856,11 @@ static int nfs4_set_client(struct nfs_server *server,
 
 	/* Allocate or find a client reference we can use */
 	clp = nfs_get_client(&cl_init);
-	if (IS_ERR(clp)) {
+	if (IS_ERR(clp))
-		error = PTR_ERR(clp);
+		return PTR_ERR(clp);
-		goto error;
-	}
 
-	if (server->nfs_client == clp) {
+	if (server->nfs_client == clp)
-		error = -ELOOP;
+		return -ELOOP;
-		goto error;
-	}
 
 	/*
 	 * Query for the lease time on clientid setup or renewal
@@ -947,11 +872,7 @@ static int nfs4_set_client(struct nfs_server *server,
 	set_bit(NFS_CS_CHECK_LEASE_TIME, &clp->cl_res_state);
 
 	server->nfs_client = clp;
-	dprintk("<-- nfs4_set_client() = 0 [new %p]\n", clp);
 	return 0;
-error:
-	dprintk("<-- nfs4_set_client() = xerror %d\n", error);
-	return error;
 }
 
 /*
@@ -982,7 +903,6 @@ struct nfs_client *nfs4_set_ds_client(struct nfs_server *mds_srv,
 		.net = mds_clp->cl_net,
 		.timeparms = &ds_timeout,
 	};
-	struct nfs_client *clp;
 	char buf[INET6_ADDRSTRLEN + 1];
 
 	if (rpc_ntop(ds_addr, buf, sizeof(buf)) <= 0)
@@ -998,10 +918,7 @@ struct nfs_client *nfs4_set_ds_client(struct nfs_server *mds_srv,
 	 * (section 13.1 RFC 5661).
 	 */
 	nfs_init_timeout_values(&ds_timeout, ds_proto, ds_timeo, ds_retrans);
-	clp = nfs_get_client(&cl_init);
+	return nfs_get_client(&cl_init);
-
-	dprintk("<-- %s %p\n", __func__, clp);
-	return clp;
 }
 EXPORT_SYMBOL_GPL(nfs4_set_ds_client);
 
@@ -1098,8 +1015,6 @@ static int nfs4_init_server(struct nfs_server *server,
 	struct rpc_timeout timeparms;
 	int error;
 
-	dprintk("--> nfs4_init_server()\n");
-
 	nfs_init_timeout_values(&timeparms, data->nfs_server.protocol,
 			data->timeo, data->retrans);
 
@@ -1127,7 +1042,7 @@ static int nfs4_init_server(struct nfs_server *server,
 			data->minorversion,
 			data->net);
 	if (error < 0)
-		goto error;
+		return error;
 
 	if (data->rsize)
 		server->rsize = nfs_block_size(data->rsize, NULL);
@@ -1138,16 +1053,10 @@ static int nfs4_init_server(struct nfs_server *server,
 	server->acregmax = data->acregmax * HZ;
 	server->acdirmin = data->acdirmin * HZ;
 	server->acdirmax = data->acdirmax * HZ;
+	server->port     = data->nfs_server.port;
 
-	server->port = data->nfs_server.port;
+	return nfs_init_server_rpcclient(server, &timeparms,
-
+					 data->selected_flavor);
-	error = nfs_init_server_rpcclient(server, &timeparms,
-					  data->selected_flavor);
-
-error:
-	/* Done */
-	dprintk("<-- nfs4_init_server() = %d\n", error);
-	return error;
 }
 
 /*
@@ -1163,8 +1072,6 @@ struct nfs_server *nfs4_create_server(struct nfs_mount_info *mount_info,
 	bool auth_probe;
 	int error;
 
-	dprintk("--> nfs4_create_server()\n");
-
 	server = nfs_alloc_server();
 	if (!server)
 		return ERR_PTR(-ENOMEM);
@@ -1180,12 +1087,10 @@ struct nfs_server *nfs4_create_server(struct nfs_mount_info *mount_info,
 	if (error < 0)
 		goto error;
 
-	dprintk("<-- nfs4_create_server() = %p\n", server);
 	return server;
 
 error:
 	nfs_free_server(server);
-	dprintk("<-- nfs4_create_server() = error %d\n", error);
 	return ERR_PTR(error);
 }
 
@@ -1200,8 +1105,6 @@ struct nfs_server *nfs4_create_referral_server(struct nfs_clone_mount *data,
 	bool auth_probe;
 	int error;
 
-	dprintk("--> nfs4_create_referral_server()\n");
-
 	server = nfs_alloc_server();
 	if (!server)
 		return ERR_PTR(-ENOMEM);
@@ -1235,12 +1138,10 @@ struct nfs_server *nfs4_create_referral_server(struct nfs_clone_mount *data,
 	if (error < 0)
 		goto error;
 
-	dprintk("<-- nfs_create_referral_server() = %p\n", server);
 	return server;
 
 error:
 	nfs_free_server(server);
-	dprintk("<-- nfs4_create_referral_server() = error %d\n", error);
 	return ERR_PTR(error);
 }
 
@@ -1300,31 +1201,16 @@ int nfs4_update_server(struct nfs_server *server, const char *hostname,
 	struct sockaddr *localaddr = (struct sockaddr *)&address;
 	int error;
 
-	dprintk("--> %s: move FSID %llx:%llx to \"%s\")\n", __func__,
-			(unsigned long long)server->fsid.major,
-			(unsigned long long)server->fsid.minor,
-			hostname);
-
 	error = rpc_switch_client_transport(clnt, &xargs, clnt->cl_timeout);
-	if (error != 0) {
+	if (error != 0)
-		dprintk("<-- %s(): rpc_switch_client_transport returned %d\n",
+		return error;
-			__func__, error);
-		goto out;
-	}
 
 	error = rpc_localaddr(clnt, localaddr, sizeof(address));
-	if (error != 0) {
+	if (error != 0)
-		dprintk("<-- %s(): rpc_localaddr returned %d\n",
+		return error;
-			__func__, error);
-		goto out;
-	}
 
-	error = -EAFNOSUPPORT;
+	if (rpc_ntop(localaddr, buf, sizeof(buf)) == 0)
-	if (rpc_ntop(localaddr, buf, sizeof(buf)) == 0) {
+		return -EAFNOSUPPORT;
-		dprintk("<-- %s(): rpc_ntop returned %d\n",
-			__func__, error);
-		goto out;
-	}
 
 	nfs_server_remove_lists(server);
 	error = nfs4_set_client(server, hostname, sap, salen, buf,
@@ -1333,21 +1219,12 @@ int nfs4_update_server(struct nfs_server *server, const char *hostname,
 	nfs_put_client(clp);
 	if (error != 0) {
 		nfs_server_insert_lists(server);
-		dprintk("<-- %s(): nfs4_set_client returned %d\n",
+		return error;
-			__func__, error);
-		goto out;
 	}
 
 	if (server->nfs_client->cl_hostname == NULL)
 		server->nfs_client->cl_hostname = kstrdup(hostname, GFP_KERNEL);
 	nfs_server_insert_lists(server);
 
-	error = nfs_probe_destination(server);
+	return nfs_probe_destination(server);
-	if (error < 0)
-		goto out;
-
-	dprintk("<-- %s() succeeded\n", __func__);
-
-out:
-	return error;
 }

fs/nfs/nfs4getroot.c

@@ -14,8 +14,6 @@ int nfs4_get_rootfh(struct nfs_server *server, struct nfs_fh *mntfh, bool auth_p
 	struct nfs_fsinfo fsinfo;
 	int ret = -ENOMEM;
 
-	dprintk("--> nfs4_get_rootfh()\n");
-
 	fsinfo.fattr = nfs_alloc_fattr();
 	if (fsinfo.fattr == NULL)
 		goto out;
@@ -38,6 +36,5 @@ int nfs4_get_rootfh(struct nfs_server *server, struct nfs_fh *mntfh, bool auth_p
 	memcpy(&server->fsid, &fsinfo.fattr->fsid, sizeof(server->fsid));
 out:
 	nfs_free_fattr(fsinfo.fattr);
-	dprintk("<-- nfs4_get_rootfh() = %d\n", ret);
 	return ret;
 }

fs/nfs/nfs4namespace.c

@@ -340,7 +340,6 @@ static struct vfsmount *nfs_follow_referral(struct dentry *dentry,
 out:
 	free_page((unsigned long) page);
 	free_page((unsigned long) page2);
-	dprintk("%s: done\n", __func__);
 	return mnt;
 }
 
@@ -358,11 +357,9 @@ static struct vfsmount *nfs_do_refmount(struct rpc_clnt *client, struct dentry *
 	int err;
 
 	/* BUG_ON(IS_ROOT(dentry)); */
-	dprintk("%s: enter\n", __func__);
-
 	page = alloc_page(GFP_KERNEL);
 	if (page == NULL)
-		goto out;
+		return mnt;
 
 	fs_locations = kmalloc(sizeof(struct nfs4_fs_locations), GFP_KERNEL);
 	if (fs_locations == NULL)
@@ -386,8 +383,6 @@ static struct vfsmount *nfs_do_refmount(struct rpc_clnt *client, struct dentry *
 out_free:
 	__free_page(page);
 	kfree(fs_locations);
-out:
-	dprintk("%s: done\n", __func__);
 	return mnt;
 }
 

fs/nfs/nfs4proc.c

@@ -698,7 +698,8 @@ static int nfs41_sequence_process(struct rpc_task *task,
 	session = slot->table->session;
 
 	if (slot->interrupted) {
-		slot->interrupted = 0;
+		if (res->sr_status != -NFS4ERR_DELAY)
+			slot->interrupted = 0;
 		interrupted = true;
 	}
 
@@ -2300,8 +2301,10 @@ static int _nfs4_proc_open(struct nfs4_opendata *data)
 		if (status != 0)
 			return status;
 	}
-	if (!(o_res->f_attr->valid & NFS_ATTR_FATTR))
+	if (!(o_res->f_attr->valid & NFS_ATTR_FATTR)) {
+		nfs4_sequence_free_slot(&o_res->seq_res);
 		nfs4_proc_getattr(server, &o_res->fh, o_res->f_attr, o_res->f_label);
+	}
 	return 0;
 }
 
@@ -3265,6 +3268,7 @@ static int _nfs4_server_capabilities(struct nfs_server *server, struct nfs_fh *f
 		.rpc_resp = &res,
 	};
 	int status;
+	int i;
 
 	bitmask[0] = FATTR4_WORD0_SUPPORTED_ATTRS |
 		     FATTR4_WORD0_FH_EXPIRE_TYPE |
@@ -3330,8 +3334,13 @@ static int _nfs4_server_capabilities(struct nfs_server *server, struct nfs_fh *f
 		server->cache_consistency_bitmask[0] &= FATTR4_WORD0_CHANGE|FATTR4_WORD0_SIZE;
 		server->cache_consistency_bitmask[1] &= FATTR4_WORD1_TIME_METADATA|FATTR4_WORD1_TIME_MODIFY;
 		server->cache_consistency_bitmask[2] = 0;
+
+		/* Avoid a regression due to buggy server */
+		for (i = 0; i < ARRAY_SIZE(res.exclcreat_bitmask); i++)
+			res.exclcreat_bitmask[i] &= res.attr_bitmask[i];
 		memcpy(server->exclcreat_bitmask, res.exclcreat_bitmask,
 			sizeof(server->exclcreat_bitmask));
+
 		server->acl_bitmask = res.acl_bitmask;
 		server->fh_expire_type = res.fh_expire_type;
 	}
@@ -4610,7 +4619,7 @@ static int nfs4_proc_pgio_rpc_prepare(struct rpc_task *task,
 		return 0;
 	if (nfs4_set_rw_stateid(&hdr->args.stateid, hdr->args.context,
 				hdr->args.lock_context,
-				hdr->rw_ops->rw_mode) == -EIO)
+				hdr->rw_mode) == -EIO)
 		return -EIO;
 	if (unlikely(test_bit(NFS_CONTEXT_BAD, &hdr->args.context->flags)))
 		return -EIO;
@@ -4804,8 +4813,10 @@ static int nfs4_proc_async_renew(struct nfs_client *clp, struct rpc_cred *cred,
 	if (!atomic_inc_not_zero(&clp->cl_count))
 		return -EIO;
 	data = kmalloc(sizeof(*data), GFP_NOFS);
-	if (data == NULL)
+	if (data == NULL) {
+		nfs_put_client(clp);
 		return -ENOMEM;
+	}
 	data->client = clp;
 	data->timestamp = jiffies;
 	return rpc_call_async(clp->cl_rpcclient, &msg, RPC_TASK_TIMEOUT,
@@ -5782,6 +5793,7 @@ struct nfs4_unlockdata {
 	struct nfs_locku_res res;
 	struct nfs4_lock_state *lsp;
 	struct nfs_open_context *ctx;
+	struct nfs_lock_context *l_ctx;
 	struct file_lock fl;
 	struct nfs_server *server;
 	unsigned long timestamp;
@@ -5806,6 +5818,7 @@ static struct nfs4_unlockdata *nfs4_alloc_unlockdata(struct file_lock *fl,
 	atomic_inc(&lsp->ls_count);
 	/* Ensure we don't close file until we're done freeing locks! */
 	p->ctx = get_nfs_open_context(ctx);
+	p->l_ctx = nfs_get_lock_context(ctx);
 	memcpy(&p->fl, fl, sizeof(p->fl));
 	p->server = NFS_SERVER(inode);
 	return p;
@@ -5816,6 +5829,7 @@ static void nfs4_locku_release_calldata(void *data)
 	struct nfs4_unlockdata *calldata = data;
 	nfs_free_seqid(calldata->arg.seqid);
 	nfs4_put_lock_state(calldata->lsp);
+	nfs_put_lock_context(calldata->l_ctx);
 	put_nfs_open_context(calldata->ctx);
 	kfree(calldata);
 }
@@ -5857,6 +5871,10 @@ static void nfs4_locku_prepare(struct rpc_task *task, void *data)
 {
 	struct nfs4_unlockdata *calldata = data;
 
+	if (test_bit(NFS_CONTEXT_UNLOCK, &calldata->l_ctx->open_context->flags) &&
+		nfs_async_iocounter_wait(task, calldata->l_ctx))
+		return;
+
 	if (nfs_wait_on_sequence(calldata->arg.seqid, task) != 0)
 		goto out_wait;
 	nfs4_stateid_copy(&calldata->arg.stateid, &calldata->lsp->ls_stateid);
@@ -5908,6 +5926,8 @@ static struct rpc_task *nfs4_do_unlck(struct file_lock *fl,
 	 * canceled lock is passed in, and it won't be an unlock.
 	 */
 	fl->fl_type = F_UNLCK;
+	if (fl->fl_flags & FL_CLOSE)
+		set_bit(NFS_CONTEXT_UNLOCK, &ctx->flags);
 
 	data = nfs4_alloc_unlockdata(fl, ctx, lsp, seqid);
 	if (data == NULL) {
@@ -6445,9 +6465,6 @@ nfs4_proc_lock(struct file *filp, int cmd, struct file_lock *request)
 	ctx = nfs_file_open_context(filp);
 	state = ctx->state;
 
-	if (request->fl_start < 0 || request->fl_end < 0)
-		return -EINVAL;
-
 	if (IS_GETLK(cmd)) {
 		if (state != NULL)
 			return nfs4_proc_getlk(state, F_GETLK, request);
@@ -6470,20 +6487,6 @@ nfs4_proc_lock(struct file *filp, int cmd, struct file_lock *request)
 	    !test_bit(NFS_STATE_POSIX_LOCKS, &state->flags))
 		return -ENOLCK;
 
-	/*
-	 * Don't rely on the VFS having checked the file open mode,
-	 * since it won't do this for flock() locks.
-	 */
-	switch (request->fl_type) {
-	case F_RDLCK:
-		if (!(filp->f_mode & FMODE_READ))
-			return -EBADF;
-		break;
-	case F_WRLCK:
-		if (!(filp->f_mode & FMODE_WRITE))
-			return -EBADF;
-	}
-
 	status = nfs4_set_lock_state(state, request);
 	if (status != 0)
 		return status;
@@ -7155,8 +7158,6 @@ int nfs4_proc_bind_one_conn_to_session(struct rpc_clnt *clnt,
 	};
 	struct rpc_task *task;
 
-	dprintk("--> %s\n", __func__);
-
 	nfs4_copy_sessionid(&args.sessionid, &clp->cl_session->sess_id);
 	if (!(clp->cl_session->flags & SESSION4_BACK_CHAN))
 		args.dir = NFS4_CDFC4_FORE;
@@ -7176,24 +7177,20 @@ int nfs4_proc_bind_one_conn_to_session(struct rpc_clnt *clnt,
 		if (memcmp(res.sessionid.data,
 		    clp->cl_session->sess_id.data, NFS4_MAX_SESSIONID_LEN)) {
 			dprintk("NFS: %s: Session ID mismatch\n", __func__);
-			status = -EIO;
+			return -EIO;
-			goto out;
 		}
 		if ((res.dir & args.dir) != res.dir || res.dir == 0) {
 			dprintk("NFS: %s: Unexpected direction from server\n",
 				__func__);
-			status = -EIO;
+			return -EIO;
-			goto out;
 		}
 		if (res.use_conn_in_rdma_mode != args.use_conn_in_rdma_mode) {
 			dprintk("NFS: %s: Server returned RDMA mode = true\n",
 				__func__);
-			status = -EIO;
+			return -EIO;
-			goto out;
 		}
 	}
-out:
+
-	dprintk("<-- %s status= %d\n", __func__, status);
 	return status;
 }
 
@@ -7459,15 +7456,16 @@ static int _nfs4_proc_exchange_id(struct nfs_client *clp, struct rpc_cred *cred,
 	};
 	struct nfs41_exchange_id_data *calldata;
 	struct rpc_task *task;
-	int status = -EIO;
+	int status;
 
 	if (!atomic_inc_not_zero(&clp->cl_count))
-		goto out;
+		return -EIO;
 
-	status = -ENOMEM;
 	calldata = kzalloc(sizeof(*calldata), GFP_NOFS);
-	if (!calldata)
+	if (!calldata) {
-		goto out;
+		nfs_put_client(clp);
+		return -ENOMEM;
+	}
 
 	if (!xprt)
 		nfs4_init_boot_verifier(clp, &verifier);
@@ -7476,10 +7474,6 @@ static int _nfs4_proc_exchange_id(struct nfs_client *clp, struct rpc_cred *cred,
 	if (status)
 		goto out_calldata;
 
-	dprintk("NFS call  exchange_id auth=%s, '%s'\n",
-		clp->cl_rpcclient->cl_auth->au_ops->au_name,
-		clp->cl_owner_id);
-
 	calldata->res.server_owner = kzalloc(sizeof(struct nfs41_server_owner),
 						GFP_NOFS);
 	status = -ENOMEM;
@@ -7545,13 +7539,6 @@ static int _nfs4_proc_exchange_id(struct nfs_client *clp, struct rpc_cred *cred,
 
 	rpc_put_task(task);
 out:
-	if (clp->cl_implid != NULL)
-		dprintk("NFS reply exchange_id: Server Implementation ID: "
-			"domain: %s, name: %s, date: %llu,%u\n",
-			clp->cl_implid->domain, clp->cl_implid->name,
-			clp->cl_implid->date.seconds,
-			clp->cl_implid->date.nseconds);
-	dprintk("NFS reply exchange_id: %d\n", status);
 	return status;
 
 out_impl_id:
@@ -7769,17 +7756,13 @@ int nfs4_proc_get_lease_time(struct nfs_client *clp, struct nfs_fsinfo *fsinfo)
 
 	nfs4_init_sequence(&args.la_seq_args, &res.lr_seq_res, 0);
 	nfs4_set_sequence_privileged(&args.la_seq_args);
-	dprintk("--> %s\n", __func__);
 	task = rpc_run_task(&task_setup);
 
 	if (IS_ERR(task))
-		status = PTR_ERR(task);
+		return PTR_ERR(task);
-	else {
-		status = task->tk_status;
-		rpc_put_task(task);
-	}
-	dprintk("<-- %s return %d\n", __func__, status);
 
+	status = task->tk_status;
+	rpc_put_task(task);
 	return status;
 }
 
@@ -8180,6 +8163,12 @@ static int nfs41_reclaim_complete_handle_errors(struct rpc_task *task, struct nf
 		/* fall through */
 	case -NFS4ERR_RETRY_UNCACHED_REP:
 		return -EAGAIN;
+	case -NFS4ERR_BADSESSION:
+	case -NFS4ERR_DEADSESSION:
+	case -NFS4ERR_CONN_NOT_BOUND_TO_SESSION:
+		nfs4_schedule_session_recovery(clp->cl_session,
+				task->tk_status);
+		break;
 	default:
 		nfs4_schedule_lease_recovery(clp);
 	}
@@ -8258,7 +8247,6 @@ static int nfs41_proc_reclaim_complete(struct nfs_client *clp,
 	if (status == 0)
 		status = task->tk_status;
 	rpc_put_task(task);
-	return 0;
 out:
 	dprintk("<-- %s status=%d\n", __func__, status);
 	return status;
@@ -8357,6 +8345,7 @@ nfs4_layoutget_handle_exception(struct rpc_task *task,
 		 */
 		pnfs_mark_layout_stateid_invalid(lo, &head);
 		spin_unlock(&inode->i_lock);
+		nfs_commit_inode(inode, 0);
 		pnfs_free_lseg_list(&head);
 		status = -EAGAIN;
 		goto out;

fs/nfs/nfs4state.c

@@ -1649,13 +1649,14 @@ static void nfs4_state_start_reclaim_reboot(struct nfs_client *clp)
 	nfs4_state_mark_reclaim_helper(clp, nfs4_state_mark_reclaim_reboot);
 }
 
