Merge git://git.linux-nfs.org/projects/trondmy/nfs-2.6

* git://git.linux-nfs.org/projects/trondmy/nfs-2.6: (53 commits)
  NFS: Fix a resolution problem with nfs_inode->cache_change_attribute
  NFS: Fix the resolution problem with nfs_inode_attrs_need_update()
  NFS: Changes to inode->i_nlinks must set the NFS_INO_INVALID_ATTR flag
  RPC/RDMA: ensure connection attempt is complete before signalling.
  RPC/RDMA: correct the reconnect timer backoff
  RPC/RDMA: optionally emit useful transport info upon connect/disconnect.
  RPC/RDMA: reformat a debug printk to keep lines together.
  RPC/RDMA: harden connection logic against missing/late rdma_cm upcalls.
  RPC/RDMA: fix connect/reconnect resource leak.
  RPC/RDMA: return a consistent error, when connect fails.
  RPC/RDMA: adhere to protocol for unpadded client trailing write chunks.
  RPC/RDMA: avoid an oops due to disconnect racing with async upcalls.
  RPC/RDMA: maintain the RPC task bytes-sent statistic.
  RPC/RDMA: suppress retransmit on RPC/RDMA clients.
  RPC/RDMA: fix connection IRD/ORD setting
  RPC/RDMA: support FRMR client memory registration.
  RPC/RDMA: check selected memory registration mode at runtime.
  RPC/RDMA: add data types and new FRMR memory registration enum.
  RPC/RDMA: refactor the inline memory registration code.
  NFS: fix nfs_parse_ip_address() corner case
  ...

fs/nfs/client.c

@@ -675,7 +675,7 @@ static int nfs_init_server(struct nfs_server *server,
 	server->nfs_client = clp;
 
 	/* Initialise the client representation from the mount data */
-	server->flags = data->flags & NFS_MOUNT_FLAGMASK;
+	server->flags = data->flags;
 
 	if (data->rsize)
 		server->rsize = nfs_block_size(data->rsize, NULL);
@@ -850,7 +850,6 @@ static struct nfs_server *nfs_alloc_server(void)
 	INIT_LIST_HEAD(&server->client_link);
 	INIT_LIST_HEAD(&server->master_link);
 
-	init_waitqueue_head(&server->active_wq);
 	atomic_set(&server->active, 0);
 
 	server->io_stats = nfs_alloc_iostats();
@@ -1073,7 +1072,7 @@ static int nfs4_init_server(struct nfs_server *server,
 		goto error;
 
 	/* Initialise the client representation from the mount data */
-	server->flags = data->flags & NFS_MOUNT_FLAGMASK;
+	server->flags = data->flags;
 	server->caps |= NFS_CAP_ATOMIC_OPEN;
 
 	if (data->rsize)

fs/nfs/dir.c

@@ -156,6 +156,7 @@ typedef struct {
 	decode_dirent_t	decode;
 	int		plus;
 	unsigned long	timestamp;
+	unsigned long	gencount;
 	int		timestamp_valid;
 } nfs_readdir_descriptor_t;
 
@@ -177,7 +178,7 @@ int nfs_readdir_filler(nfs_readdir_descriptor_t *desc, struct page *page)
 	struct file	*file = desc->file;
 	struct inode	*inode = file->f_path.dentry->d_inode;
 	struct rpc_cred	*cred = nfs_file_cred(file);
-	unsigned long	timestamp;
+	unsigned long	timestamp, gencount;
 	int		error;
 
 	dfprintk(DIRCACHE, "NFS: %s: reading cookie %Lu into page %lu\n",
@@ -186,6 +187,7 @@ int nfs_readdir_filler(nfs_readdir_descriptor_t *desc, struct page *page)
 
  again:
 	timestamp = jiffies;
+	gencount = nfs_inc_attr_generation_counter();
 	error = NFS_PROTO(inode)->readdir(file->f_path.dentry, cred, desc->entry->cookie, page,
 					  NFS_SERVER(inode)->dtsize, desc->plus);
 	if (error < 0) {
@@ -199,6 +201,7 @@ int nfs_readdir_filler(nfs_readdir_descriptor_t *desc, struct page *page)
 		goto error;
 	}
 	desc->timestamp = timestamp;
+	desc->gencount = gencount;
 	desc->timestamp_valid = 1;
 	SetPageUptodate(page);
 	/* Ensure consistent page alignment of the data.
@@ -224,9 +227,10 @@ int dir_decode(nfs_readdir_descriptor_t *desc)
 	if (IS_ERR(p))
 		return PTR_ERR(p);
 	desc->ptr = p;
-	if (desc->timestamp_valid)
+	if (desc->timestamp_valid) {
 		desc->entry->fattr->time_start = desc->timestamp;
-	else
+		desc->entry->fattr->gencount = desc->gencount;
+	} else
 		desc->entry->fattr->valid &= ~NFS_ATTR_FATTR;
 	return 0;
 }
@@ -471,7 +475,7 @@ int uncached_readdir(nfs_readdir_descriptor_t *desc, void *dirent,
 	struct rpc_cred	*cred = nfs_file_cred(file);
 	struct page	*page = NULL;
 	int		status;
-	unsigned long	timestamp;
+	unsigned long	timestamp, gencount;
 
