NFS client updates for Linux 5.20

Highlights include:
 
 Stable fixes:
 - pNFS/flexfiles: Fix infinite looping when the RDMA connection errors out
 
 Bugfixes:
 - NFS: fix port value parsing
 - SUNRPC: Reinitialise the backchannel request buffers before reuse
 - SUNRPC: fix expiry of auth creds
 - NFSv4: Fix races in the legacy idmapper upcall
 - NFS: O_DIRECT fixes from Jeff Layton
 - NFSv4.1: Fix OP_SEQUENCE error handling
 - SUNRPC: Fix an RPC/RDMA performance regression
 - NFS: Fix case insensitive renames
 - NFSv4/pnfs: Fix a use-after-free bug in open
 - NFSv4.1: RECLAIM_COMPLETE must handle EACCES
 
 Features:
 - NFSv4.1: session trunking enhancements
 - NFSv4.2: READ_PLUS performance optimisations
 - NFS: relax the rules for rsize/wsize mount options
 - NFS: don't unhash dentry during unlink/rename
 - SUNRPC: Fail faster on bad verifier
 - NFS/SUNRPC: Various tracing improvements
 -----BEGIN PGP SIGNATURE-----
 
 iQIzBAABCAAdFiEESQctxSBg8JpV8KqEZwvnipYKAPIFAmLzy24ACgkQZwvnipYK
 APJJvhAAnVmv3jeOLjwDm+3X9xBevTq8PXAikWVeJgbSKCdjfqXO1J+XF49MpxXl
 N8PQZMqnwWkF3WvhvYobblwGSl6gJ3RnhVuTgdv4jiPl0ZWyS3XngJY0dCTnNgdL
 5O9AjoOMtnGVZwN5j8agymA8f2TUcel5mED6sAk10t2zDZY7VxuQqjp0m696jYjF
 0PKBmxoC4+6tXtYJS7d2PGRCTjfEUx2BLnnGuLOKsB7X8f63XmxJiu8/AIiY7spr
 M/M+BjAF3ok86dT1LjGlkvSNp23H70Wmsv98udfvxIWBm1l4972oCR/CS+kh3/mU
 dYM+oQ3JxxgKEN7Fdak+zU/+qma9q5z2rPFpSIT1fMEuaKN/7H2cbiHPi5RnEBLa
 AHWilX/lWBIMnZJZd9g3yYcGe6E/pkT6TqW5JY+2510koyfNER4IismAWMx2iYKU
 D0WSZOkmEBS/OYZxpTnqGwvS4L1szo9DN3c+yG2KXLifnmVPpjXZO25wahqSuUo3
 V6eYUCXRJmVg+IuXGsMNdrjxGYxD12xChoYzx5RlXls2lwHGeZr+iG3aL3+XayHa
 I1Kji3500UmfEOUUUr4UiQ428dOdL3QqNzVzdymN8Vh4d7v64LUL0GSseY+10Xrs
 xcbR6l/hwjBIo+I1Bi2mmv3W10tKErFy9eBIKzql3D6VHg7ESOo=
 =U00h
 -----END PGP SIGNATURE-----

Merge tag 'nfs-for-5.20-1' of git://git.linux-nfs.org/projects/trondmy/linux-nfs

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

  Stable fixes:
   - pNFS/flexfiles: Fix infinite looping when the RDMA connection
     errors out

  Bugfixes:
   - NFS: fix port value parsing
   - SUNRPC: Reinitialise the backchannel request buffers before reuse
   - SUNRPC: fix expiry of auth creds
   - NFSv4: Fix races in the legacy idmapper upcall
   - NFS: O_DIRECT fixes from Jeff Layton
   - NFSv4.1: Fix OP_SEQUENCE error handling
   - SUNRPC: Fix an RPC/RDMA performance regression
   - NFS: Fix case insensitive renames
   - NFSv4/pnfs: Fix a use-after-free bug in open
   - NFSv4.1: RECLAIM_COMPLETE must handle EACCES

  Features:
   - NFSv4.1: session trunking enhancements
   - NFSv4.2: READ_PLUS performance optimisations
   - NFS: relax the rules for rsize/wsize mount options
   - NFS: don't unhash dentry during unlink/rename
   - SUNRPC: Fail faster on bad verifier
   - NFS/SUNRPC: Various tracing improvements"

* tag 'nfs-for-5.20-1' of git://git.linux-nfs.org/projects/trondmy/linux-nfs: (46 commits)
  NFS: Improve readpage/writepage tracing
  NFS: Improve O_DIRECT tracing
  NFS: Improve write error tracing
  NFS: don't unhash dentry during unlink/rename
  NFSv4/pnfs: Fix a use-after-free bug in open
  NFS: nfs_async_write_reschedule_io must not recurse into the writeback code
  SUNRPC: Don't reuse bvec on retransmission of the request
  SUNRPC: Reinitialise the backchannel request buffers before reuse
  NFSv4.1: RECLAIM_COMPLETE must handle EACCES
  NFSv4.1 probe offline transports for trunking on session creation
  SUNRPC create a function that probes only offline transports
  SUNRPC export xprt_iter_rewind function
  SUNRPC restructure rpc_clnt_setup_test_and_add_xprt
  NFSv4.1 remove xprt from xprt_switch if session trunking test fails
  SUNRPC create an rpc function that allows xprt removal from rpc_clnt
  SUNRPC enable back offline transports in trunking discovery
  SUNRPC create an iterator to list only OFFLINE xprts
  NFSv4.1 offline trunkable transports on DESTROY_SESSION
  SUNRPC add function to offline remove trunkable transports
  SUNRPC expose functions for offline remote xprt functionality
  ...

fs/nfs/blocklayout/dev.c

@@ -301,18 +301,14 @@ bl_validate_designator(struct pnfs_block_volume *v)
 	}
 }
 
-/*
- * Try to open the udev path for the WWN.  At least on Debian the udev
- * by-id path will always point to the dm-multipath device if one exists.
- */
 static struct block_device *
-bl_open_udev_path(struct pnfs_block_volume *v)
+bl_open_path(struct pnfs_block_volume *v, const char *prefix)
 {
 	struct block_device *bdev;
 	const char *devname;
 
-	devname = kasprintf(GFP_KERNEL, "/dev/disk/by-id/wwn-0x%*phN",
+	devname = kasprintf(GFP_KERNEL, "/dev/disk/by-id/%s%*phN",
-				v->scsi.designator_len, v->scsi.designator);
+			prefix, v->scsi.designator_len, v->scsi.designator);
 	if (!devname)
 		return ERR_PTR(-ENOMEM);
 
@@ -326,28 +322,6 @@ bl_open_udev_path(struct pnfs_block_volume *v)
 	return bdev;
 }
 
-/*
- * Try to open the RH/Fedora specific dm-mpath udev path for this WWN, as the
- * wwn- links will only point to the first discovered SCSI device there.
- */
-static struct block_device *
-bl_open_dm_mpath_udev_path(struct pnfs_block_volume *v)
-{
-	struct block_device *bdev;
-	const char *devname;
-
-	devname = kasprintf(GFP_KERNEL,
-			"/dev/disk/by-id/dm-uuid-mpath-%d%*phN",
-			v->scsi.designator_type,
-			v->scsi.designator_len, v->scsi.designator);
-	if (!devname)
-		return ERR_PTR(-ENOMEM);
-
-	bdev = blkdev_get_by_path(devname, FMODE_READ | FMODE_WRITE, NULL);
-	kfree(devname);
-	return bdev;
-}
-
 static int
 bl_parse_scsi(struct nfs_server *server, struct pnfs_block_dev *d,
 		struct pnfs_block_volume *volumes, int idx, gfp_t gfp_mask)
@@ -360,9 +334,15 @@ bl_parse_scsi(struct nfs_server *server, struct pnfs_block_dev *d,
 	if (!bl_validate_designator(v))
 		return -EINVAL;
 
-	bdev = bl_open_dm_mpath_udev_path(v);
+	/*
+	 * Try to open the RH/Fedora specific dm-mpath udev path first, as the
+	 * wwn- links will only point to the first discovered SCSI device there.
+	 * On other distributions like Debian, the default SCSI by-id path will
+	 * point to the dm-multipath device if one exists.
+	 */
+	bdev = bl_open_path(v, "dm-uuid-mpath-0x");
 	if (IS_ERR(bdev))
-		bdev = bl_open_udev_path(v);
+		bdev = bl_open_path(v, "wwn-0x");
 	if (IS_ERR(bdev))
 		return PTR_ERR(bdev);
 	d->bdev = bdev;

fs/nfs/client.c

@@ -708,9 +708,9 @@ static int nfs_init_server(struct nfs_server *server,
 	}
 
 	if (ctx->rsize)
-		server->rsize = nfs_block_size(ctx->rsize, NULL);
+		server->rsize = nfs_io_size(ctx->rsize, clp->cl_proto);
 	if (ctx->wsize)
-		server->wsize = nfs_block_size(ctx->wsize, NULL);
+		server->wsize = nfs_io_size(ctx->wsize, clp->cl_proto);
 
 	server->acregmin = ctx->acregmin * HZ;
 	server->acregmax = ctx->acregmax * HZ;
@@ -755,18 +755,19 @@ error:
 static void nfs_server_set_fsinfo(struct nfs_server *server,
 				  struct nfs_fsinfo *fsinfo)
 {
+	struct nfs_client *clp = server->nfs_client;
 	unsigned long max_rpc_payload, raw_max_rpc_payload;
 
 	/* Work out a lot of parameters */
 	if (server->rsize == 0)
-		server->rsize = nfs_block_size(fsinfo->rtpref, NULL);
+		server->rsize = nfs_io_size(fsinfo->rtpref, clp->cl_proto);
 	if (server->wsize == 0)
-		server->wsize = nfs_block_size(fsinfo->wtpref, NULL);
+		server->wsize = nfs_io_size(fsinfo->wtpref, clp->cl_proto);
 
 	if (fsinfo->rtmax >= 512 && server->rsize > fsinfo->rtmax)
-		server->rsize = nfs_block_size(fsinfo->rtmax, NULL);
+		server->rsize = nfs_io_size(fsinfo->rtmax, clp->cl_proto);
 	if (fsinfo->wtmax >= 512 && server->wsize > fsinfo->wtmax)
-		server->wsize = nfs_block_size(fsinfo->wtmax, NULL);
+		server->wsize = nfs_io_size(fsinfo->wtmax, clp->cl_proto);
 
 	raw_max_rpc_payload = rpc_max_payload(server->client);
 	max_rpc_payload = nfs_block_size(raw_max_rpc_payload, NULL);

fs/nfs/dir.c

@@ -1084,7 +1084,7 @@ static void nfs_do_filldir(struct nfs_readdir_descriptor *desc,
 	struct nfs_cache_array *array;
 	unsigned int i;
 
-	array = kmap(desc->page);
+	array = kmap_local_page(desc->page);
 	for (i = desc->cache_entry_index; i < array->size; i++) {
 		struct nfs_cache_array_entry *ent;
 
@@ -1110,7 +1110,7 @@ static void nfs_do_filldir(struct nfs_readdir_descriptor *desc,
 	if (array->page_is_eof)
 		desc->eof = !desc->eob;
 
-	kunmap(desc->page);
+	kunmap_local(array);
 	dfprintk(DIRCACHE, "NFS: nfs_do_filldir() filling ended @ cookie %llu\n",
 			(unsigned long long)desc->dir_cookie);
 }
@@ -1739,6 +1739,10 @@ nfs_do_lookup_revalidate(struct inode *dir, struct dentry *dentry,
 		goto out_bad;
 	}
 
+	if ((flags & LOOKUP_RENAME_TARGET) && d_count(dentry) < 2 &&
+	    nfs_server_capable(dir, NFS_CAP_CASE_INSENSITIVE))
+		goto out_bad;
+
 	if (nfs_verifier_is_delegated(dentry))
 		return nfs_lookup_revalidate_delegated(dir, dentry, inode);
 
@@ -1778,6 +1782,8 @@ __nfs_lookup_revalidate(struct dentry *dentry, unsigned int flags,
 	int ret;
 
 	if (flags & LOOKUP_RCU) {
+		if (dentry->d_fsdata == NFS_FSDATA_BLOCKED)
+			return -ECHILD;
 		parent = READ_ONCE(dentry->d_parent);
 		dir = d_inode_rcu(parent);
 		if (!dir)
@@ -1786,6 +1792,9 @@ __nfs_lookup_revalidate(struct dentry *dentry, unsigned int flags,
 		if (parent != READ_ONCE(dentry->d_parent))
 			return -ECHILD;
 	} else {
+		/* Wait for unlink to complete */
+		wait_var_event(&dentry->d_fsdata,
+			       dentry->d_fsdata != NFS_FSDATA_BLOCKED);
 		parent = dget_parent(dentry);
 		ret = reval(d_inode(parent), dentry, flags);
 		dput(parent);
@@ -2454,7 +2463,6 @@ out:
 int nfs_unlink(struct inode *dir, struct dentry *dentry)
 {
 	int error;
-	int need_rehash = 0;
 
 	dfprintk(VFS, "NFS: unlink(%s/%lu, %pd)\n", dir->i_sb->s_id,
 		dir->i_ino, dentry);
@@ -2469,15 +2477,25 @@ int nfs_unlink(struct inode *dir, struct dentry *dentry)
 		error = nfs_sillyrename(dir, dentry);
 		goto out;
 	}
-	if (!d_unhashed(dentry)) {
+	/* We must prevent any concurrent open until the unlink
-		__d_drop(dentry);
+	 * completes.  ->d_revalidate will wait for ->d_fsdata
-		need_rehash = 1;
+	 * to clear.  We set it here to ensure no lookup succeeds until
-	}
+	 * the unlink is complete on the server.
+	 */
+	error = -ETXTBSY;
+	if (WARN_ON(dentry->d_flags & DCACHE_NFSFS_RENAMED) ||
+	    WARN_ON(dentry->d_fsdata == NFS_FSDATA_BLOCKED))
+		goto out;
+	if (dentry->d_fsdata)
+		/* old devname */
+		kfree(dentry->d_fsdata);
+	dentry->d_fsdata = NFS_FSDATA_BLOCKED;
+
 	spin_unlock(&dentry->d_lock);
 	error = nfs_safe_remove(dentry);
 	nfs_dentry_remove_handle_error(dir, dentry, error);
-	if (need_rehash)
+	dentry->d_fsdata = NULL;
-		d_rehash(dentry);
+	wake_up_var(&dentry->d_fsdata);
 out:
 	trace_nfs_unlink_exit(dir, dentry, error);
 	return error;
@@ -2584,6 +2602,15 @@ nfs_link(struct dentry *old_dentry, struct inode *dir, struct dentry *dentry)
 }
 EXPORT_SYMBOL_GPL(nfs_link);
 