-static void nfs4_reclaim_complete(struct nfs_client *clp,
+static int nfs4_reclaim_complete(struct nfs_client *clp,
 				 const struct nfs4_state_recovery_ops *ops,
 				 struct rpc_cred *cred)
 {
 	/* Notify the server we're done reclaiming our state */
 	if (ops->reclaim_complete)
-		(void)ops->reclaim_complete(clp, cred);
+		return ops->reclaim_complete(clp, cred);
+	return 0;
 }
 
 static void nfs4_clear_reclaim_server(struct nfs_server *server)
@@ -1702,13 +1703,16 @@ static void nfs4_state_end_reclaim_reboot(struct nfs_client *clp)
 {
 	const struct nfs4_state_recovery_ops *ops;
 	struct rpc_cred *cred;
+	int err;
 
 	if (!nfs4_state_clear_reclaim_reboot(clp))
 		return;
 	ops = clp->cl_mvops->reboot_recovery_ops;
 	cred = nfs4_get_clid_cred(clp);
-	nfs4_reclaim_complete(clp, ops, cred);
+	err = nfs4_reclaim_complete(clp, ops, cred);
 	put_rpccred(cred);
+	if (err == -NFS4ERR_CONN_NOT_BOUND_TO_SESSION)
+		set_bit(NFS4CLNT_RECLAIM_REBOOT, &clp->cl_state);
 }
 
 static void nfs4_state_start_reclaim_nograce(struct nfs_client *clp)

fs/nfs/nfs4xdr.c

@@ -1000,8 +1000,9 @@ static void encode_nfs4_verifier(struct xdr_stream *xdr, const nfs4_verifier *ve
 
 static void encode_attrs(struct xdr_stream *xdr, const struct iattr *iap,
 				const struct nfs4_label *label,
+				const umode_t *umask,
 				const struct nfs_server *server,
-				bool excl_check, const umode_t *umask)
+				const uint32_t attrmask[])
 {
 	char owner_name[IDMAP_NAMESZ];
 	char owner_group[IDMAP_NAMESZ];
@@ -1016,22 +1017,20 @@ static void encode_attrs(struct xdr_stream *xdr, const struct iattr *iap,
 	/*
 	 * We reserve enough space to write the entire attribute buffer at once.
 	 */
-	if (iap->ia_valid & ATTR_SIZE) {
+	if ((iap->ia_valid & ATTR_SIZE) && (attrmask[0] & FATTR4_WORD0_SIZE)) {
 		bmval[0] |= FATTR4_WORD0_SIZE;
 		len += 8;
 	}
-	if (!(server->attr_bitmask[2] & FATTR4_WORD2_MODE_UMASK))
-		umask = NULL;
 	if (iap->ia_valid & ATTR_MODE) {
-		if (umask) {
+		if (umask && (attrmask[2] & FATTR4_WORD2_MODE_UMASK)) {
 			bmval[2] |= FATTR4_WORD2_MODE_UMASK;
 			len += 8;
-		} else {
+		} else if (attrmask[1] & FATTR4_WORD1_MODE) {
 			bmval[1] |= FATTR4_WORD1_MODE;
 			len += 4;
 		}
 	}
-	if (iap->ia_valid & ATTR_UID) {
+	if ((iap->ia_valid & ATTR_UID) && (attrmask[1] & FATTR4_WORD1_OWNER)) {
 		owner_namelen = nfs_map_uid_to_name(server, iap->ia_uid, owner_name, IDMAP_NAMESZ);
 		if (owner_namelen < 0) {
 			dprintk("nfs: couldn't resolve uid %d to string\n",
@@ -1044,7 +1043,8 @@ static void encode_attrs(struct xdr_stream *xdr, const struct iattr *iap,
 		bmval[1] |= FATTR4_WORD1_OWNER;
 		len += 4 + (XDR_QUADLEN(owner_namelen) << 2);
 	}
-	if (iap->ia_valid & ATTR_GID) {
+	if ((iap->ia_valid & ATTR_GID) &&
+	   (attrmask[1] & FATTR4_WORD1_OWNER_GROUP)) {
 		owner_grouplen = nfs_map_gid_to_group(server, iap->ia_gid, owner_group, IDMAP_NAMESZ);
 		if (owner_grouplen < 0) {
 			dprintk("nfs: couldn't resolve gid %d to string\n",
@@ -1056,32 +1056,26 @@ static void encode_attrs(struct xdr_stream *xdr, const struct iattr *iap,
 		bmval[1] |= FATTR4_WORD1_OWNER_GROUP;
 		len += 4 + (XDR_QUADLEN(owner_grouplen) << 2);
 	}
-	if (iap->ia_valid & ATTR_ATIME_SET) {
+	if (attrmask[1] & FATTR4_WORD1_TIME_ACCESS_SET) {
-		bmval[1] |= FATTR4_WORD1_TIME_ACCESS_SET;
+		if (iap->ia_valid & ATTR_ATIME_SET) {
-		len += 16;
+			bmval[1] |= FATTR4_WORD1_TIME_ACCESS_SET;
-	} else if (iap->ia_valid & ATTR_ATIME) {
+			len += 16;
-		bmval[1] |= FATTR4_WORD1_TIME_ACCESS_SET;
+		} else if (iap->ia_valid & ATTR_ATIME) {
-		len += 4;
+			bmval[1] |= FATTR4_WORD1_TIME_ACCESS_SET;
+			len += 4;
+		}
 	}
-	if (iap->ia_valid & ATTR_MTIME_SET) {
+	if (attrmask[1] & FATTR4_WORD1_TIME_MODIFY_SET) {
-		bmval[1] |= FATTR4_WORD1_TIME_MODIFY_SET;
+		if (iap->ia_valid & ATTR_MTIME_SET) {
-		len += 16;
+			bmval[1] |= FATTR4_WORD1_TIME_MODIFY_SET;
-	} else if (iap->ia_valid & ATTR_MTIME) {
+			len += 16;
-		bmval[1] |= FATTR4_WORD1_TIME_MODIFY_SET;
+		} else if (iap->ia_valid & ATTR_MTIME) {
-		len += 4;
+			bmval[1] |= FATTR4_WORD1_TIME_MODIFY_SET;
+			len += 4;
+		}
 	}
 
-	if (excl_check) {
+	if (label && (attrmask[2] & FATTR4_WORD2_SECURITY_LABEL)) {
-		const u32 *excl_bmval = server->exclcreat_bitmask;
-		bmval[0] &= excl_bmval[0];
-		bmval[1] &= excl_bmval[1];
-		bmval[2] &= excl_bmval[2];
-
-		if (!(excl_bmval[2] & FATTR4_WORD2_SECURITY_LABEL))
-			label = NULL;
-	}
-
-	if (label) {
 		len += 4 + 4 + 4 + (XDR_QUADLEN(label->len) << 2);
 		bmval[2] |= FATTR4_WORD2_SECURITY_LABEL;
 	}
@@ -1188,8 +1182,8 @@ static void encode_create(struct xdr_stream *xdr, const struct nfs4_create_arg *
 	}
 
 	encode_string(xdr, create->name->len, create->name->name);
-	encode_attrs(xdr, create->attrs, create->label, create->server, false,
+	encode_attrs(xdr, create->attrs, create->label, &create->umask,
-		     &create->umask);
+			create->server, create->server->attr_bitmask);
 }
 
 static void encode_getattr_one(struct xdr_stream *xdr, uint32_t bitmap, struct compound_hdr *hdr)
@@ -1409,13 +1403,13 @@ static inline void encode_createmode(struct xdr_stream *xdr, const struct nfs_op
 	switch(arg->createmode) {
 	case NFS4_CREATE_UNCHECKED:
 		*p = cpu_to_be32(NFS4_CREATE_UNCHECKED);
-		encode_attrs(xdr, arg->u.attrs, arg->label, arg->server, false,
+		encode_attrs(xdr, arg->u.attrs, arg->label, &arg->umask,
-			     &arg->umask);
+				arg->server, arg->server->attr_bitmask);
 		break;
 	case NFS4_CREATE_GUARDED:
 		*p = cpu_to_be32(NFS4_CREATE_GUARDED);
-		encode_attrs(xdr, arg->u.attrs, arg->label, arg->server, false,
+		encode_attrs(xdr, arg->u.attrs, arg->label, &arg->umask,
-			     &arg->umask);
+				arg->server, arg->server->attr_bitmask);
 		break;
 	case NFS4_CREATE_EXCLUSIVE:
 		*p = cpu_to_be32(NFS4_CREATE_EXCLUSIVE);
@@ -1424,8 +1418,8 @@ static inline void encode_createmode(struct xdr_stream *xdr, const struct nfs_op
 	case NFS4_CREATE_EXCLUSIVE4_1:
 		*p = cpu_to_be32(NFS4_CREATE_EXCLUSIVE4_1);
 		encode_nfs4_verifier(xdr, &arg->u.verifier);
-		encode_attrs(xdr, arg->u.attrs, arg->label, arg->server, true,
+		encode_attrs(xdr, arg->u.attrs, arg->label, &arg->umask,
-			     &arg->umask);
+				arg->server, arg->server->exclcreat_bitmask);
 	}
 }
 
@@ -1681,7 +1675,8 @@ static void encode_setattr(struct xdr_stream *xdr, const struct nfs_setattrargs
 {
 	encode_op_hdr(xdr, OP_SETATTR, decode_setattr_maxsz, hdr);
 	encode_nfs4_stateid(xdr, &arg->stateid);
-	encode_attrs(xdr, arg->iap, arg->label, server, false, NULL);
+	encode_attrs(xdr, arg->iap, arg->label, NULL, server,
+			server->attr_bitmask);
 }
 
 static void encode_setclientid(struct xdr_stream *xdr, const struct nfs4_setclientid *setclientid, struct compound_hdr *hdr)
@@ -2005,16 +2000,10 @@ encode_layoutcommit(struct xdr_stream *xdr,
 	*p++ = cpu_to_be32(0); /* Never send time_modify_changed */
 	*p++ = cpu_to_be32(NFS_SERVER(args->inode)->pnfs_curr_ld->id);/* type */
 
-	if (NFS_SERVER(inode)->pnfs_curr_ld->encode_layoutcommit) {
+	encode_uint32(xdr, args->layoutupdate_len);
-		NFS_SERVER(inode)->pnfs_curr_ld->encode_layoutcommit(
+	if (args->layoutupdate_pages)
-			NFS_I(inode)->layout, xdr, args);
+		xdr_write_pages(xdr, args->layoutupdate_pages, 0,
-	} else {
+				args->layoutupdate_len);
-		encode_uint32(xdr, args->layoutupdate_len);
-		if (args->layoutupdate_pages) {
-			xdr_write_pages(xdr, args->layoutupdate_pages, 0,
-					args->layoutupdate_len);
-		}
-	}
 
 	return 0;
 }
@@ -2024,7 +2013,6 @@ encode_layoutreturn(struct xdr_stream *xdr,
 		    const struct nfs4_layoutreturn_args *args,
 		    struct compound_hdr *hdr)
 {
-	const struct pnfs_layoutdriver_type *lr_ops = NFS_SERVER(args->inode)->pnfs_curr_ld;
 	__be32 *p;
 
 	encode_op_hdr(xdr, OP_LAYOUTRETURN, decode_layoutreturn_maxsz, hdr);
@@ -2041,8 +2029,6 @@ encode_layoutreturn(struct xdr_stream *xdr,
 	spin_unlock(&args->inode->i_lock);
 	if (args->ld_private->ops && args->ld_private->ops->encode)
 		args->ld_private->ops->encode(xdr, args, args->ld_private);
-	else if (lr_ops->encode_layoutreturn)
-		lr_ops->encode_layoutreturn(xdr, args);
 	else
 		encode_uint32(xdr, 0);
 }
@@ -5579,6 +5565,8 @@ static int decode_op_map(struct xdr_stream *xdr, struct nfs4_op_map *op_map)
 	unsigned int i;
 
 	p = xdr_inline_decode(xdr, 4);
+	if (!p)
+		return -EIO;
 	bitmap_words = be32_to_cpup(p++);
 	if (bitmap_words > NFS4_OP_MAP_NUM_WORDS)
 		return -EIO;

fs/nfs/objlayout/Kbuild

@@ -1,5 +0,0 @@
-#
-# Makefile for the pNFS Objects Layout Driver kernel module
-#
-objlayoutdriver-y := objio_osd.o pnfs_osd_xdr_cli.o objlayout.o
-obj-$(CONFIG_PNFS_OBJLAYOUT) += objlayoutdriver.o