 	dfprintk(DIRCACHE, "NFS: uncached_readdir() searching for cookie %Lu\n",
 			(unsigned long long)*desc->dir_cookie);
@@ -482,6 +486,7 @@ int uncached_readdir(nfs_readdir_descriptor_t *desc, void *dirent,
 		goto out;
 	}
 	timestamp = jiffies;
+	gencount = nfs_inc_attr_generation_counter();
 	status = NFS_PROTO(inode)->readdir(file->f_path.dentry, cred,
 						*desc->dir_cookie, page,
 						NFS_SERVER(inode)->dtsize,
@@ -490,6 +495,7 @@ int uncached_readdir(nfs_readdir_descriptor_t *desc, void *dirent,
 	desc->ptr = kmap(page);		/* matching kunmap in nfs_do_filldir */
 	if (status >= 0) {
 		desc->timestamp = timestamp;
+		desc->gencount = gencount;
 		desc->timestamp_valid = 1;
 		if ((status = dir_decode(desc)) == 0)
 			desc->entry->prev_cookie = *desc->dir_cookie;
@@ -655,7 +661,7 @@ static int nfs_fsync_dir(struct file *filp, struct dentry *dentry, int datasync)
  */
 void nfs_force_lookup_revalidate(struct inode *dir)
 {
-	NFS_I(dir)->cache_change_attribute = jiffies;
+	NFS_I(dir)->cache_change_attribute++;
 }
 
 /*
@@ -667,6 +673,8 @@ static int nfs_check_verifier(struct inode *dir, struct dentry *dentry)
 {
 	if (IS_ROOT(dentry))
 		return 1;
+	if (NFS_SERVER(dir)->flags & NFS_MOUNT_LOOKUP_CACHE_NONE)
+		return 0;
 	if (!nfs_verify_change_attribute(dir, dentry->d_time))
 		return 0;
 	/* Revalidate nfsi->cache_change_attribute before we declare a match */
@@ -750,6 +758,8 @@ int nfs_neg_need_reval(struct inode *dir, struct dentry *dentry,
 	/* Don't revalidate a negative dentry if we're creating a new file */
 	if (nd != NULL && nfs_lookup_check_intent(nd, LOOKUP_CREATE) != 0)
 		return 0;
+	if (NFS_SERVER(dir)->flags & NFS_MOUNT_LOOKUP_CACHE_NONEG)
+		return 1;
 	return !nfs_check_verifier(dir, dentry);
 }
 

fs/nfs/file.c

@@ -188,13 +188,16 @@ static loff_t nfs_file_llseek(struct file *filp, loff_t offset, int origin)
 	/* origin == SEEK_END => we must revalidate the cached file length */
 	if (origin == SEEK_END) {
 		struct inode *inode = filp->f_mapping->host;
+
 		int retval = nfs_revalidate_file_size(inode, filp);
 		if (retval < 0)
 			return (loff_t)retval;
-	}
-	lock_kernel();	/* BKL needed? */
+
+		spin_lock(&inode->i_lock);
+		loff = generic_file_llseek_unlocked(filp, offset, origin);
+		spin_unlock(&inode->i_lock);
+	} else
 		loff = generic_file_llseek_unlocked(filp, offset, origin);
-	unlock_kernel();
 	return loff;
 }
 
@@ -699,13 +702,6 @@ static int nfs_flock(struct file *filp, int cmd, struct file_lock *fl)
 			filp->f_path.dentry->d_name.name,
 			fl->fl_type, fl->fl_flags);
 
-	/*
-	 * No BSD flocks over NFS allowed.
-	 * Note: we could try to fake a POSIX lock request here by
-	 * using ((u32) filp | 0x80000000) or some such as the pid.
-	 * Not sure whether that would be unique, though, or whether
-	 * that would break in other places.
-	 */
 	if (!(fl->fl_flags & FL_FLOCK))
 		return -ENOLCK;
 

fs/nfs/inode.c

@@ -305,8 +305,7 @@ nfs_fhget(struct super_block *sb, struct nfs_fh *fh, struct nfs_fattr *fattr)
 			init_special_inode(inode, inode->i_mode, fattr->rdev);
 