+static void
+nfs_unblock_rename(struct rpc_task *task, struct nfs_renamedata *data)
+{
+	struct dentry *new_dentry = data->new_dentry;
+
+	new_dentry->d_fsdata = NULL;
+	wake_up_var(&new_dentry->d_fsdata);
+}
+
 /*
  * RENAME
  * FIXME: Some nfsds, like the Linux user space nfsd, may generate a
@@ -2614,8 +2641,9 @@ int nfs_rename(struct user_namespace *mnt_userns, struct inode *old_dir,
 {
 	struct inode *old_inode = d_inode(old_dentry);
 	struct inode *new_inode = d_inode(new_dentry);
-	struct dentry *dentry = NULL, *rehash = NULL;
+	struct dentry *dentry = NULL;
 	struct rpc_task *task;
+	bool must_unblock = false;
 	int error = -EBUSY;
 
 	if (flags)
@@ -2633,18 +2661,27 @@ int nfs_rename(struct user_namespace *mnt_userns, struct inode *old_dir,
 	 * the new target.
 	 */
 	if (new_inode && !S_ISDIR(new_inode->i_mode)) {
-		/*
+		/* We must prevent any concurrent open until the unlink
-		 * To prevent any new references to the target during the
+		 * completes.  ->d_revalidate will wait for ->d_fsdata
-		 * rename, we unhash the dentry in advance.
+		 * to clear.  We set it here to ensure no lookup succeeds until
+		 * the unlink is complete on the server.
 		 */
-		if (!d_unhashed(new_dentry)) {
+		error = -ETXTBSY;
-			d_drop(new_dentry);
+		if (WARN_ON(new_dentry->d_flags & DCACHE_NFSFS_RENAMED) ||
-			rehash = new_dentry;
+		    WARN_ON(new_dentry->d_fsdata == NFS_FSDATA_BLOCKED))
+			goto out;
+		if (new_dentry->d_fsdata) {
+			/* old devname */
+			kfree(new_dentry->d_fsdata);
+			new_dentry->d_fsdata = NULL;
 		}
 
+		spin_lock(&new_dentry->d_lock);
 		if (d_count(new_dentry) > 2) {
 			int err;
 
+			spin_unlock(&new_dentry->d_lock);
+
 			/* copy the target dentry's name */
 			dentry = d_alloc(new_dentry->d_parent,
 					 &new_dentry->d_name);
@@ -2657,14 +2694,19 @@ int nfs_rename(struct user_namespace *mnt_userns, struct inode *old_dir,
 				goto out;
 
 			new_dentry = dentry;
-			rehash = NULL;
 			new_inode = NULL;
+		} else {
+			new_dentry->d_fsdata = NFS_FSDATA_BLOCKED;
+			must_unblock = true;
+			spin_unlock(&new_dentry->d_lock);
 		}
+
 	}
 
 	if (S_ISREG(old_inode->i_mode))
 		nfs_sync_inode(old_inode);
-	task = nfs_async_rename(old_dir, new_dir, old_dentry, new_dentry, NULL);
+	task = nfs_async_rename(old_dir, new_dir, old_dentry, new_dentry,
+				must_unblock ? nfs_unblock_rename : NULL);
 	if (IS_ERR(task)) {
 		error = PTR_ERR(task);
 		goto out;
@@ -2688,8 +2730,6 @@ int nfs_rename(struct user_namespace *mnt_userns, struct inode *old_dir,
 		spin_unlock(&old_inode->i_lock);
 	}
 out:
-	if (rehash)
-		d_rehash(rehash);
 	trace_nfs_rename_exit(old_dir, old_dentry,
 			new_dir, new_dentry, error);
 	if (!error) {

fs/nfs/direct.c

@@ -60,44 +60,12 @@
 #include "iostat.h"
 #include "pnfs.h"
 #include "fscache.h"
+#include "nfstrace.h"
 
 #define NFSDBG_FACILITY		NFSDBG_VFS
 
 static struct kmem_cache *nfs_direct_cachep;
 
-struct nfs_direct_req {
-	struct kref		kref;		/* release manager */
-
-	/* I/O parameters */
-	struct nfs_open_context	*ctx;		/* file open context info */
-	struct nfs_lock_context *l_ctx;		/* Lock context info */
-	struct kiocb *		iocb;		/* controlling i/o request */
-	struct inode *		inode;		/* target file of i/o */
-
-	/* completion state */
-	atomic_t		io_count;	/* i/os we're waiting for */
-	spinlock_t		lock;		/* protect completion state */
-
-	loff_t			io_start;	/* Start offset for I/O */
-	ssize_t			count,		/* bytes actually processed */
-				max_count,	/* max expected count */
-				bytes_left,	/* bytes left to be sent */
-				error;		/* any reported error */
-	struct completion	completion;	/* wait for i/o completion */
-
-	/* commit state */
-	struct nfs_mds_commit_info mds_cinfo;	/* Storage for cinfo */
-	struct pnfs_ds_commit_info ds_cinfo;	/* Storage for cinfo */
-	struct work_struct	work;
-	int			flags;
-	/* for write */
-#define NFS_ODIRECT_DO_COMMIT		(1)	/* an unstable reply was received */
-#define NFS_ODIRECT_RESCHED_WRITES	(2)	/* write verification failed */
-	/* for read */
-#define NFS_ODIRECT_SHOULD_DIRTY	(3)	/* dirty user-space page after read */
-#define NFS_ODIRECT_DONE		INT_MAX	/* write verification failed */
-};
-
 static const struct nfs_pgio_completion_ops nfs_direct_write_completion_ops;
 static const struct nfs_commit_completion_ops nfs_direct_commit_completion_ops;
 static void nfs_direct_write_complete(struct nfs_direct_req *dreq);
@@ -594,14 +562,17 @@ static void nfs_direct_commit_complete(struct nfs_commit_data *data)
 	struct nfs_page *req;
 	int status = data->task.tk_status;
 
+	trace_nfs_direct_commit_complete(dreq);
+
 	if (status < 0) {
 		/* Errors in commit are fatal */
 		dreq->error = status;
 		dreq->max_count = 0;
 		dreq->count = 0;
 		dreq->flags = NFS_ODIRECT_DONE;
-	} else if (dreq->flags == NFS_ODIRECT_DONE)
+	} else {
 		status = dreq->error;
+	}
 
 	nfs_init_cinfo_from_dreq(&cinfo, dreq);
 
@@ -630,6 +601,8 @@ static void nfs_direct_resched_write(struct nfs_commit_info *cinfo,
 {
 	struct nfs_direct_req *dreq = cinfo->dreq;
 
+	trace_nfs_direct_resched_write(dreq);
+
 	spin_lock(&dreq->lock);
 	if (dreq->flags != NFS_ODIRECT_DONE)
 		dreq->flags = NFS_ODIRECT_RESCHED_WRITES;
@@ -694,6 +667,7 @@ static void nfs_direct_write_schedule_work(struct work_struct *work)
 
 static void nfs_direct_write_complete(struct nfs_direct_req *dreq)
 {
+	trace_nfs_direct_write_complete(dreq);
 	queue_work(nfsiod_workqueue, &dreq->work); /* Calls nfs_direct_write_schedule_work */
 }
 
@@ -704,6 +678,8 @@ static void nfs_direct_write_completion(struct nfs_pgio_header *hdr)
 	struct nfs_page *req = nfs_list_entry(hdr->pages.next);
 	int flags = NFS_ODIRECT_DONE;
 
+	trace_nfs_direct_write_completion(dreq);
+
 	nfs_init_cinfo_from_dreq(&cinfo, dreq);
 
 	spin_lock(&dreq->lock);
@@ -713,7 +689,7 @@ static void nfs_direct_write_completion(struct nfs_pgio_header *hdr)
 	}
 
 	nfs_direct_count_bytes(dreq, hdr);
-	if (hdr->good_bytes != 0 && nfs_write_need_commit(hdr)) {
+	if (test_bit(NFS_IOHDR_UNSTABLE_WRITES, &hdr->flags)) {
 		if (!dreq->flags)
 			dreq->flags = NFS_ODIRECT_DO_COMMIT;
 		flags = dreq->flags;
@@ -758,6 +734,8 @@ static void nfs_direct_write_reschedule_io(struct nfs_pgio_header *hdr)
 {
 	struct nfs_direct_req *dreq = hdr->dreq;
 
+	trace_nfs_direct_write_reschedule_io(dreq);
+
 	spin_lock(&dreq->lock);
 	if (dreq->error == 0) {
 		dreq->flags = NFS_ODIRECT_RESCHED_WRITES;
@@ -798,6 +776,8 @@ static ssize_t nfs_direct_write_schedule_iovec(struct nfs_direct_req *dreq,
 	size_t requested_bytes = 0;
 	size_t wsize = max_t(size_t, NFS_SERVER(inode)->wsize, PAGE_SIZE);
 
+	trace_nfs_direct_write_schedule_iovec(dreq);
+
 	nfs_pageio_init_write(&desc, inode, ioflags, false,
 			      &nfs_direct_write_completion_ops);
 	desc.pg_dreq = dreq;

fs/nfs/file.c

@@ -661,8 +661,6 @@ ssize_t nfs_file_write(struct kiocb *iocb, struct iov_iter *from)
 		result = filemap_fdatawait_range(file->f_mapping,
 						 iocb->ki_pos - written,
 						 iocb->ki_pos - 1);
-		if (result < 0)
-			goto out;
 	}
 	result = generic_write_sync(iocb, written);
 	if (result < 0)

fs/nfs/filelayout/filelayout.c

@@ -181,6 +181,8 @@ static int filelayout_async_handle_error(struct rpc_task *task,
 	case -EIO:
 	case -ETIMEDOUT:
 	case -EPIPE:
+	case -EPROTO:
+	case -ENODEV:
 		dprintk("%s DS connection error %d\n", __func__,
 			task->tk_status);
 		nfs4_mark_deviceid_unavailable(devid);

fs/nfs/flexfilelayout/flexfilelayout.c

@@ -1131,6 +1131,8 @@ static int ff_layout_async_handle_error_v4(struct rpc_task *task,
 	case -EIO:
 	case -ETIMEDOUT:
 	case -EPIPE:
+	case -EPROTO:
+	case -ENODEV:
 		dprintk("%s DS connection error %d\n", __func__,
 			task->tk_status);
 		nfs4_delete_deviceid(devid->ld, devid->nfs_client,
@@ -1236,6 +1238,8 @@ static void ff_layout_io_track_ds_error(struct pnfs_layout_segment *lseg,
 		case -ENOBUFS:
 		case -EPIPE:
 		case -EPERM:
+		case -EPROTO:
+		case -ENODEV:
 			*op_status = status = NFS4ERR_NXIO;
 			break;
 		case -EACCES:

fs/nfs/flexfilelayout/flexfilelayoutdev.c

@@ -113,8 +113,10 @@ nfs4_ff_alloc_deviceid_node(struct nfs_server *server, struct pnfs_device *pdev,
 			goto out_err_drain_dsaddrs;
 		ds_versions[i].version = be32_to_cpup(p++);
 		ds_versions[i].minor_version = be32_to_cpup(p++);
-		ds_versions[i].rsize = nfs_block_size(be32_to_cpup(p++), NULL);
+		ds_versions[i].rsize = nfs_io_size(be32_to_cpup(p++),
-		ds_versions[i].wsize = nfs_block_size(be32_to_cpup(p++), NULL);
+						   server->nfs_client->cl_proto);
+		ds_versions[i].wsize = nfs_io_size(be32_to_cpup(p++),
+						   server->nfs_client->cl_proto);
 		ds_versions[i].tightly_coupled = be32_to_cpup(p);
 
 		if (ds_versions[i].rsize > NFS_MAX_FILE_IO_SIZE)

fs/nfs/fs_context.c

@@ -21,6 +21,8 @@
 #include "nfs.h"
 #include "internal.h"
 
+#include "nfstrace.h"
+
 #define NFSDBG_FACILITY		NFSDBG_MOUNT
 
 #if IS_ENABLED(CONFIG_NFS_V3)
@@ -284,7 +286,6 @@ static int nfs_verify_server_address(struct sockaddr *addr)
 	}
 	}
 
-	dfprintk(MOUNT, "NFS: Invalid IP address specified\n");
 	return 0;
 }
 
@@ -378,7 +379,7 @@ static int nfs_parse_security_flavors(struct fs_context *fc,
 	char *string = param->string, *p;
 	int ret;
 
-	dfprintk(MOUNT, "NFS: parsing %s=%s option\n", param->key, param->string);
+	trace_nfs_mount_assign(param->key, string);
 
 	while ((p = strsep(&string, ":")) != NULL) {
 		if (!*p)
@@ -480,11 +481,11 @@ static int nfs_fs_context_parse_param(struct fs_context *fc,
 	unsigned int len;
 	int ret, opt;
 
-	dfprintk(MOUNT, "NFS:   parsing nfs mount option '%s'\n", param->key);
+	trace_nfs_mount_option(param);
 
 	opt = fs_parse(fc, nfs_fs_parameters, param, &result);
 	if (opt < 0)
-		return ctx->sloppy ? 1 : opt;
+		return (opt == -ENOPARAM && ctx->sloppy) ? 1 : opt;
 
 	if (fc->security)
 		ctx->has_sec_mnt_opts = 1;
@@ -683,6 +684,7 @@ static int nfs_fs_context_parse_param(struct fs_context *fc,
 			return ret;
 		break;
 	case Opt_vers:
+		trace_nfs_mount_assign(param->key, param->string);
 		ret = nfs_parse_version_string(fc, param->string);
 		if (ret < 0)
 			return ret;
@@ -694,6 +696,7 @@ static int nfs_fs_context_parse_param(struct fs_context *fc,
 		break;
 
 	case Opt_proto:
+		trace_nfs_mount_assign(param->key, param->string);
 		protofamily = AF_INET;
 		switch (lookup_constant(nfs_xprt_protocol_tokens, param->string, -1)) {
 		case Opt_xprt_udp6:
@@ -729,6 +732,7 @@ static int nfs_fs_context_parse_param(struct fs_context *fc,
 		break;
 
 	case Opt_mountproto:
+		trace_nfs_mount_assign(param->key, param->string);
 		mountfamily = AF_INET;
 		switch (lookup_constant(nfs_xprt_protocol_tokens, param->string, -1)) {
 		case Opt_xprt_udp6:
@@ -751,6 +755,7 @@ static int nfs_fs_context_parse_param(struct fs_context *fc,
 		break;
 
 	case Opt_addr:
+		trace_nfs_mount_assign(param->key, param->string);
 		len = rpc_pton(fc->net_ns, param->string, param->size,
 			       &ctx->nfs_server.address,
 			       sizeof(ctx->nfs_server._address));
@@ -759,16 +764,19 @@ static int nfs_fs_context_parse_param(struct fs_context *fc,
 		ctx->nfs_server.addrlen = len;
 		break;
 	case Opt_clientaddr:
+		trace_nfs_mount_assign(param->key, param->string);
 		kfree(ctx->client_address);
 		ctx->client_address = param->string;
 		param->string = NULL;
 		break;
 	case Opt_mounthost:
+		trace_nfs_mount_assign(param->key, param->string);
 		kfree(ctx->mount_server.hostname);
 		ctx->mount_server.hostname = param->string;
 		param->string = NULL;
 		break;
 	case Opt_mountaddr:
+		trace_nfs_mount_assign(param->key, param->string);
 		len = rpc_pton(fc->net_ns, param->string, param->size,
 			       &ctx->mount_server.address,
 			       sizeof(ctx->mount_server._address));
@@ -846,7 +854,6 @@ static int nfs_fs_context_parse_param(struct fs_context *fc,
 		 */
 	case Opt_sloppy:
 		ctx->sloppy = true;
-		dfprintk(MOUNT, "NFS:   relaxing parsing rules\n");
 		break;
 	}
 
@@ -879,10 +886,8 @@ static int nfs_parse_source(struct fs_context *fc,
 	size_t len;
 	const char *end;
 
-	if (unlikely(!dev_name || !*dev_name)) {
+	if (unlikely(!dev_name || !*dev_name))
-		dfprintk(MOUNT, "NFS: device name not specified\n");
 		return -EINVAL;
-	}
 