fs/nfs/objlayout/objio_osd.c

@@ -1,675 +0,0 @@
-/*
- *  pNFS Objects layout implementation over open-osd initiator library
- *
- *  Copyright (C) 2009 Panasas Inc. [year of first publication]
- *  All rights reserved.
- *
- *  Benny Halevy <bhalevy@panasas.com>
- *  Boaz Harrosh <ooo@electrozaur.com>
- *
- *  This program is free software; you can redistribute it and/or modify
- *  it under the terms of the GNU General Public License version 2
- *  See the file COPYING included with this distribution for more details.
- *
- *  Redistribution and use in source and binary forms, with or without
- *  modification, are permitted provided that the following conditions
- *  are met:
- *
- *  1. Redistributions of source code must retain the above copyright
- *     notice, this list of conditions and the following disclaimer.
- *  2. Redistributions in binary form must reproduce the above copyright
- *     notice, this list of conditions and the following disclaimer in the
- *     documentation and/or other materials provided with the distribution.
- *  3. Neither the name of the Panasas company nor the names of its
- *     contributors may be used to endorse or promote products derived
- *     from this software without specific prior written permission.
- *
- *  THIS SOFTWARE IS PROVIDED ``AS IS'' AND ANY EXPRESS OR IMPLIED
- *  WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
- *  MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
- *  DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
- *  FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
- *  CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
- *  SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR
- *  BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
- *  LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
- *  NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
- *  SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
- */
-
-#include <linux/module.h>
-#include <scsi/osd_ore.h>
-
-#include "objlayout.h"
-#include "../internal.h"
-
-#define NFSDBG_FACILITY         NFSDBG_PNFS_LD
-
-struct objio_dev_ent {
-	struct nfs4_deviceid_node id_node;
-	struct ore_dev od;
-};
-
-static void
-objio_free_deviceid_node(struct nfs4_deviceid_node *d)
-{
-	struct objio_dev_ent *de = container_of(d, struct objio_dev_ent, id_node);
-
-	dprintk("%s: free od=%p\n", __func__, de->od.od);
-	osduld_put_device(de->od.od);
-	kfree_rcu(d, rcu);
-}
-
-struct objio_segment {
-	struct pnfs_layout_segment lseg;
-
-	struct ore_layout layout;
-	struct ore_components oc;
-};
-
-static inline struct objio_segment *
-OBJIO_LSEG(struct pnfs_layout_segment *lseg)
-{
-	return container_of(lseg, struct objio_segment, lseg);
-}
-
-struct objio_state {
-	/* Generic layer */
-	struct objlayout_io_res oir;
-
-	bool sync;
-	/*FIXME: Support for extra_bytes at ore_get_rw_state() */
-	struct ore_io_state *ios;
-};
-
-/* Send and wait for a get_device_info of devices in the layout,
-   then look them up with the osd_initiator library */
-struct nfs4_deviceid_node *
-objio_alloc_deviceid_node(struct nfs_server *server, struct pnfs_device *pdev,
-			gfp_t gfp_flags)
-{
-	struct pnfs_osd_deviceaddr *deviceaddr;
-	struct objio_dev_ent *ode = NULL;
-	struct osd_dev *od;
-	struct osd_dev_info odi;
-	bool retry_flag = true;
-	__be32 *p;
-	int err;
-
-	deviceaddr = kzalloc(sizeof(*deviceaddr), gfp_flags);
-	if (!deviceaddr)
-		return NULL;
-
-	p = page_address(pdev->pages[0]);
-	pnfs_osd_xdr_decode_deviceaddr(deviceaddr, p);
-
-	odi.systemid_len = deviceaddr->oda_systemid.len;
-	if (odi.systemid_len > sizeof(odi.systemid)) {
-		dprintk("%s: odi.systemid_len > sizeof(systemid=%zd)\n",
-			__func__, sizeof(odi.systemid));
-		err = -EINVAL;
-		goto out;
-	} else if (odi.systemid_len)
-		memcpy(odi.systemid, deviceaddr->oda_systemid.data,
-		       odi.systemid_len);
-	odi.osdname_len	 = deviceaddr->oda_osdname.len;
-	odi.osdname	 = (u8 *)deviceaddr->oda_osdname.data;
-
-	if (!odi.osdname_len && !odi.systemid_len) {
-		dprintk("%s: !odi.osdname_len && !odi.systemid_len\n",
-			__func__);
-		err = -ENODEV;
-		goto out;
-	}
-
-retry_lookup:
-	od = osduld_info_lookup(&odi);
-	if (IS_ERR(od)) {
-		err = PTR_ERR(od);
-		dprintk("%s: osduld_info_lookup => %d\n", __func__, err);
-		if (err == -ENODEV && retry_flag) {
-			err = objlayout_autologin(deviceaddr);
-			if (likely(!err)) {
-				retry_flag = false;
-				goto retry_lookup;
-			}
-		}
-		goto out;
-	}
-
-	dprintk("Adding new dev_id(%llx:%llx)\n",
-		_DEVID_LO(&pdev->dev_id), _DEVID_HI(&pdev->dev_id));
-
-	ode = kzalloc(sizeof(*ode), gfp_flags);
-	if (!ode) {
-		dprintk("%s: -ENOMEM od=%p\n", __func__, od);
-		goto out;
-	}
-
-	nfs4_init_deviceid_node(&ode->id_node, server, &pdev->dev_id);
-	kfree(deviceaddr);
-
-	ode->od.od = od;
-	return &ode->id_node;
-
-out:
-	kfree(deviceaddr);
-	return NULL;
-}
-
-static void copy_single_comp(struct ore_components *oc, unsigned c,
-			     struct pnfs_osd_object_cred *src_comp)
-{
-	struct ore_comp *ocomp = &oc->comps[c];
-
-	WARN_ON(src_comp->oc_cap_key.cred_len > 0); /* libosd is NO_SEC only */
-	WARN_ON(src_comp->oc_cap.cred_len > sizeof(ocomp->cred));
-
-	ocomp->obj.partition = src_comp->oc_object_id.oid_partition_id;
-	ocomp->obj.id = src_comp->oc_object_id.oid_object_id;
-
-	memcpy(ocomp->cred, src_comp->oc_cap.cred, sizeof(ocomp->cred));
-}
-
-static int __alloc_objio_seg(unsigned numdevs, gfp_t gfp_flags,
-		       struct objio_segment **pseg)
-{
-/*	This is the in memory structure of the objio_segment
- *
- *	struct __alloc_objio_segment {
- *		struct objio_segment olseg;
- *		struct ore_dev *ods[numdevs];
- *		struct ore_comp	comps[numdevs];
- *	} *aolseg;
- *	NOTE: The code as above compiles and runs perfectly. It is elegant,
- *	type safe and compact. At some Past time Linus has decided he does not
- *	like variable length arrays, For the sake of this principal we uglify
- *	the code as below.
- */
-	struct objio_segment *lseg;
-	size_t lseg_size = sizeof(*lseg) +
-			numdevs * sizeof(lseg->oc.ods[0]) +
-			numdevs * sizeof(*lseg->oc.comps);
-
-	lseg = kzalloc(lseg_size, gfp_flags);
-	if (unlikely(!lseg)) {
-		dprintk("%s: Failed allocation numdevs=%d size=%zd\n", __func__,
-			numdevs, lseg_size);
-		return -ENOMEM;
-	}
-
-	lseg->oc.numdevs = numdevs;
-	lseg->oc.single_comp = EC_MULTPLE_COMPS;
-	lseg->oc.ods = (void *)(lseg + 1);
-	lseg->oc.comps = (void *)(lseg->oc.ods + numdevs);
-
-	*pseg = lseg;
-	return 0;
-}
-
-int objio_alloc_lseg(struct pnfs_layout_segment **outp,
-	struct pnfs_layout_hdr *pnfslay,
-	struct pnfs_layout_range *range,
-	struct xdr_stream *xdr,
-	gfp_t gfp_flags)
-{
-	struct nfs_server *server = NFS_SERVER(pnfslay->plh_inode);
-	struct objio_segment *objio_seg;
-	struct pnfs_osd_xdr_decode_layout_iter iter;
-	struct pnfs_osd_layout layout;
-	struct pnfs_osd_object_cred src_comp;
-	unsigned cur_comp;
-	int err;
-
-	err = pnfs_osd_xdr_decode_layout_map(&layout, &iter, xdr);
-	if (unlikely(err))
-		return err;
-
-	err = __alloc_objio_seg(layout.olo_num_comps, gfp_flags, &objio_seg);
-	if (unlikely(err))
-		return err;
-
-	objio_seg->layout.stripe_unit = layout.olo_map.odm_stripe_unit;
-	objio_seg->layout.group_width = layout.olo_map.odm_group_width;
-	objio_seg->layout.group_depth = layout.olo_map.odm_group_depth;
-	objio_seg->layout.mirrors_p1 = layout.olo_map.odm_mirror_cnt + 1;
-	objio_seg->layout.raid_algorithm = layout.olo_map.odm_raid_algorithm;
-
-	err = ore_verify_layout(layout.olo_map.odm_num_comps,
-					  &objio_seg->layout);
-	if (unlikely(err))
-		goto err;
-
-	objio_seg->oc.first_dev = layout.olo_comps_index;
-	cur_comp = 0;
-	while (pnfs_osd_xdr_decode_layout_comp(&src_comp, &iter, xdr, &err)) {
-		struct nfs4_deviceid_node *d;
-		struct objio_dev_ent *ode;
-
-		copy_single_comp(&objio_seg->oc, cur_comp, &src_comp);
-
-		d = nfs4_find_get_deviceid(server,
-				&src_comp.oc_object_id.oid_device_id,
-				pnfslay->plh_lc_cred, gfp_flags);
-		if (!d) {
-			err = -ENXIO;
-			goto err;
-		}
-
-		ode = container_of(d, struct objio_dev_ent, id_node);
-		objio_seg->oc.ods[cur_comp++] = &ode->od;
-	}
-	/* pnfs_osd_xdr_decode_layout_comp returns false on error */
-	if (unlikely(err))
-		goto err;
-
-	*outp = &objio_seg->lseg;
-	return 0;
-
-err:
-	kfree(objio_seg);
-	dprintk("%s: Error: return %d\n", __func__, err);
-	*outp = NULL;
-	return err;
-}
-
-void objio_free_lseg(struct pnfs_layout_segment *lseg)
-{
-	int i;
-	struct objio_segment *objio_seg = OBJIO_LSEG(lseg);
-
-	for (i = 0; i < objio_seg->oc.numdevs; i++) {
-		struct ore_dev *od = objio_seg->oc.ods[i];
-		struct objio_dev_ent *ode;
-
-		if (!od)
-			break;
-		ode = container_of(od, typeof(*ode), od);
-		nfs4_put_deviceid_node(&ode->id_node);
-	}
-	kfree(objio_seg);
-}
-
-static int
-objio_alloc_io_state(struct pnfs_layout_hdr *pnfs_layout_type, bool is_reading,
-	struct pnfs_layout_segment *lseg, struct page **pages, unsigned pgbase,
-	loff_t offset, size_t count, void *rpcdata, gfp_t gfp_flags,
-	struct objio_state **outp)
-{
-	struct objio_segment *objio_seg = OBJIO_LSEG(lseg);
-	struct ore_io_state *ios;
-	int ret;
-	struct __alloc_objio_state {
-		struct objio_state objios;
-		struct pnfs_osd_ioerr ioerrs[objio_seg->oc.numdevs];
-	} *aos;
-
-	aos = kzalloc(sizeof(*aos), gfp_flags);
-	if (unlikely(!aos))
-		return -ENOMEM;
-
-	objlayout_init_ioerrs(&aos->objios.oir, objio_seg->oc.numdevs,
-			aos->ioerrs, rpcdata, pnfs_layout_type);
-
-	ret = ore_get_rw_state(&objio_seg->layout, &objio_seg->oc, is_reading,
-			       offset, count, &ios);
-	if (unlikely(ret)) {
-		kfree(aos);
-		return ret;
-	}
-
-	ios->pages = pages;
-	ios->pgbase = pgbase;
-	ios->private = aos;
-	BUG_ON(ios->nr_pages > (pgbase + count + PAGE_SIZE - 1) >> PAGE_SHIFT);
-
-	aos->objios.sync = 0;
-	aos->objios.ios = ios;
-	*outp = &aos->objios;
-	return 0;
-}
-
-void objio_free_result(struct objlayout_io_res *oir)
-{
-	struct objio_state *objios = container_of(oir, struct objio_state, oir);
-
-	ore_put_io_state(objios->ios);
-	kfree(objios);
-}
-
-static enum pnfs_osd_errno osd_pri_2_pnfs_err(enum osd_err_priority oep)
-{
-	switch (oep) {
-	case OSD_ERR_PRI_NO_ERROR:
-		return (enum pnfs_osd_errno)0;
-
-	case OSD_ERR_PRI_CLEAR_PAGES:
-		BUG_ON(1);
-		return 0;
-
-	case OSD_ERR_PRI_RESOURCE:
-		return PNFS_OSD_ERR_RESOURCE;
-	case OSD_ERR_PRI_BAD_CRED:
-		return PNFS_OSD_ERR_BAD_CRED;
-	case OSD_ERR_PRI_NO_ACCESS:
-		return PNFS_OSD_ERR_NO_ACCESS;
-	case OSD_ERR_PRI_UNREACHABLE:
-		return PNFS_OSD_ERR_UNREACHABLE;
-	case OSD_ERR_PRI_NOT_FOUND:
-		return PNFS_OSD_ERR_NOT_FOUND;
-	case OSD_ERR_PRI_NO_SPACE:
-		return PNFS_OSD_ERR_NO_SPACE;
-	default:
-		WARN_ON(1);
-		/* fallthrough */
-	case OSD_ERR_PRI_EIO:
-		return PNFS_OSD_ERR_EIO;
-	}
-}
-
-static void __on_dev_error(struct ore_io_state *ios,
-	struct ore_dev *od, unsigned dev_index, enum osd_err_priority oep,
-	u64 dev_offset, u64  dev_len)
-{
-	struct objio_state *objios = ios->private;
-	struct pnfs_osd_objid pooid;
-	struct objio_dev_ent *ode = container_of(od, typeof(*ode), od);
-	/* FIXME: what to do with more-then-one-group layouts. We need to
-	 * translate from ore_io_state index to oc->comps index
-	 */
-	unsigned comp = dev_index;
-
-	pooid.oid_device_id = ode->id_node.deviceid;
-	pooid.oid_partition_id = ios->oc->comps[comp].obj.partition;
-	pooid.oid_object_id = ios->oc->comps[comp].obj.id;
-
-	objlayout_io_set_result(&objios->oir, comp,
-				&pooid, osd_pri_2_pnfs_err(oep),
-				dev_offset, dev_len, !ios->reading);
-}
-
-/*
- * read
- */
-static void _read_done(struct ore_io_state *ios, void *private)
-{
-	struct objio_state *objios = private;
-	ssize_t status;
-	int ret = ore_check_io(ios, &__on_dev_error);
-
-	/* FIXME: _io_free(ios) can we dealocate the libosd resources; */
-
-	if (likely(!ret))
-		status = ios->length;
-	else
-		status = ret;
-
-	objlayout_read_done(&objios->oir, status, objios->sync);
-}
-
-int objio_read_pagelist(struct nfs_pgio_header *hdr)
-{
-	struct objio_state *objios;
-	int ret;
-
-	ret = objio_alloc_io_state(NFS_I(hdr->inode)->layout, true,
-			hdr->lseg, hdr->args.pages, hdr->args.pgbase,
-			hdr->args.offset, hdr->args.count, hdr,
-			GFP_KERNEL, &objios);
-	if (unlikely(ret))
-		return ret;
-
-	objios->ios->done = _read_done;
-	dprintk("%s: offset=0x%llx length=0x%x\n", __func__,
-		hdr->args.offset, hdr->args.count);
-	ret = ore_read(objios->ios);
-	if (unlikely(ret))
-		objio_free_result(&objios->oir);
-	return ret;
-}
-
-/*
- * write
- */
-static void _write_done(struct ore_io_state *ios, void *private)
-{
-	struct objio_state *objios = private;
-	ssize_t status;
-	int ret = ore_check_io(ios, &__on_dev_error);
-
-	/* FIXME: _io_free(ios) can we dealocate the libosd resources; */
-
-	if (likely(!ret)) {
-		/* FIXME: should be based on the OSD's persistence model
-		 * See OSD2r05 Section 4.13 Data persistence model */
-		objios->oir.committed = NFS_FILE_SYNC;
-		status = ios->length;
-	} else {
-		status = ret;
-	}
-
-	objlayout_write_done(&objios->oir, status, objios->sync);
-}
-
-static struct page *__r4w_get_page(void *priv, u64 offset, bool *uptodate)
-{
-	struct objio_state *objios = priv;
-	struct nfs_pgio_header *hdr = objios->oir.rpcdata;
-	struct address_space *mapping = hdr->inode->i_mapping;
-	pgoff_t index = offset / PAGE_SIZE;
-	struct page *page;
-	loff_t i_size = i_size_read(hdr->inode);
-
-	if (offset >= i_size) {
-		*uptodate = true;
-		dprintk("%s: g_zero_page index=0x%lx\n", __func__, index);
-		return ZERO_PAGE(0);
-	}
-
-	page = find_get_page(mapping, index);
-	if (!page) {
-		page = find_or_create_page(mapping, index, GFP_NOFS);
-		if (unlikely(!page)) {
-			dprintk("%s: grab_cache_page Failed index=0x%lx\n",
-				__func__, index);
-			return NULL;
-		}
-		unlock_page(page);
-	}
-	*uptodate = PageUptodate(page);
-	dprintk("%s: index=0x%lx uptodate=%d\n", __func__, index, *uptodate);
-	return page;
-}
-
-static void __r4w_put_page(void *priv, struct page *page)
-{
-	dprintk("%s: index=0x%lx\n", __func__,
-		(page == ZERO_PAGE(0)) ? -1UL : page->index);
-	if (ZERO_PAGE(0) != page)
-		put_page(page);
-	return;
-}
-
-static const struct _ore_r4w_op _r4w_op = {
-	.get_page = &__r4w_get_page,
-	.put_page = &__r4w_put_page,
-};
-
-int objio_write_pagelist(struct nfs_pgio_header *hdr, int how)
-{
-	struct objio_state *objios;
-	int ret;
-
-	ret = objio_alloc_io_state(NFS_I(hdr->inode)->layout, false,
-			hdr->lseg, hdr->args.pages, hdr->args.pgbase,
-			hdr->args.offset, hdr->args.count, hdr, GFP_NOFS,
-			&objios);
-	if (unlikely(ret))
-		return ret;
-
-	objios->sync = 0 != (how & FLUSH_SYNC);
-	objios->ios->r4w = &_r4w_op;
-
-	if (!objios->sync)
-		objios->ios->done = _write_done;
-
-	dprintk("%s: offset=0x%llx length=0x%x\n", __func__,
-		hdr->args.offset, hdr->args.count);
-	ret = ore_write(objios->ios);
-	if (unlikely(ret)) {
-		objio_free_result(&objios->oir);
-		return ret;
-	}
-
-	if (objios->sync)
-		_write_done(objios->ios, objios);
-
-	return 0;
-}
-
-/*
- * Return 0 if @req cannot be coalesced into @pgio, otherwise return the number
- * of bytes (maximum @req->wb_bytes) that can be coalesced.
- */
-static size_t objio_pg_test(struct nfs_pageio_descriptor *pgio,
-			  struct nfs_page *prev, struct nfs_page *req)
-{
-	struct nfs_pgio_mirror *mirror = nfs_pgio_current_mirror(pgio);
-	unsigned int size;
-
-	size = pnfs_generic_pg_test(pgio, prev, req);
-
-	if (!size || mirror->pg_count + req->wb_bytes >
-	    (unsigned long)pgio->pg_layout_private)
-		return 0;
-
-	return min(size, req->wb_bytes);
-}
-
-static void objio_init_read(struct nfs_pageio_descriptor *pgio, struct nfs_page *req)
-{
-	pnfs_generic_pg_init_read(pgio, req);
-	if (unlikely(pgio->pg_lseg == NULL))
-		return; /* Not pNFS */
-
-	pgio->pg_layout_private = (void *)
-				OBJIO_LSEG(pgio->pg_lseg)->layout.max_io_length;
-}
-
-static bool aligned_on_raid_stripe(u64 offset, struct ore_layout *layout,
-				   unsigned long *stripe_end)
-{
-	u32 stripe_off;
-	unsigned stripe_size;
-
-	if (layout->raid_algorithm == PNFS_OSD_RAID_0)
-		return true;
-
-	stripe_size = layout->stripe_unit *
-				(layout->group_width - layout->parity);
-
-	div_u64_rem(offset, stripe_size, &stripe_off);
-	if (!stripe_off)
-		return true;
-
-	*stripe_end = stripe_size - stripe_off;
-	return false;
-}
-
-static void objio_init_write(struct nfs_pageio_descriptor *pgio, struct nfs_page *req)
-{
-	unsigned long stripe_end = 0;
-	u64 wb_size;
-
-	if (pgio->pg_dreq == NULL)
-		wb_size = i_size_read(pgio->pg_inode) - req_offset(req);
-	else
-		wb_size = nfs_dreq_bytes_left(pgio->pg_dreq);
-
-	pnfs_generic_pg_init_write(pgio, req, wb_size);
-	if (unlikely(pgio->pg_lseg == NULL))
-		return; /* Not pNFS */
-
-	if (req->wb_offset ||
-	    !aligned_on_raid_stripe(req->wb_index * PAGE_SIZE,
-			       &OBJIO_LSEG(pgio->pg_lseg)->layout,
-			       &stripe_end)) {
-		pgio->pg_layout_private = (void *)stripe_end;
-	} else {
-		pgio->pg_layout_private = (void *)
-				OBJIO_LSEG(pgio->pg_lseg)->layout.max_io_length;
-	}
-}
-
-static const struct nfs_pageio_ops objio_pg_read_ops = {
-	.pg_init = objio_init_read,
-	.pg_test = objio_pg_test,
-	.pg_doio = pnfs_generic_pg_readpages,
-	.pg_cleanup = pnfs_generic_pg_cleanup,
-};
-
-static const struct nfs_pageio_ops objio_pg_write_ops = {
-	.pg_init = objio_init_write,
-	.pg_test = objio_pg_test,
-	.pg_doio = pnfs_generic_pg_writepages,
-	.pg_cleanup = pnfs_generic_pg_cleanup,
-};
-
-static struct pnfs_layoutdriver_type objlayout_type = {
-	.id = LAYOUT_OSD2_OBJECTS,
-	.name = "LAYOUT_OSD2_OBJECTS",
-	.flags                   = PNFS_LAYOUTRET_ON_SETATTR |
-				   PNFS_LAYOUTRET_ON_ERROR,
-
-	.max_deviceinfo_size	 = PAGE_SIZE,
-	.owner		       	 = THIS_MODULE,
-	.alloc_layout_hdr        = objlayout_alloc_layout_hdr,
-	.free_layout_hdr         = objlayout_free_layout_hdr,
-
-	.alloc_lseg              = objlayout_alloc_lseg,
-	.free_lseg               = objlayout_free_lseg,
-
-	.read_pagelist           = objlayout_read_pagelist,
-	.write_pagelist          = objlayout_write_pagelist,
-	.pg_read_ops             = &objio_pg_read_ops,
-	.pg_write_ops            = &objio_pg_write_ops,
-
-	.sync			 = pnfs_generic_sync,
-
-	.free_deviceid_node	 = objio_free_deviceid_node,
-
-	.encode_layoutcommit	 = objlayout_encode_layoutcommit,
-	.encode_layoutreturn     = objlayout_encode_layoutreturn,
-};
-
-MODULE_DESCRIPTION("pNFS Layout Driver for OSD2 objects");
-MODULE_AUTHOR("Benny Halevy <bhalevy@panasas.com>");
-MODULE_LICENSE("GPL");
-
-static int __init
-objlayout_init(void)
-{
-	int ret = pnfs_register_layoutdriver(&objlayout_type);
-
-	if (ret)
-		printk(KERN_INFO
-			"NFS: %s: Registering OSD pNFS Layout Driver failed: error=%d\n",
-			__func__, ret);
-	else
-		printk(KERN_INFO "NFS: %s: Registered OSD pNFS Layout Driver\n",
-			__func__);
-	return ret;
-}
-
-static void __exit
-objlayout_exit(void)
-{
-	pnfs_unregister_layoutdriver(&objlayout_type);
-	printk(KERN_INFO "NFS: %s: Unregistered OSD pNFS Layout Driver\n",
-	       __func__);
-}
-
-MODULE_ALIAS("nfs-layouttype4-2");
-
-module_init(objlayout_init);
-module_exit(objlayout_exit);