 		nfsi->read_cache_jiffies = fattr->time_start;
-		nfsi->last_updated = now;
-		nfsi->cache_change_attribute = now;
+		nfsi->attr_gencount = fattr->gencount;
 		inode->i_atime = fattr->atime;
 		inode->i_mtime = fattr->mtime;
 		inode->i_ctime = fattr->ctime;
@@ -453,6 +452,7 @@ static int nfs_vmtruncate(struct inode * inode, loff_t offset)
 void nfs_setattr_update_inode(struct inode *inode, struct iattr *attr)
 {
 	if ((attr->ia_valid & (ATTR_MODE|ATTR_UID|ATTR_GID)) != 0) {
+		spin_lock(&inode->i_lock);
 		if ((attr->ia_valid & ATTR_MODE) != 0) {
 			int mode = attr->ia_mode & S_IALLUGO;
 			mode |= inode->i_mode & ~S_IALLUGO;
@@ -462,7 +462,6 @@ void nfs_setattr_update_inode(struct inode *inode, struct iattr *attr)
 			inode->i_uid = attr->ia_uid;
 		if ((attr->ia_valid & ATTR_GID) != 0)
 			inode->i_gid = attr->ia_gid;
-		spin_lock(&inode->i_lock);
 		NFS_I(inode)->cache_validity |= NFS_INO_INVALID_ACCESS|NFS_INO_INVALID_ACL;
 		spin_unlock(&inode->i_lock);
 	}
@@ -472,37 +471,6 @@ void nfs_setattr_update_inode(struct inode *inode, struct iattr *attr)
 	}
 }
 
-static int nfs_wait_schedule(void *word)
-{
-	if (signal_pending(current))
-		return -ERESTARTSYS;
-	schedule();
-	return 0;
-}
-
-/*
- * Wait for the inode to get unlocked.
- */
-static int nfs_wait_on_inode(struct inode *inode)
-{
-	struct nfs_inode *nfsi = NFS_I(inode);
-	int error;
-
-	error = wait_on_bit_lock(&nfsi->flags, NFS_INO_REVALIDATING,
-					nfs_wait_schedule, TASK_KILLABLE);
-
-	return error;
-}
-
-static void nfs_wake_up_inode(struct inode *inode)
-{
-	struct nfs_inode *nfsi = NFS_I(inode);
-
-	clear_bit(NFS_INO_REVALIDATING, &nfsi->flags);
-	smp_mb__after_clear_bit();
-	wake_up_bit(&nfsi->flags, NFS_INO_REVALIDATING);
-}
-
 int nfs_getattr(struct vfsmount *mnt, struct dentry *dentry, struct kstat *stat)
 {
 	struct inode *inode = dentry->d_inode;
@@ -697,20 +665,15 @@ __nfs_revalidate_inode(struct nfs_server *server, struct inode *inode)
 	dfprintk(PAGECACHE, "NFS: revalidating (%s/%Ld)\n",
 		inode->i_sb->s_id, (long long)NFS_FILEID(inode));
 
-	nfs_inc_stats(inode, NFSIOS_INODEREVALIDATE);
 	if (is_bad_inode(inode))
- 		goto out_nowait;
+		goto out;
 	if (NFS_STALE(inode))
- 		goto out_nowait;
-
-	status = nfs_wait_on_inode(inode);
-	if (status < 0)
 		goto out;
 
-	status = -ESTALE;
 	if (NFS_STALE(inode))
 		goto out;
 
+	nfs_inc_stats(inode, NFSIOS_INODEREVALIDATE);
 	status = NFS_PROTO(inode)->getattr(server, NFS_FH(inode), &fattr);
 	if (status != 0) {
 		dfprintk(PAGECACHE, "nfs_revalidate_inode: (%s/%Ld) getattr failed, error=%d\n",
@@ -724,16 +687,13 @@ __nfs_revalidate_inode(struct nfs_server *server, struct inode *inode)
 		goto out;
 	}
 
-	spin_lock(&inode->i_lock);
-	status = nfs_update_inode(inode, &fattr);
+	status = nfs_refresh_inode(inode, &fattr);
 	if (status) {
-		spin_unlock(&inode->i_lock);
 		dfprintk(PAGECACHE, "nfs_revalidate_inode: (%s/%Ld) refresh failed, error=%d\n",
 			 inode->i_sb->s_id,
 			 (long long)NFS_FILEID(inode), status);
 		goto out;
 	}
-	spin_unlock(&inode->i_lock);
 
 	if (nfsi->cache_validity & NFS_INO_INVALID_ACL)
 		nfs_zap_acl_cache(inode);
@@ -743,9 +703,6 @@ __nfs_revalidate_inode(struct nfs_server *server, struct inode *inode)
 		(long long)NFS_FILEID(inode));
 
  out:
-	nfs_wake_up_inode(inode);
-
- out_nowait:
 	return status;
 }
 
@@ -908,9 +865,6 @@ static int nfs_check_inode_attributes(struct inode *inode, struct nfs_fattr *fat
 		return -EIO;
 	}
 