 	/* Is the host name protected with square brakcets? */
 	if (*dev_name == '[') {
@@ -922,7 +927,7 @@ static int nfs_parse_source(struct fs_context *fc,
 	if (!ctx->nfs_server.export_path)
 		goto out_nomem;
 
-	dfprintk(MOUNT, "NFS: MNTPATH: '%s'\n", ctx->nfs_server.export_path);
+	trace_nfs_mount_path(ctx->nfs_server.export_path);
 	return 0;
 
 out_bad_devname:
@@ -1116,7 +1121,6 @@ out_no_sec:
 	return nfs_invalf(fc, "NFS: nfs_mount_data version supports only AUTH_SYS");
 
 out_nomem:
-	dfprintk(MOUNT, "NFS: not enough memory to handle mount options");
 	return -ENOMEM;
 
 out_no_address:
@@ -1248,7 +1252,7 @@ static int nfs4_parse_monolithic(struct fs_context *fc,
 	if (IS_ERR(c))
 		return PTR_ERR(c);
 	ctx->nfs_server.export_path = c;
-	dfprintk(MOUNT, "NFS: MNTPATH: '%s'\n", c);
+	trace_nfs_mount_path(c);
 
 	c = strndup_user(data->client_addr.data, 16);
 	if (IS_ERR(c))

fs/nfs/internal.h

@@ -706,6 +706,24 @@ unsigned long nfs_block_size(unsigned long bsize, unsigned char *nrbitsp)
 	return nfs_block_bits(bsize, nrbitsp);
 }
 
+/*
+ * Compute and set NFS server rsize / wsize
+ */
+static inline
+unsigned long nfs_io_size(unsigned long iosize, enum xprt_transports proto)
+{
+	if (iosize < NFS_MIN_FILE_IO_SIZE)
+		iosize = NFS_DEF_FILE_IO_SIZE;
+	else if (iosize >= NFS_MAX_FILE_IO_SIZE)
+		iosize = NFS_MAX_FILE_IO_SIZE;
+	else
+		iosize = iosize & PAGE_MASK;
+
+	if (proto == XPRT_TRANSPORT_UDP)
+		return nfs_block_bits(iosize, NULL);
+	return iosize;
+}
+
 /*
  * Determine the maximum file size for a superblock
  */
@@ -861,3 +879,36 @@ static inline void nfs_set_port(struct sockaddr *sap, int *port,
 
 	rpc_set_port(sap, *port);
 }
+
+struct nfs_direct_req {
+	struct kref		kref;		/* release manager */
+
+	/* I/O parameters */
+	struct nfs_open_context	*ctx;		/* file open context info */
+	struct nfs_lock_context *l_ctx;		/* Lock context info */
+	struct kiocb *		iocb;		/* controlling i/o request */
+	struct inode *		inode;		/* target file of i/o */
+
+	/* completion state */
+	atomic_t		io_count;	/* i/os we're waiting for */
+	spinlock_t		lock;		/* protect completion state */
+
+	loff_t			io_start;	/* Start offset for I/O */
+	ssize_t			count,		/* bytes actually processed */
+				max_count,	/* max expected count */
+				bytes_left,	/* bytes left to be sent */
+				error;		/* any reported error */
+	struct completion	completion;	/* wait for i/o completion */
+
+	/* commit state */
+	struct nfs_mds_commit_info mds_cinfo;	/* Storage for cinfo */
+	struct pnfs_ds_commit_info ds_cinfo;	/* Storage for cinfo */
+	struct work_struct	work;
+	int			flags;
+	/* for write */
+#define NFS_ODIRECT_DO_COMMIT		(1)	/* an unstable reply was received */
+#define NFS_ODIRECT_RESCHED_WRITES	(2)	/* write verification failed */
+	/* for read */
+#define NFS_ODIRECT_SHOULD_DIRTY	(3)	/* dirty user-space page after read */
+#define NFS_ODIRECT_DONE		INT_MAX	/* write verification failed */
+};

fs/nfs/nfs3client.c

@@ -108,7 +108,6 @@ struct nfs_client *nfs3_set_ds_client(struct nfs_server *mds_srv,
 	if (mds_srv->flags & NFS_MOUNT_NORESVPORT)
 		__set_bit(NFS_CS_NORESVPORT, &cl_init.init_flags);
 
-	__set_bit(NFS_CS_NOPING, &cl_init.init_flags);
 	__set_bit(NFS_CS_DS, &cl_init.init_flags);
 
 	/* Use the MDS nfs_client cl_ipaddr. */

fs/nfs/nfs42xdr.c

@@ -1025,73 +1025,84 @@ static int decode_deallocate(struct xdr_stream *xdr, struct nfs42_falloc_res *re
 	return decode_op_hdr(xdr, OP_DEALLOCATE);
 }
 
-static int decode_read_plus_data(struct xdr_stream *xdr,
+struct read_plus_segment {
-				 struct nfs_pgio_args *args,
+	enum data_content4 type;
-				 struct nfs_pgio_res *res)
-{
-	uint32_t count, recvd;
 	uint64_t offset;
+	union {
+		struct {
+			uint64_t length;
+		} hole;
+
+		struct {
+			uint32_t length;
+			unsigned int from;
+		} data;
+	};
+};
+
+static inline uint64_t read_plus_segment_length(struct read_plus_segment *seg)
+{
+	return seg->type == NFS4_CONTENT_DATA ? seg->data.length : seg->hole.length;
+}
+
+static int decode_read_plus_segment(struct xdr_stream *xdr,
+				    struct read_plus_segment *seg)
+{
 	__be32 *p;
 
-	p = xdr_inline_decode(xdr, 8 + 4);
+	p = xdr_inline_decode(xdr, 4);
 	if (!p)
-		return 1;
+		return -EIO;
+	seg->type = be32_to_cpup(p++);
 
-	p = xdr_decode_hyper(p, &offset);
+	p = xdr_inline_decode(xdr, seg->type == NFS4_CONTENT_DATA ? 12 : 16);
-	count = be32_to_cpup(p);
+	if (!p)
-	recvd = xdr_align_data(xdr, res->count, xdr_align_size(count));
+		return -EIO;
-	if (recvd > count)
+	p = xdr_decode_hyper(p, &seg->offset);
-		recvd = count;
+
-	if (res->count + recvd > args->count) {
+	if (seg->type == NFS4_CONTENT_DATA) {
-		if (args->count > res->count)
+		struct xdr_buf buf;
-			res->count += args->count - res->count;
+		uint32_t len = be32_to_cpup(p);
-		return 1;
+
-	}
+		seg->data.length = len;
-	res->count += recvd;
+		seg->data.from = xdr_stream_pos(xdr);
-	if (count > recvd)
+
-		return 1;
+		if (!xdr_stream_subsegment(xdr, &buf, xdr_align_size(len)))
+			return -EIO;
+	} else if (seg->type == NFS4_CONTENT_HOLE) {
+		xdr_decode_hyper(p, &seg->hole.length);
+	} else
+		return -EINVAL;
 	return 0;
 }
 
-static int decode_read_plus_hole(struct xdr_stream *xdr,
+static int process_read_plus_segment(struct xdr_stream *xdr,
-				 struct nfs_pgio_args *args,
+				     struct nfs_pgio_args *args,
-				 struct nfs_pgio_res *res, uint32_t *eof)
+				     struct nfs_pgio_res *res,
+				     struct read_plus_segment *seg)
 {
-	uint64_t offset, length, recvd;
+	unsigned long offset = seg->offset;
-	__be32 *p;
+	unsigned long length = read_plus_segment_length(seg);
+	unsigned int bufpos;
 
-	p = xdr_inline_decode(xdr, 8 + 8);
+	if (offset + length < args->offset)
-	if (!p)
+		return 0;
-		return 1;
+	else if (offset > args->offset + args->count) {
-
+		res->eof = 0;
-	p = xdr_decode_hyper(p, &offset);
+		return 0;
-	p = xdr_decode_hyper(p, &length);
+	} else if (offset < args->offset) {
-	if (offset != args->offset + res->count) {
+		length -= (args->offset - offset);
-		/* Server returned an out-of-sequence extent */
+		offset = args->offset;
-		if (offset > args->offset + res->count ||
+	} else if (offset + length > args->offset + args->count) {
-		    offset + length < args->offset + res->count) {
+		length = (args->offset + args->count) - offset;
-			dprintk("NFS: server returned out of sequence extent: "
+		res->eof = 0;
-				"offset/size = %llu/%llu != expected %llu\n",
-				(unsigned long long)offset,
-				(unsigned long long)length,
-				(unsigned long long)(args->offset +
-						     res->count));
-			return 1;
-		}
-		length -= args->offset + res->count - offset;
 	}
-	if (length + res->count > args->count) {
-		*eof = 0;
-		if (unlikely(res->count >= args->count))
-			return 1;
-		length = args->count - res->count;
-	}
-	recvd = xdr_expand_hole(xdr, res->count, length);
-	res->count += recvd;
 
-	if (recvd < length)
+	bufpos = xdr->buf->head[0].iov_len + (offset - args->offset);
-		return 1;
+	if (seg->type == NFS4_CONTENT_HOLE)
-	return 0;
+		return xdr_stream_zero(xdr, bufpos, length);
+	else
+		return xdr_stream_move_subsegment(xdr, seg->data.from, bufpos, length);
 }
 
 static int decode_read_plus(struct xdr_stream *xdr, struct nfs_pgio_res *res)
@@ -1099,8 +1110,10 @@ static int decode_read_plus(struct xdr_stream *xdr, struct nfs_pgio_res *res)
 	struct nfs_pgio_header *hdr =
 		container_of(res, struct nfs_pgio_header, res);
 	struct nfs_pgio_args *args = &hdr->args;
-	uint32_t eof, segments, type;
+	uint32_t segments;
+	struct read_plus_segment *segs;
 	int status, i;
+	char scratch_buf[16];
 	__be32 *p;
 
 	status = decode_op_hdr(xdr, OP_READ_PLUS);
@@ -1112,38 +1125,31 @@ static int decode_read_plus(struct xdr_stream *xdr, struct nfs_pgio_res *res)
 		return -EIO;
 
 	res->count = 0;
-	eof = be32_to_cpup(p++);
+	res->eof = be32_to_cpup(p++);
 	segments = be32_to_cpup(p++);
 	if (segments == 0)
-		goto out;
+		return status;
 
+	segs = kmalloc_array(segments, sizeof(*segs), GFP_KERNEL);
+	if (!segs)
+		return -ENOMEM;
+
+	xdr_set_scratch_buffer(xdr, &scratch_buf, 32);
+	status = -EIO;
 	for (i = 0; i < segments; i++) {
-		p = xdr_inline_decode(xdr, 4);
+		status = decode_read_plus_segment(xdr, &segs[i]);
-		if (!p)
-			goto early_out;
-
-		type = be32_to_cpup(p++);
-		if (type == NFS4_CONTENT_DATA)
-			status = decode_read_plus_data(xdr, args, res);
-		else if (type == NFS4_CONTENT_HOLE)
-			status = decode_read_plus_hole(xdr, args, res, &eof);
-		else
-			return -EINVAL;
-
 		if (status < 0)
-			return status;
+			goto out;
-		if (status > 0)
-			goto early_out;
 	}
 
+	xdr_set_pagelen(xdr, xdr_align_size(args->count));
+	for (i = segments; i > 0; i--)
+		res->count += process_read_plus_segment(xdr, args, res, &segs[i-1]);
+	status = 0;
+
 out:
-	res->eof = eof;
+	kfree(segs);
-	return 0;
+	return status;
-early_out:
-	if (unlikely(!i))
-		return -EIO;
-	res->eof = 0;
-	return 0;
 }
 
 static int decode_seek(struct xdr_stream *xdr, struct nfs42_seek_res *res)

fs/nfs/nfs4client.c

@@ -1161,9 +1161,9 @@ static int nfs4_init_server(struct nfs_server *server, struct fs_context *fc)
 		return error;
 
 	if (ctx->rsize)
-		server->rsize = nfs_block_size(ctx->rsize, NULL);
+		server->rsize = nfs_io_size(ctx->rsize, server->nfs_client->cl_proto);
 	if (ctx->wsize)
-		server->wsize = nfs_block_size(ctx->wsize, NULL);
+		server->wsize = nfs_io_size(ctx->wsize, server->nfs_client->cl_proto);
 
 	server->acregmin = ctx->acregmin * HZ;
 	server->acregmax = ctx->acregmax * HZ;

fs/nfs/nfs4idmap.c

@@ -561,22 +561,20 @@ nfs_idmap_prepare_pipe_upcall(struct idmap *idmap,
 	return true;
 }
 
-static void
+static void nfs_idmap_complete_pipe_upcall(struct idmap_legacy_upcalldata *data,
-nfs_idmap_complete_pipe_upcall_locked(struct idmap *idmap, int ret)
+					   int ret)
 {
-	struct key *authkey = idmap->idmap_upcall_data->authkey;
+	complete_request_key(data->authkey, ret);
-
+	key_put(data->authkey);
-	kfree(idmap->idmap_upcall_data);
+	kfree(data);
-	idmap->idmap_upcall_data = NULL;
-	complete_request_key(authkey, ret);
-	key_put(authkey);
 }
 
-static void
+static void nfs_idmap_abort_pipe_upcall(struct idmap *idmap,
-nfs_idmap_abort_pipe_upcall(struct idmap *idmap, int ret)
+					struct idmap_legacy_upcalldata *data,
+					int ret)
 {
-	if (idmap->idmap_upcall_data != NULL)
+	if (cmpxchg(&idmap->idmap_upcall_data, data, NULL) == data)
-		nfs_idmap_complete_pipe_upcall_locked(idmap, ret);
+		nfs_idmap_complete_pipe_upcall(data, ret);
 }
 
 static int nfs_idmap_legacy_upcall(struct key *authkey, void *aux)
@@ -613,7 +611,7 @@ static int nfs_idmap_legacy_upcall(struct key *authkey, void *aux)
 
 	ret = rpc_queue_upcall(idmap->idmap_pipe, msg);
 	if (ret < 0)
-		nfs_idmap_abort_pipe_upcall(idmap, ret);
+		nfs_idmap_abort_pipe_upcall(idmap, data, ret);
 
 	return ret;
 out2:
@@ -669,6 +667,7 @@ idmap_pipe_downcall(struct file *filp, const char __user *src, size_t mlen)
 	struct request_key_auth *rka;
 	struct rpc_inode *rpci = RPC_I(file_inode(filp));
 	struct idmap *idmap = (struct idmap *)rpci->private;
+	struct idmap_legacy_upcalldata *data;
 	struct key *authkey;
 	struct idmap_msg im;
 	size_t namelen_in;
@@ -678,10 +677,11 @@ idmap_pipe_downcall(struct file *filp, const char __user *src, size_t mlen)
 	 * will have been woken up and someone else may now have used
 	 * idmap_key_cons - so after this point we may no longer touch it.
 	 */
-	if (idmap->idmap_upcall_data == NULL)
+	data = xchg(&idmap->idmap_upcall_data, NULL);
+	if (data == NULL)
 		goto out_noupcall;
 
-	authkey = idmap->idmap_upcall_data->authkey;
+	authkey = data->authkey;
 	rka = get_request_key_auth(authkey);
 
 	if (mlen != sizeof(im)) {
@@ -703,18 +703,17 @@ idmap_pipe_downcall(struct file *filp, const char __user *src, size_t mlen)
 	if (namelen_in == 0 || namelen_in == IDMAP_NAMESZ) {
 		ret = -EINVAL;
 		goto out;
-}
+	}
 