fs/nfs/objlayout/objlayout.c

@@ -1,706 +0,0 @@
-/*
- *  pNFS Objects layout driver high level definitions
- *
- *  Copyright (C) 2007 Panasas Inc. [year of first publication]
- *  All rights reserved.
- *
- *  Benny Halevy <bhalevy@panasas.com>
- *  Boaz Harrosh <ooo@electrozaur.com>
- *
- *  This program is free software; you can redistribute it and/or modify
- *  it under the terms of the GNU General Public License version 2
- *  See the file COPYING included with this distribution for more details.
- *
- *  Redistribution and use in source and binary forms, with or without
- *  modification, are permitted provided that the following conditions
- *  are met:
- *
- *  1. Redistributions of source code must retain the above copyright
- *     notice, this list of conditions and the following disclaimer.
- *  2. Redistributions in binary form must reproduce the above copyright
- *     notice, this list of conditions and the following disclaimer in the
- *     documentation and/or other materials provided with the distribution.
- *  3. Neither the name of the Panasas company nor the names of its
- *     contributors may be used to endorse or promote products derived
- *     from this software without specific prior written permission.
- *
- *  THIS SOFTWARE IS PROVIDED ``AS IS'' AND ANY EXPRESS OR IMPLIED
- *  WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
- *  MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
- *  DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
- *  FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
- *  CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
- *  SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR
- *  BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
- *  LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
- *  NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
- *  SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
- */
-
-#include <linux/kmod.h>
-#include <linux/moduleparam.h>
-#include <linux/ratelimit.h>
-#include <scsi/osd_initiator.h>
-#include "objlayout.h"
-
-#define NFSDBG_FACILITY         NFSDBG_PNFS_LD
-/*
- * Create a objlayout layout structure for the given inode and return it.
- */
-struct pnfs_layout_hdr *
-objlayout_alloc_layout_hdr(struct inode *inode, gfp_t gfp_flags)
-{
-	struct objlayout *objlay;
-
-	objlay = kzalloc(sizeof(struct objlayout), gfp_flags);
-	if (!objlay)
-		return NULL;
-	spin_lock_init(&objlay->lock);
-	INIT_LIST_HEAD(&objlay->err_list);
-	dprintk("%s: Return %p\n", __func__, objlay);
-	return &objlay->pnfs_layout;
-}
-
-/*
- * Free an objlayout layout structure
- */
-void
-objlayout_free_layout_hdr(struct pnfs_layout_hdr *lo)
-{
-	struct objlayout *objlay = OBJLAYOUT(lo);
-
-	dprintk("%s: objlay %p\n", __func__, objlay);
-
-	WARN_ON(!list_empty(&objlay->err_list));
-	kfree(objlay);
-}
-
-/*
- * Unmarshall layout and store it in pnfslay.
- */
-struct pnfs_layout_segment *
-objlayout_alloc_lseg(struct pnfs_layout_hdr *pnfslay,
-		     struct nfs4_layoutget_res *lgr,
-		     gfp_t gfp_flags)
-{
-	int status = -ENOMEM;
-	struct xdr_stream stream;
-	struct xdr_buf buf = {
-		.pages =  lgr->layoutp->pages,
-		.page_len =  lgr->layoutp->len,
-		.buflen =  lgr->layoutp->len,
-		.len = lgr->layoutp->len,
-	};
-	struct page *scratch;
-	struct pnfs_layout_segment *lseg;
-
-	dprintk("%s: Begin pnfslay %p\n", __func__, pnfslay);
-
-	scratch = alloc_page(gfp_flags);
-	if (!scratch)
-		goto err_nofree;
-
-	xdr_init_decode(&stream, &buf, NULL);
-	xdr_set_scratch_buffer(&stream, page_address(scratch), PAGE_SIZE);
-
-	status = objio_alloc_lseg(&lseg, pnfslay, &lgr->range, &stream, gfp_flags);
-	if (unlikely(status)) {
-		dprintk("%s: objio_alloc_lseg Return err %d\n", __func__,
-			status);
-		goto err;
-	}
-
-	__free_page(scratch);
-
-	dprintk("%s: Return %p\n", __func__, lseg);
-	return lseg;
-
-err:
-	__free_page(scratch);
-err_nofree:
-	dprintk("%s: Err Return=>%d\n", __func__, status);
-	return ERR_PTR(status);
-}
-
-/*
- * Free a layout segement
- */
-void
-objlayout_free_lseg(struct pnfs_layout_segment *lseg)
-{
-	dprintk("%s: freeing layout segment %p\n", __func__, lseg);
-
-	if (unlikely(!lseg))
-		return;
-
-	objio_free_lseg(lseg);
-}
-
-/*
- * I/O Operations
- */
-static inline u64
-end_offset(u64 start, u64 len)
-{
-	u64 end;
-
-	end = start + len;
-	return end >= start ? end : NFS4_MAX_UINT64;
-}
-
-static void _fix_verify_io_params(struct pnfs_layout_segment *lseg,
-			   struct page ***p_pages, unsigned *p_pgbase,
-			   u64 offset, unsigned long count)
-{
-	u64 lseg_end_offset;
-
-	BUG_ON(offset < lseg->pls_range.offset);
-	lseg_end_offset = end_offset(lseg->pls_range.offset,
-				     lseg->pls_range.length);
-	BUG_ON(offset >= lseg_end_offset);
-	WARN_ON(offset + count > lseg_end_offset);
-
-	if (*p_pgbase > PAGE_SIZE) {
-		dprintk("%s: pgbase(0x%x) > PAGE_SIZE\n", __func__, *p_pgbase);
-		*p_pages += *p_pgbase >> PAGE_SHIFT;
-		*p_pgbase &= ~PAGE_MASK;
-	}
-}
-
-/*
- * I/O done common code
- */
-static void
-objlayout_iodone(struct objlayout_io_res *oir)
-{
-	if (likely(oir->status >= 0)) {
-		objio_free_result(oir);
-	} else {
-		struct objlayout *objlay = oir->objlay;
-
-		spin_lock(&objlay->lock);
-		objlay->delta_space_valid = OBJ_DSU_INVALID;
-		list_add(&objlay->err_list, &oir->err_list);
-		spin_unlock(&objlay->lock);
-	}
-}
-
-/*
- * objlayout_io_set_result - Set an osd_error code on a specific osd comp.
- *
- * The @index component IO failed (error returned from target). Register
- * the error for later reporting at layout-return.
- */
-void
-objlayout_io_set_result(struct objlayout_io_res *oir, unsigned index,
-			struct pnfs_osd_objid *pooid, int osd_error,
-			u64 offset, u64 length, bool is_write)
-{
-	struct pnfs_osd_ioerr *ioerr = &oir->ioerrs[index];
-
-	BUG_ON(index >= oir->num_comps);
-	if (osd_error) {
-		ioerr->oer_component = *pooid;
-		ioerr->oer_comp_offset = offset;
-		ioerr->oer_comp_length = length;
-		ioerr->oer_iswrite = is_write;
-		ioerr->oer_errno = osd_error;
-
-		dprintk("%s: err[%d]: errno=%d is_write=%d dev(%llx:%llx) "
-			"par=0x%llx obj=0x%llx offset=0x%llx length=0x%llx\n",
-			__func__, index, ioerr->oer_errno,
-			ioerr->oer_iswrite,
-			_DEVID_LO(&ioerr->oer_component.oid_device_id),
-			_DEVID_HI(&ioerr->oer_component.oid_device_id),
-			ioerr->oer_component.oid_partition_id,
-			ioerr->oer_component.oid_object_id,
-			ioerr->oer_comp_offset,
-			ioerr->oer_comp_length);
-	} else {
-		/* User need not call if no error is reported */
-		ioerr->oer_errno = 0;
-	}
-}
-
-/* Function scheduled on rpc workqueue to call ->nfs_readlist_complete().
- * This is because the osd completion is called with ints-off from
- * the block layer
- */
-static void _rpc_read_complete(struct work_struct *work)
-{
-	struct rpc_task *task;
-	struct nfs_pgio_header *hdr;
-
-	dprintk("%s enter\n", __func__);
-	task = container_of(work, struct rpc_task, u.tk_work);
-	hdr = container_of(task, struct nfs_pgio_header, task);
-
-	pnfs_ld_read_done(hdr);
-}
-
-void
-objlayout_read_done(struct objlayout_io_res *oir, ssize_t status, bool sync)
-{
-	struct nfs_pgio_header *hdr = oir->rpcdata;
-
-	oir->status = hdr->task.tk_status = status;
-	if (status >= 0)
-		hdr->res.count = status;
-	else
-		hdr->pnfs_error = status;
-	objlayout_iodone(oir);
-	/* must not use oir after this point */
-
-	dprintk("%s: Return status=%zd eof=%d sync=%d\n", __func__,
-		status, hdr->res.eof, sync);
-
-	if (sync)
-		pnfs_ld_read_done(hdr);
-	else {
-		INIT_WORK(&hdr->task.u.tk_work, _rpc_read_complete);
-		schedule_work(&hdr->task.u.tk_work);
-	}
-}
-
-/*
- * Perform sync or async reads.
- */
-enum pnfs_try_status
-objlayout_read_pagelist(struct nfs_pgio_header *hdr)
-{
-	struct inode *inode = hdr->inode;
-	loff_t offset = hdr->args.offset;
-	size_t count = hdr->args.count;
-	int err;
-	loff_t eof;
-
-	eof = i_size_read(inode);
-	if (unlikely(offset + count > eof)) {
-		if (offset >= eof) {
-			err = 0;
-			hdr->res.count = 0;
-			hdr->res.eof = 1;
-			/*FIXME: do we need to call pnfs_ld_read_done() */
-			goto out;
-		}
-		count = eof - offset;
-	}
-
-	hdr->res.eof = (offset + count) >= eof;
-	_fix_verify_io_params(hdr->lseg, &hdr->args.pages,
-			      &hdr->args.pgbase,
-			      hdr->args.offset, hdr->args.count);
-
-	dprintk("%s: inode(%lx) offset 0x%llx count 0x%zx eof=%d\n",
-		__func__, inode->i_ino, offset, count, hdr->res.eof);
-
-	err = objio_read_pagelist(hdr);
- out:
-	if (unlikely(err)) {
-		hdr->pnfs_error = err;
-		dprintk("%s: Returned Error %d\n", __func__, err);
-		return PNFS_NOT_ATTEMPTED;
-	}
-	return PNFS_ATTEMPTED;
-}
-
-/* Function scheduled on rpc workqueue to call ->nfs_writelist_complete().
- * This is because the osd completion is called with ints-off from
- * the block layer
- */
-static void _rpc_write_complete(struct work_struct *work)
-{
-	struct rpc_task *task;
-	struct nfs_pgio_header *hdr;
-
-	dprintk("%s enter\n", __func__);
-	task = container_of(work, struct rpc_task, u.tk_work);
-	hdr = container_of(task, struct nfs_pgio_header, task);
-
-	pnfs_ld_write_done(hdr);
-}
-
-void
-objlayout_write_done(struct objlayout_io_res *oir, ssize_t status, bool sync)
-{
-	struct nfs_pgio_header *hdr = oir->rpcdata;
-
-	oir->status = hdr->task.tk_status = status;
-	if (status >= 0) {
-		hdr->res.count = status;
-		hdr->verf.committed = oir->committed;
-	} else {
-		hdr->pnfs_error = status;
-	}
-	objlayout_iodone(oir);
-	/* must not use oir after this point */
-
-	dprintk("%s: Return status %zd committed %d sync=%d\n", __func__,
-		status, hdr->verf.committed, sync);
-
-	if (sync)
-		pnfs_ld_write_done(hdr);
-	else {
-		INIT_WORK(&hdr->task.u.tk_work, _rpc_write_complete);
-		schedule_work(&hdr->task.u.tk_work);
-	}
-}
-
-/*
- * Perform sync or async writes.
- */
-enum pnfs_try_status
-objlayout_write_pagelist(struct nfs_pgio_header *hdr, int how)
-{
-	int err;
-
-	_fix_verify_io_params(hdr->lseg, &hdr->args.pages,
-			      &hdr->args.pgbase,
-			      hdr->args.offset, hdr->args.count);
-
-	err = objio_write_pagelist(hdr, how);
-	if (unlikely(err)) {
-		hdr->pnfs_error = err;
-		dprintk("%s: Returned Error %d\n", __func__, err);
-		return PNFS_NOT_ATTEMPTED;
-	}
-	return PNFS_ATTEMPTED;
-}
-
-void
-objlayout_encode_layoutcommit(struct pnfs_layout_hdr *pnfslay,
-			      struct xdr_stream *xdr,
-			      const struct nfs4_layoutcommit_args *args)
-{
-	struct objlayout *objlay = OBJLAYOUT(pnfslay);
-	struct pnfs_osd_layoutupdate lou;
-	__be32 *start;
-
-	dprintk("%s: Begin\n", __func__);
-
-	spin_lock(&objlay->lock);
-	lou.dsu_valid = (objlay->delta_space_valid == OBJ_DSU_VALID);
-	lou.dsu_delta = objlay->delta_space_used;
-	objlay->delta_space_used = 0;
-	objlay->delta_space_valid = OBJ_DSU_INIT;
-	lou.olu_ioerr_flag = !list_empty(&objlay->err_list);
-	spin_unlock(&objlay->lock);
-
-	start = xdr_reserve_space(xdr, 4);
-
-	BUG_ON(pnfs_osd_xdr_encode_layoutupdate(xdr, &lou));
-
-	*start = cpu_to_be32((xdr->p - start - 1) * 4);
-
-	dprintk("%s: Return delta_space_used %lld err %d\n", __func__,
-		lou.dsu_delta, lou.olu_ioerr_flag);
-}
-
-static int
-err_prio(u32 oer_errno)
-{
-	switch (oer_errno) {
-	case 0:
-		return 0;
-
-	case PNFS_OSD_ERR_RESOURCE:
-		return OSD_ERR_PRI_RESOURCE;
-	case PNFS_OSD_ERR_BAD_CRED:
-		return OSD_ERR_PRI_BAD_CRED;
-	case PNFS_OSD_ERR_NO_ACCESS:
-		return OSD_ERR_PRI_NO_ACCESS;
-	case PNFS_OSD_ERR_UNREACHABLE:
-		return OSD_ERR_PRI_UNREACHABLE;
-	case PNFS_OSD_ERR_NOT_FOUND:
-		return OSD_ERR_PRI_NOT_FOUND;
-	case PNFS_OSD_ERR_NO_SPACE:
-		return OSD_ERR_PRI_NO_SPACE;
-	default:
-		WARN_ON(1);
-		/* fallthrough */
-	case PNFS_OSD_ERR_EIO:
-		return OSD_ERR_PRI_EIO;
-	}
-}
-
-static void
-merge_ioerr(struct pnfs_osd_ioerr *dest_err,
-	    const struct pnfs_osd_ioerr *src_err)
-{
-	u64 dest_end, src_end;
-
-	if (!dest_err->oer_errno) {
-		*dest_err = *src_err;
-		/* accumulated device must be blank */
-		memset(&dest_err->oer_component.oid_device_id, 0,
-			sizeof(dest_err->oer_component.oid_device_id));
-
-		return;
-	}
-
-	if (dest_err->oer_component.oid_partition_id !=
-				src_err->oer_component.oid_partition_id)
-		dest_err->oer_component.oid_partition_id = 0;
-
-	if (dest_err->oer_component.oid_object_id !=
-				src_err->oer_component.oid_object_id)
-		dest_err->oer_component.oid_object_id = 0;
-
-	if (dest_err->oer_comp_offset > src_err->oer_comp_offset)
-		dest_err->oer_comp_offset = src_err->oer_comp_offset;
-
-	dest_end = end_offset(dest_err->oer_comp_offset,
-			      dest_err->oer_comp_length);
-	src_end =  end_offset(src_err->oer_comp_offset,
-			      src_err->oer_comp_length);
-	if (dest_end < src_end)
-		dest_end = src_end;
-
-	dest_err->oer_comp_length = dest_end - dest_err->oer_comp_offset;
-
-	if ((src_err->oer_iswrite == dest_err->oer_iswrite) &&
-	    (err_prio(src_err->oer_errno) > err_prio(dest_err->oer_errno))) {
-			dest_err->oer_errno = src_err->oer_errno;
-	} else if (src_err->oer_iswrite) {
-		dest_err->oer_iswrite = true;
-		dest_err->oer_errno = src_err->oer_errno;
-	}
-}
-
-static void
-encode_accumulated_error(struct objlayout *objlay, __be32 *p)
-{
-	struct objlayout_io_res *oir, *tmp;
-	struct pnfs_osd_ioerr accumulated_err = {.oer_errno = 0};
-
-	list_for_each_entry_safe(oir, tmp, &objlay->err_list, err_list) {
-		unsigned i;
-
-		for (i = 0; i < oir->num_comps; i++) {
-			struct pnfs_osd_ioerr *ioerr = &oir->ioerrs[i];
-
-			if (!ioerr->oer_errno)
-				continue;
-
-			printk(KERN_ERR "NFS: %s: err[%d]: errno=%d "
-				"is_write=%d dev(%llx:%llx) par=0x%llx "
-				"obj=0x%llx offset=0x%llx length=0x%llx\n",
-				__func__, i, ioerr->oer_errno,
-				ioerr->oer_iswrite,
-				_DEVID_LO(&ioerr->oer_component.oid_device_id),
-				_DEVID_HI(&ioerr->oer_component.oid_device_id),
-				ioerr->oer_component.oid_partition_id,
-				ioerr->oer_component.oid_object_id,
-				ioerr->oer_comp_offset,
-				ioerr->oer_comp_length);
-
-			merge_ioerr(&accumulated_err, ioerr);
-		}
-		list_del(&oir->err_list);
-		objio_free_result(oir);
-	}
-
-	pnfs_osd_xdr_encode_ioerr(p, &accumulated_err);
-}
-
-void
-objlayout_encode_layoutreturn(struct xdr_stream *xdr,
-			      const struct nfs4_layoutreturn_args *args)
-{
-	struct pnfs_layout_hdr *pnfslay = args->layout;
-	struct objlayout *objlay = OBJLAYOUT(pnfslay);
-	struct objlayout_io_res *oir, *tmp;
-	__be32 *start;
-
-	dprintk("%s: Begin\n", __func__);
-	start = xdr_reserve_space(xdr, 4);
-	BUG_ON(!start);
-
-	spin_lock(&objlay->lock);
-
-	list_for_each_entry_safe(oir, tmp, &objlay->err_list, err_list) {
-		__be32 *last_xdr = NULL, *p;
-		unsigned i;
-		int res = 0;
-
-		for (i = 0; i < oir->num_comps; i++) {
-			struct pnfs_osd_ioerr *ioerr = &oir->ioerrs[i];
-
-			if (!ioerr->oer_errno)
-				continue;
-
-			dprintk("%s: err[%d]: errno=%d is_write=%d "
-				"dev(%llx:%llx) par=0x%llx obj=0x%llx "
-				"offset=0x%llx length=0x%llx\n",
-				__func__, i, ioerr->oer_errno,
-				ioerr->oer_iswrite,
-				_DEVID_LO(&ioerr->oer_component.oid_device_id),
-				_DEVID_HI(&ioerr->oer_component.oid_device_id),
-				ioerr->oer_component.oid_partition_id,
-				ioerr->oer_component.oid_object_id,
-				ioerr->oer_comp_offset,
-				ioerr->oer_comp_length);
-
-			p = pnfs_osd_xdr_ioerr_reserve_space(xdr);
-			if (unlikely(!p)) {
-				res = -E2BIG;
-				break; /* accumulated_error */
-			}
-
-			last_xdr = p;
-			pnfs_osd_xdr_encode_ioerr(p, &oir->ioerrs[i]);
-		}
-
-		/* TODO: use xdr_write_pages */
-		if (unlikely(res)) {
-			/* no space for even one error descriptor */
-			BUG_ON(!last_xdr);
-
-			/* we've encountered a situation with lots and lots of
-			 * errors and no space to encode them all. Use the last
-			 * available slot to report the union of all the
-			 * remaining errors.
-			 */
-			encode_accumulated_error(objlay, last_xdr);
-			goto loop_done;
-		}
-		list_del(&oir->err_list);
-		objio_free_result(oir);
-	}
-loop_done:
-	spin_unlock(&objlay->lock);
-
-	*start = cpu_to_be32((xdr->p - start - 1) * 4);
-	dprintk("%s: Return\n", __func__);
-}
-
-enum {
-	OBJLAYOUT_MAX_URI_LEN = 256, OBJLAYOUT_MAX_OSDNAME_LEN = 64,
-	OBJLAYOUT_MAX_SYSID_HEX_LEN = OSD_SYSTEMID_LEN * 2 + 1,
-	OSD_LOGIN_UPCALL_PATHLEN  = 256
-};
-
-static char osd_login_prog[OSD_LOGIN_UPCALL_PATHLEN] = "/sbin/osd_login";
-
-module_param_string(osd_login_prog, osd_login_prog, sizeof(osd_login_prog),
-		    0600);
-MODULE_PARM_DESC(osd_login_prog, "Path to the osd_login upcall program");
-
-struct __auto_login {
-	char uri[OBJLAYOUT_MAX_URI_LEN];
-	char osdname[OBJLAYOUT_MAX_OSDNAME_LEN];
-	char systemid_hex[OBJLAYOUT_MAX_SYSID_HEX_LEN];
-};
-
-static int __objlayout_upcall(struct __auto_login *login)
-{
-	static char *envp[] = { "HOME=/",
-		"TERM=linux",
-		"PATH=/sbin:/usr/sbin:/bin:/usr/bin",
-		NULL
-	};
-	char *argv[8];
-	int ret;
-
-	if (unlikely(!osd_login_prog[0])) {
-		dprintk("%s: osd_login_prog is disabled\n", __func__);
-		return -EACCES;
-	}
-
-	dprintk("%s uri: %s\n", __func__, login->uri);
-	dprintk("%s osdname %s\n", __func__, login->osdname);
-	dprintk("%s systemid_hex %s\n", __func__, login->systemid_hex);
-
-	argv[0] = (char *)osd_login_prog;
-	argv[1] = "-u";
-	argv[2] = login->uri;
-	argv[3] = "-o";
-	argv[4] = login->osdname;
-	argv[5] = "-s";
-	argv[6] = login->systemid_hex;
-	argv[7] = NULL;
-
-	ret = call_usermodehelper(argv[0], argv, envp, UMH_WAIT_PROC);
-	/*
-	 * Disable the upcall mechanism if we're getting an ENOENT or
-	 * EACCES error. The admin can re-enable it on the fly by using
-	 * sysfs to set the objlayoutdriver.osd_login_prog module parameter once
-	 * the problem has been fixed.
-	 */
-	if (ret == -ENOENT || ret == -EACCES) {
-		printk(KERN_ERR "PNFS-OBJ: %s was not found please set "
-			"objlayoutdriver.osd_login_prog kernel parameter!\n",
-			osd_login_prog);
-		osd_login_prog[0] = '\0';
-	}
-	dprintk("%s %s return value: %d\n", __func__, osd_login_prog, ret);
-
-	return ret;
-}
-
-/* Assume dest is all zeros */
-static void __copy_nfsS_and_zero_terminate(struct nfs4_string s,
-					   char *dest, int max_len,
-					   const char *var_name)
-{
-	if (!s.len)
-		return;
-
-	if (s.len >= max_len) {
-		pr_warn_ratelimited(
-			"objlayout_autologin: %s: s.len(%d) >= max_len(%d)",
-			var_name, s.len, max_len);
-		s.len = max_len - 1; /* space for null terminator */
-	}
-
-	memcpy(dest, s.data, s.len);
-}
-
-/* Assume sysid is all zeros */
-static void _sysid_2_hex(struct nfs4_string s,
-		  char sysid[OBJLAYOUT_MAX_SYSID_HEX_LEN])
-{
-	int i;
-	char *cur;
-
-	if (!s.len)
-		return;
-
-	if (s.len != OSD_SYSTEMID_LEN) {
-		pr_warn_ratelimited(
-		    "objlayout_autologin: systemid_len(%d) != OSD_SYSTEMID_LEN",
-		    s.len);
-		if (s.len > OSD_SYSTEMID_LEN)
-			s.len = OSD_SYSTEMID_LEN;
-	}
-
-	cur = sysid;
-	for (i = 0; i < s.len; i++)
-		cur = hex_byte_pack(cur, s.data[i]);
-}
-
-int objlayout_autologin(struct pnfs_osd_deviceaddr *deviceaddr)
-{
-	int rc;
-	struct __auto_login login;
-
-	if (!deviceaddr->oda_targetaddr.ota_netaddr.r_addr.len)
-		return -ENODEV;
-
-	memset(&login, 0, sizeof(login));
-	__copy_nfsS_and_zero_terminate(
-		deviceaddr->oda_targetaddr.ota_netaddr.r_addr,
-		login.uri, sizeof(login.uri), "URI");
-
-	__copy_nfsS_and_zero_terminate(
-		deviceaddr->oda_osdname,
-		login.osdname, sizeof(login.osdname), "OSDNAME");
-
-	_sysid_2_hex(deviceaddr->oda_systemid, login.systemid_hex);
-
-	rc = __objlayout_upcall(&login);
-	if (rc > 0) /* script returns positive values */
-		rc = -ENODEV;
-
-	return rc;
-}

fs/nfs/objlayout/objlayout.h

@@ -1,183 +0,0 @@
-/*
- *  Data types and function declerations for interfacing with the
- *  pNFS standard object layout driver.
- *
- *  Copyright (C) 2007 Panasas Inc. [year of first publication]
- *  All rights reserved.
- *
- *  Benny Halevy <bhalevy@panasas.com>
- *  Boaz Harrosh <ooo@electrozaur.com>
- *
- *  This program is free software; you can redistribute it and/or modify
- *  it under the terms of the GNU General Public License version 2
- *  See the file COPYING included with this distribution for more details.
- *
- *  Redistribution and use in source and binary forms, with or without
- *  modification, are permitted provided that the following conditions
- *  are met:
- *
- *  1. Redistributions of source code must retain the above copyright
- *     notice, this list of conditions and the following disclaimer.
- *  2. Redistributions in binary form must reproduce the above copyright
- *     notice, this list of conditions and the following disclaimer in the
- *     documentation and/or other materials provided with the distribution.
- *  3. Neither the name of the Panasas company nor the names of its
- *     contributors may be used to endorse or promote products derived
- *     from this software without specific prior written permission.
- *
- *  THIS SOFTWARE IS PROVIDED ``AS IS'' AND ANY EXPRESS OR IMPLIED
- *  WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
- *  MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
- *  DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
- *  FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
- *  CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
- *  SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR
- *  BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
- *  LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
- *  NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
- *  SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
- */
-
-#ifndef _OBJLAYOUT_H
-#define _OBJLAYOUT_H
-
-#include <linux/nfs_fs.h>
-#include <linux/pnfs_osd_xdr.h>
-#include "../pnfs.h"
-
-/*
- * per-inode layout
- */
-struct objlayout {
-	struct pnfs_layout_hdr pnfs_layout;
-
-	 /* for layout_commit */
-	enum osd_delta_space_valid_enum {
-		OBJ_DSU_INIT = 0,
-		OBJ_DSU_VALID,
-		OBJ_DSU_INVALID,
-	} delta_space_valid;
-	s64 delta_space_used;  /* consumed by write ops */
-
-	 /* for layout_return */
-	spinlock_t lock;
-	struct list_head err_list;
-};
-
-static inline struct objlayout *
-OBJLAYOUT(struct pnfs_layout_hdr *lo)
-{
-	return container_of(lo, struct objlayout, pnfs_layout);
-}
-
-/*
- * per-I/O operation state
- * embedded in objects provider io_state data structure
- */
-struct objlayout_io_res {
-	struct objlayout *objlay;
-
-	void *rpcdata;
-	int status;             /* res */
-	int committed;          /* res */
-
-	/* Error reporting (layout_return) */
-	struct list_head err_list;
-	unsigned num_comps;
-	/* Pointer to array of error descriptors of size num_comps.
-	 * It should contain as many entries as devices in the osd_layout
-	 * that participate in the I/O. It is up to the io_engine to allocate
-	 * needed space and set num_comps.
-	 */
-	struct pnfs_osd_ioerr *ioerrs;
-};
-
-static inline
-void objlayout_init_ioerrs(struct objlayout_io_res *oir, unsigned num_comps,
-			struct pnfs_osd_ioerr *ioerrs, void *rpcdata,
-			struct pnfs_layout_hdr *pnfs_layout_type)
-{
-	oir->objlay = OBJLAYOUT(pnfs_layout_type);
-	oir->rpcdata = rpcdata;
-	INIT_LIST_HEAD(&oir->err_list);
-	oir->num_comps = num_comps;
-	oir->ioerrs = ioerrs;
-}
-
-/*
- * Raid engine I/O API
- */
-extern int objio_alloc_lseg(struct pnfs_layout_segment **outp,
-	struct pnfs_layout_hdr *pnfslay,
-	struct pnfs_layout_range *range,
-	struct xdr_stream *xdr,
-	gfp_t gfp_flags);
-extern void objio_free_lseg(struct pnfs_layout_segment *lseg);
-
-/* objio_free_result will free these @oir structs received from
- * objlayout_{read,write}_done
- */
-extern void objio_free_result(struct objlayout_io_res *oir);
-
-extern int objio_read_pagelist(struct nfs_pgio_header *rdata);
-extern int objio_write_pagelist(struct nfs_pgio_header *wdata, int how);
-
-/*
- * callback API
- */
-extern void objlayout_io_set_result(struct objlayout_io_res *oir,
-			unsigned index, struct pnfs_osd_objid *pooid,
-			int osd_error, u64 offset, u64 length, bool is_write);
-
-static inline void
-objlayout_add_delta_space_used(struct objlayout *objlay, s64 space_used)
-{
-	/* If one of the I/Os errored out and the delta_space_used was
-	 * invalid we render the complete report as invalid. Protocol mandate
-	 * the DSU be accurate or not reported.
-	 */
-	spin_lock(&objlay->lock);
-	if (objlay->delta_space_valid != OBJ_DSU_INVALID) {
-		objlay->delta_space_valid = OBJ_DSU_VALID;
-		objlay->delta_space_used += space_used;
-	}
-	spin_unlock(&objlay->lock);
-}
-
-extern void objlayout_read_done(struct objlayout_io_res *oir,
-				ssize_t status, bool sync);
-extern void objlayout_write_done(struct objlayout_io_res *oir,
-				 ssize_t status, bool sync);
-
-/*
- * exported generic objects function vectors
- */
-
-extern struct pnfs_layout_hdr *objlayout_alloc_layout_hdr(struct inode *, gfp_t gfp_flags);
-extern void objlayout_free_layout_hdr(struct pnfs_layout_hdr *);
-
-extern struct pnfs_layout_segment *objlayout_alloc_lseg(
-	struct pnfs_layout_hdr *,
-	struct nfs4_layoutget_res *,
-	gfp_t gfp_flags);
-extern void objlayout_free_lseg(struct pnfs_layout_segment *);
-
-extern enum pnfs_try_status objlayout_read_pagelist(
-	struct nfs_pgio_header *);
-
-extern enum pnfs_try_status objlayout_write_pagelist(
-	struct nfs_pgio_header *,
-	int how);
-
-extern void objlayout_encode_layoutcommit(
-	struct pnfs_layout_hdr *,
-	struct xdr_stream *,
-	const struct nfs4_layoutcommit_args *);
-
-extern void objlayout_encode_layoutreturn(
-	struct xdr_stream *,
-	const struct nfs4_layoutreturn_args *);
-
-extern int objlayout_autologin(struct pnfs_osd_deviceaddr *deviceaddr);
-
-#endif /* _OBJLAYOUT_H */