-	/* Do atomic weak cache consistency updates */
-	nfs_wcc_update_inode(inode, fattr);
-
 	if ((fattr->valid & NFS_ATTR_FATTR_V4) != 0 &&
 			nfsi->change_attr != fattr->change_attr)
 		invalid |= NFS_INO_INVALID_ATTR|NFS_INO_REVAL_PAGECACHE;
@@ -939,15 +893,81 @@ static int nfs_check_inode_attributes(struct inode *inode, struct nfs_fattr *fat
 
 	if (invalid != 0)
 		nfsi->cache_validity |= invalid;
-	else
-		nfsi->cache_validity &= ~(NFS_INO_INVALID_ATTR
-				| NFS_INO_INVALID_ATIME
-				| NFS_INO_REVAL_PAGECACHE);
 
 	nfsi->read_cache_jiffies = fattr->time_start;
 	return 0;
 }
 
+static int nfs_ctime_need_update(const struct inode *inode, const struct nfs_fattr *fattr)
+{
+	return timespec_compare(&fattr->ctime, &inode->i_ctime) > 0;
+}
+
+static int nfs_size_need_update(const struct inode *inode, const struct nfs_fattr *fattr)
+{
+	return nfs_size_to_loff_t(fattr->size) > i_size_read(inode);
+}
+
+static unsigned long nfs_attr_generation_counter;
+
+static unsigned long nfs_read_attr_generation_counter(void)
+{
+	smp_rmb();
+	return nfs_attr_generation_counter;
+}
+
+unsigned long nfs_inc_attr_generation_counter(void)
+{
+	unsigned long ret;
+	smp_rmb();
+	ret = ++nfs_attr_generation_counter;
+	smp_wmb();
+	return ret;
+}
+
+void nfs_fattr_init(struct nfs_fattr *fattr)
+{
+	fattr->valid = 0;
+	fattr->time_start = jiffies;
+	fattr->gencount = nfs_inc_attr_generation_counter();
+}
+
+/**
+ * nfs_inode_attrs_need_update - check if the inode attributes need updating
+ * @inode - pointer to inode
+ * @fattr - attributes
+ *
+ * Attempt to divine whether or not an RPC call reply carrying stale
+ * attributes got scheduled after another call carrying updated ones.
+ *
+ * To do so, the function first assumes that a more recent ctime means
+ * that the attributes in fattr are newer, however it also attempt to
+ * catch the case where ctime either didn't change, or went backwards
+ * (if someone reset the clock on the server) by looking at whether
+ * or not this RPC call was started after the inode was last updated.
+ * Note also the check for wraparound of 'attr_gencount'
+ *
+ * The function returns 'true' if it thinks the attributes in 'fattr' are
+ * more recent than the ones cached in the inode.
+ *
+ */
+static int nfs_inode_attrs_need_update(const struct inode *inode, const struct nfs_fattr *fattr)
+{
+	const struct nfs_inode *nfsi = NFS_I(inode);
+
+	return ((long)fattr->gencount - (long)nfsi->attr_gencount) > 0 ||
+		nfs_ctime_need_update(inode, fattr) ||
+		nfs_size_need_update(inode, fattr) ||
+		((long)nfsi->attr_gencount - (long)nfs_read_attr_generation_counter() > 0);
+}
+
+static int nfs_refresh_inode_locked(struct inode *inode, struct nfs_fattr *fattr)
+{
+	if (nfs_inode_attrs_need_update(inode, fattr))
+		return nfs_update_inode(inode, fattr);
+	return nfs_check_inode_attributes(inode, fattr);
+}
+
 /**
  * nfs_refresh_inode - try to update the inode attribute cache
  * @inode - pointer to inode
@@ -960,21 +980,28 @@ static int nfs_check_inode_attributes(struct inode *inode, struct nfs_fattr *fat
  */
 int nfs_refresh_inode(struct inode *inode, struct nfs_fattr *fattr)
 {
-	struct nfs_inode *nfsi = NFS_I(inode);
 	int status;
 
 	if ((fattr->valid & NFS_ATTR_FATTR) == 0)
 		return 0;
 	spin_lock(&inode->i_lock);
-	if (time_after(fattr->time_start, nfsi->last_updated))
-		status = nfs_update_inode(inode, fattr);
-	else
-		status = nfs_check_inode_attributes(inode, fattr);
-
+	status = nfs_refresh_inode_locked(inode, fattr);
 	spin_unlock(&inode->i_lock);
 	return status;
 }
 
+static int nfs_post_op_update_inode_locked(struct inode *inode, struct nfs_fattr *fattr)
+{
+	struct nfs_inode *nfsi = NFS_I(inode);
+
+	nfsi->cache_validity |= NFS_INO_INVALID_ATTR|NFS_INO_REVAL_PAGECACHE;
+	if (S_ISDIR(inode->i_mode))
+		nfsi->cache_validity |= NFS_INO_INVALID_DATA;
+	if ((fattr->valid & NFS_ATTR_FATTR) == 0)
+		return 0;
+	return nfs_refresh_inode_locked(inode, fattr);
+}
+
 /**
  * nfs_post_op_update_inode - try to update the inode attribute cache
  * @inode - pointer to inode
@@ -991,14 +1018,12 @@ int nfs_refresh_inode(struct inode *inode, struct nfs_fattr *fattr)
  */
 int nfs_post_op_update_inode(struct inode *inode, struct nfs_fattr *fattr)
 {
-	struct nfs_inode *nfsi = NFS_I(inode);
+	int status;
 
 	spin_lock(&inode->i_lock);
-	nfsi->cache_validity |= NFS_INO_INVALID_ATTR|NFS_INO_REVAL_PAGECACHE;
-	if (S_ISDIR(inode->i_mode))
-		nfsi->cache_validity |= NFS_INO_INVALID_DATA;
+	status = nfs_post_op_update_inode_locked(inode, fattr);
 	spin_unlock(&inode->i_lock);
-	return nfs_refresh_inode(inode, fattr);
+	return status;
 }
 