-	ret = nfs_idmap_read_and_verify_message(&im,
+	ret = nfs_idmap_read_and_verify_message(&im, &data->idmap_msg,
-			&idmap->idmap_upcall_data->idmap_msg,
+						rka->target_key, authkey);
-			rka->target_key, authkey);
 	if (ret >= 0) {
 		key_set_timeout(rka->target_key, nfs_idmap_cache_timeout);
 		ret = mlen;
 	}
 
 out:
-	nfs_idmap_complete_pipe_upcall_locked(idmap, ret);
+	nfs_idmap_complete_pipe_upcall(data, ret);
 out_noupcall:
 	return ret;
 }
@@ -728,7 +727,7 @@ idmap_pipe_destroy_msg(struct rpc_pipe_msg *msg)
 	struct idmap *idmap = data->idmap;
 
 	if (msg->errno)
-		nfs_idmap_abort_pipe_upcall(idmap, msg->errno);
+		nfs_idmap_abort_pipe_upcall(idmap, data, msg->errno);
 }
 
 static void
@@ -736,8 +735,11 @@ idmap_release_pipe(struct inode *inode)
 {
 	struct rpc_inode *rpci = RPC_I(inode);
 	struct idmap *idmap = (struct idmap *)rpci->private;
+	struct idmap_legacy_upcalldata *data;
 
-	nfs_idmap_abort_pipe_upcall(idmap, -EPIPE);
+	data = xchg(&idmap->idmap_upcall_data, NULL);
+	if (data)
+		nfs_idmap_complete_pipe_upcall(data, -EPIPE);
 }
 
 int nfs_map_name_to_uid(const struct nfs_server *server, const char *name, size_t namelen, kuid_t *uid)

fs/nfs/nfs4proc.c

@@ -784,10 +784,9 @@ static void nfs4_slot_sequence_record_sent(struct nfs4_slot *slot,
 	if ((s32)(seqnr - slot->seq_nr_highest_sent) > 0)
 		slot->seq_nr_highest_sent = seqnr;
 }
-static void nfs4_slot_sequence_acked(struct nfs4_slot *slot,
+static void nfs4_slot_sequence_acked(struct nfs4_slot *slot, u32 seqnr)
-		u32 seqnr)
 {
-	slot->seq_nr_highest_sent = seqnr;
+	nfs4_slot_sequence_record_sent(slot, seqnr);
 	slot->seq_nr_last_acked = seqnr;
 }
 
@@ -854,7 +853,6 @@ static int nfs41_sequence_process(struct rpc_task *task,
 			__func__,
 			slot->slot_nr,
 			slot->seq_nr);
-		nfs4_slot_sequence_acked(slot, slot->seq_nr);
 		goto out_retry;
 	case -NFS4ERR_RETRY_UNCACHED_REP:
 	case -NFS4ERR_SEQ_FALSE_RETRY:
@@ -3098,12 +3096,13 @@ static int _nfs4_open_and_get_state(struct nfs4_opendata *opendata,
 	}
 
 out:
-	if (opendata->lgp) {
+	if (!opendata->cancelled) {
-		nfs4_lgopen_release(opendata->lgp);
+		if (opendata->lgp) {
-		opendata->lgp = NULL;
+			nfs4_lgopen_release(opendata->lgp);
-	}
+			opendata->lgp = NULL;
-	if (!opendata->cancelled)
+		}
 		nfs4_sequence_free_slot(&opendata->o_res.seq_res);
+	}
 	return ret;
 }
 
@@ -8924,6 +8923,9 @@ void nfs4_test_session_trunk(struct rpc_clnt *clnt, struct rpc_xprt *xprt,
 
 	if (status == 0)
 		rpc_clnt_xprt_switch_add_xprt(clnt, xprt);
+	else if (rpc_clnt_xprt_switch_has_addr(clnt,
+				(struct sockaddr *)&xprt->addr))
+		rpc_clnt_xprt_switch_remove_xprt(clnt, xprt);
 
 	rpc_put_task(task);
 }
@@ -9248,6 +9250,13 @@ int nfs4_proc_create_session(struct nfs_client *clp, const struct cred *cred)
 	int status;
 	unsigned *ptr;
 	struct nfs4_session *session = clp->cl_session;
+	struct nfs4_add_xprt_data xprtdata = {
+		.clp = clp,
+	};
+	struct rpc_add_xprt_test rpcdata = {
+		.add_xprt_test = clp->cl_mvops->session_trunk,
+		.data = &xprtdata,
+	};
 
 	dprintk("--> %s clp=%p session=%p\n", __func__, clp, session);
 
@@ -9264,6 +9273,7 @@ int nfs4_proc_create_session(struct nfs_client *clp, const struct cred *cred)
 	ptr = (unsigned *)&session->sess_id.data[0];
 	dprintk("%s client>seqid %d sessionid %u:%u:%u:%u\n", __func__,
 		clp->cl_seqid, ptr[0], ptr[1], ptr[2], ptr[3]);
+	rpc_clnt_probe_trunked_xprts(clp->cl_rpcclient, &rpcdata);
 out:
 	return status;
 }
@@ -9293,6 +9303,7 @@ int nfs4_proc_destroy_session(struct nfs4_session *session,
 	if (status)
 		dprintk("NFS: Got error %d from the server on DESTROY_SESSION. "
 			"Session has been destroyed regardless...\n", status);
+	rpc_clnt_manage_trunked_xprts(session->clp->cl_rpcclient);
 	return status;
 }
 
@@ -9477,6 +9488,9 @@ static int nfs41_reclaim_complete_handle_errors(struct rpc_task *task, struct nf
 		rpc_delay(task, NFS4_POLL_RETRY_MAX);
 		fallthrough;
 	case -NFS4ERR_RETRY_UNCACHED_REP:
+	case -EACCES:
+		dprintk("%s: failed to reclaim complete error %d for server %s, retrying\n",
+			__func__, task->tk_status, clp->cl_hostname);
 		return -EAGAIN;
 	case -NFS4ERR_BADSESSION:
 	case -NFS4ERR_DEADSESSION:

fs/nfs/nfstrace.h

@@ -1137,7 +1137,7 @@ TRACE_EVENT(nfs_readpage_done,
 			__field(u32, arg_count)
 			__field(u32, res_count)
 			__field(bool, eof)
-			__field(int, status)
+			__field(int, error)
 		),
 
 		TP_fast_assign(
@@ -1146,7 +1146,7 @@ TRACE_EVENT(nfs_readpage_done,
 			const struct nfs_fh *fh = hdr->args.fh ?
 						  hdr->args.fh : &nfsi->fh;
 
-			__entry->status = task->tk_status;
+			__entry->error = task->tk_status;
 			__entry->offset = hdr->args.offset;
 			__entry->arg_count = hdr->args.count;
 			__entry->res_count = hdr->res.count;
@@ -1157,14 +1157,13 @@ TRACE_EVENT(nfs_readpage_done,
 		),
 
 		TP_printk(
-			"fileid=%02x:%02x:%llu fhandle=0x%08x "
+			"error=%d fileid=%02x:%02x:%llu fhandle=0x%08x "
-			"offset=%lld count=%u res=%u status=%d%s",
+			"offset=%lld count=%u res=%u%s", __entry->error,
 			MAJOR(__entry->dev), MINOR(__entry->dev),
 			(unsigned long long)__entry->fileid,
 			__entry->fhandle,
 			(long long)__entry->offset, __entry->arg_count,
-			__entry->res_count, __entry->status,
+			__entry->res_count, __entry->eof ? " eof" : ""
-			__entry->eof ? " eof" : ""
 		)
 );
 
@@ -1184,7 +1183,7 @@ TRACE_EVENT(nfs_readpage_short,
 			__field(u32, arg_count)
 			__field(u32, res_count)
 			__field(bool, eof)
-			__field(int, status)
+			__field(int, error)
 		),
 
 		TP_fast_assign(
@@ -1193,7 +1192,7 @@ TRACE_EVENT(nfs_readpage_short,
 			const struct nfs_fh *fh = hdr->args.fh ?
 						  hdr->args.fh : &nfsi->fh;
 
-			__entry->status = task->tk_status;
+			__entry->error = task->tk_status;
 			__entry->offset = hdr->args.offset;
 			__entry->arg_count = hdr->args.count;
 			__entry->res_count = hdr->res.count;
@@ -1204,14 +1203,13 @@ TRACE_EVENT(nfs_readpage_short,
 		),
 
 		TP_printk(
-			"fileid=%02x:%02x:%llu fhandle=0x%08x "
+			"error=%d fileid=%02x:%02x:%llu fhandle=0x%08x "
-			"offset=%lld count=%u res=%u status=%d%s",
+			"offset=%lld count=%u res=%u%s", __entry->error,
 			MAJOR(__entry->dev), MINOR(__entry->dev),
 			(unsigned long long)__entry->fileid,
 			__entry->fhandle,
 			(long long)__entry->offset, __entry->arg_count,
-			__entry->res_count, __entry->status,
+			__entry->res_count, __entry->eof ? " eof" : ""
-			__entry->eof ? " eof" : ""
 		)
 );
 
@@ -1323,7 +1321,7 @@ TRACE_EVENT(nfs_pgio_error,
 		__field(u32, arg_count)
 		__field(u32, res_count)
 		__field(loff_t, pos)
-		__field(int, status)
+		__field(int, error)
 	),
 
 	TP_fast_assign(
@@ -1332,7 +1330,7 @@ TRACE_EVENT(nfs_pgio_error,
 		const struct nfs_fh *fh = hdr->args.fh ?
 					  hdr->args.fh : &nfsi->fh;
 
-		__entry->status = error;
+		__entry->error = error;
 		__entry->offset = hdr->args.offset;
 		__entry->arg_count = hdr->args.count;
 		__entry->res_count = hdr->res.count;
@@ -1341,12 +1339,12 @@ TRACE_EVENT(nfs_pgio_error,
 		__entry->fhandle = nfs_fhandle_hash(fh);
 	),
 
-	TP_printk("fileid=%02x:%02x:%llu fhandle=0x%08x "
+	TP_printk("error=%d fileid=%02x:%02x:%llu fhandle=0x%08x "
-		  "offset=%lld count=%u res=%u pos=%llu status=%d",
+		  "offset=%lld count=%u res=%u pos=%llu", __entry->error,
 		MAJOR(__entry->dev), MINOR(__entry->dev),
 		(unsigned long long)__entry->fileid, __entry->fhandle,
 		(long long)__entry->offset, __entry->arg_count, __entry->res_count,
-		__entry->pos, __entry->status
+		__entry->pos
 	)
 );
 
@@ -1406,7 +1404,7 @@ TRACE_EVENT(nfs_writeback_done,
 			__field(loff_t, offset)
 			__field(u32, arg_count)
 			__field(u32, res_count)
-			__field(int, status)
+			__field(int, error)
 			__field(unsigned long, stable)
 			__array(char, verifier, NFS4_VERIFIER_SIZE)
 		),
@@ -1418,7 +1416,7 @@ TRACE_EVENT(nfs_writeback_done,
 						  hdr->args.fh : &nfsi->fh;
 			const struct nfs_writeverf *verf = hdr->res.verf;
 
-			__entry->status = task->tk_status;
+			__entry->error = task->tk_status;
 			__entry->offset = hdr->args.offset;
 			__entry->arg_count = hdr->args.count;
 			__entry->res_count = hdr->res.count;
@@ -1432,14 +1430,14 @@ TRACE_EVENT(nfs_writeback_done,
 		),
 
 		TP_printk(
-			"fileid=%02x:%02x:%llu fhandle=0x%08x "
+			"error=%d fileid=%02x:%02x:%llu fhandle=0x%08x "
-			"offset=%lld count=%u res=%u status=%d stable=%s "
+			"offset=%lld count=%u res=%u stable=%s "
-			"verifier=%s",
+			"verifier=%s", __entry->error,
 			MAJOR(__entry->dev), MINOR(__entry->dev),
 			(unsigned long long)__entry->fileid,
 			__entry->fhandle,
 			(long long)__entry->offset, __entry->arg_count,
-			__entry->res_count, __entry->status,
+			__entry->res_count,
 			show_nfs_stable_how(__entry->stable),
 			show_nfs4_verifier(__entry->verifier)
 		)
@@ -1447,44 +1445,50 @@ TRACE_EVENT(nfs_writeback_done,
 
 DECLARE_EVENT_CLASS(nfs_page_error_class,
 		TP_PROTO(
+			const struct inode *inode,
 			const struct nfs_page *req,
 			int error
 		),
 
-		TP_ARGS(req, error),
+		TP_ARGS(inode, req, error),
 
 		TP_STRUCT__entry(
-			__field(const void *, req)
+			__field(dev_t, dev)
-			__field(pgoff_t, index)
+			__field(u32, fhandle)
-			__field(unsigned int, offset)
+			__field(u64, fileid)
-			__field(unsigned int, pgbase)
+			__field(loff_t, offset)
-			__field(unsigned int, bytes)
+			__field(unsigned int, count)
 			__field(int, error)
 		),
 
 		TP_fast_assign(
-			__entry->req = req;
+			const struct nfs_inode *nfsi = NFS_I(inode);
-			__entry->index = req->wb_index;
+			__entry->dev = inode->i_sb->s_dev;
-			__entry->offset = req->wb_offset;
+			__entry->fileid = nfsi->fileid;
-			__entry->pgbase = req->wb_pgbase;
+			__entry->fhandle = nfs_fhandle_hash(&nfsi->fh);
-			__entry->bytes = req->wb_bytes;
+			__entry->offset = req_offset(req);
+			__entry->count = req->wb_bytes;
 			__entry->error = error;
 		),
 
 		TP_printk(
-			"req=%p index=%lu offset=%u pgbase=%u bytes=%u error=%d",
+			"error=%d fileid=%02x:%02x:%llu fhandle=0x%08x "
-			__entry->req, __entry->index, __entry->offset,
+			"offset=%lld count=%u", __entry->error,
-			__entry->pgbase, __entry->bytes, __entry->error
+			MAJOR(__entry->dev), MINOR(__entry->dev),
+			(unsigned long long)__entry->fileid,
+			__entry->fhandle, __entry->offset,
+			__entry->count
 		)
 );
 
 #define DEFINE_NFS_PAGEERR_EVENT(name) \
 	DEFINE_EVENT(nfs_page_error_class, name, \
 			TP_PROTO( \
+				const struct inode *inode, \
 				const struct nfs_page *req, \
 				int error \
 			), \
-			TP_ARGS(req, error))
+			TP_ARGS(inode, req, error))
 
 DEFINE_NFS_PAGEERR_EVENT(nfs_write_error);
 DEFINE_NFS_PAGEERR_EVENT(nfs_comp_error);
@@ -1541,7 +1545,7 @@ TRACE_EVENT(nfs_commit_done,
 			__field(u32, fhandle)
 			__field(u64, fileid)
 			__field(loff_t, offset)
-			__field(int, status)
+			__field(int, error)
 			__field(unsigned long, stable)
 			__array(char, verifier, NFS4_VERIFIER_SIZE)
 		),
@@ -1553,7 +1557,7 @@ TRACE_EVENT(nfs_commit_done,
 						  data->args.fh : &nfsi->fh;
 			const struct nfs_writeverf *verf = data->res.verf;
 
-			__entry->status = task->tk_status;
+			__entry->error = task->tk_status;
 			__entry->offset = data->args.offset;
 			__entry->stable = verf->committed;
 			memcpy(__entry->verifier,
@@ -1565,17 +1569,83 @@ TRACE_EVENT(nfs_commit_done,
 		),
 