fs/nfs/objlayout/pnfs_osd_xdr_cli.c

@@ -1,415 +0,0 @@
-/*
- *  Object-Based pNFS Layout XDR layer
- *
- *  Copyright (C) 2007 Panasas Inc. [year of first publication]
- *  All rights reserved.
- *
- *  Benny Halevy <bhalevy@panasas.com>
- *  Boaz Harrosh <ooo@electrozaur.com>
- *
- *  This program is free software; you can redistribute it and/or modify
- *  it under the terms of the GNU General Public License version 2
- *  See the file COPYING included with this distribution for more details.
- *
- *  Redistribution and use in source and binary forms, with or without
- *  modification, are permitted provided that the following conditions
- *  are met:
- *
- *  1. Redistributions of source code must retain the above copyright
- *     notice, this list of conditions and the following disclaimer.
- *  2. Redistributions in binary form must reproduce the above copyright
- *     notice, this list of conditions and the following disclaimer in the
- *     documentation and/or other materials provided with the distribution.
- *  3. Neither the name of the Panasas company nor the names of its
- *     contributors may be used to endorse or promote products derived
- *     from this software without specific prior written permission.
- *
- *  THIS SOFTWARE IS PROVIDED ``AS IS'' AND ANY EXPRESS OR IMPLIED
- *  WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
- *  MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
- *  DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
- *  FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
- *  CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
- *  SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR
- *  BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
- *  LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
- *  NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
- *  SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
- */
-
-#include <linux/pnfs_osd_xdr.h>
-
-#define NFSDBG_FACILITY         NFSDBG_PNFS_LD
-
-/*
- * The following implementation is based on RFC5664
- */
-
-/*
- * struct pnfs_osd_objid {
- *	struct nfs4_deviceid	oid_device_id;
- *	u64			oid_partition_id;
- *	u64			oid_object_id;
- * }; // xdr size 32 bytes
- */
-static __be32 *
-_osd_xdr_decode_objid(__be32 *p, struct pnfs_osd_objid *objid)
-{
-	p = xdr_decode_opaque_fixed(p, objid->oid_device_id.data,
-				    sizeof(objid->oid_device_id.data));
-
-	p = xdr_decode_hyper(p, &objid->oid_partition_id);
-	p = xdr_decode_hyper(p, &objid->oid_object_id);
-	return p;
-}
-/*
- * struct pnfs_osd_opaque_cred {
- *	u32 cred_len;
- *	void *cred;
- * }; // xdr size [variable]
- * The return pointers are from the xdr buffer
- */
-static int
-_osd_xdr_decode_opaque_cred(struct pnfs_osd_opaque_cred *opaque_cred,
-			    struct xdr_stream *xdr)
-{
-	__be32 *p = xdr_inline_decode(xdr, 1);
-
-	if (!p)
-		return -EINVAL;
-
-	opaque_cred->cred_len = be32_to_cpu(*p++);
-
-	p = xdr_inline_decode(xdr, opaque_cred->cred_len);
-	if (!p)
-		return -EINVAL;
-
-	opaque_cred->cred = p;
-	return 0;
-}
-
-/*
- * struct pnfs_osd_object_cred {
- *	struct pnfs_osd_objid		oc_object_id;
- *	u32				oc_osd_version;
- *	u32				oc_cap_key_sec;
- *	struct pnfs_osd_opaque_cred	oc_cap_key
- *	struct pnfs_osd_opaque_cred	oc_cap;
- * }; // xdr size 32 + 4 + 4 + [variable] + [variable]
- */
-static int
-_osd_xdr_decode_object_cred(struct pnfs_osd_object_cred *comp,
-			    struct xdr_stream *xdr)
-{
-	__be32 *p = xdr_inline_decode(xdr, 32 + 4 + 4);
-	int ret;
-
-	if (!p)
-		return -EIO;
-
-	p = _osd_xdr_decode_objid(p, &comp->oc_object_id);
-	comp->oc_osd_version = be32_to_cpup(p++);
-	comp->oc_cap_key_sec = be32_to_cpup(p);
-
-	ret = _osd_xdr_decode_opaque_cred(&comp->oc_cap_key, xdr);
-	if (unlikely(ret))
-		return ret;
-
-	ret = _osd_xdr_decode_opaque_cred(&comp->oc_cap, xdr);
-	return ret;
-}
-
-/*
- * struct pnfs_osd_data_map {
- *	u32	odm_num_comps;
- *	u64	odm_stripe_unit;
- *	u32	odm_group_width;
- *	u32	odm_group_depth;
- *	u32	odm_mirror_cnt;
- *	u32	odm_raid_algorithm;
- * }; // xdr size 4 + 8 + 4 + 4 + 4 + 4
- */
-static inline int
-_osd_data_map_xdr_sz(void)
-{
-	return 4 + 8 + 4 + 4 + 4 + 4;
-}
-
-static __be32 *
-_osd_xdr_decode_data_map(__be32 *p, struct pnfs_osd_data_map *data_map)
-{
-	data_map->odm_num_comps = be32_to_cpup(p++);
-	p = xdr_decode_hyper(p, &data_map->odm_stripe_unit);
-	data_map->odm_group_width = be32_to_cpup(p++);
-	data_map->odm_group_depth = be32_to_cpup(p++);
-	data_map->odm_mirror_cnt = be32_to_cpup(p++);
-	data_map->odm_raid_algorithm = be32_to_cpup(p++);
-	dprintk("%s: odm_num_comps=%u odm_stripe_unit=%llu odm_group_width=%u "
-		"odm_group_depth=%u odm_mirror_cnt=%u odm_raid_algorithm=%u\n",
-		__func__,
-		data_map->odm_num_comps,
-		(unsigned long long)data_map->odm_stripe_unit,
-		data_map->odm_group_width,
-		data_map->odm_group_depth,
-		data_map->odm_mirror_cnt,
-		data_map->odm_raid_algorithm);
-	return p;
-}
-
-int pnfs_osd_xdr_decode_layout_map(struct pnfs_osd_layout *layout,
-	struct pnfs_osd_xdr_decode_layout_iter *iter, struct xdr_stream *xdr)
-{
-	__be32 *p;
-
-	memset(iter, 0, sizeof(*iter));
-
-	p = xdr_inline_decode(xdr, _osd_data_map_xdr_sz() + 4 + 4);
-	if (unlikely(!p))
-		return -EINVAL;
-
-	p = _osd_xdr_decode_data_map(p, &layout->olo_map);
-	layout->olo_comps_index = be32_to_cpup(p++);
-	layout->olo_num_comps = be32_to_cpup(p++);
-	dprintk("%s: olo_comps_index=%d olo_num_comps=%d\n", __func__,
-		layout->olo_comps_index, layout->olo_num_comps);
-
-	iter->total_comps = layout->olo_num_comps;
-	return 0;
-}
-
-bool pnfs_osd_xdr_decode_layout_comp(struct pnfs_osd_object_cred *comp,
-	struct pnfs_osd_xdr_decode_layout_iter *iter, struct xdr_stream *xdr,
-	int *err)
-{
-	BUG_ON(iter->decoded_comps > iter->total_comps);
-	if (iter->decoded_comps == iter->total_comps)
-		return false;
-
-	*err = _osd_xdr_decode_object_cred(comp, xdr);
-	if (unlikely(*err)) {
-		dprintk("%s: _osd_xdr_decode_object_cred=>%d decoded_comps=%d "
-			"total_comps=%d\n", __func__, *err,
-			iter->decoded_comps, iter->total_comps);
-		return false; /* stop the loop */
-	}
-	dprintk("%s: dev(%llx:%llx) par=0x%llx obj=0x%llx "
-		"key_len=%u cap_len=%u\n",
-		__func__,
-		_DEVID_LO(&comp->oc_object_id.oid_device_id),
-		_DEVID_HI(&comp->oc_object_id.oid_device_id),
-		comp->oc_object_id.oid_partition_id,
-		comp->oc_object_id.oid_object_id,
-		comp->oc_cap_key.cred_len, comp->oc_cap.cred_len);
-
-	iter->decoded_comps++;
-	return true;
-}
-
-/*
- * Get Device Information Decoding
- *
- * Note: since Device Information is currently done synchronously, all
- *       variable strings fields are left inside the rpc buffer and are only
- *       pointed to by the pnfs_osd_deviceaddr members. So the read buffer
- *       should not be freed while the returned information is in use.
- */
-/*
- *struct nfs4_string {
- *	unsigned int len;
- *	char *data;
- *}; // size [variable]
- * NOTE: Returned string points to inside the XDR buffer
- */
-static __be32 *
-__read_u8_opaque(__be32 *p, struct nfs4_string *str)
-{
-	str->len = be32_to_cpup(p++);
-	str->data = (char *)p;
-
-	p += XDR_QUADLEN(str->len);
-	return p;
-}
-
-/*
- * struct pnfs_osd_targetid {
- *	u32			oti_type;
- *	struct nfs4_string	oti_scsi_device_id;
- * };// size 4 + [variable]
- */
-static __be32 *
-__read_targetid(__be32 *p, struct pnfs_osd_targetid* targetid)
-{
-	u32 oti_type;
-
-	oti_type = be32_to_cpup(p++);
-	targetid->oti_type = oti_type;
-
-	switch (oti_type) {
-	case OBJ_TARGET_SCSI_NAME:
-	case OBJ_TARGET_SCSI_DEVICE_ID:
-		p = __read_u8_opaque(p, &targetid->oti_scsi_device_id);
-	}
-
-	return p;
-}
-
-/*
- * struct pnfs_osd_net_addr {
- *	struct nfs4_string	r_netid;
- *	struct nfs4_string	r_addr;
- * };
- */
-static __be32 *
-__read_net_addr(__be32 *p, struct pnfs_osd_net_addr* netaddr)
-{
-	p = __read_u8_opaque(p, &netaddr->r_netid);
-	p = __read_u8_opaque(p, &netaddr->r_addr);
-
-	return p;
-}
-
-/*
- * struct pnfs_osd_targetaddr {
- *	u32				ota_available;
- *	struct pnfs_osd_net_addr	ota_netaddr;
- * };
- */
-static __be32 *
-__read_targetaddr(__be32 *p, struct pnfs_osd_targetaddr *targetaddr)
-{
-	u32 ota_available;
-
-	ota_available = be32_to_cpup(p++);
-	targetaddr->ota_available = ota_available;
-
-	if (ota_available)
-		p = __read_net_addr(p, &targetaddr->ota_netaddr);
-
-
-	return p;
-}
-
-/*
- * struct pnfs_osd_deviceaddr {
- *	struct pnfs_osd_targetid	oda_targetid;
- *	struct pnfs_osd_targetaddr	oda_targetaddr;
- *	u8				oda_lun[8];
- *	struct nfs4_string		oda_systemid;
- *	struct pnfs_osd_object_cred	oda_root_obj_cred;
- *	struct nfs4_string		oda_osdname;
- * };
- */
-
-/* We need this version for the pnfs_osd_xdr_decode_deviceaddr which does
- * not have an xdr_stream
- */
-static __be32 *
-__read_opaque_cred(__be32 *p,
-			      struct pnfs_osd_opaque_cred *opaque_cred)
-{
-	opaque_cred->cred_len = be32_to_cpu(*p++);
-	opaque_cred->cred = p;
-	return p + XDR_QUADLEN(opaque_cred->cred_len);
-}
-
-static __be32 *
-__read_object_cred(__be32 *p, struct pnfs_osd_object_cred *comp)
-{
-	p = _osd_xdr_decode_objid(p, &comp->oc_object_id);
-	comp->oc_osd_version = be32_to_cpup(p++);
-	comp->oc_cap_key_sec = be32_to_cpup(p++);
-
-	p = __read_opaque_cred(p, &comp->oc_cap_key);
-	p = __read_opaque_cred(p, &comp->oc_cap);
-	return p;
-}
-
-void pnfs_osd_xdr_decode_deviceaddr(
-	struct pnfs_osd_deviceaddr *deviceaddr, __be32 *p)
-{
-	p = __read_targetid(p, &deviceaddr->oda_targetid);
-
-	p = __read_targetaddr(p, &deviceaddr->oda_targetaddr);
-
-	p = xdr_decode_opaque_fixed(p, deviceaddr->oda_lun,
-				    sizeof(deviceaddr->oda_lun));
-
-	p = __read_u8_opaque(p, &deviceaddr->oda_systemid);
-
-	p = __read_object_cred(p, &deviceaddr->oda_root_obj_cred);
-
-	p = __read_u8_opaque(p, &deviceaddr->oda_osdname);
-
-	/* libosd likes this terminated in dbg. It's last, so no problems */
-	deviceaddr->oda_osdname.data[deviceaddr->oda_osdname.len] = 0;
-}
-
-/*
- * struct pnfs_osd_layoutupdate {
- *	u32	dsu_valid;
- *	s64	dsu_delta;
- *	u32	olu_ioerr_flag;
- * }; xdr size 4 + 8 + 4
- */
-int
-pnfs_osd_xdr_encode_layoutupdate(struct xdr_stream *xdr,
-				 struct pnfs_osd_layoutupdate *lou)
-{
-	__be32 *p = xdr_reserve_space(xdr,  4 + 8 + 4);
-
-	if (!p)
-		return -E2BIG;
-
-	*p++ = cpu_to_be32(lou->dsu_valid);
-	if (lou->dsu_valid)
-		p = xdr_encode_hyper(p, lou->dsu_delta);
-	*p++ = cpu_to_be32(lou->olu_ioerr_flag);
-	return 0;
-}
-
-/*
- * struct pnfs_osd_objid {
- *	struct nfs4_deviceid	oid_device_id;
- *	u64			oid_partition_id;
- *	u64			oid_object_id;
- * }; // xdr size 32 bytes
- */
-static inline __be32 *
-pnfs_osd_xdr_encode_objid(__be32 *p, struct pnfs_osd_objid *object_id)
-{
-	p = xdr_encode_opaque_fixed(p, &object_id->oid_device_id.data,
-				    sizeof(object_id->oid_device_id.data));
-	p = xdr_encode_hyper(p, object_id->oid_partition_id);
-	p = xdr_encode_hyper(p, object_id->oid_object_id);
-
-	return p;
-}
-
-/*
- * struct pnfs_osd_ioerr {
- *	struct pnfs_osd_objid	oer_component;
- *	u64			oer_comp_offset;
- *	u64			oer_comp_length;
- *	u32			oer_iswrite;
- *	u32			oer_errno;
- * }; // xdr size 32 + 24 bytes
- */
-void pnfs_osd_xdr_encode_ioerr(__be32 *p, struct pnfs_osd_ioerr *ioerr)
-{
-	p = pnfs_osd_xdr_encode_objid(p, &ioerr->oer_component);
-	p = xdr_encode_hyper(p, ioerr->oer_comp_offset);
-	p = xdr_encode_hyper(p, ioerr->oer_comp_length);
-	*p++ = cpu_to_be32(ioerr->oer_iswrite);
-	*p   = cpu_to_be32(ioerr->oer_errno);
-}
-
-__be32 *pnfs_osd_xdr_ioerr_reserve_space(struct xdr_stream *xdr)
-{
-	__be32 *p;
-
-	p = xdr_reserve_space(xdr, 32 + 24);
-	if (unlikely(!p))
-		dprintk("%s: out of xdr space\n", __func__);
-
-	return p;
-}

fs/nfs/pagelist.c

@@ -29,19 +29,6 @@
 static struct kmem_cache *nfs_page_cachep;
 static const struct rpc_call_ops nfs_pgio_common_ops;
 
-static bool nfs_pgarray_set(struct nfs_page_array *p, unsigned int pagecount)
-{
-	p->npages = pagecount;
-	if (pagecount <= ARRAY_SIZE(p->page_array))
-		p->pagevec = p->page_array;
-	else {
-		p->pagevec = kcalloc(pagecount, sizeof(struct page *), GFP_KERNEL);
-		if (!p->pagevec)
-			p->npages = 0;
-	}
-	return p->pagevec != NULL;
-}
-
 struct nfs_pgio_mirror *
 nfs_pgio_current_mirror(struct nfs_pageio_descriptor *desc)
 {
@@ -115,6 +102,35 @@ nfs_iocounter_wait(struct nfs_lock_context *l_ctx)
 			TASK_KILLABLE);
 }
 
+/**
+ * nfs_async_iocounter_wait - wait on a rpc_waitqueue for I/O
+ * to complete
+ * @task: the rpc_task that should wait
+ * @l_ctx: nfs_lock_context with io_counter to check
+ *
+ * Returns true if there is outstanding I/O to wait on and the
+ * task has been put to sleep.
+ */
+bool
+nfs_async_iocounter_wait(struct rpc_task *task, struct nfs_lock_context *l_ctx)
+{
+	struct inode *inode = d_inode(l_ctx->open_context->dentry);
+	bool ret = false;
+
+	if (atomic_read(&l_ctx->io_count) > 0) {
+		rpc_sleep_on(&NFS_SERVER(inode)->uoc_rpcwaitq, task, NULL);
+		ret = true;
+	}
+
+	if (atomic_read(&l_ctx->io_count) == 0) {
+		rpc_wake_up_queued_task(&NFS_SERVER(inode)->uoc_rpcwaitq, task);
+		ret = false;
+	}
+
+	return ret;
+}
+EXPORT_SYMBOL_GPL(nfs_async_iocounter_wait);
+
 /*
  * nfs_page_group_lock - lock the head of the page group
  * @req - request in group that is to be locked
@@ -398,8 +414,11 @@ static void nfs_clear_request(struct nfs_page *req)
 		req->wb_page = NULL;
 	}
 	if (l_ctx != NULL) {
-		if (atomic_dec_and_test(&l_ctx->io_count))
+		if (atomic_dec_and_test(&l_ctx->io_count)) {
 			wake_up_atomic_t(&l_ctx->io_count);
+			if (test_bit(NFS_CONTEXT_UNLOCK, &ctx->flags))
+				rpc_wake_up(&NFS_SERVER(d_inode(ctx->dentry))->uoc_rpcwaitq);
+		}
 		nfs_put_lock_context(l_ctx);
 		req->wb_lock_context = NULL;
 	}
@@ -677,7 +696,8 @@ void nfs_pageio_init(struct nfs_pageio_descriptor *desc,
 		     const struct nfs_pgio_completion_ops *compl_ops,
 		     const struct nfs_rw_ops *rw_ops,
 		     size_t bsize,
-		     int io_flags)
+		     int io_flags,
+		     gfp_t gfp_flags)
 {
 	struct nfs_pgio_mirror *new;
 	int i;
@@ -701,7 +721,7 @@ void nfs_pageio_init(struct nfs_pageio_descriptor *desc,
 		/* until we have a request, we don't have an lseg and no
 		 * idea how many mirrors there will be */
 		new = kcalloc(NFS_PAGEIO_DESCRIPTOR_MIRROR_MAX,
-			      sizeof(struct nfs_pgio_mirror), GFP_KERNEL);
+			      sizeof(struct nfs_pgio_mirror), gfp_flags);
 		desc->pg_mirrors_dynamic = new;
 		desc->pg_mirrors = new;
 