 /**
@@ -1014,6 +1039,15 @@ int nfs_post_op_update_inode(struct inode *inode, struct nfs_fattr *fattr)
  */
 int nfs_post_op_update_inode_force_wcc(struct inode *inode, struct nfs_fattr *fattr)
 {
+	int status;
+
+	spin_lock(&inode->i_lock);
+	/* Don't do a WCC update if these attributes are already stale */
+	if ((fattr->valid & NFS_ATTR_FATTR) == 0 ||
+			!nfs_inode_attrs_need_update(inode, fattr)) {
+		fattr->valid &= ~(NFS_ATTR_WCC_V4|NFS_ATTR_WCC);
+		goto out_noforce;
+	}
 	if ((fattr->valid & NFS_ATTR_FATTR_V4) != 0 &&
 			(fattr->valid & NFS_ATTR_WCC_V4) == 0) {
 		fattr->pre_change_attr = NFS_I(inode)->change_attr;
@@ -1026,7 +1060,10 @@ int nfs_post_op_update_inode_force_wcc(struct inode *inode, struct nfs_fattr *fa
 		fattr->pre_size = i_size_read(inode);
 		fattr->valid |= NFS_ATTR_WCC;
 	}
-	return nfs_post_op_update_inode(inode, fattr);
+out_noforce:
+	status = nfs_post_op_update_inode_locked(inode, fattr);
+	spin_unlock(&inode->i_lock);
+	return status;
 }
 
 /*
@@ -1092,7 +1129,7 @@ static int nfs_update_inode(struct inode *inode, struct nfs_fattr *fattr)
 		}
 		/* If ctime has changed we should definitely clear access+acl caches */
 		if (!timespec_equal(&inode->i_ctime, &fattr->ctime))
-			invalid |= NFS_INO_INVALID_ACCESS|NFS_INO_INVALID_ACL;
+			invalid |= NFS_INO_INVALID_ATTR|NFS_INO_INVALID_ACCESS|NFS_INO_INVALID_ACL;
 	} else if (nfsi->change_attr != fattr->change_attr) {
 		dprintk("NFS: change_attr change on server for file %s/%ld\n",
 				inode->i_sb->s_id, inode->i_ino);
@@ -1126,6 +1163,9 @@ static int nfs_update_inode(struct inode *inode, struct nfs_fattr *fattr)
 	    inode->i_gid != fattr->gid)
 		invalid |= NFS_INO_INVALID_ATTR|NFS_INO_INVALID_ACCESS|NFS_INO_INVALID_ACL;
 
+	if (inode->i_nlink != fattr->nlink)
+		invalid |= NFS_INO_INVALID_ATTR;
+
 	inode->i_mode = fattr->mode;
 	inode->i_nlink = fattr->nlink;
 	inode->i_uid = fattr->uid;
@@ -1145,18 +1185,13 @@ static int nfs_update_inode(struct inode *inode, struct nfs_fattr *fattr)
 		nfs_inc_stats(inode, NFSIOS_ATTRINVALIDATE);
 		nfsi->attrtimeo = NFS_MINATTRTIMEO(inode);
 		nfsi->attrtimeo_timestamp = now;
-		nfsi->last_updated = now;
+		nfsi->attr_gencount = nfs_inc_attr_generation_counter();
 	} else {
 		if (!time_in_range(now, nfsi->attrtimeo_timestamp, nfsi->attrtimeo_timestamp + nfsi->attrtimeo)) {
 			if ((nfsi->attrtimeo <<= 1) > NFS_MAXATTRTIMEO(inode))
 				nfsi->attrtimeo = NFS_MAXATTRTIMEO(inode);
 			nfsi->attrtimeo_timestamp = now;
 		}
-		/*
-		 * Avoid jiffy wraparound issues with nfsi->last_updated
-		 */
-		if (!time_in_range(nfsi->last_updated, nfsi->read_cache_jiffies, now))
-			nfsi->last_updated = nfsi->read_cache_jiffies;
 	}
 	invalid &= ~NFS_INO_INVALID_ATTR;
 	/* Don't invalidate the data if we were to blame */

fs/nfs/internal.h

@@ -153,6 +153,7 @@ extern void nfs4_clear_inode(struct inode *);
 void nfs_zap_acl_cache(struct inode *inode);
 