 		TP_printk(
-			"fileid=%02x:%02x:%llu fhandle=0x%08x "
+			"error=%d fileid=%02x:%02x:%llu fhandle=0x%08x "
-			"offset=%lld status=%d stable=%s verifier=%s",
+			"offset=%lld stable=%s verifier=%s", __entry->error,
 			MAJOR(__entry->dev), MINOR(__entry->dev),
 			(unsigned long long)__entry->fileid,
 			__entry->fhandle,
-			(long long)__entry->offset, __entry->status,
+			(long long)__entry->offset,
 			show_nfs_stable_how(__entry->stable),
 			show_nfs4_verifier(__entry->verifier)
 		)
 );
 
+#define nfs_show_direct_req_flags(v) \
+	__print_flags(v, "|", \
+			{ NFS_ODIRECT_DO_COMMIT, "DO_COMMIT" }, \
+			{ NFS_ODIRECT_RESCHED_WRITES, "RESCHED_WRITES" }, \
+			{ NFS_ODIRECT_SHOULD_DIRTY, "SHOULD DIRTY" }, \
+			{ NFS_ODIRECT_DONE, "DONE" } )
+
+DECLARE_EVENT_CLASS(nfs_direct_req_class,
+		TP_PROTO(
+			const struct nfs_direct_req *dreq
+		),
+
+		TP_ARGS(dreq),
+
+		TP_STRUCT__entry(
+			__field(dev_t, dev)
+			__field(u64, fileid)
+			__field(u32, fhandle)
+			__field(loff_t, offset)
+			__field(ssize_t, count)
+			__field(ssize_t, bytes_left)
+			__field(ssize_t, error)
+			__field(int, flags)
+		),
+
+		TP_fast_assign(
+			const struct inode *inode = dreq->inode;
+			const struct nfs_inode *nfsi = NFS_I(inode);
+			const struct nfs_fh *fh = &nfsi->fh;
+
+			__entry->dev = inode->i_sb->s_dev;
+			__entry->fileid = nfsi->fileid;
+			__entry->fhandle = nfs_fhandle_hash(fh);
+			__entry->offset = dreq->io_start;
+			__entry->count = dreq->count;
+			__entry->bytes_left = dreq->bytes_left;
+			__entry->error = dreq->error;
+			__entry->flags = dreq->flags;
+		),
+
+		TP_printk(
+			"error=%zd fileid=%02x:%02x:%llu fhandle=0x%08x "
+			"offset=%lld count=%zd bytes_left=%zd flags=%s",
+			__entry->error, MAJOR(__entry->dev),
+			MINOR(__entry->dev),
+			(unsigned long long)__entry->fileid,
+			__entry->fhandle, __entry->offset,
+			__entry->count, __entry->bytes_left,
+			nfs_show_direct_req_flags(__entry->flags)
+		)
+);
+
+#define DEFINE_NFS_DIRECT_REQ_EVENT(name) \
+	DEFINE_EVENT(nfs_direct_req_class, name, \
+			TP_PROTO( \
+				const struct nfs_direct_req *dreq \
+			), \
+			TP_ARGS(dreq))
+
+DEFINE_NFS_DIRECT_REQ_EVENT(nfs_direct_commit_complete);
+DEFINE_NFS_DIRECT_REQ_EVENT(nfs_direct_resched_write);
+DEFINE_NFS_DIRECT_REQ_EVENT(nfs_direct_write_complete);
+DEFINE_NFS_DIRECT_REQ_EVENT(nfs_direct_write_completion);
+DEFINE_NFS_DIRECT_REQ_EVENT(nfs_direct_write_schedule_iovec);
+DEFINE_NFS_DIRECT_REQ_EVENT(nfs_direct_write_reschedule_io);
+
 TRACE_EVENT(nfs_fh_to_dentry,
 		TP_PROTO(
 			const struct super_block *sb,
@@ -1609,6 +1679,65 @@ TRACE_EVENT(nfs_fh_to_dentry,
 		)
 );
 
+TRACE_EVENT(nfs_mount_assign,
+	TP_PROTO(
+		const char *option,
+		const char *value
+	),
+
+	TP_ARGS(option, value),
+
+	TP_STRUCT__entry(
+		__string(option, option)
+		__string(value, value)
+	),
+
+	TP_fast_assign(
+		__assign_str(option, option);
+		__assign_str(value, value);
+	),
+
+	TP_printk("option %s=%s",
+		__get_str(option), __get_str(value)
+	)
+);
+
+TRACE_EVENT(nfs_mount_option,
+	TP_PROTO(
+		const struct fs_parameter *param
+	),
+
+	TP_ARGS(param),
+
+	TP_STRUCT__entry(
+		__string(option, param->key)
+	),
+
+	TP_fast_assign(
+		__assign_str(option, param->key);
+	),
+
+	TP_printk("option %s", __get_str(option))
+);
+
+TRACE_EVENT(nfs_mount_path,
+	TP_PROTO(
+		const char *path
+	),
+
+	TP_ARGS(path),
+
+	TP_STRUCT__entry(
+		__string(path, path)
+	),
+
+	TP_fast_assign(
+		__assign_str(path, path);
+	),
+
+	TP_printk("path='%s'", __get_str(path))
+);
+
 DECLARE_EVENT_CLASS(nfs_xdr_event,
 		TP_PROTO(
 			const struct xdr_stream *xdr,

fs/nfs/write.c

@@ -592,7 +592,8 @@ nfs_lock_and_join_requests(struct page *page)
 
 static void nfs_write_error(struct nfs_page *req, int error)
 {
-	trace_nfs_write_error(req, error);
+	trace_nfs_write_error(page_file_mapping(req->wb_page)->host, req,
+			      error);
 	nfs_mapping_set_error(req->wb_page, error);
 	nfs_inode_remove_request(req);
 	nfs_end_page_writeback(req);
@@ -1000,7 +1001,7 @@ static void nfs_write_completion(struct nfs_pgio_header *hdr)
 		nfs_list_remove_request(req);
 		if (test_bit(NFS_IOHDR_ERROR, &hdr->flags) &&
 		    (hdr->good_bytes < bytes)) {
-			trace_nfs_comp_error(req, hdr->error);
+			trace_nfs_comp_error(hdr->inode, req, hdr->error);
 			nfs_mapping_set_error(req->wb_page, hdr->error);
 			goto remove_req;
 		}
@@ -1444,8 +1445,6 @@ static void nfs_async_write_error(struct list_head *head, int error)
 static void nfs_async_write_reschedule_io(struct nfs_pgio_header *hdr)
 {
 	nfs_async_write_error(&hdr->pages, 0);
-	filemap_fdatawrite_range(hdr->inode->i_mapping, hdr->args.offset,
-			hdr->args.offset + hdr->args.count - 1);
 }
 
 static const struct nfs_pgio_completion_ops nfs_async_write_completion_ops = {
@@ -1576,25 +1575,37 @@ static int nfs_writeback_done(struct rpc_task *task,
 	nfs_add_stats(inode, NFSIOS_SERVERWRITTENBYTES, hdr->res.count);
 	trace_nfs_writeback_done(task, hdr);
 
-	if (hdr->res.verf->committed < hdr->args.stable &&
+	if (task->tk_status >= 0) {
-	    task->tk_status >= 0) {
+		enum nfs3_stable_how committed = hdr->res.verf->committed;
-		/* We tried a write call, but the server did not
-		 * commit data to stable storage even though we
-		 * requested it.
-		 * Note: There is a known bug in Tru64 < 5.0 in which
-		 *	 the server reports NFS_DATA_SYNC, but performs
-		 *	 NFS_FILE_SYNC. We therefore implement this checking
-		 *	 as a dprintk() in order to avoid filling syslog.
-		 */
-		static unsigned long    complain;
 
-		/* Note this will print the MDS for a DS write */
+		if (committed == NFS_UNSTABLE) {
-		if (time_before(complain, jiffies)) {
+			/*
-			dprintk("NFS:       faulty NFS server %s:"
+			 * We have some uncommitted data on the server at
-				" (committed = %d) != (stable = %d)\n",
+			 * this point, so ensure that we keep track of that
-				NFS_SERVER(inode)->nfs_client->cl_hostname,
+			 * fact irrespective of what later writes do.
-				hdr->res.verf->committed, hdr->args.stable);
+			 */
-			complain = jiffies + 300 * HZ;
+			set_bit(NFS_IOHDR_UNSTABLE_WRITES, &hdr->flags);
+		}
+
+		if (committed < hdr->args.stable) {
+			/* We tried a write call, but the server did not
+			 * commit data to stable storage even though we
+			 * requested it.
+			 * Note: There is a known bug in Tru64 < 5.0 in which
+			 *	 the server reports NFS_DATA_SYNC, but performs
+			 *	 NFS_FILE_SYNC. We therefore implement this checking
+			 *	 as a dprintk() in order to avoid filling syslog.
+			 */
+			static unsigned long    complain;
+
+			/* Note this will print the MDS for a DS write */
+			if (time_before(complain, jiffies)) {
+				dprintk("NFS:       faulty NFS server %s:"
+					" (committed = %d) != (stable = %d)\n",
+					NFS_SERVER(inode)->nfs_client->cl_hostname,
+					committed, hdr->args.stable);
+				complain = jiffies + 300 * HZ;
+			}
 		}
 	}
 
@@ -1872,7 +1883,8 @@ static void nfs_commit_release_pages(struct nfs_commit_data *data)
 			(long long)req_offset(req));
 		if (status < 0) {
 			if (req->wb_page) {
-				trace_nfs_commit_error(req, status);
+				trace_nfs_commit_error(data->inode, req,
+						       status);
 				nfs_mapping_set_error(req->wb_page, status);
 				nfs_inode_remove_request(req);
 			}

include/linux/nfs_fs.h

@@ -617,6 +617,15 @@ nfs_fileid_to_ino_t(u64 fileid)
 
 #define NFS_JUKEBOX_RETRY_TIME (5 * HZ)
 
+/* We need to block new opens while a file is being unlinked.
+ * If it is opened *before* we decide to unlink, we will silly-rename
+ * instead. If it is opened *after*, then we need to create or will fail.
+ * If we allow the two to race, we could end up with a file that is open
+ * but deleted on the server resulting in ESTALE.
+ * So use ->d_fsdata to record when the unlink is happening
+ * and block dentry revalidation while it is set.
+ */
+#define NFS_FSDATA_BLOCKED ((void*)1)
 
 # undef ifdebug
 # ifdef NFS_DEBUG

include/linux/nfs_page.h

@@ -202,8 +202,7 @@ nfs_list_entry(struct list_head *head)
 	return list_entry(head, struct nfs_page, wb_list);
 }
 
-static inline
+static inline loff_t req_offset(const struct nfs_page *req)
-loff_t req_offset(struct nfs_page *req)
 {
 	return (((loff_t)req->wb_index) << PAGE_SHIFT) + req->wb_offset;
 }

include/linux/nfs_xdr.h

@@ -1600,6 +1600,7 @@ enum {
 	NFS_IOHDR_STAT,
 	NFS_IOHDR_RESEND_PNFS,
 	NFS_IOHDR_RESEND_MDS,
+	NFS_IOHDR_UNSTABLE_WRITES,
 };
 
 struct nfs_io_completion;

include/linux/sunrpc/clnt.h

@@ -234,13 +234,18 @@ int		rpc_clnt_setup_test_and_add_xprt(struct rpc_clnt *,
 			struct rpc_xprt_switch *,
 			struct rpc_xprt *,
 			void *);
+void		rpc_clnt_manage_trunked_xprts(struct rpc_clnt *);
+void		rpc_clnt_probe_trunked_xprts(struct rpc_clnt *,
+			struct rpc_add_xprt_test *);
 
 const char *rpc_proc_name(const struct rpc_task *task);
 
 void rpc_clnt_xprt_switch_put(struct rpc_clnt *);
 void rpc_clnt_xprt_switch_add_xprt(struct rpc_clnt *, struct rpc_xprt *);
+void rpc_clnt_xprt_switch_remove_xprt(struct rpc_clnt *, struct rpc_xprt *);
 bool rpc_clnt_xprt_switch_has_addr(struct rpc_clnt *clnt,
 			const struct sockaddr *sap);
+void rpc_clnt_xprt_set_online(struct rpc_clnt *clnt, struct rpc_xprt *xprt);
 void rpc_cleanup_clids(void);
 
 static inline int rpc_reply_expected(struct rpc_task *task)

include/linux/sunrpc/sched.h

@@ -61,8 +61,6 @@ struct rpc_task {
 		struct rpc_wait		tk_wait;	/* RPC wait */
 	} u;
 
-	int			tk_rpc_status;	/* Result of last RPC operation */
-
 	/*
 	 * RPC call state
 	 */
@@ -82,6 +80,8 @@ struct rpc_task {
 	ktime_t			tk_start;	/* RPC task init timestamp */
 
 	pid_t			tk_owner;	/* Process id for batching tasks */
+
+	int			tk_rpc_status;	/* Result of last RPC operation */
 	unsigned short		tk_flags;	/* misc flags */
 	unsigned short		tk_timeouts;	/* maj timeouts */
 

include/linux/sunrpc/xdr.h

@@ -258,10 +258,13 @@ extern __be32 *xdr_inline_decode(struct xdr_stream *xdr, size_t nbytes);
 extern unsigned int xdr_read_pages(struct xdr_stream *xdr, unsigned int len);
 extern void xdr_enter_page(struct xdr_stream *xdr, unsigned int len);
 extern int xdr_process_buf(const struct xdr_buf *buf, unsigned int offset, unsigned int len, int (*actor)(struct scatterlist *, void *), void *data);
-extern unsigned int xdr_align_data(struct xdr_stream *, unsigned int offset, unsigned int length);
+extern void xdr_set_pagelen(struct xdr_stream *, unsigned int len);
-extern unsigned int xdr_expand_hole(struct xdr_stream *, unsigned int offset, unsigned int length);
 extern bool xdr_stream_subsegment(struct xdr_stream *xdr, struct xdr_buf *subbuf,
 				  unsigned int len);
+extern unsigned int xdr_stream_move_subsegment(struct xdr_stream *xdr, unsigned int offset,
+					       unsigned int target, unsigned int length);
+extern unsigned int xdr_stream_zero(struct xdr_stream *xdr, unsigned int offset,
+				    unsigned int length);
 
 /**
  * xdr_set_scratch_buffer - Attach a scratch buffer for decoding data.

include/linux/sunrpc/xprt.h

@@ -144,7 +144,8 @@ struct rpc_xprt_ops {
 	unsigned short	(*get_srcport)(struct rpc_xprt *xprt);
 	int		(*buf_alloc)(struct rpc_task *task);
 	void		(*buf_free)(struct rpc_task *task);
-	int		(*prepare_request)(struct rpc_rqst *req);
+	int		(*prepare_request)(struct rpc_rqst *req,
+					   struct xdr_buf *buf);
 	int		(*send_request)(struct rpc_rqst *req);
 	void		(*wait_for_reply_request)(struct rpc_task *task);
 	void		(*timer)(struct rpc_xprt *xprt, struct rpc_task *task);
@@ -505,4 +506,7 @@ static inline int xprt_test_and_set_binding(struct rpc_xprt *xprt)
 	return test_and_set_bit(XPRT_BINDING, &xprt->state);
 }
 