@@ -754,13 +774,24 @@ int nfs_generic_pgio(struct nfs_pageio_descriptor *desc,
 				*last_page;
 	struct list_head *head = &mirror->pg_list;
 	struct nfs_commit_info cinfo;
+	struct nfs_page_array *pg_array = &hdr->page_array;
 	unsigned int pagecount, pageused;
+	gfp_t gfp_flags = GFP_KERNEL;
 
 	pagecount = nfs_page_array_len(mirror->pg_base, mirror->pg_count);
-	if (!nfs_pgarray_set(&hdr->page_array, pagecount)) {
+
-		nfs_pgio_error(hdr);
+	if (pagecount <= ARRAY_SIZE(pg_array->page_array))
-		desc->pg_error = -ENOMEM;
+		pg_array->pagevec = pg_array->page_array;
-		return desc->pg_error;
+	else {
+		if (hdr->rw_mode == FMODE_WRITE)
+			gfp_flags = GFP_NOIO;
+		pg_array->pagevec = kcalloc(pagecount, sizeof(struct page *), gfp_flags);
+		if (!pg_array->pagevec) {
+			pg_array->npages = 0;
+			nfs_pgio_error(hdr);
+			desc->pg_error = -ENOMEM;
+			return desc->pg_error;
+		}
 	}
 
 	nfs_init_cinfo(&cinfo, desc->pg_inode, desc->pg_dreq);
@@ -1256,8 +1287,10 @@ void nfs_pageio_cond_complete(struct nfs_pageio_descriptor *desc, pgoff_t index)
 		mirror = &desc->pg_mirrors[midx];
 		if (!list_empty(&mirror->pg_list)) {
 			prev = nfs_list_entry(mirror->pg_list.prev);
-			if (index != prev->wb_index + 1)
+			if (index != prev->wb_index + 1) {
-				nfs_pageio_complete_mirror(desc, midx);
+				nfs_pageio_complete(desc);
+				break;
+			}
 		}
 	}
 }

fs/nfs/pnfs.c

@@ -322,9 +322,15 @@ pnfs_set_plh_return_info(struct pnfs_layout_hdr *lo, enum pnfs_iomode iomode,
 static void
 pnfs_clear_layoutreturn_info(struct pnfs_layout_hdr *lo)
 {
+	struct pnfs_layout_segment *lseg;
 	lo->plh_return_iomode = 0;
 	lo->plh_return_seq = 0;
 	clear_bit(NFS_LAYOUT_RETURN_REQUESTED, &lo->plh_flags);
+	list_for_each_entry(lseg, &lo->plh_segs, pls_list) {
+		if (!test_bit(NFS_LSEG_LAYOUTRETURN, &lseg->pls_flags))
+			continue;
+		pnfs_set_plh_return_info(lo, lseg->pls_range.iomode, 0);
+	}
 }
 
 static void pnfs_clear_layoutreturn_waitbit(struct pnfs_layout_hdr *lo)
@@ -367,9 +373,9 @@ pnfs_mark_layout_stateid_invalid(struct pnfs_layout_hdr *lo,
 	struct pnfs_layout_segment *lseg, *next;
 
 	set_bit(NFS_LAYOUT_INVALID_STID, &lo->plh_flags);
-	pnfs_clear_layoutreturn_info(lo);
 	list_for_each_entry_safe(lseg, next, &lo->plh_segs, pls_list)
 		pnfs_clear_lseg_state(lseg, lseg_list);
+	pnfs_clear_layoutreturn_info(lo);
 	pnfs_free_returned_lsegs(lo, lseg_list, &range, 0);
 	if (test_bit(NFS_LAYOUT_RETURN, &lo->plh_flags) &&
 	    !test_and_set_bit(NFS_LAYOUT_RETURN_LOCK, &lo->plh_flags))
@@ -563,7 +569,6 @@ pnfs_put_lseg_locked(struct pnfs_layout_segment *lseg)
 		}
 	}
 }
-EXPORT_SYMBOL_GPL(pnfs_put_lseg_locked);
 
 /*
  * is l2 fully contained in l1?
@@ -728,6 +733,7 @@ pnfs_destroy_layout(struct nfs_inode *nfsi)
 		pnfs_layout_clear_fail_bit(lo, NFS_LAYOUT_RW_FAILED);
 		spin_unlock(&nfsi->vfs_inode.i_lock);
 		pnfs_free_lseg_list(&tmp_list);
+		nfs_commit_inode(&nfsi->vfs_inode, 0);
 		pnfs_put_layout_hdr(lo);
 	} else
 		spin_unlock(&nfsi->vfs_inode.i_lock);
@@ -1209,7 +1215,6 @@ out:
 	dprintk("<-- %s status: %d\n", __func__, status);
 	return status;
 }
-EXPORT_SYMBOL_GPL(_pnfs_return_layout);
 
 int
 pnfs_commit_and_return_layout(struct inode *inode)
@@ -1991,6 +1996,8 @@ out_forget:
 	spin_unlock(&ino->i_lock);
 	lseg->pls_layout = lo;
 	NFS_SERVER(ino)->pnfs_curr_ld->free_lseg(lseg);
+	if (!pnfs_layout_is_valid(lo))
+		nfs_commit_inode(ino, 0);
 	return ERR_PTR(-EAGAIN);
 }
 
@@ -2051,9 +2058,11 @@ void pnfs_error_mark_layout_for_return(struct inode *inode,
 	bool return_now = false;
 
 	spin_lock(&inode->i_lock);
+	if (!pnfs_layout_is_valid(lo)) {
+		spin_unlock(&inode->i_lock);
+		return;
+	}
 	pnfs_set_plh_return_info(lo, range.iomode, 0);
-	/* Block LAYOUTGET */
-	set_bit(NFS_LAYOUT_RETURN, &lo->plh_flags);
 	/*
 	 * mark all matching lsegs so that we are sure to have no live
 	 * segments at hand when sending layoutreturn. See pnfs_put_lseg()
@@ -2074,11 +2083,23 @@ void pnfs_error_mark_layout_for_return(struct inode *inode,
 }
 EXPORT_SYMBOL_GPL(pnfs_error_mark_layout_for_return);
 
+void
+pnfs_generic_pg_check_layout(struct nfs_pageio_descriptor *pgio)
+{
+	if (pgio->pg_lseg == NULL ||
+	    test_bit(NFS_LSEG_VALID, &pgio->pg_lseg->pls_flags))
+		return;
+	pnfs_put_lseg(pgio->pg_lseg);
+	pgio->pg_lseg = NULL;
+}
+EXPORT_SYMBOL_GPL(pnfs_generic_pg_check_layout);
+
 void
 pnfs_generic_pg_init_read(struct nfs_pageio_descriptor *pgio, struct nfs_page *req)
 {
 	u64 rd_size = req->wb_bytes;
 
+	pnfs_generic_pg_check_layout(pgio);
 	if (pgio->pg_lseg == NULL) {
 		if (pgio->pg_dreq == NULL)
 			rd_size = i_size_read(pgio->pg_inode) - req_offset(req);
@@ -2109,6 +2130,7 @@ void
 pnfs_generic_pg_init_write(struct nfs_pageio_descriptor *pgio,
 			   struct nfs_page *req, u64 wb_size)
 {
+	pnfs_generic_pg_check_layout(pgio);
 	if (pgio->pg_lseg == NULL) {
 		pgio->pg_lseg = pnfs_update_layout(pgio->pg_inode,
 						   req->wb_context,
@@ -2277,8 +2299,20 @@ pnfs_do_write(struct nfs_pageio_descriptor *desc,
 	enum pnfs_try_status trypnfs;
 
 	trypnfs = pnfs_try_to_write_data(hdr, call_ops, lseg, how);
-	if (trypnfs == PNFS_NOT_ATTEMPTED)
+	switch (trypnfs) {
+	case PNFS_NOT_ATTEMPTED:
 		pnfs_write_through_mds(desc, hdr);
+	case PNFS_ATTEMPTED:
+		break;
+	case PNFS_TRY_AGAIN:
+		/* cleanup hdr and prepare to redo pnfs */
+		if (!test_and_set_bit(NFS_IOHDR_REDO, &hdr->flags)) {
+			struct nfs_pgio_mirror *mirror = nfs_pgio_current_mirror(desc);
+			list_splice_init(&hdr->pages, &mirror->pg_list);
+			mirror->pg_recoalesce = 1;
+		}
+		hdr->mds_ops->rpc_release(hdr);
+	}
 }
 
 static void pnfs_writehdr_free(struct nfs_pgio_header *hdr)
@@ -2408,10 +2442,20 @@ pnfs_do_read(struct nfs_pageio_descriptor *desc, struct nfs_pgio_header *hdr)
 	enum pnfs_try_status trypnfs;
 
 	trypnfs = pnfs_try_to_read_data(hdr, call_ops, lseg);
-	if (trypnfs == PNFS_TRY_AGAIN)
+	switch (trypnfs) {
-		pnfs_read_resend_pnfs(hdr);
+	case PNFS_NOT_ATTEMPTED:
-	if (trypnfs == PNFS_NOT_ATTEMPTED || hdr->task.tk_status)
 		pnfs_read_through_mds(desc, hdr);
+	case PNFS_ATTEMPTED:
+		break;
+	case PNFS_TRY_AGAIN:
+		/* cleanup hdr and prepare to redo pnfs */
+		if (!test_and_set_bit(NFS_IOHDR_REDO, &hdr->flags)) {
+			struct nfs_pgio_mirror *mirror = nfs_pgio_current_mirror(desc);
+			list_splice_init(&hdr->pages, &mirror->pg_list);
+			mirror->pg_recoalesce = 1;
+		}
+		hdr->mds_ops->rpc_release(hdr);
+	}
 }
 
 static void pnfs_readhdr_free(struct nfs_pgio_header *hdr)

fs/nfs/pnfs.h

@@ -173,14 +173,9 @@ struct pnfs_layoutdriver_type {
 			gfp_t gfp_flags);
 
 	int (*prepare_layoutreturn) (struct nfs4_layoutreturn_args *);
-	void (*encode_layoutreturn) (struct xdr_stream *xdr,
-				     const struct nfs4_layoutreturn_args *args);
 
 	void (*cleanup_layoutcommit) (struct nfs4_layoutcommit_data *data);
 	int (*prepare_layoutcommit) (struct nfs4_layoutcommit_args *args);
-	void (*encode_layoutcommit) (struct pnfs_layout_hdr *lo,
-				     struct xdr_stream *xdr,
-				     const struct nfs4_layoutcommit_args *args);
 	int (*prepare_layoutstats) (struct nfs42_layoutstat_args *args);
 };
 
@@ -239,6 +234,7 @@ void pnfs_put_lseg_locked(struct pnfs_layout_segment *lseg);
 
 void set_pnfs_layoutdriver(struct nfs_server *, const struct nfs_fh *, struct nfs_fsinfo *);
 void unset_pnfs_layoutdriver(struct nfs_server *);
+void pnfs_generic_pg_check_layout(struct nfs_pageio_descriptor *pgio);
 void pnfs_generic_pg_init_read(struct nfs_pageio_descriptor *, struct nfs_page *);
 int pnfs_generic_pg_readpages(struct nfs_pageio_descriptor *desc);
 void pnfs_generic_pg_init_write(struct nfs_pageio_descriptor *pgio,

fs/nfs/pnfs_nfs.c

@@ -217,7 +217,14 @@ pnfs_generic_alloc_ds_commits(struct nfs_commit_info *cinfo,
 	for (i = 0; i < fl_cinfo->nbuckets; i++, bucket++) {
 		if (list_empty(&bucket->committing))
 			continue;
-		data = nfs_commitdata_alloc();
+		/*
+		 * If the layout segment is invalid, then let
+		 * pnfs_generic_retry_commit() clean up the bucket.
+		 */
+		if (bucket->clseg && !pnfs_is_valid_lseg(bucket->clseg) &&
+		    !test_bit(NFS_LSEG_LAYOUTRETURN, &bucket->clseg->pls_flags))
+			break;
+		data = nfs_commitdata_alloc(false);
 		if (!data)
 			break;
 		data->ds_commit_index = i;
@@ -283,16 +290,10 @@ pnfs_generic_commit_pagelist(struct inode *inode, struct list_head *mds_pages,
 	unsigned int nreq = 0;
 
 	if (!list_empty(mds_pages)) {
-		data = nfs_commitdata_alloc();
+		data = nfs_commitdata_alloc(true);
-		if (data != NULL) {
+		data->ds_commit_index = -1;
-			data->ds_commit_index = -1;
+		list_add(&data->pages, &list);
-			list_add(&data->pages, &list);
+		nreq++;
-			nreq++;
-		} else {
-			nfs_retry_commit(mds_pages, NULL, cinfo, 0);
-			pnfs_generic_retry_commit(cinfo, 0);
-			return -ENOMEM;
-		}
 	}
 
 	nreq += pnfs_generic_alloc_ds_commits(cinfo, &list);
@@ -619,7 +620,6 @@ void nfs4_pnfs_v3_ds_connect_unload(void)
 		get_v3_ds_connect = NULL;
 	}
 }
-EXPORT_SYMBOL_GPL(nfs4_pnfs_v3_ds_connect_unload);
 
 static int _nfs4_pnfs_v3_ds_connect(struct nfs_server *mds_srv,
 				 struct nfs4_pnfs_ds *ds,

fs/nfs/proc.c

@@ -638,7 +638,7 @@ nfs_proc_lock(struct file *filp, int cmd, struct file_lock *fl)
 {
 	struct inode *inode = file_inode(filp);
 
-	return nlmclnt_proc(NFS_SERVER(inode)->nlm_host, cmd, fl);
+	return nlmclnt_proc(NFS_SERVER(inode)->nlm_host, cmd, fl, NULL);
 }
 
 /* Helper functions for NFS lock bounds checking */

fs/nfs/read.c

@@ -35,7 +35,11 @@ static struct kmem_cache *nfs_rdata_cachep;
 
 static struct nfs_pgio_header *nfs_readhdr_alloc(void)
 {
-	return kmem_cache_zalloc(nfs_rdata_cachep, GFP_KERNEL);
+	struct nfs_pgio_header *p = kmem_cache_zalloc(nfs_rdata_cachep, GFP_KERNEL);
+
+	if (p)
+		p->rw_mode = FMODE_READ;
+	return p;
 }
 
 static void nfs_readhdr_free(struct nfs_pgio_header *rhdr)
@@ -64,7 +68,7 @@ void nfs_pageio_init_read(struct nfs_pageio_descriptor *pgio,
 		pg_ops = server->pnfs_curr_ld->pg_read_ops;
 #endif
 	nfs_pageio_init(pgio, inode, pg_ops, compl_ops, &nfs_rw_read_ops,
-			server->rsize, 0);
+			server->rsize, 0, GFP_KERNEL);
 }
 EXPORT_SYMBOL_GPL(nfs_pageio_init_read);
 
@@ -451,7 +455,6 @@ void nfs_destroy_readpagecache(void)
 }
 
 static const struct nfs_rw_ops nfs_rw_read_ops = {
-	.rw_mode		= FMODE_READ,
 	.rw_alloc_header	= nfs_readhdr_alloc,
 	.rw_free_header		= nfs_readhdr_free,
 	.rw_done		= nfs_readpage_done,

fs/nfs/write.c

@@ -60,14 +60,28 @@ static mempool_t *nfs_wdata_mempool;
 static struct kmem_cache *nfs_cdata_cachep;
 static mempool_t *nfs_commit_mempool;
 
-struct nfs_commit_data *nfs_commitdata_alloc(void)
+struct nfs_commit_data *nfs_commitdata_alloc(bool never_fail)
 {
-	struct nfs_commit_data *p = mempool_alloc(nfs_commit_mempool, GFP_NOIO);
+	struct nfs_commit_data *p;
 
-	if (p) {
+	if (never_fail)
-		memset(p, 0, sizeof(*p));
+		p = mempool_alloc(nfs_commit_mempool, GFP_NOIO);
-		INIT_LIST_HEAD(&p->pages);
+	else {
+		/* It is OK to do some reclaim, not no safe to wait
+		 * for anything to be returned to the pool.
+		 * mempool_alloc() cannot handle that particular combination,
+		 * so we need two separate attempts.
+		 */
+		p = mempool_alloc(nfs_commit_mempool, GFP_NOWAIT);
+		if (!p)
+			p = kmem_cache_alloc(nfs_cdata_cachep, GFP_NOIO |
+					     __GFP_NOWARN | __GFP_NORETRY);
+		if (!p)
+			return NULL;
 	}
+
+	memset(p, 0, sizeof(*p));
+	INIT_LIST_HEAD(&p->pages);
 	return p;
 }
 EXPORT_SYMBOL_GPL(nfs_commitdata_alloc);
@@ -82,8 +96,10 @@ static struct nfs_pgio_header *nfs_writehdr_alloc(void)
 {
 	struct nfs_pgio_header *p = mempool_alloc(nfs_wdata_mempool, GFP_NOIO);
 
-	if (p)
+	if (p) {
 		memset(p, 0, sizeof(*p));
+		p->rw_mode = FMODE_WRITE;
+	}
 	return p;
 }
 
@@ -547,9 +563,21 @@ static void nfs_write_error_remove_page(struct nfs_page *req)
 {
 	nfs_unlock_request(req);
 	nfs_end_page_writeback(req);
-	nfs_release_request(req);
 	generic_error_remove_page(page_file_mapping(req->wb_page),
 				  req->wb_page);
+	nfs_release_request(req);
+}
+
+static bool
+nfs_error_is_fatal_on_server(int err)
+{
+	switch (err) {
+	case 0:
+	case -ERESTARTSYS:
+	case -EINTR:
+		return false;
+	}
+	return nfs_error_is_fatal(err);
 }
 
 /*
@@ -557,8 +585,7 @@ static void nfs_write_error_remove_page(struct nfs_page *req)
  * May return an error if the user signalled nfs_wait_on_request().
  */
 static int nfs_page_async_flush(struct nfs_pageio_descriptor *pgio,
-				struct page *page, bool nonblock,
+				struct page *page, bool nonblock)
-				bool launder)
 {
 	struct nfs_page *req;
 	int ret = 0;
@@ -574,19 +601,19 @@ static int nfs_page_async_flush(struct nfs_pageio_descriptor *pgio,
 	WARN_ON_ONCE(test_bit(PG_CLEAN, &req->wb_flags));
 
 	ret = 0;
+	/* If there is a fatal error that covers this write, just exit */
+	if (nfs_error_is_fatal_on_server(req->wb_context->error))
+		goto out_launder;
+
 	if (!nfs_pageio_add_request(pgio, req)) {
 		ret = pgio->pg_error;
 		/*
-		 * Remove the problematic req upon fatal errors
+		 * Remove the problematic req upon fatal errors on the server
-		 * in launder case, while other dirty pages can
-		 * still be around until they get flushed.
 		 */
 		if (nfs_error_is_fatal(ret)) {
 			nfs_context_set_write_error(req->wb_context, ret);
-			if (launder) {
+			if (nfs_error_is_fatal_on_server(ret))
-				nfs_write_error_remove_page(req);
+				goto out_launder;
-				goto out;
-			}
 		}
 		nfs_redirty_request(req);
 		ret = -EAGAIN;
@@ -595,16 +622,18 @@ static int nfs_page_async_flush(struct nfs_pageio_descriptor *pgio,
 				NFSIOS_WRITEPAGES, 1);
 out:
 	return ret;
+out_launder:
+	nfs_write_error_remove_page(req);
+	return ret;
 }
 
 static int nfs_do_writepage(struct page *page, struct writeback_control *wbc,
-			    struct nfs_pageio_descriptor *pgio, bool launder)
+			    struct nfs_pageio_descriptor *pgio)
 {
 	int ret;
 
 	nfs_pageio_cond_complete(pgio, page_index(page));
-	ret = nfs_page_async_flush(pgio, page, wbc->sync_mode == WB_SYNC_NONE,
+	ret = nfs_page_async_flush(pgio, page, wbc->sync_mode == WB_SYNC_NONE);
-				   launder);
 	if (ret == -EAGAIN) {
 		redirty_page_for_writepage(wbc, page);
 		ret = 0;
@@ -616,8 +645,7 @@ static int nfs_do_writepage(struct page *page, struct writeback_control *wbc,
  * Write an mmapped page to the server.
  */
 static int nfs_writepage_locked(struct page *page,
-				struct writeback_control *wbc,
+				struct writeback_control *wbc)
-				bool launder)
 {
 	struct nfs_pageio_descriptor pgio;
 	struct inode *inode = page_file_mapping(page)->host;
@@ -626,7 +654,7 @@ static int nfs_writepage_locked(struct page *page,
 	nfs_inc_stats(inode, NFSIOS_VFSWRITEPAGE);
 	nfs_pageio_init_write(&pgio, inode, 0,
 				false, &nfs_async_write_completion_ops);
-	err = nfs_do_writepage(page, wbc, &pgio, launder);
+	err = nfs_do_writepage(page, wbc, &pgio);
 	nfs_pageio_complete(&pgio);
 	if (err < 0)
 		return err;
@@ -639,7 +667,7 @@ int nfs_writepage(struct page *page, struct writeback_control *wbc)
 {
 	int ret;
 
-	ret = nfs_writepage_locked(page, wbc, false);
+	ret = nfs_writepage_locked(page, wbc);
 	unlock_page(page);
 	return ret;
 }
@@ -648,7 +676,7 @@ static int nfs_writepages_callback(struct page *page, struct writeback_control *
 {
 	int ret;
 
-	ret = nfs_do_writepage(page, wbc, data, false);
+	ret = nfs_do_writepage(page, wbc, data);
 	unlock_page(page);
 	return ret;
 }
@@ -1367,7 +1395,7 @@ void nfs_pageio_init_write(struct nfs_pageio_descriptor *pgio,
 		pg_ops = server->pnfs_curr_ld->pg_write_ops;
 #endif
 	nfs_pageio_init(pgio, inode, pg_ops, compl_ops, &nfs_rw_write_ops,
-			server->wsize, ioflags);
+			server->wsize, ioflags, GFP_NOIO);
 }
 EXPORT_SYMBOL_GPL(nfs_pageio_init_write);
 