 /* super.c */
+void nfs_parse_ip_address(char *, size_t, struct sockaddr *, size_t *);
 extern struct file_system_type nfs_xdev_fs_type;
 #ifdef CONFIG_NFS_V4
 extern struct file_system_type nfs4_xdev_fs_type;
@@ -163,8 +164,8 @@ extern struct rpc_stat nfs_rpcstat;
 
 extern int __init register_nfs_fs(void);
 extern void __exit unregister_nfs_fs(void);
-extern void nfs_sb_active(struct nfs_server *server);
-extern void nfs_sb_deactive(struct nfs_server *server);
+extern void nfs_sb_active(struct super_block *sb);
+extern void nfs_sb_deactive(struct super_block *sb);
 
 /* namespace.c */
 extern char *nfs_path(const char *base,
@@ -276,3 +277,23 @@ unsigned int nfs_page_array_len(unsigned int base, size_t len)
 		PAGE_SIZE - 1) >> PAGE_SHIFT;
 }
 
+#define IPV6_SCOPE_DELIMITER	'%'
+
+/*
+ * Set the port number in an address.  Be agnostic about the address
+ * family.
+ */
+static inline void nfs_set_port(struct sockaddr *sap, unsigned short port)
+{
+	struct sockaddr_in *ap = (struct sockaddr_in *)sap;
+	struct sockaddr_in6 *ap6 = (struct sockaddr_in6 *)sap;
+
+	switch (sap->sa_family) {
+	case AF_INET:
+		ap->sin_port = htons(port);
+		break;
+	case AF_INET6:
+		ap6->sin6_port = htons(port);
+		break;
+	}
+}

fs/nfs/mount_clnt.c

@@ -14,6 +14,7 @@
 #include <linux/sunrpc/clnt.h>
 #include <linux/sunrpc/sched.h>
 #include <linux/nfs_fs.h>
+#include "internal.h"
 
 #ifdef RPC_DEBUG
 # define NFSDBG_FACILITY	NFSDBG_MOUNT
@@ -98,7 +99,7 @@ int nfs_mount(struct sockaddr *addr, size_t len, char *hostname, char *path,
 
 out_mnt_err:
 	dprintk("NFS: MNT server returned result %d\n", result.status);
-	status = -EACCES;
+	status = nfs_stat_to_errno(result.status);
 	goto out;
 }
 

fs/nfs/namespace.c

@@ -105,7 +105,10 @@ static void * nfs_follow_mountpoint(struct dentry *dentry, struct nameidata *nd)
 
 	dprintk("--> nfs_follow_mountpoint()\n");
 
-	BUG_ON(IS_ROOT(dentry));
+	err = -ESTALE;
+	if (IS_ROOT(dentry))
+		goto out_err;
+
 	dprintk("%s: enter\n", __func__);
 	dput(nd->path.dentry);
 	nd->path.dentry = dget(dentry);
@@ -189,7 +192,7 @@ static struct vfsmount *nfs_do_clone_mount(struct nfs_server *server,
 					   struct nfs_clone_mount *mountdata)
 {
 #ifdef CONFIG_NFS_V4
-	struct vfsmount *mnt = NULL;
+	struct vfsmount *mnt = ERR_PTR(-EINVAL);
 	switch (server->nfs_client->rpc_ops->version) {
 		case 2:
 		case 3:

fs/nfs/nfs3acl.c

@@ -229,6 +229,7 @@ struct posix_acl *nfs3_proc_getacl(struct inode *inode, int type)
 
 	dprintk("NFS call getacl\n");
 	msg.rpc_proc = &server->client_acl->cl_procinfo[ACLPROC3_GETACL];
+	nfs_fattr_init(&fattr);
 	status = rpc_call_sync(server->client_acl, &msg, 0);
 	dprintk("NFS reply getacl: %d\n", status);
 
@@ -322,6 +323,7 @@ static int nfs3_proc_setacls(struct inode *inode, struct posix_acl *acl,
 
 	dprintk("NFS call setacl\n");
 	msg.rpc_proc = &server->client_acl->cl_procinfo[ACLPROC3_SETACL];
+	nfs_fattr_init(&fattr);
 	status = rpc_call_sync(server->client_acl, &msg, 0);
 	nfs_access_zap_cache(inode);
 	nfs_zap_acl_cache(inode);

fs/nfs/nfs3proc.c

@@ -699,7 +699,7 @@ nfs3_proc_statfs(struct nfs_server *server, struct nfs_fh *fhandle,
 }
 
 static int
-nfs3_proc_fsinfo(struct nfs_server *server, struct nfs_fh *fhandle,
+do_proc_fsinfo(struct rpc_clnt *client, struct nfs_fh *fhandle,
 		 struct nfs_fsinfo *info)
 {
 	struct rpc_message msg = {
@@ -711,11 +711,27 @@ nfs3_proc_fsinfo(struct nfs_server *server, struct nfs_fh *fhandle,
 
 	dprintk("NFS call  fsinfo\n");
 	nfs_fattr_init(info->fattr);
-	status = rpc_call_sync(server->nfs_client->cl_rpcclient, &msg, 0);
+	status = rpc_call_sync(client, &msg, 0);
 	dprintk("NFS reply fsinfo: %d\n", status);
 	return status;
 }
 