+void xprt_set_offline_locked(struct rpc_xprt *xprt, struct rpc_xprt_switch *xps);
+void xprt_set_online_locked(struct rpc_xprt *xprt, struct rpc_xprt_switch *xps);
+void xprt_delete_locked(struct rpc_xprt *xprt, struct rpc_xprt_switch *xps);
 #endif /* _LINUX_SUNRPC_XPRT_H */

include/linux/sunrpc/xprtmultipath.h

@@ -55,7 +55,7 @@ extern void rpc_xprt_switch_set_roundrobin(struct rpc_xprt_switch *xps);
 extern void rpc_xprt_switch_add_xprt(struct rpc_xprt_switch *xps,
 		struct rpc_xprt *xprt);
 extern void rpc_xprt_switch_remove_xprt(struct rpc_xprt_switch *xps,
-		struct rpc_xprt *xprt);
+		struct rpc_xprt *xprt, bool offline);
 
 extern void xprt_iter_init(struct rpc_xprt_iter *xpi,
 		struct rpc_xprt_switch *xps);
@@ -63,8 +63,13 @@ extern void xprt_iter_init(struct rpc_xprt_iter *xpi,
 extern void xprt_iter_init_listall(struct rpc_xprt_iter *xpi,
 		struct rpc_xprt_switch *xps);
 
+extern void xprt_iter_init_listoffline(struct rpc_xprt_iter *xpi,
+		struct rpc_xprt_switch *xps);
+
 extern void xprt_iter_destroy(struct rpc_xprt_iter *xpi);
 
+extern void xprt_iter_rewind(struct rpc_xprt_iter *xpi);
+
 extern struct rpc_xprt_switch *xprt_iter_xchg_switch(
 		struct rpc_xprt_iter *xpi,
 		struct rpc_xprt_switch *newswitch);

include/trace/events/sunrpc.h

@@ -1266,6 +1266,26 @@ TRACE_EVENT(xprt_reserve,
 	)
 );
 
+TRACE_EVENT(xs_data_ready,
+	TP_PROTO(
+		const struct rpc_xprt *xprt
+	),
+
+	TP_ARGS(xprt),
+
+	TP_STRUCT__entry(
+		__string(addr, xprt->address_strings[RPC_DISPLAY_ADDR])
+		__string(port, xprt->address_strings[RPC_DISPLAY_PORT])
+	),
+
+	TP_fast_assign(
+		__assign_str(addr, xprt->address_strings[RPC_DISPLAY_ADDR]);
+		__assign_str(port, xprt->address_strings[RPC_DISPLAY_PORT]);
+	),
+
+	TP_printk("peer=[%s]:%s", __get_str(addr), __get_str(port))
+);
+
 TRACE_EVENT(xs_stream_read_data,
 	TP_PROTO(struct rpc_xprt *xprt, ssize_t err, size_t total),
 

net/sunrpc/auth.c

@@ -445,7 +445,7 @@ rpcauth_prune_expired(struct list_head *free, int nr_to_scan)
 		 * Enforce a 60 second garbage collection moratorium
 		 * Note that the cred_unused list must be time-ordered.
 		 */
-		if (!time_in_range(cred->cr_expire, expired, jiffies))
+		if (time_in_range(cred->cr_expire, expired, jiffies))
 			continue;
 		if (!rpcauth_unhash_cred(cred))
 			continue;

net/sunrpc/auth_gss/auth_gss.c

@@ -1340,14 +1340,11 @@ gss_hash_cred(struct auth_cred *acred, unsigned int hashbits)
 /*
  * Lookup RPCSEC_GSS cred for the current process
  */
-static struct rpc_cred *
+static struct rpc_cred *gss_lookup_cred(struct rpc_auth *auth,
-gss_lookup_cred(struct rpc_auth *auth, struct auth_cred *acred, int flags)
+					struct auth_cred *acred, int flags)
 {
-	gfp_t gfp = GFP_KERNEL;
+	return rpcauth_lookup_credcache(auth, acred, flags,
-
+					rpc_task_gfp_mask());
-	if (flags & RPCAUTH_LOOKUP_ASYNC)
-		gfp = GFP_NOWAIT | __GFP_NOWARN;
-	return rpcauth_lookup_credcache(auth, acred, flags, gfp);
 }
 
 static struct rpc_cred *

net/sunrpc/backchannel_rqst.c

@@ -50,6 +50,17 @@ static void xprt_free_allocation(struct rpc_rqst *req)
 	kfree(req);
 }
 
+static void xprt_bc_reinit_xdr_buf(struct xdr_buf *buf)
+{
+	buf->head[0].iov_len = PAGE_SIZE;
+	buf->tail[0].iov_len = 0;
+	buf->pages = NULL;
+	buf->page_len = 0;
+	buf->flags = 0;
+	buf->len = 0;
+	buf->buflen = PAGE_SIZE;
+}
+
 static int xprt_alloc_xdr_buf(struct xdr_buf *buf, gfp_t gfp_flags)
 {
 	struct page *page;
@@ -278,6 +289,9 @@ void xprt_free_bc_rqst(struct rpc_rqst *req)
 	 */
 	spin_lock_bh(&xprt->bc_pa_lock);
 	if (xprt_need_to_requeue(xprt)) {
+		xprt_bc_reinit_xdr_buf(&req->rq_snd_buf);
+		xprt_bc_reinit_xdr_buf(&req->rq_rcv_buf);
+		req->rq_rcv_buf.len = PAGE_SIZE;
 		list_add_tail(&req->rq_bc_pa_list, &xprt->bc_pa_list);
 		xprt->bc_alloc_count++;
 		atomic_inc(&xprt->bc_slot_count);

net/sunrpc/clnt.c

@@ -786,7 +786,8 @@ out_revert:
 EXPORT_SYMBOL_GPL(rpc_switch_client_transport);
 
 static
-int rpc_clnt_xprt_iter_init(struct rpc_clnt *clnt, struct rpc_xprt_iter *xpi)
+int _rpc_clnt_xprt_iter_init(struct rpc_clnt *clnt, struct rpc_xprt_iter *xpi,
+			     void func(struct rpc_xprt_iter *xpi, struct rpc_xprt_switch *xps))
 {
 	struct rpc_xprt_switch *xps;
 
@@ -795,11 +796,24 @@ int rpc_clnt_xprt_iter_init(struct rpc_clnt *clnt, struct rpc_xprt_iter *xpi)
 	rcu_read_unlock();
 	if (xps == NULL)
 		return -EAGAIN;
-	xprt_iter_init_listall(xpi, xps);
+	func(xpi, xps);
 	xprt_switch_put(xps);
 	return 0;
 }
 
+static
+int rpc_clnt_xprt_iter_init(struct rpc_clnt *clnt, struct rpc_xprt_iter *xpi)
+{
+	return _rpc_clnt_xprt_iter_init(clnt, xpi, xprt_iter_init_listall);
+}
+
+static
+int rpc_clnt_xprt_iter_offline_init(struct rpc_clnt *clnt,
+				    struct rpc_xprt_iter *xpi)
+{
+	return _rpc_clnt_xprt_iter_init(clnt, xpi, xprt_iter_init_listoffline);
+}
+
 /**
  * rpc_clnt_iterate_for_each_xprt - Apply a function to all transports
  * @clnt: pointer to client
@@ -1856,7 +1870,6 @@ rpc_xdr_encode(struct rpc_task *task)
 	req->rq_snd_buf.head[0].iov_len = 0;
 	xdr_init_encode(&xdr, &req->rq_snd_buf,
 			req->rq_snd_buf.head[0].iov_base, req);
-	xdr_free_bvec(&req->rq_snd_buf);
 	if (rpc_encode_header(task, &xdr))
 		return;
 
@@ -2137,7 +2150,8 @@ call_connect_status(struct rpc_task *task)
 				xprt_release(task);
 				value = atomic_long_dec_return(&xprt->queuelen);
 				if (value == 0)
-					rpc_xprt_switch_remove_xprt(xps, saved);
+					rpc_xprt_switch_remove_xprt(xps, saved,
+								    true);
 				xprt_put(saved);
 				task->tk_xprt = NULL;
 				task->tk_action = call_start;
@@ -2650,7 +2664,7 @@ out_unparsable:
 
 out_verifier:
 	trace_rpc_bad_verifier(task);
-	goto out_garbage;
+	goto out_err;
 
 out_msg_denied:
 	error = -EACCES;
@@ -2866,6 +2880,30 @@ success:
 }
 EXPORT_SYMBOL_GPL(rpc_clnt_test_and_add_xprt);
 
+static int rpc_clnt_add_xprt_helper(struct rpc_clnt *clnt,
+				    struct rpc_xprt *xprt,
+				    struct rpc_add_xprt_test *data)
+{
+	struct rpc_task *task;
+	int status = -EADDRINUSE;
+
+	/* Test the connection */
+	task = rpc_call_null_helper(clnt, xprt, NULL, 0, NULL, NULL);
+	if (IS_ERR(task))
+		return PTR_ERR(task);
+
+	status = task->tk_status;
+	rpc_put_task(task);
+
+	if (status < 0)
+		return status;
+
+	/* rpc_xprt_switch and rpc_xprt are deferrenced by add_xprt_test() */
+	data->add_xprt_test(clnt, xprt, data->data);
+
+	return 0;
+}
+
 /**
  * rpc_clnt_setup_test_and_add_xprt()
  *
@@ -2889,8 +2927,6 @@ int rpc_clnt_setup_test_and_add_xprt(struct rpc_clnt *clnt,
 				     struct rpc_xprt *xprt,
 				     void *data)
 {
-	struct rpc_task *task;
-	struct rpc_add_xprt_test *xtest = (struct rpc_add_xprt_test *)data;
 	int status = -EADDRINUSE;
 
 	xprt = xprt_get(xprt);
@@ -2899,31 +2935,19 @@ int rpc_clnt_setup_test_and_add_xprt(struct rpc_clnt *clnt,
 	if (rpc_xprt_switch_has_addr(xps, (struct sockaddr *)&xprt->addr))
 		goto out_err;
 
-	/* Test the connection */
+	status = rpc_clnt_add_xprt_helper(clnt, xprt, data);
-	task = rpc_call_null_helper(clnt, xprt, NULL, 0, NULL, NULL);
-	if (IS_ERR(task)) {
-		status = PTR_ERR(task);
-		goto out_err;
-	}
-	status = task->tk_status;
-	rpc_put_task(task);
-
 	if (status < 0)
 		goto out_err;
 
-	/* rpc_xprt_switch and rpc_xprt are deferrenced by add_xprt_test() */
+	status = 1;
-	xtest->add_xprt_test(clnt, xprt, xtest->data);
-
-	xprt_put(xprt);
-	xprt_switch_put(xps);
-
-	/* so that rpc_clnt_add_xprt does not call rpc_xprt_switch_add_xprt */
-	return 1;
 out_err:
 	xprt_put(xprt);
 	xprt_switch_put(xps);
-	pr_info("RPC:   rpc_clnt_test_xprt failed: %d addr %s not added\n",
+	if (status < 0)
-		status, xprt->address_strings[RPC_DISPLAY_ADDR]);
+		pr_info("RPC:   rpc_clnt_test_xprt failed: %d addr %s not "
+			"added\n", status,
+			xprt->address_strings[RPC_DISPLAY_ADDR]);
+	/* so that rpc_clnt_add_xprt does not call rpc_xprt_switch_add_xprt */
 	return status;
 }
 EXPORT_SYMBOL_GPL(rpc_clnt_setup_test_and_add_xprt);
@@ -3000,6 +3024,110 @@ out_put_switch:
 }
 EXPORT_SYMBOL_GPL(rpc_clnt_add_xprt);
 
+static int rpc_xprt_probe_trunked(struct rpc_clnt *clnt,
+				  struct rpc_xprt *xprt,
+				  struct rpc_add_xprt_test *data)
+{
+	struct rpc_xprt_switch *xps;
+	struct rpc_xprt *main_xprt;
+	int status = 0;
+
+	xprt_get(xprt);
+
+	rcu_read_lock();
+	main_xprt = xprt_get(rcu_dereference(clnt->cl_xprt));
+	xps = xprt_switch_get(rcu_dereference(clnt->cl_xpi.xpi_xpswitch));
+	status = rpc_cmp_addr_port((struct sockaddr *)&xprt->addr,
+				   (struct sockaddr *)&main_xprt->addr);
+	rcu_read_unlock();
+	xprt_put(main_xprt);
+	if (status || !test_bit(XPRT_OFFLINE, &xprt->state))
+		goto out;
+
+	status = rpc_clnt_add_xprt_helper(clnt, xprt, data);
+out:
+	xprt_put(xprt);
+	xprt_switch_put(xps);
+	return status;
+}
+
+/* rpc_clnt_probe_trunked_xprt -- probe offlined transport for session trunking
+ * @clnt rpc_clnt structure
+ *
+ * For each offlined transport found in the rpc_clnt structure call
+ * the function rpc_xprt_probe_trunked() which will determine if this
+ * transport still belongs to the trunking group.
+ */
+void rpc_clnt_probe_trunked_xprts(struct rpc_clnt *clnt,
+				  struct rpc_add_xprt_test *data)
+{
+	struct rpc_xprt_iter xpi;
+	int ret;
+
+	ret = rpc_clnt_xprt_iter_offline_init(clnt, &xpi);
+	if (ret)
+		return;
+	for (;;) {
+		struct rpc_xprt *xprt = xprt_iter_get_next(&xpi);
+
+		if (!xprt)
+			break;
+		ret = rpc_xprt_probe_trunked(clnt, xprt, data);
+		xprt_put(xprt);
+		if (ret < 0)
+			break;
+		xprt_iter_rewind(&xpi);
+	}
+	xprt_iter_destroy(&xpi);
+}
+EXPORT_SYMBOL_GPL(rpc_clnt_probe_trunked_xprts);
+
+static int rpc_xprt_offline(struct rpc_clnt *clnt,
+			    struct rpc_xprt *xprt,
+			    void *data)
+{
+	struct rpc_xprt *main_xprt;
+	struct rpc_xprt_switch *xps;
+	int err = 0;
+
+	xprt_get(xprt);
+
+	rcu_read_lock();
+	main_xprt = xprt_get(rcu_dereference(clnt->cl_xprt));
+	xps = xprt_switch_get(rcu_dereference(clnt->cl_xpi.xpi_xpswitch));
+	err = rpc_cmp_addr_port((struct sockaddr *)&xprt->addr,
+				(struct sockaddr *)&main_xprt->addr);
+	rcu_read_unlock();
+	xprt_put(main_xprt);
+	if (err)
+		goto out;
+
+	if (wait_on_bit_lock(&xprt->state, XPRT_LOCKED, TASK_KILLABLE)) {
+		err = -EINTR;
+		goto out;
+	}
+	xprt_set_offline_locked(xprt, xps);
+
+	xprt_release_write(xprt, NULL);
+out:
+	xprt_put(xprt);
+	xprt_switch_put(xps);
+	return err;
+}
+
+/* rpc_clnt_manage_trunked_xprts -- offline trunked transports
+ * @clnt rpc_clnt structure
+ *
+ * For each active transport found in the rpc_clnt structure call
+ * the function rpc_xprt_offline() which will identify trunked transports
+ * and will mark them offline.
+ */
+void rpc_clnt_manage_trunked_xprts(struct rpc_clnt *clnt)
+{
+	rpc_clnt_iterate_for_each_xprt(clnt, rpc_xprt_offline, NULL);
+}
+EXPORT_SYMBOL_GPL(rpc_clnt_manage_trunked_xprts);
+
 struct connect_timeout_data {
 	unsigned long connect_timeout;
 	unsigned long reconnect_timeout;
@@ -3042,8 +3170,22 @@ void rpc_clnt_xprt_switch_put(struct rpc_clnt *clnt)
 }
 EXPORT_SYMBOL_GPL(rpc_clnt_xprt_switch_put);
 