@@ -1704,51 +1732,15 @@ nfs_commit_list(struct inode *inode, struct list_head *head, int how,
 	if (list_empty(head))
 		return 0;
 
-	data = nfs_commitdata_alloc();
+	data = nfs_commitdata_alloc(true);
-
-	if (!data)
-		goto out_bad;
 
 	/* Set up the argument struct */
 	nfs_init_commit(data, head, NULL, cinfo);
 	atomic_inc(&cinfo->mds->rpcs_out);
 	return nfs_initiate_commit(NFS_CLIENT(inode), data, NFS_PROTO(inode),
 				   data->mds_ops, how, 0);
- out_bad:
-	nfs_retry_commit(head, NULL, cinfo, 0);
-	return -ENOMEM;
 }
 
-int nfs_commit_file(struct file *file, struct nfs_write_verifier *verf)
-{
-	struct inode *inode = file_inode(file);
-	struct nfs_open_context *open;
-	struct nfs_commit_info cinfo;
-	struct nfs_page *req;
-	int ret;
-
-	open = get_nfs_open_context(nfs_file_open_context(file));
-	req  = nfs_create_request(open, NULL, NULL, 0, i_size_read(inode));
-	if (IS_ERR(req)) {
-		ret = PTR_ERR(req);
-		goto out_put;
-	}
-
-	nfs_init_cinfo_from_inode(&cinfo, inode);
-
-	memcpy(&req->wb_verf, verf, sizeof(struct nfs_write_verifier));
-	nfs_request_add_commit_list(req, &cinfo);
-	ret = nfs_commit_inode(inode, FLUSH_SYNC);
-	if (ret > 0)
-		ret = 0;
-
-	nfs_free_request(req);
-out_put:
-	put_nfs_open_context(open);
-	return ret;
-}
-EXPORT_SYMBOL_GPL(nfs_commit_file);
-
 /*
  * COMMIT call returned
  */
@@ -1985,7 +1977,7 @@ int nfs_wb_page_cancel(struct inode *inode, struct page *page)
 /*
  * Write back all requests on one page - we do this before reading it.
  */
-int nfs_wb_single_page(struct inode *inode, struct page *page, bool launder)
+int nfs_wb_page(struct inode *inode, struct page *page)
 {
 	loff_t range_start = page_file_offset(page);
 	loff_t range_end = range_start + (loff_t)(PAGE_SIZE - 1);
@@ -2002,7 +1994,7 @@ int nfs_wb_single_page(struct inode *inode, struct page *page, bool launder)
 	for (;;) {
 		wait_on_page_writeback(page);
 		if (clear_page_dirty_for_io(page)) {
-			ret = nfs_writepage_locked(page, &wbc, launder);
+			ret = nfs_writepage_locked(page, &wbc);
 			if (ret < 0)
 				goto out_error;
 			continue;
@@ -2107,7 +2099,6 @@ void nfs_destroy_writepagecache(void)
 }
 
 static const struct nfs_rw_ops nfs_rw_write_ops = {
-	.rw_mode		= FMODE_WRITE,
 	.rw_alloc_header	= nfs_writehdr_alloc,
 	.rw_free_header		= nfs_writehdr_free,
 	.rw_done		= nfs_writeback_done,

include/linux/fs.h

@@ -909,6 +909,8 @@ static inline struct file *get_file(struct file *f)
 #define FL_OFDLCK	1024	/* lock is "owned" by struct file */
 #define FL_LAYOUT	2048	/* outstanding pNFS layout */
 
+#define FL_CLOSE_POSIX (FL_POSIX | FL_CLOSE)
+
 /*
  * Special return value from posix_lock_file() and vfs_lock_file() for
  * asynchronous locking.

include/linux/lockd/bind.h

@@ -18,6 +18,7 @@
 
 /* Dummy declarations */
 struct svc_rqst;
+struct rpc_task;
 
 /*
  * This is the set of functions for lockd->nfsd communication
@@ -43,6 +44,7 @@ struct nlmclnt_initdata {
 	u32			nfs_version;
 	int			noresvport;
 	struct net		*net;
+	const struct nlmclnt_operations	*nlmclnt_ops;
 };
 
 /*
@@ -52,8 +54,26 @@ struct nlmclnt_initdata {
 extern struct nlm_host *nlmclnt_init(const struct nlmclnt_initdata *nlm_init);
 extern void	nlmclnt_done(struct nlm_host *host);
 
-extern int	nlmclnt_proc(struct nlm_host *host, int cmd,
+/*
-					struct file_lock *fl);
+ * NLM client operations provide a means to modify RPC processing of NLM
+ * requests.  Callbacks receive a pointer to data passed into the call to
+ * nlmclnt_proc().
+ */
+struct nlmclnt_operations {
+	/* Called on successful allocation of nlm_rqst, use for allocation or
+	 * reference counting. */
+	void (*nlmclnt_alloc_call)(void *);
+
+	/* Called in rpc_task_prepare for unlock.  A return value of true
+	 * indicates the callback has put the task to sleep on a waitqueue
+	 * and NLM should not call rpc_call_start(). */
+	bool (*nlmclnt_unlock_prepare)(struct rpc_task*, void *);
+
+	/* Called when the nlm_rqst is freed, callbacks should clean up here */
+	void (*nlmclnt_release_call)(void *);
+};
+
+extern int	nlmclnt_proc(struct nlm_host *host, int cmd, struct file_lock *fl, void *data);
 extern int	lockd_up(struct net *net);
 extern void	lockd_down(struct net *net);
 

include/linux/lockd/lockd.h

@@ -69,6 +69,7 @@ struct nlm_host {
 	char			*h_addrbuf;	/* address eyecatcher */
 	struct net		*net;		/* host net */
 	char			nodename[UNX_MAXNODENAME + 1];
+	const struct nlmclnt_operations	*h_nlmclnt_ops;	/* Callback ops for NLM users */
 };
 
 /*
@@ -142,6 +143,7 @@ struct nlm_rqst {
 	struct nlm_block *	a_block;
 	unsigned int		a_retries;	/* Retry count */
 	u8			a_owner[NLMCLNT_OHSIZE];
+	void *	a_callback_data; /* sent to nlmclnt_operations callbacks */
 };
 
 /*

include/linux/nfs_fs.h

@@ -76,6 +76,7 @@ struct nfs_open_context {
 #define NFS_CONTEXT_ERROR_WRITE		(0)
 #define NFS_CONTEXT_RESEND_WRITES	(1)
 #define NFS_CONTEXT_BAD			(2)
+#define NFS_CONTEXT_UNLOCK	(3)
 	int error;
 
 	struct list_head list;
@@ -499,24 +500,12 @@ extern int  nfs_updatepage(struct file *, struct page *, unsigned int, unsigned
  */
 extern int nfs_sync_inode(struct inode *inode);
 extern int nfs_wb_all(struct inode *inode);
-extern int nfs_wb_single_page(struct inode *inode, struct page *page, bool launder);
+extern int nfs_wb_page(struct inode *inode, struct page *page);
 extern int nfs_wb_page_cancel(struct inode *inode, struct page* page);
 extern int  nfs_commit_inode(struct inode *, int);
-extern struct nfs_commit_data *nfs_commitdata_alloc(void);
+extern struct nfs_commit_data *nfs_commitdata_alloc(bool never_fail);
 extern void nfs_commit_free(struct nfs_commit_data *data);
 
-static inline int
-nfs_wb_launder_page(struct inode *inode, struct page *page)
-{
-	return nfs_wb_single_page(inode, page, true);
-}
-
-static inline int
-nfs_wb_page(struct inode *inode, struct page *page)
-{
-	return nfs_wb_single_page(inode, page, false);
-}
-
 static inline int
 nfs_have_writebacks(struct inode *inode)
 {

include/linux/nfs_fs_sb.h

@@ -221,6 +221,7 @@ struct nfs_server {
 	u32			mountd_version;
 	unsigned short		mountd_port;
 	unsigned short		mountd_protocol;
+	struct rpc_wait_queue	uoc_rpcwaitq;
 };
 
 /* Server capabilities */

include/linux/nfs_page.h

@@ -64,7 +64,6 @@ struct nfs_pageio_ops {
 };
 
 struct nfs_rw_ops {
-	const fmode_t rw_mode;
 	struct nfs_pgio_header *(*rw_alloc_header)(void);
 	void (*rw_free_header)(struct nfs_pgio_header *);
 	int  (*rw_done)(struct rpc_task *, struct nfs_pgio_header *,
@@ -124,7 +123,8 @@ extern	void nfs_pageio_init(struct nfs_pageio_descriptor *desc,
 			     const struct nfs_pgio_completion_ops *compl_ops,
 			     const struct nfs_rw_ops *rw_ops,
 			     size_t bsize,
-			     int how);
+			     int how,
+			     gfp_t gfp_flags);
 extern	int nfs_pageio_add_request(struct nfs_pageio_descriptor *,
 				   struct nfs_page *);
 extern  int nfs_pageio_resend(struct nfs_pageio_descriptor *,
@@ -141,6 +141,7 @@ extern int nfs_page_group_lock(struct nfs_page *, bool);
 extern void nfs_page_group_lock_wait(struct nfs_page *);
 extern void nfs_page_group_unlock(struct nfs_page *);
 extern bool nfs_page_group_sync_on_bit(struct nfs_page *, unsigned int);
+extern bool nfs_async_iocounter_wait(struct rpc_task *, struct nfs_lock_context *);
 
 /*
  * Lock the page of an asynchronous request

include/linux/nfs_xdr.h

@@ -1383,6 +1383,7 @@ struct nfs42_copy_res {
 	struct nfs42_write_res		write_res;
 	bool				consecutive;
 	bool				synchronous;
+	struct nfs_commitres		commit_res;
 };
 
 struct nfs42_seek_args {
@@ -1427,6 +1428,7 @@ struct nfs_pgio_header {
 	struct list_head	pages;
 	struct nfs_page		*req;
 	struct nfs_writeverf	verf;		/* Used for writes */
+	fmode_t			rw_mode;
 	struct pnfs_layout_segment *lseg;
 	loff_t			io_start;
 	const struct rpc_call_ops *mds_ops;
@@ -1550,6 +1552,7 @@ struct nfs_rpc_ops {
 	const struct inode_operations *dir_inode_ops;
 	const struct inode_operations *file_inode_ops;
 	const struct file_operations *file_ops;
+	const struct nlmclnt_operations *nlmclnt_ops;
 
 	int	(*getroot) (struct nfs_server *, struct nfs_fh *,
 			    struct nfs_fsinfo *);

net/sunrpc/clnt.c

@@ -1042,8 +1042,6 @@ struct rpc_task *rpc_run_task(const struct rpc_task_setup *task_setup_data)
 	struct rpc_task *task;
 
 	task = rpc_new_task(task_setup_data);
-	if (IS_ERR(task))
-		goto out;
 
 	rpc_task_set_client(task, task_setup_data->rpc_client);
 	rpc_task_set_rpc_message(task, task_setup_data->rpc_message);
@@ -1053,7 +1051,6 @@ struct rpc_task *rpc_run_task(const struct rpc_task_setup *task_setup_data)
 
 	atomic_inc(&task->tk_count);
 	rpc_execute(task);
-out:
 	return task;
 }
 EXPORT_SYMBOL_GPL(rpc_run_task);
@@ -1140,10 +1137,6 @@ struct rpc_task *rpc_run_bc_task(struct rpc_rqst *req)
 	 * Create an rpc_task to send the data
 	 */
 	task = rpc_new_task(&task_setup_data);
-	if (IS_ERR(task)) {
-		xprt_free_bc_request(req);
-		goto out;
-	}
 	task->tk_rqstp = req;
 
 	/*
@@ -1158,7 +1151,6 @@ struct rpc_task *rpc_run_bc_task(struct rpc_rqst *req)
 	WARN_ON_ONCE(atomic_read(&task->tk_count) != 2);
 	rpc_execute(task);
 
-out:
 	dprintk("RPC: rpc_run_bc_task: task= %p\n", task);
 	return task;
 }

net/sunrpc/sched.c

@@ -965,11 +965,6 @@ struct rpc_task *rpc_new_task(const struct rpc_task_setup *setup_data)
 
 	if (task == NULL) {
 		task = rpc_alloc_task();
-		if (task == NULL) {
-			rpc_release_calldata(setup_data->callback_ops,
-					setup_data->callback_data);
-			return ERR_PTR(-ENOMEM);
-		}
 		flags = RPC_TASK_DYNAMIC;
 	}
 

net/sunrpc/xdr.c

@@ -807,7 +807,7 @@ void xdr_init_decode(struct xdr_stream *xdr, struct xdr_buf *buf, __be32 *p)
 EXPORT_SYMBOL_GPL(xdr_init_decode);
 
 /**
- * xdr_init_decode - Initialize an xdr_stream for decoding data.
+ * xdr_init_decode_pages - Initialize an xdr_stream for decoding into pages
  * @xdr: pointer to xdr_stream struct
  * @buf: pointer to XDR buffer from which to decode data
  * @pages: list of pages to decode into

net/sunrpc/xprt.c

@@ -651,6 +651,7 @@ void xprt_force_disconnect(struct rpc_xprt *xprt)
 	xprt_wake_pending_tasks(xprt, -EAGAIN);
 	spin_unlock_bh(&xprt->transport_lock);
 }
+EXPORT_SYMBOL_GPL(xprt_force_disconnect);
 
 /**
  * xprt_conditional_disconnect - force a transport to disconnect

net/sunrpc/xprtrdma/rpc_rdma.c

@@ -494,7 +494,7 @@ rpcrdma_prepare_hdr_sge(struct rpcrdma_ia *ia, struct rpcrdma_req *req,
 	}
 	sge->length = len;
 
-	ib_dma_sync_single_for_device(ia->ri_device, sge->addr,
+	ib_dma_sync_single_for_device(rdmab_device(rb), sge->addr,
 				      sge->length, DMA_TO_DEVICE);
 	req->rl_send_wr.num_sge++;
 	return true;
@@ -523,7 +523,7 @@ rpcrdma_prepare_msg_sges(struct rpcrdma_ia *ia, struct rpcrdma_req *req,
 	sge[sge_no].addr = rdmab_addr(rb);
 	sge[sge_no].length = xdr->head[0].iov_len;
 	sge[sge_no].lkey = rdmab_lkey(rb);
-	ib_dma_sync_single_for_device(device, sge[sge_no].addr,
+	ib_dma_sync_single_for_device(rdmab_device(rb), sge[sge_no].addr,
 				      sge[sge_no].length, DMA_TO_DEVICE);
 
 	/* If there is a Read chunk, the page list is being handled
@@ -781,9 +781,11 @@ rpcrdma_marshal_req(struct rpc_rqst *rqst)
 	return 0;
 
 out_err:
-	pr_err("rpcrdma: rpcrdma_marshal_req failed, status %ld\n",
+	if (PTR_ERR(iptr) != -ENOBUFS) {
-	       PTR_ERR(iptr));
+		pr_err("rpcrdma: rpcrdma_marshal_req failed, status %ld\n",
-	r_xprt->rx_stats.failed_marshal_count++;
+		       PTR_ERR(iptr));
+		r_xprt->rx_stats.failed_marshal_count++;
+	}
 	return PTR_ERR(iptr);
 }
 

net/sunrpc/xprtrdma/transport.c

@@ -66,8 +66,8 @@ static unsigned int xprt_rdma_slot_table_entries = RPCRDMA_DEF_SLOT_TABLE;
 unsigned int xprt_rdma_max_inline_read = RPCRDMA_DEF_INLINE;
 static unsigned int xprt_rdma_max_inline_write = RPCRDMA_DEF_INLINE;
 static unsigned int xprt_rdma_inline_write_padding;
-static unsigned int xprt_rdma_memreg_strategy = RPCRDMA_FRMR;
+unsigned int xprt_rdma_memreg_strategy		= RPCRDMA_FRMR;
-		int xprt_rdma_pad_optimize = 0;
+int xprt_rdma_pad_optimize;
 
 #if IS_ENABLED(CONFIG_SUNRPC_DEBUG)
 
@@ -396,7 +396,7 @@ xprt_setup_rdma(struct xprt_create *args)
 
 	new_xprt = rpcx_to_rdmax(xprt);
 
-	rc = rpcrdma_ia_open(new_xprt, sap, xprt_rdma_memreg_strategy);
+	rc = rpcrdma_ia_open(new_xprt, sap);
 	if (rc)
 		goto out1;
 
@@ -457,19 +457,33 @@ out1:
 	return ERR_PTR(rc);
 }
 
-/*
+/**
- * Close a connection, during shutdown or timeout/reconnect
+ * xprt_rdma_close - Close down RDMA connection
+ * @xprt: generic transport to be closed
+ *
+ * Called during transport shutdown reconnect, or device
+ * removal. Caller holds the transport's write lock.
  */
 static void
 xprt_rdma_close(struct rpc_xprt *xprt)
 {
 	struct rpcrdma_xprt *r_xprt = rpcx_to_rdmax(xprt);
+	struct rpcrdma_ep *ep = &r_xprt->rx_ep;
+	struct rpcrdma_ia *ia = &r_xprt->rx_ia;
 
-	dprintk("RPC:       %s: closing\n", __func__);
+	dprintk("RPC:       %s: closing xprt %p\n", __func__, xprt);
-	if (r_xprt->rx_ep.rep_connected > 0)
+
+	if (test_and_clear_bit(RPCRDMA_IAF_REMOVING, &ia->ri_flags)) {
+		xprt_clear_connected(xprt);
+		rpcrdma_ia_remove(ia);
+		return;
+	}
+	if (ep->rep_connected == -ENODEV)
+		return;
+	if (ep->rep_connected > 0)
 		xprt->reestablish_timeout = 0;
 	xprt_disconnect_done(xprt);
-	rpcrdma_ep_disconnect(&r_xprt->rx_ep, &r_xprt->rx_ia);
+	rpcrdma_ep_disconnect(ep, ia);
 }
 
 static void
@@ -484,6 +498,27 @@ xprt_rdma_set_port(struct rpc_xprt *xprt, u16 port)
 	dprintk("RPC:       %s: %u\n", __func__, port);
 }
 
+/**
+ * xprt_rdma_timer - invoked when an RPC times out
+ * @xprt: controlling RPC transport
+ * @task: RPC task that timed out
+ *
+ * Invoked when the transport is still connected, but an RPC
+ * retransmit timeout occurs.
+ *
+ * Since RDMA connections don't have a keep-alive, forcibly
+ * disconnect and retry to connect. This drives full
+ * detection of the network path, and retransmissions of
+ * all pending RPCs.
+ */
+static void
+xprt_rdma_timer(struct rpc_xprt *xprt, struct rpc_task *task)
+{
+	dprintk("RPC: %5u %s: xprt = %p\n", task->tk_pid, __func__, xprt);
+
+	xprt_force_disconnect(xprt);
+}
+
 static void
 xprt_rdma_connect(struct rpc_xprt *xprt, struct rpc_task *task)
 {
@@ -659,6 +694,8 @@ xprt_rdma_free(struct rpc_task *task)
  * xprt_rdma_send_request - marshal and send an RPC request
  * @task: RPC task with an RPC message in rq_snd_buf
  *
+ * Caller holds the transport's write lock.
+ *
  * Return values:
  *        0:	The request has been sent
  * ENOTCONN:	Caller needs to invoke connect logic then call again
@@ -685,6 +722,9 @@ xprt_rdma_send_request(struct rpc_task *task)
 	struct rpcrdma_xprt *r_xprt = rpcx_to_rdmax(xprt);
 	int rc = 0;
 
+	if (!xprt_connected(xprt))
+		goto drop_connection;
+
 	/* On retransmit, remove any previously registered chunks */
 	if (unlikely(!list_empty(&req->rl_registered)))
 		r_xprt->rx_ia.ri_ops->ro_unmap_safe(r_xprt, req, false);
@@ -776,6 +816,7 @@ static struct rpc_xprt_ops xprt_rdma_procs = {
 	.alloc_slot		= xprt_alloc_slot,
 	.release_request	= xprt_release_rqst_cong,       /* ditto */
 	.set_retrans_timeout	= xprt_set_retrans_timeout_def, /* ditto */
+	.timer			= xprt_rdma_timer,
 	.rpcbind		= rpcb_getport_async,	/* sunrpc/rpcb_clnt.c */
 	.set_port		= xprt_rdma_set_port,
 	.connect		= xprt_rdma_connect,

net/sunrpc/xprtrdma/verbs.c

@@ -53,7 +53,7 @@
 #include <linux/sunrpc/addr.h>
 #include <linux/sunrpc/svc_rdma.h>
 #include <asm/bitops.h>
-#include <linux/module.h> /* try_module_get()/module_put() */
+
 #include <rdma/ib_cm.h>
 
 #include "xprt_rdma.h"
@@ -69,8 +69,11 @@
 /*
  * internal functions
  */
+static void rpcrdma_create_mrs(struct rpcrdma_xprt *r_xprt);
+static void rpcrdma_destroy_mrs(struct rpcrdma_buffer *buf);
+static void rpcrdma_dma_unmap_regbuf(struct rpcrdma_regbuf *rb);
 
-static struct workqueue_struct *rpcrdma_receive_wq;
+static struct workqueue_struct *rpcrdma_receive_wq __read_mostly;
 
 int
 rpcrdma_alloc_wq(void)
@@ -180,7 +183,7 @@ rpcrdma_wc_receive(struct ib_cq *cq, struct ib_wc *wc)
 	rep->rr_wc_flags = wc->wc_flags;
 	rep->rr_inv_rkey = wc->ex.invalidate_rkey;
 
-	ib_dma_sync_single_for_cpu(rep->rr_device,
+	ib_dma_sync_single_for_cpu(rdmab_device(rep->rr_rdmabuf),
 				   rdmab_addr(rep->rr_rdmabuf),
 				   rep->rr_len, DMA_FROM_DEVICE);
 