+/*
+ * Bare-bones access to fsinfo: this is for nfs_get_root/nfs_get_sb via
+ * nfs_create_server
+ */
+static int
+nfs3_proc_fsinfo(struct nfs_server *server, struct nfs_fh *fhandle,
+		   struct nfs_fsinfo *info)
+{
+	int	status;
+
+	status = do_proc_fsinfo(server->client, fhandle, info);
+	if (status && server->nfs_client->cl_rpcclient != server->client)
+		status = do_proc_fsinfo(server->nfs_client->cl_rpcclient, fhandle, info);
+	return status;
+}
+
 static int
 nfs3_proc_pathconf(struct nfs_server *server, struct nfs_fh *fhandle,
 		   struct nfs_pathconf *info)

fs/nfs/nfs4namespace.c

@@ -93,21 +93,52 @@ static int nfs4_validate_fspath(const struct vfsmount *mnt_parent,
 	return 0;
 }
 
-/*
- * Check if the string represents a "valid" IPv4 address
- */
-static inline int valid_ipaddr4(const char *buf)
+static struct vfsmount *try_location(struct nfs_clone_mount *mountdata,
+				     char *page, char *page2,
+				     const struct nfs4_fs_location *location)
 {
-	int rc, count, in[4];
-
-	rc = sscanf(buf, "%d.%d.%d.%d", &in[0], &in[1], &in[2], &in[3]);
-	if (rc != 4)
-		return -EINVAL;
-	for (count = 0; count < 4; count++) {
-		if (in[count] > 255)
-			return -EINVAL;
+	struct vfsmount *mnt = ERR_PTR(-ENOENT);
+	char *mnt_path;
+	int page2len;
+	unsigned int s;
+
+	mnt_path = nfs4_pathname_string(&location->rootpath, page2, PAGE_SIZE);
+	if (IS_ERR(mnt_path))
+		return mnt;
+	mountdata->mnt_path = mnt_path;
+	page2 += strlen(mnt_path) + 1;
+	page2len = PAGE_SIZE - strlen(mnt_path) - 1;
+
+	for (s = 0; s < location->nservers; s++) {
+		const struct nfs4_string *buf = &location->servers[s];
+		struct sockaddr_storage addr;
+
+		if (buf->len <= 0 || buf->len >= PAGE_SIZE)
+			continue;
+
+		mountdata->addr = (struct sockaddr *)&addr;
+
+		if (memchr(buf->data, IPV6_SCOPE_DELIMITER, buf->len))
+			continue;
+		nfs_parse_ip_address(buf->data, buf->len,
+				mountdata->addr, &mountdata->addrlen);
+		if (mountdata->addr->sa_family == AF_UNSPEC)
+			continue;
+		nfs_set_port(mountdata->addr, NFS_PORT);
+
+		strncpy(page2, buf->data, page2len);
+		page2[page2len] = '\0';
+		mountdata->hostname = page2;
+
+		snprintf(page, PAGE_SIZE, "%s:%s",
+				mountdata->hostname,
+				mountdata->mnt_path);
+
+		mnt = vfs_kern_mount(&nfs4_referral_fs_type, 0, page, mountdata);
+		if (!IS_ERR(mnt))
+			break;
 	}
-	return 0;
+	return mnt;
 }
 
 /**
@@ -128,7 +159,6 @@ static struct vfsmount *nfs_follow_referral(const struct vfsmount *mnt_parent,
 		.authflavor = NFS_SB(mnt_parent->mnt_sb)->client->cl_auth->au_flavor,
 	};
 	char *page = NULL, *page2 = NULL;
-	unsigned int s;
 	int loc, error;
 
 	if (locations == NULL || locations->nlocations <= 0)
@@ -152,54 +182,17 @@ static struct vfsmount *nfs_follow_referral(const struct vfsmount *mnt_parent,
 		goto out;
 	}
 
-	loc = 0;
-	while (loc < locations->nlocations && IS_ERR(mnt)) {
+	for (loc = 0; loc < locations->nlocations; loc++) {
 		const struct nfs4_fs_location *location = &locations->locations[loc];
-		char *mnt_path;
 
 		if (location == NULL || location->nservers <= 0 ||
-		    location->rootpath.ncomponents == 0) {
-			loc++;
-			continue;
-		}
-
-		mnt_path = nfs4_pathname_string(&location->rootpath, page2, PAGE_SIZE);
-		if (IS_ERR(mnt_path)) {
-			loc++;
+		    location->rootpath.ncomponents == 0)
 			continue;
-		}
-		mountdata.mnt_path = mnt_path;
 