+void rpc_clnt_xprt_set_online(struct rpc_clnt *clnt, struct rpc_xprt *xprt)
+{
+	struct rpc_xprt_switch *xps;
+
+	rcu_read_lock();
+	xps = rcu_dereference(clnt->cl_xpi.xpi_xpswitch);
+	rcu_read_unlock();
+	xprt_set_online_locked(xprt, xps);
+}
+
 void rpc_clnt_xprt_switch_add_xprt(struct rpc_clnt *clnt, struct rpc_xprt *xprt)
 {
+	if (rpc_clnt_xprt_switch_has_addr(clnt,
+		(const struct sockaddr *)&xprt->addr)) {
+		return rpc_clnt_xprt_set_online(clnt, xprt);
+	}
 	rcu_read_lock();
 	rpc_xprt_switch_add_xprt(rcu_dereference(clnt->cl_xpi.xpi_xpswitch),
 				 xprt);
@@ -3051,6 +3193,19 @@ void rpc_clnt_xprt_switch_add_xprt(struct rpc_clnt *clnt, struct rpc_xprt *xprt)
 }
 EXPORT_SYMBOL_GPL(rpc_clnt_xprt_switch_add_xprt);
 
+void rpc_clnt_xprt_switch_remove_xprt(struct rpc_clnt *clnt, struct rpc_xprt *xprt)
+{
+	struct rpc_xprt_switch *xps;
+
+	rcu_read_lock();
+	xps = rcu_dereference(clnt->cl_xpi.xpi_xpswitch);
+	rpc_xprt_switch_remove_xprt(rcu_dereference(clnt->cl_xpi.xpi_xpswitch),
+				    xprt, 0);
+	xps->xps_nunique_destaddr_xprts--;
+	rcu_read_unlock();
+}
+EXPORT_SYMBOL_GPL(rpc_clnt_xprt_switch_remove_xprt);
+
 bool rpc_clnt_xprt_switch_has_addr(struct rpc_clnt *clnt,
 				   const struct sockaddr *sap)
 {

net/sunrpc/sched.c

@@ -63,6 +63,7 @@ gfp_t rpc_task_gfp_mask(void)
 		return GFP_KERNEL | __GFP_NORETRY | __GFP_NOWARN;
 	return GFP_KERNEL;
 }
+EXPORT_SYMBOL_GPL(rpc_task_gfp_mask);
 
 unsigned long
 rpc_task_timeout(const struct rpc_task *task)

net/sunrpc/sysfs.c

@@ -314,32 +314,14 @@ static ssize_t rpc_sysfs_xprt_state_change(struct kobject *kobj,
 		goto release_tasks;
 	}
 	if (offline) {
-		if (!test_and_set_bit(XPRT_OFFLINE, &xprt->state)) {
+		xprt_set_offline_locked(xprt, xps);
-			spin_lock(&xps->xps_lock);
-			xps->xps_nactive--;
-			spin_unlock(&xps->xps_lock);
-		}
 	} else if (online) {
-		if (test_and_clear_bit(XPRT_OFFLINE, &xprt->state)) {
+		xprt_set_online_locked(xprt, xps);
-			spin_lock(&xps->xps_lock);
-			xps->xps_nactive++;
-			spin_unlock(&xps->xps_lock);
-		}
 	} else if (remove) {
-		if (test_bit(XPRT_OFFLINE, &xprt->state)) {
+		if (test_bit(XPRT_OFFLINE, &xprt->state))
-			if (!test_and_set_bit(XPRT_REMOVE, &xprt->state)) {
+			xprt_delete_locked(xprt, xps);
-				xprt_force_disconnect(xprt);
+		else
-				if (test_bit(XPRT_CONNECTED, &xprt->state)) {
-					if (!xprt->sending.qlen &&
-					    !xprt->pending.qlen &&
-					    !xprt->backlog.qlen &&
-					    !atomic_long_read(&xprt->queuelen))
-						rpc_xprt_switch_remove_xprt(xps, xprt);
-				}
-			}
-		} else {
 			count = -EINVAL;
-		}
 	}
 
 release_tasks:

net/sunrpc/xdr.c

@@ -775,6 +775,34 @@ static void xdr_buf_pages_shift_left(const struct xdr_buf *buf,
 	xdr_buf_tail_copy_left(buf, 0, len - buf->page_len, shift);
 }
 
+static void xdr_buf_head_shift_left(const struct xdr_buf *buf,
+				    unsigned int base, unsigned int len,
+				    unsigned int shift)
+{
+	const struct kvec *head = buf->head;
+	unsigned int bytes;
+
+	if (!shift || !len)
+		return;
+
+	if (shift > base) {
+		bytes = (shift - base);
+		if (bytes >= len)
+			return;
+		base += bytes;
+		len -= bytes;
+	}
+
+	if (base < head->iov_len) {
+		bytes = min_t(unsigned int, len, head->iov_len - base);
+		memmove(head->iov_base + (base - shift),
+			head->iov_base + base, bytes);
+		base += bytes;
+		len -= bytes;
+	}
+	xdr_buf_pages_shift_left(buf, base - head->iov_len, len, shift);
+}
+
 /**
  * xdr_shrink_bufhead
  * @buf: xdr_buf
@@ -1472,71 +1500,35 @@ unsigned int xdr_read_pages(struct xdr_stream *xdr, unsigned int len)
 }
 EXPORT_SYMBOL_GPL(xdr_read_pages);
 
-unsigned int xdr_align_data(struct xdr_stream *xdr, unsigned int offset,
+/**
-			    unsigned int length)
+ * xdr_set_pagelen - Sets the length of the XDR pages
+ * @xdr: pointer to xdr_stream struct
+ * @len: new length of the XDR page data
+ *
+ * Either grows or shrinks the length of the xdr pages by setting pagelen to
+ * @len bytes. When shrinking, any extra data is moved into buf->tail, whereas
+ * when growing any data beyond the current pointer is moved into the tail.
+ *
+ * Returns True if the operation was successful, and False otherwise.
+ */
+void xdr_set_pagelen(struct xdr_stream *xdr, unsigned int len)
 {
 	struct xdr_buf *buf = xdr->buf;
-	unsigned int from, bytes, len;
+	size_t remaining = xdr_stream_remaining(xdr);
-	unsigned int shift;
+	size_t base = 0;
 
-	xdr_realign_pages(xdr);
+	if (len < buf->page_len) {
-	from = xdr_page_pos(xdr);
+		base = buf->page_len - len;
-
+		xdr_shrink_pagelen(buf, len);
-	if (from >= buf->page_len + buf->tail->iov_len)
+	} else {
-		return 0;
+		xdr_buf_head_shift_right(buf, xdr_stream_pos(xdr),
-	if (from + buf->head->iov_len >= buf->len)
+					 buf->page_len, remaining);
-		return 0;
+		if (len > buf->page_len)
-
+			xdr_buf_try_expand(buf, len - buf->page_len);
-	len = buf->len - buf->head->iov_len;
+	}
-
+	xdr_set_tail_base(xdr, base, remaining);
-	/* We only shift data left! */
-	if (WARN_ONCE(from < offset, "SUNRPC: misaligned data src=%u dst=%u\n",
-		      from, offset))
-		return 0;
-	if (WARN_ONCE(offset > buf->page_len,
-		      "SUNRPC: buffer overflow. offset=%u, page_len=%u\n",
-		      offset, buf->page_len))
-		return 0;
-
-	/* Move page data to the left */
-	shift = from - offset;
-	xdr_buf_pages_shift_left(buf, from, len, shift);
-
-	bytes = xdr_stream_remaining(xdr);
-	if (length > bytes)
-		length = bytes;
-	bytes -= length;
-
-	xdr->buf->len -= shift;
-	xdr_set_page(xdr, offset + length, bytes);
-	return length;
 }
-EXPORT_SYMBOL_GPL(xdr_align_data);
+EXPORT_SYMBOL_GPL(xdr_set_pagelen);
-
-unsigned int xdr_expand_hole(struct xdr_stream *xdr, unsigned int offset,
-			     unsigned int length)
-{
-	struct xdr_buf *buf = xdr->buf;
-	unsigned int from, to, shift;
-
-	xdr_realign_pages(xdr);
-	from = xdr_page_pos(xdr);
-	to = xdr_align_size(offset + length);
-
-	/* Could the hole be behind us? */
-	if (to > from) {
-		unsigned int buflen = buf->len - buf->head->iov_len;
-		shift = to - from;
-		xdr_buf_try_expand(buf, shift);
-		xdr_buf_pages_shift_right(buf, from, buflen, shift);
-		xdr_set_page(xdr, to, xdr_stream_remaining(xdr));
-	} else if (to != from)
-		xdr_align_data(xdr, to, 0);
-	xdr_buf_pages_zero(buf, offset, length);
-
-	return length;
-}
-EXPORT_SYMBOL_GPL(xdr_expand_hole);
 
 /**
  * xdr_enter_page - decode data from the XDR page
@@ -1680,6 +1672,60 @@ bool xdr_stream_subsegment(struct xdr_stream *xdr, struct xdr_buf *subbuf,
 }
 EXPORT_SYMBOL_GPL(xdr_stream_subsegment);
 
+/**
+ * xdr_stream_move_subsegment - Move part of a stream to another position
+ * @xdr: the source xdr_stream
+ * @offset: the source offset of the segment
+ * @target: the target offset of the segment
+ * @length: the number of bytes to move
+ *
+ * Moves @length bytes from @offset to @target in the xdr_stream, overwriting
+ * anything in its space. Returns the number of bytes in the segment.
+ */
+unsigned int xdr_stream_move_subsegment(struct xdr_stream *xdr, unsigned int offset,
+					unsigned int target, unsigned int length)
+{
+	struct xdr_buf buf;
+	unsigned int shift;
+
+	if (offset < target) {
+		shift = target - offset;
+		if (xdr_buf_subsegment(xdr->buf, &buf, offset, shift + length) < 0)
+			return 0;
+		xdr_buf_head_shift_right(&buf, 0, length, shift);
+	} else if (offset > target) {
+		shift = offset - target;
+		if (xdr_buf_subsegment(xdr->buf, &buf, target, shift + length) < 0)
+			return 0;
+		xdr_buf_head_shift_left(&buf, shift, length, shift);
+	}
+	return length;
+}
+EXPORT_SYMBOL_GPL(xdr_stream_move_subsegment);
+
+/**
+ * xdr_stream_zero - zero out a portion of an xdr_stream
+ * @xdr: an xdr_stream to zero out
+ * @offset: the starting point in the stream
+ * @length: the number of bytes to zero
+ */
+unsigned int xdr_stream_zero(struct xdr_stream *xdr, unsigned int offset,
+			     unsigned int length)
+{
+	struct xdr_buf buf;
+
+	if (xdr_buf_subsegment(xdr->buf, &buf, offset, length) < 0)
+		return 0;
+	if (buf.head[0].iov_len)
+		xdr_buf_iov_zero(buf.head, 0, buf.head[0].iov_len);
+	if (buf.page_len > 0)
+		xdr_buf_pages_zero(&buf, 0, buf.page_len);
+	if (buf.tail[0].iov_len)
+		xdr_buf_iov_zero(buf.tail, 0, buf.tail[0].iov_len);
+	return length;
+}
+EXPORT_SYMBOL_GPL(xdr_stream_zero);
+
 /**
  * xdr_buf_trim - lop at most "len" bytes off the end of "buf"
  * @buf: buf to be trimmed

net/sunrpc/xprt.c

@@ -73,7 +73,7 @@ static void	xprt_init(struct rpc_xprt *xprt, struct net *net);
 static __be32	xprt_alloc_xid(struct rpc_xprt *xprt);
 static void	xprt_destroy(struct rpc_xprt *xprt);
 static void	xprt_request_init(struct rpc_task *task);
-static int	xprt_request_prepare(struct rpc_rqst *req);
+static int	xprt_request_prepare(struct rpc_rqst *req, struct xdr_buf *buf);
 
 static DEFINE_SPINLOCK(xprt_list_lock);
 static LIST_HEAD(xprt_list);
@@ -1149,7 +1149,7 @@ xprt_request_enqueue_receive(struct rpc_task *task)
 	if (!xprt_request_need_enqueue_receive(task, req))
 		return 0;
 
-	ret = xprt_request_prepare(task->tk_rqstp);
+	ret = xprt_request_prepare(task->tk_rqstp, &req->rq_rcv_buf);
 	if (ret)
 		return ret;
 	spin_lock(&xprt->queue_lock);
@@ -1179,8 +1179,11 @@ xprt_request_dequeue_receive_locked(struct rpc_task *task)
 {
 	struct rpc_rqst *req = task->tk_rqstp;
 
-	if (test_and_clear_bit(RPC_TASK_NEED_RECV, &task->tk_runstate))
+	if (test_and_clear_bit(RPC_TASK_NEED_RECV, &task->tk_runstate)) {
 		xprt_request_rb_remove(req->rq_xprt, req);
+		xdr_free_bvec(&req->rq_rcv_buf);
+		req->rq_private_buf.bvec = NULL;
+	}
 }
 
 /**
@@ -1336,8 +1339,14 @@ xprt_request_enqueue_transmit(struct rpc_task *task)
 {
 	struct rpc_rqst *pos, *req = task->tk_rqstp;
 	struct rpc_xprt *xprt = req->rq_xprt;
+	int ret;
 
 	if (xprt_request_need_enqueue_transmit(task, req)) {
+		ret = xprt_request_prepare(task->tk_rqstp, &req->rq_snd_buf);
+		if (ret) {
+			task->tk_status = ret;
+			return;
+		}
 		req->rq_bytes_sent = 0;
 		spin_lock(&xprt->queue_lock);
 		/*
@@ -1397,6 +1406,7 @@ xprt_request_dequeue_transmit_locked(struct rpc_task *task)
 	} else
 		list_del(&req->rq_xmit2);
 	atomic_long_dec(&req->rq_xprt->xmit_queuelen);
+	xdr_free_bvec(&req->rq_snd_buf);
 }
 
 /**
@@ -1433,8 +1443,6 @@ xprt_request_dequeue_xprt(struct rpc_task *task)
 	    test_bit(RPC_TASK_NEED_RECV, &task->tk_runstate) ||
 	    xprt_is_pinned_rqst(req)) {
 		spin_lock(&xprt->queue_lock);
-		xprt_request_dequeue_transmit_locked(task);
-		xprt_request_dequeue_receive_locked(task);
 		while (xprt_is_pinned_rqst(req)) {
 			set_bit(RPC_TASK_MSG_PIN_WAIT, &task->tk_runstate);
 			spin_unlock(&xprt->queue_lock);
@@ -1442,6 +1450,8 @@ xprt_request_dequeue_xprt(struct rpc_task *task)
 			spin_lock(&xprt->queue_lock);
 			clear_bit(RPC_TASK_MSG_PIN_WAIT, &task->tk_runstate);
 		}
+		xprt_request_dequeue_transmit_locked(task);
+		xprt_request_dequeue_receive_locked(task);
 		spin_unlock(&xprt->queue_lock);
 	}
 }
@@ -1449,18 +1459,19 @@ xprt_request_dequeue_xprt(struct rpc_task *task)
 /**
  * xprt_request_prepare - prepare an encoded request for transport
  * @req: pointer to rpc_rqst
+ * @buf: pointer to send/rcv xdr_buf
  *
  * Calls into the transport layer to do whatever is needed to prepare
  * the request for transmission or receive.
  * Returns error, or zero.
  */
 static int
-xprt_request_prepare(struct rpc_rqst *req)
+xprt_request_prepare(struct rpc_rqst *req, struct xdr_buf *buf)
 {
 	struct rpc_xprt *xprt = req->rq_xprt;
 
 	if (xprt->ops->prepare_request)
-		return xprt->ops->prepare_request(req);
+		return xprt->ops->prepare_request(req, buf);
 	return 0;
 }
 