@@ -262,6 +265,21 @@ rpcrdma_conn_upcall(struct rdma_cm_id *id, struct rdma_cm_event *event)
 			__func__, ep);
 		complete(&ia->ri_done);
 		break;
+	case RDMA_CM_EVENT_DEVICE_REMOVAL:
+#if IS_ENABLED(CONFIG_SUNRPC_DEBUG)
+		pr_info("rpcrdma: removing device for %pIS:%u\n",
+			sap, rpc_get_port(sap));
+#endif
+		set_bit(RPCRDMA_IAF_REMOVING, &ia->ri_flags);
+		ep->rep_connected = -ENODEV;
+		xprt_force_disconnect(&xprt->rx_xprt);
+		wait_for_completion(&ia->ri_remove_done);
+
+		ia->ri_id = NULL;
+		ia->ri_pd = NULL;
+		ia->ri_device = NULL;
+		/* Return 1 to ensure the core destroys the id. */
+		return 1;
 	case RDMA_CM_EVENT_ESTABLISHED:
 		connstate = 1;
 		ib_query_qp(ia->ri_id->qp, attr,
@@ -291,9 +309,6 @@ rpcrdma_conn_upcall(struct rdma_cm_id *id, struct rdma_cm_event *event)
 		goto connected;
 	case RDMA_CM_EVENT_DISCONNECTED:
 		connstate = -ECONNABORTED;
-		goto connected;
-	case RDMA_CM_EVENT_DEVICE_REMOVAL:
-		connstate = -ENODEV;
 connected:
 		dprintk("RPC:       %s: %sconnected\n",
 					__func__, connstate > 0 ? "" : "dis");
@@ -329,14 +344,6 @@ connected:
 	return 0;
 }
 
-static void rpcrdma_destroy_id(struct rdma_cm_id *id)
-{
-	if (id) {
-		module_put(id->device->owner);
-		rdma_destroy_id(id);
-	}
-}
-
 static struct rdma_cm_id *
 rpcrdma_create_id(struct rpcrdma_xprt *xprt,
 			struct rpcrdma_ia *ia, struct sockaddr *addr)
@@ -346,6 +353,7 @@ rpcrdma_create_id(struct rpcrdma_xprt *xprt,
 	int rc;
 
 	init_completion(&ia->ri_done);
+	init_completion(&ia->ri_remove_done);
 
 	id = rdma_create_id(&init_net, rpcrdma_conn_upcall, xprt, RDMA_PS_TCP,
 			    IB_QPT_RC);
@@ -370,16 +378,6 @@ rpcrdma_create_id(struct rpcrdma_xprt *xprt,
 		goto out;
 	}
 
-	/* FIXME:
-	 * Until xprtrdma supports DEVICE_REMOVAL, the provider must
-	 * be pinned while there are active NFS/RDMA mounts to prevent
-	 * hangs and crashes at umount time.
-	 */
-	if (!ia->ri_async_rc && !try_module_get(id->device->owner)) {
-		dprintk("RPC:       %s: Failed to get device module\n",
-			__func__);
-		ia->ri_async_rc = -ENODEV;
-	}
 	rc = ia->ri_async_rc;
 	if (rc)
 		goto out;
@@ -389,21 +387,20 @@ rpcrdma_create_id(struct rpcrdma_xprt *xprt,
 	if (rc) {
 		dprintk("RPC:       %s: rdma_resolve_route() failed %i\n",
 			__func__, rc);
-		goto put;
+		goto out;
 	}
 	rc = wait_for_completion_interruptible_timeout(&ia->ri_done, wtimeout);
 	if (rc < 0) {
 		dprintk("RPC:       %s: wait() exited: %i\n",
 			__func__, rc);
-		goto put;
+		goto out;
 	}
 	rc = ia->ri_async_rc;
 	if (rc)
-		goto put;
+		goto out;
 
 	return id;
-put:
+
-	module_put(id->device->owner);
 out:
 	rdma_destroy_id(id);
 	return ERR_PTR(rc);
@@ -413,13 +410,16 @@ out:
  * Exported functions.
  */
 
-/*
+/**
- * Open and initialize an Interface Adapter.
+ * rpcrdma_ia_open - Open and initialize an Interface Adapter.
- *  o initializes fields of struct rpcrdma_ia, including
+ * @xprt: controlling transport
- *    interface and provider attributes and protection zone.
+ * @addr: IP address of remote peer
+ *
+ * Returns 0 on success, negative errno if an appropriate
+ * Interface Adapter could not be found and opened.
  */
 int
-rpcrdma_ia_open(struct rpcrdma_xprt *xprt, struct sockaddr *addr, int memreg)
+rpcrdma_ia_open(struct rpcrdma_xprt *xprt, struct sockaddr *addr)
 {
 	struct rpcrdma_ia *ia = &xprt->rx_ia;
 	int rc;
@@ -427,7 +427,7 @@ rpcrdma_ia_open(struct rpcrdma_xprt *xprt, struct sockaddr *addr, int memreg)
 	ia->ri_id = rpcrdma_create_id(xprt, ia, addr);
 	if (IS_ERR(ia->ri_id)) {
 		rc = PTR_ERR(ia->ri_id);
-		goto out1;
+		goto out_err;
 	}
 	ia->ri_device = ia->ri_id->device;
 
@@ -435,10 +435,10 @@ rpcrdma_ia_open(struct rpcrdma_xprt *xprt, struct sockaddr *addr, int memreg)
 	if (IS_ERR(ia->ri_pd)) {
 		rc = PTR_ERR(ia->ri_pd);
 		pr_err("rpcrdma: ib_alloc_pd() returned %d\n", rc);
-		goto out2;
+		goto out_err;
 	}
 
-	switch (memreg) {
+	switch (xprt_rdma_memreg_strategy) {
 	case RPCRDMA_FRMR:
 		if (frwr_is_supported(ia)) {
 			ia->ri_ops = &rpcrdma_frwr_memreg_ops;
@@ -452,28 +452,73 @@ rpcrdma_ia_open(struct rpcrdma_xprt *xprt, struct sockaddr *addr, int memreg)
 		}
 		/*FALLTHROUGH*/
 	default:
-		pr_err("rpcrdma: Unsupported memory registration mode: %d\n",
+		pr_err("rpcrdma: Device %s does not support memreg mode %d\n",
-		       memreg);
+		       ia->ri_device->name, xprt_rdma_memreg_strategy);
 		rc = -EINVAL;
-		goto out3;
+		goto out_err;
 	}
 
 	return 0;
 
-out3:
+out_err:
-	ib_dealloc_pd(ia->ri_pd);
+	rpcrdma_ia_close(ia);
-	ia->ri_pd = NULL;
-out2:
-	rpcrdma_destroy_id(ia->ri_id);
-	ia->ri_id = NULL;
-out1:
 	return rc;
 }
 
-/*
+/**
- * Clean up/close an IA.
+ * rpcrdma_ia_remove - Handle device driver unload
- *   o if event handles and PD have been initialized, free them.
+ * @ia: interface adapter being removed
- *   o close the IA
+ *
+ * Divest transport H/W resources associated with this adapter,
+ * but allow it to be restored later.
+ */
+void
+rpcrdma_ia_remove(struct rpcrdma_ia *ia)
+{
+	struct rpcrdma_xprt *r_xprt = container_of(ia, struct rpcrdma_xprt,
+						   rx_ia);
+	struct rpcrdma_ep *ep = &r_xprt->rx_ep;
+	struct rpcrdma_buffer *buf = &r_xprt->rx_buf;
+	struct rpcrdma_req *req;
+	struct rpcrdma_rep *rep;
+
+	cancel_delayed_work_sync(&buf->rb_refresh_worker);
+
+	/* This is similar to rpcrdma_ep_destroy, but:
+	 * - Don't cancel the connect worker.
+	 * - Don't call rpcrdma_ep_disconnect, which waits
+	 *   for another conn upcall, which will deadlock.
+	 * - rdma_disconnect is unneeded, the underlying
+	 *   connection is already gone.
+	 */
+	if (ia->ri_id->qp) {
+		ib_drain_qp(ia->ri_id->qp);
+		rdma_destroy_qp(ia->ri_id);
+		ia->ri_id->qp = NULL;
+	}
+	ib_free_cq(ep->rep_attr.recv_cq);
+	ib_free_cq(ep->rep_attr.send_cq);
+
+	/* The ULP is responsible for ensuring all DMA
+	 * mappings and MRs are gone.
+	 */
+	list_for_each_entry(rep, &buf->rb_recv_bufs, rr_list)
+		rpcrdma_dma_unmap_regbuf(rep->rr_rdmabuf);
+	list_for_each_entry(req, &buf->rb_allreqs, rl_all) {
+		rpcrdma_dma_unmap_regbuf(req->rl_rdmabuf);
+		rpcrdma_dma_unmap_regbuf(req->rl_sendbuf);
+		rpcrdma_dma_unmap_regbuf(req->rl_recvbuf);
+	}
+	rpcrdma_destroy_mrs(buf);
+
+	/* Allow waiters to continue */
+	complete(&ia->ri_remove_done);
+}
+
+/**
+ * rpcrdma_ia_close - Clean up/close an IA.
+ * @ia: interface adapter to close
+ *
  */
 void
 rpcrdma_ia_close(struct rpcrdma_ia *ia)
@@ -482,13 +527,15 @@ rpcrdma_ia_close(struct rpcrdma_ia *ia)
 	if (ia->ri_id != NULL && !IS_ERR(ia->ri_id)) {
 		if (ia->ri_id->qp)
 			rdma_destroy_qp(ia->ri_id);
-		rpcrdma_destroy_id(ia->ri_id);
+		rdma_destroy_id(ia->ri_id);
-		ia->ri_id = NULL;
 	}
+	ia->ri_id = NULL;
+	ia->ri_device = NULL;
 
 	/* If the pd is still busy, xprtrdma missed freeing a resource */
 	if (ia->ri_pd && !IS_ERR(ia->ri_pd))
 		ib_dealloc_pd(ia->ri_pd);
+	ia->ri_pd = NULL;
 }
 
 /*
@@ -646,6 +693,99 @@ rpcrdma_ep_destroy(struct rpcrdma_ep *ep, struct rpcrdma_ia *ia)
 	ib_free_cq(ep->rep_attr.send_cq);
 }
 
+/* Re-establish a connection after a device removal event.
+ * Unlike a normal reconnection, a fresh PD and a new set
+ * of MRs and buffers is needed.
+ */
+static int
+rpcrdma_ep_recreate_xprt(struct rpcrdma_xprt *r_xprt,
+			 struct rpcrdma_ep *ep, struct rpcrdma_ia *ia)
+{
+	struct sockaddr *sap = (struct sockaddr *)&r_xprt->rx_data.addr;
+	int rc, err;
+
+	pr_info("%s: r_xprt = %p\n", __func__, r_xprt);
+
+	rc = -EHOSTUNREACH;
+	if (rpcrdma_ia_open(r_xprt, sap))
+		goto out1;
+
+	rc = -ENOMEM;
+	err = rpcrdma_ep_create(ep, ia, &r_xprt->rx_data);
+	if (err) {
+		pr_err("rpcrdma: rpcrdma_ep_create returned %d\n", err);
+		goto out2;
+	}
+
+	rc = -ENETUNREACH;
+	err = rdma_create_qp(ia->ri_id, ia->ri_pd, &ep->rep_attr);
+	if (err) {
+		pr_err("rpcrdma: rdma_create_qp returned %d\n", err);
+		goto out3;
+	}
+
+	rpcrdma_create_mrs(r_xprt);
+	return 0;
+
+out3:
+	rpcrdma_ep_destroy(ep, ia);
+out2:
+	rpcrdma_ia_close(ia);
+out1:
+	return rc;
+}
+
+static int
+rpcrdma_ep_reconnect(struct rpcrdma_xprt *r_xprt, struct rpcrdma_ep *ep,
+		     struct rpcrdma_ia *ia)
+{
+	struct sockaddr *sap = (struct sockaddr *)&r_xprt->rx_data.addr;
+	struct rdma_cm_id *id, *old;
+	int err, rc;
+
+	dprintk("RPC:       %s: reconnecting...\n", __func__);
+
+	rpcrdma_ep_disconnect(ep, ia);
+
+	rc = -EHOSTUNREACH;
+	id = rpcrdma_create_id(r_xprt, ia, sap);
+	if (IS_ERR(id))
+		goto out;
+
+	/* As long as the new ID points to the same device as the
+	 * old ID, we can reuse the transport's existing PD and all
+	 * previously allocated MRs. Also, the same device means
+	 * the transport's previous DMA mappings are still valid.
+	 *
+	 * This is a sanity check only. There should be no way these
+	 * point to two different devices here.
+	 */
+	old = id;
+	rc = -ENETUNREACH;
+	if (ia->ri_device != id->device) {
+		pr_err("rpcrdma: can't reconnect on different device!\n");
+		goto out_destroy;
+	}
+
+	err = rdma_create_qp(id, ia->ri_pd, &ep->rep_attr);
+	if (err) {
+		dprintk("RPC:       %s: rdma_create_qp returned %d\n",
+			__func__, err);
+		goto out_destroy;
+	}
+
+	/* Atomically replace the transport's ID and QP. */
+	rc = 0;
+	old = ia->ri_id;
+	ia->ri_id = id;
+	rdma_destroy_qp(old);
+
+out_destroy:
+	rdma_destroy_id(old);
+out:
+	return rc;
+}
+
 /*
  * Connect unconnected endpoint.
  */
@@ -654,61 +794,30 @@ rpcrdma_ep_connect(struct rpcrdma_ep *ep, struct rpcrdma_ia *ia)
 {
 	struct rpcrdma_xprt *r_xprt = container_of(ia, struct rpcrdma_xprt,
 						   rx_ia);
-	struct rdma_cm_id *id, *old;
-	struct sockaddr *sap;
 	unsigned int extras;
-	int rc = 0;
+	int rc;
 
-	if (ep->rep_connected != 0) {
 retry:
-		dprintk("RPC:       %s: reconnecting...\n", __func__);
+	switch (ep->rep_connected) {
-
+	case 0:
-		rpcrdma_ep_disconnect(ep, ia);
-
-		sap = (struct sockaddr *)&r_xprt->rx_data.addr;
-		id = rpcrdma_create_id(r_xprt, ia, sap);
-		if (IS_ERR(id)) {
-			rc = -EHOSTUNREACH;
-			goto out;
-		}
-		/* TEMP TEMP TEMP - fail if new device:
-		 * Deregister/remarshal *all* requests!
-		 * Close and recreate adapter, pd, etc!
-		 * Re-determine all attributes still sane!
-		 * More stuff I haven't thought of!
-		 * Rrrgh!
-		 */
-		if (ia->ri_device != id->device) {
-			printk("RPC:       %s: can't reconnect on "
-				"different device!\n", __func__);
-			rpcrdma_destroy_id(id);
-			rc = -ENETUNREACH;
-			goto out;
-		}
-		/* END TEMP */
-		rc = rdma_create_qp(id, ia->ri_pd, &ep->rep_attr);
-		if (rc) {
-			dprintk("RPC:       %s: rdma_create_qp failed %i\n",
-				__func__, rc);
-			rpcrdma_destroy_id(id);
-			rc = -ENETUNREACH;
-			goto out;
-		}
-
-		old = ia->ri_id;
-		ia->ri_id = id;
-
-		rdma_destroy_qp(old);
-		rpcrdma_destroy_id(old);
-	} else {
 		dprintk("RPC:       %s: connecting...\n", __func__);
 		rc = rdma_create_qp(ia->ri_id, ia->ri_pd, &ep->rep_attr);
 		if (rc) {
 			dprintk("RPC:       %s: rdma_create_qp failed %i\n",
 				__func__, rc);
-			/* do not update ep->rep_connected */
+			rc = -ENETUNREACH;
-			return -ENETUNREACH;
+			goto out_noupdate;
 		}
+		break;
+	case -ENODEV:
+		rc = rpcrdma_ep_recreate_xprt(r_xprt, ep, ia);
+		if (rc)
+			goto out_noupdate;
+		break;
+	default:
+		rc = rpcrdma_ep_reconnect(r_xprt, ep, ia);
+		if (rc)
+			goto out;
 	}
 
 	ep->rep_connected = 0;
@@ -736,6 +845,8 @@ retry:
 out:
 	if (rc)
 		ep->rep_connected = rc;
+
+out_noupdate:
 	return rc;
 }
 
@@ -878,7 +989,6 @@ struct rpcrdma_rep *
 rpcrdma_create_rep(struct rpcrdma_xprt *r_xprt)
 {
 	struct rpcrdma_create_data_internal *cdata = &r_xprt->rx_data;
-	struct rpcrdma_ia *ia = &r_xprt->rx_ia;
 	struct rpcrdma_rep *rep;
 	int rc;
 
@@ -894,7 +1004,6 @@ rpcrdma_create_rep(struct rpcrdma_xprt *r_xprt)
 		goto out_free;
 	}
 
-	rep->rr_device = ia->ri_device;
 	rep->rr_cqe.done = rpcrdma_wc_receive;
 	rep->rr_rxprt = r_xprt;
 	INIT_WORK(&rep->rr_work, rpcrdma_reply_handler);
@@ -1037,6 +1146,7 @@ void
 rpcrdma_buffer_destroy(struct rpcrdma_buffer *buf)
 {
 	cancel_delayed_work_sync(&buf->rb_recovery_worker);
+	cancel_delayed_work_sync(&buf->rb_refresh_worker);
 
 	while (!list_empty(&buf->rb_recv_bufs)) {
 		struct rpcrdma_rep *rep;
@@ -1081,7 +1191,8 @@ rpcrdma_get_mw(struct rpcrdma_xprt *r_xprt)
 
 out_nomws:
 	dprintk("RPC:       %s: no MWs available\n", __func__);
-	schedule_delayed_work(&buf->rb_refresh_worker, 0);
+	if (r_xprt->rx_ep.rep_connected != -ENODEV)
+		schedule_delayed_work(&buf->rb_refresh_worker, 0);
 
 	/* Allow the reply handler and refresh worker to run */
 	cond_resched();
@@ -1231,17 +1342,19 @@ rpcrdma_alloc_regbuf(size_t size, enum dma_data_direction direction,
 bool
 __rpcrdma_dma_map_regbuf(struct rpcrdma_ia *ia, struct rpcrdma_regbuf *rb)
 {
+	struct ib_device *device = ia->ri_device;
+
 	if (rb->rg_direction == DMA_NONE)
 		return false;
 
-	rb->rg_iov.addr = ib_dma_map_single(ia->ri_device,
+	rb->rg_iov.addr = ib_dma_map_single(device,
 					    (void *)rb->rg_base,
 					    rdmab_length(rb),
 					    rb->rg_direction);
-	if (ib_dma_mapping_error(ia->ri_device, rdmab_addr(rb)))
+	if (ib_dma_mapping_error(device, rdmab_addr(rb)))
 		return false;
 
-	rb->rg_device = ia->ri_device;
+	rb->rg_device = device;
 	rb->rg_iov.lkey = ia->ri_pd->local_dma_lkey;
 	return true;
 }

net/sunrpc/xprtrdma/xprt_rdma.h

@@ -69,6 +69,7 @@ struct rpcrdma_ia {
 	struct rdma_cm_id 	*ri_id;
 	struct ib_pd		*ri_pd;
 	struct completion	ri_done;
+	struct completion	ri_remove_done;
 	int			ri_async_rc;
 	unsigned int		ri_max_segs;
 	unsigned int		ri_max_frmr_depth;
@@ -78,10 +79,15 @@ struct rpcrdma_ia {
 	bool			ri_reminv_expected;
 	bool			ri_implicit_roundup;
 	enum ib_mr_type		ri_mrtype;
+	unsigned long		ri_flags;
 	struct ib_qp_attr	ri_qp_attr;
 	struct ib_qp_init_attr	ri_qp_init_attr;
 };
 
+enum {
+	RPCRDMA_IAF_REMOVING = 0,
+};
+
 /*
  * RDMA Endpoint -- one per transport instance
  */
@@ -164,6 +170,12 @@ rdmab_to_msg(struct rpcrdma_regbuf *rb)
 	return (struct rpcrdma_msg *)rb->rg_base;
 }
 
+static inline struct ib_device *
+rdmab_device(struct rpcrdma_regbuf *rb)
+{
+	return rb->rg_device;
+}
+
 #define RPCRDMA_DEF_GFP		(GFP_NOIO | __GFP_NOWARN)
 
 /* To ensure a transport can always make forward progress,
@@ -209,7 +221,6 @@ struct rpcrdma_rep {
 	unsigned int		rr_len;
 	int			rr_wc_flags;
 	u32			rr_inv_rkey;
-	struct ib_device	*rr_device;
 	struct rpcrdma_xprt	*rr_rxprt;
 	struct work_struct	rr_work;
 	struct list_head	rr_list;
@@ -380,7 +391,6 @@ struct rpcrdma_buffer {
 	spinlock_t		rb_mwlock;	/* protect rb_mws list */
 	struct list_head	rb_mws;
 	struct list_head	rb_all;
-	char			*rb_pool;
 
 	spinlock_t		rb_lock;	/* protect buf lists */
 	int			rb_send_count, rb_recv_count;
@@ -497,10 +507,16 @@ struct rpcrdma_xprt {
  * Default is 0, see sysctl entry and rpc_rdma.c rpcrdma_convert_iovs() */
 extern int xprt_rdma_pad_optimize;
 
+/* This setting controls the hunt for a supported memory
+ * registration strategy.
+ */
+extern unsigned int xprt_rdma_memreg_strategy;
+
 /*
  * Interface Adapter calls - xprtrdma/verbs.c
  */
-int rpcrdma_ia_open(struct rpcrdma_xprt *, struct sockaddr *, int);
+int rpcrdma_ia_open(struct rpcrdma_xprt *xprt, struct sockaddr *addr);
+void rpcrdma_ia_remove(struct rpcrdma_ia *ia);
 void rpcrdma_ia_close(struct rpcrdma_ia *);
 bool frwr_is_supported(struct rpcrdma_ia *);
 bool fmr_is_supported(struct rpcrdma_ia *);