@@ -1961,8 +1972,6 @@ void xprt_release(struct rpc_task *task)
 	spin_unlock(&xprt->transport_lock);
 	if (req->rq_buffer)
 		xprt->ops->buf_free(task);
-	xdr_free_bvec(&req->rq_rcv_buf);
-	xdr_free_bvec(&req->rq_snd_buf);
 	if (req->rq_cred != NULL)
 		put_rpccred(req->rq_cred);
 	if (req->rq_release_snd_buf)
@@ -2152,3 +2161,35 @@ void xprt_put(struct rpc_xprt *xprt)
 		kref_put(&xprt->kref, xprt_destroy_kref);
 }
 EXPORT_SYMBOL_GPL(xprt_put);
+
+void xprt_set_offline_locked(struct rpc_xprt *xprt, struct rpc_xprt_switch *xps)
+{
+	if (!test_and_set_bit(XPRT_OFFLINE, &xprt->state)) {
+		spin_lock(&xps->xps_lock);
+		xps->xps_nactive--;
+		spin_unlock(&xps->xps_lock);
+	}
+}
+
+void xprt_set_online_locked(struct rpc_xprt *xprt, struct rpc_xprt_switch *xps)
+{
+	if (test_and_clear_bit(XPRT_OFFLINE, &xprt->state)) {
+		spin_lock(&xps->xps_lock);
+		xps->xps_nactive++;
+		spin_unlock(&xps->xps_lock);
+	}
+}
+
+void xprt_delete_locked(struct rpc_xprt *xprt, struct rpc_xprt_switch *xps)
+{
+	if (test_and_set_bit(XPRT_REMOVE, &xprt->state))
+		return;
+
+	xprt_force_disconnect(xprt);
+	if (!test_bit(XPRT_CONNECTED, &xprt->state))
+		return;
+
+	if (!xprt->sending.qlen && !xprt->pending.qlen &&
+	    !xprt->backlog.qlen && !atomic_long_read(&xprt->queuelen))
+		rpc_xprt_switch_remove_xprt(xps, xprt, true);
+}

net/sunrpc/xprtmultipath.c

@@ -27,6 +27,7 @@ typedef struct rpc_xprt *(*xprt_switch_find_xprt_t)(struct rpc_xprt_switch *xps,
 static const struct rpc_xprt_iter_ops rpc_xprt_iter_singular;
 static const struct rpc_xprt_iter_ops rpc_xprt_iter_roundrobin;
 static const struct rpc_xprt_iter_ops rpc_xprt_iter_listall;
+static const struct rpc_xprt_iter_ops rpc_xprt_iter_listoffline;
 
 static void xprt_switch_add_xprt_locked(struct rpc_xprt_switch *xps,
 		struct rpc_xprt *xprt)
@@ -61,11 +62,11 @@ void rpc_xprt_switch_add_xprt(struct rpc_xprt_switch *xps,
 }
 
 static void xprt_switch_remove_xprt_locked(struct rpc_xprt_switch *xps,
-		struct rpc_xprt *xprt)
+		struct rpc_xprt *xprt, bool offline)
 {
 	if (unlikely(xprt == NULL))
 		return;
-	if (!test_bit(XPRT_OFFLINE, &xprt->state))
+	if (!test_bit(XPRT_OFFLINE, &xprt->state) && offline)
 		xps->xps_nactive--;
 	xps->xps_nxprts--;
 	if (xps->xps_nxprts == 0)
@@ -78,14 +79,15 @@ static void xprt_switch_remove_xprt_locked(struct rpc_xprt_switch *xps,
  * rpc_xprt_switch_remove_xprt - Removes an rpc_xprt from a rpc_xprt_switch
  * @xps: pointer to struct rpc_xprt_switch
  * @xprt: pointer to struct rpc_xprt
+ * @offline: indicates if the xprt that's being removed is in an offline state
  *
  * Removes xprt from the list of struct rpc_xprt in xps.
  */
 void rpc_xprt_switch_remove_xprt(struct rpc_xprt_switch *xps,
-		struct rpc_xprt *xprt)
+		struct rpc_xprt *xprt, bool offline)
 {
 	spin_lock(&xps->xps_lock);
-	xprt_switch_remove_xprt_locked(xps, xprt);
+	xprt_switch_remove_xprt_locked(xps, xprt, offline);
 	spin_unlock(&xps->xps_lock);
 	xprt_put(xprt);
 }
@@ -154,7 +156,7 @@ static void xprt_switch_free_entries(struct rpc_xprt_switch *xps)
 
 		xprt = list_first_entry(&xps->xps_xprt_list,
 				struct rpc_xprt, xprt_switch);
-		xprt_switch_remove_xprt_locked(xps, xprt);
+		xprt_switch_remove_xprt_locked(xps, xprt, true);
 		spin_unlock(&xps->xps_lock);
 		xprt_put(xprt);
 		spin_lock(&xps->xps_lock);
@@ -248,6 +250,18 @@ struct rpc_xprt *xprt_switch_find_first_entry(struct list_head *head)
 	return NULL;
 }
 
+static
+struct rpc_xprt *xprt_switch_find_first_entry_offline(struct list_head *head)
+{
+	struct rpc_xprt *pos;
+
+	list_for_each_entry_rcu(pos, head, xprt_switch) {
+		if (!xprt_is_active(pos))
+			return pos;
+	}
+	return NULL;
+}
+
 static
 struct rpc_xprt *xprt_iter_first_entry(struct rpc_xprt_iter *xpi)
 {
@@ -259,8 +273,9 @@ struct rpc_xprt *xprt_iter_first_entry(struct rpc_xprt_iter *xpi)
 }
 
 static
-struct rpc_xprt *xprt_switch_find_current_entry(struct list_head *head,
+struct rpc_xprt *_xprt_switch_find_current_entry(struct list_head *head,
-		const struct rpc_xprt *cur)
+						 const struct rpc_xprt *cur,
+						 bool find_active)
 {
 	struct rpc_xprt *pos;
 	bool found = false;
@@ -268,14 +283,25 @@ struct rpc_xprt *xprt_switch_find_current_entry(struct list_head *head,
 	list_for_each_entry_rcu(pos, head, xprt_switch) {
 		if (cur == pos)
 			found = true;
-		if (found && xprt_is_active(pos))
+		if (found && ((find_active && xprt_is_active(pos)) ||
+			      (!find_active && xprt_is_active(pos))))
 			return pos;
 	}
 	return NULL;
 }
 
 static
-struct rpc_xprt *xprt_iter_current_entry(struct rpc_xprt_iter *xpi)
+struct rpc_xprt *xprt_switch_find_current_entry(struct list_head *head,
+						const struct rpc_xprt *cur)
+{
+	return _xprt_switch_find_current_entry(head, cur, true);
+}
+
+static
+struct rpc_xprt * _xprt_iter_current_entry(struct rpc_xprt_iter *xpi,
+		struct rpc_xprt *first_entry(struct list_head *head),
+		struct rpc_xprt *current_entry(struct list_head *head,
+					       const struct rpc_xprt *cur))
 {
 	struct rpc_xprt_switch *xps = rcu_dereference(xpi->xpi_xpswitch);
 	struct list_head *head;
@@ -284,8 +310,30 @@ struct rpc_xprt *xprt_iter_current_entry(struct rpc_xprt_iter *xpi)
 		return NULL;
 	head = &xps->xps_xprt_list;
 	if (xpi->xpi_cursor == NULL || xps->xps_nxprts < 2)
-		return xprt_switch_find_first_entry(head);
+		return first_entry(head);
-	return xprt_switch_find_current_entry(head, xpi->xpi_cursor);
+	return current_entry(head, xpi->xpi_cursor);
+}
+
+static
+struct rpc_xprt *xprt_iter_current_entry(struct rpc_xprt_iter *xpi)
+{
+	return _xprt_iter_current_entry(xpi, xprt_switch_find_first_entry,
+			xprt_switch_find_current_entry);
+}
+
+static
+struct rpc_xprt *xprt_switch_find_current_entry_offline(struct list_head *head,
+		const struct rpc_xprt *cur)
+{
+	return _xprt_switch_find_current_entry(head, cur, false);
+}
+
+static
+struct rpc_xprt *xprt_iter_current_entry_offline(struct rpc_xprt_iter *xpi)
+{
+	return _xprt_iter_current_entry(xpi,
+			xprt_switch_find_first_entry_offline,
+			xprt_switch_find_current_entry_offline);
 }
 
 bool rpc_xprt_switch_has_addr(struct rpc_xprt_switch *xps,
@@ -310,7 +358,7 @@ bool rpc_xprt_switch_has_addr(struct rpc_xprt_switch *xps,
 
 static
 struct rpc_xprt *xprt_switch_find_next_entry(struct list_head *head,
-		const struct rpc_xprt *cur)
+		const struct rpc_xprt *cur, bool check_active)
 {
 	struct rpc_xprt *pos, *prev = NULL;
 	bool found = false;
@@ -318,7 +366,12 @@ struct rpc_xprt *xprt_switch_find_next_entry(struct list_head *head,
 	list_for_each_entry_rcu(pos, head, xprt_switch) {
 		if (cur == prev)
 			found = true;
-		if (found && xprt_is_active(pos))
+		/* for request to return active transports return only
+		 * active, for request to return offline transports
+		 * return only offline
+		 */
+		if (found && ((check_active && xprt_is_active(pos)) ||
+			      (!check_active && !xprt_is_active(pos))))
 			return pos;
 		prev = pos;
 	}
@@ -355,7 +408,7 @@ struct rpc_xprt *__xprt_switch_find_next_entry_roundrobin(struct list_head *head
 {
 	struct rpc_xprt *ret;
 
-	ret = xprt_switch_find_next_entry(head, cur);
+	ret = xprt_switch_find_next_entry(head, cur, true);
 	if (ret != NULL)
 		return ret;
 	return xprt_switch_find_first_entry(head);
@@ -397,7 +450,14 @@ static
 struct rpc_xprt *xprt_switch_find_next_entry_all(struct rpc_xprt_switch *xps,
 		const struct rpc_xprt *cur)
 {
-	return xprt_switch_find_next_entry(&xps->xps_xprt_list, cur);
+	return xprt_switch_find_next_entry(&xps->xps_xprt_list, cur, true);
+}
+
+static
+struct rpc_xprt *xprt_switch_find_next_entry_offline(struct rpc_xprt_switch *xps,
+		const struct rpc_xprt *cur)
+{
+	return xprt_switch_find_next_entry(&xps->xps_xprt_list, cur, false);
 }
 
 static
@@ -407,6 +467,13 @@ struct rpc_xprt *xprt_iter_next_entry_all(struct rpc_xprt_iter *xpi)
 			xprt_switch_find_next_entry_all);
 }
 
+static
+struct rpc_xprt *xprt_iter_next_entry_offline(struct rpc_xprt_iter *xpi)
+{
+	return xprt_iter_next_entry_multiple(xpi,
+			xprt_switch_find_next_entry_offline);
+}
+
 /*
  * xprt_iter_rewind - Resets the xprt iterator
  * @xpi: pointer to rpc_xprt_iter
@@ -414,7 +481,6 @@ struct rpc_xprt *xprt_iter_next_entry_all(struct rpc_xprt_iter *xpi)
  * Resets xpi to ensure that it points to the first entry in the list
  * of transports.
  */
-static
 void xprt_iter_rewind(struct rpc_xprt_iter *xpi)
 {
 	rcu_read_lock();
@@ -460,6 +526,12 @@ void xprt_iter_init_listall(struct rpc_xprt_iter *xpi,
 	__xprt_iter_init(xpi, xps, &rpc_xprt_iter_listall);
 }
 
+void xprt_iter_init_listoffline(struct rpc_xprt_iter *xpi,
+		struct rpc_xprt_switch *xps)
+{
+	__xprt_iter_init(xpi, xps, &rpc_xprt_iter_listoffline);
+}
+
 /**
  * xprt_iter_xchg_switch - Atomically swap out the rpc_xprt_switch
  * @xpi: pointer to rpc_xprt_iter
@@ -574,3 +646,10 @@ const struct rpc_xprt_iter_ops rpc_xprt_iter_listall = {
 	.xpi_xprt = xprt_iter_current_entry,
 	.xpi_next = xprt_iter_next_entry_all,
 };
+
+static
+const struct rpc_xprt_iter_ops rpc_xprt_iter_listoffline = {
+	.xpi_rewind = xprt_iter_default_rewind,
+	.xpi_xprt = xprt_iter_current_entry_offline,
+	.xpi_next = xprt_iter_next_entry_offline,
+};

net/sunrpc/xprtrdma/transport.c

@@ -571,11 +571,7 @@ xprt_rdma_allocate(struct rpc_task *task)
 	struct rpc_rqst *rqst = task->tk_rqstp;
 	struct rpcrdma_xprt *r_xprt = rpcx_to_rdmax(rqst->rq_xprt);
 	struct rpcrdma_req *req = rpcr_to_rdmar(rqst);
-	gfp_t flags;
+	gfp_t flags = rpc_task_gfp_mask();
-
-	flags = RPCRDMA_DEF_GFP;
-	if (RPC_IS_ASYNC(task))
-		flags = GFP_NOWAIT | __GFP_NOWARN;
 
 	if (!rpcrdma_check_regbuf(r_xprt, req->rl_sendbuf, rqst->rq_callsize,
 				  flags))

net/sunrpc/xprtsock.c

@@ -822,17 +822,9 @@ static int xs_stream_nospace(struct rpc_rqst *req, bool vm_wait)
 	return ret;
 }
 
-static int
+static int xs_stream_prepare_request(struct rpc_rqst *req, struct xdr_buf *buf)
-xs_stream_prepare_request(struct rpc_rqst *req)
 {
-	gfp_t gfp = rpc_task_gfp_mask();
+	return xdr_alloc_bvec(buf, rpc_task_gfp_mask());
-	int ret;
-
-	ret = xdr_alloc_bvec(&req->rq_snd_buf, gfp);
-	if (ret < 0)
-		return ret;
-	xdr_free_bvec(&req->rq_rcv_buf);
-	return xdr_alloc_bvec(&req->rq_rcv_buf, gfp);
 }
 
 /*
@@ -1378,7 +1370,7 @@ static void xs_udp_data_receive_workfn(struct work_struct *work)
 }
 
 /**
- * xs_data_ready - "data ready" callback for UDP sockets
+ * xs_data_ready - "data ready" callback for sockets
  * @sk: socket with data to read
  *
  */
@@ -1386,11 +1378,13 @@ static void xs_data_ready(struct sock *sk)
 {
 	struct rpc_xprt *xprt;
 
-	dprintk("RPC:       xs_data_ready...\n");
 	xprt = xprt_from_sock(sk);
 	if (xprt != NULL) {
 		struct sock_xprt *transport = container_of(xprt,
 				struct sock_xprt, xprt);
+
+		trace_xs_data_ready(xprt);
+
 		transport->old_data_ready(sk);
 		/* Any data means we had a useful conversation, so
 		 * then we don't need to delay the next reconnect