-		s = 0;
-		while (s < location->nservers) {
-			struct sockaddr_in addr = {
-				.sin_family	= AF_INET,
-				.sin_port	= htons(NFS_PORT),
-			};
-
-			if (location->servers[s].len <= 0 ||
-			    valid_ipaddr4(location->servers[s].data) < 0) {
-				s++;
-				continue;
-			}
-
-			mountdata.hostname = location->servers[s].data;
-			addr.sin_addr.s_addr = in_aton(mountdata.hostname),
-			mountdata.addr = (struct sockaddr *)&addr;
-			mountdata.addrlen = sizeof(addr);
-
-			snprintf(page, PAGE_SIZE, "%s:%s",
-					mountdata.hostname,
-					mountdata.mnt_path);
-
-			mnt = vfs_kern_mount(&nfs4_referral_fs_type, 0, page, &mountdata);
-			if (!IS_ERR(mnt)) {
+		mnt = try_location(&mountdata, page, page2, location);
+		if (!IS_ERR(mnt))
 			break;
 	}
-			s++;
-		}
-		loc++;
-	}
 
 out:
 	free_page((unsigned long) page);

fs/nfs/proc.c

@@ -65,6 +65,9 @@ nfs_proc_get_root(struct nfs_server *server, struct nfs_fh *fhandle,
 
 	dprintk("%s: call getattr\n", __func__);
 	nfs_fattr_init(fattr);
+	status = rpc_call_sync(server->client, &msg, 0);
+	/* Retry with default authentication if different */
+	if (status && server->nfs_client->cl_rpcclient != server->client)
 		status = rpc_call_sync(server->nfs_client->cl_rpcclient, &msg, 0);
 	dprintk("%s: reply getattr: %d\n", __func__, status);
 	if (status)
@@ -72,6 +75,9 @@ nfs_proc_get_root(struct nfs_server *server, struct nfs_fh *fhandle,
 	dprintk("%s: call statfs\n", __func__);
 	msg.rpc_proc = &nfs_procedures[NFSPROC_STATFS];
 	msg.rpc_resp = &fsinfo;
+	status = rpc_call_sync(server->client, &msg, 0);
+	/* Retry with default authentication if different */
+	if (status && server->nfs_client->cl_rpcclient != server->client)
 		status = rpc_call_sync(server->nfs_client->cl_rpcclient, &msg, 0);
 	dprintk("%s: reply statfs: %d\n", __func__, status);
 	if (status)

fs/nfs/super.c

@@ -91,6 +91,7 @@ enum {
 	/* Mount options that take string arguments */
 	Opt_sec, Opt_proto, Opt_mountproto, Opt_mounthost,
 	Opt_addr, Opt_mountaddr, Opt_clientaddr,
+	Opt_lookupcache,
 
 	/* Special mount options */
 	Opt_userspace, Opt_deprecated, Opt_sloppy,
@@ -154,6 +155,8 @@ static const match_table_t nfs_mount_option_tokens = {
 	{ Opt_mounthost, "mounthost=%s" },
 	{ Opt_mountaddr, "mountaddr=%s" },
 
+	{ Opt_lookupcache, "lookupcache=%s" },
+
 	{ Opt_err, NULL }
 };
 
@@ -200,6 +203,22 @@ static const match_table_t nfs_secflavor_tokens = {
 	{ Opt_sec_err, NULL }
 };
 
+enum {
+	Opt_lookupcache_all, Opt_lookupcache_positive,
+	Opt_lookupcache_none,
+
+	Opt_lookupcache_err
+};
+
+static match_table_t nfs_lookupcache_tokens = {
+	{ Opt_lookupcache_all, "all" },
+	{ Opt_lookupcache_positive, "pos" },
+	{ Opt_lookupcache_positive, "positive" },
+	{ Opt_lookupcache_none, "none" },
+
+	{ Opt_lookupcache_err, NULL }
+};
+
 
 static void nfs_umount_begin(struct super_block *);
 static int  nfs_statfs(struct dentry *, struct kstatfs *);
@@ -209,7 +228,6 @@ static int nfs_get_sb(struct file_system_type *, int, const char *, void *, stru
 static int nfs_xdev_get_sb(struct file_system_type *fs_type,
 		int flags, const char *dev_name, void *raw_data, struct vfsmount *mnt);
 static void nfs_kill_super(struct super_block *);
-static void nfs_put_super(struct super_block *);
 static int nfs_remount(struct super_block *sb, int *flags, char *raw_data);
 
 static struct file_system_type nfs_fs_type = {
@@ -232,7 +250,6 @@ static const struct super_operations nfs_sops = {
 	.alloc_inode	= nfs_alloc_inode,
 	.destroy_inode	= nfs_destroy_inode,
 	.write_inode	= nfs_write_inode,
-	.put_super	= nfs_put_super,
 	.statfs		= nfs_statfs,
 	.clear_inode	= nfs_clear_inode,
 	.umount_begin	= nfs_umount_begin,
@@ -337,26 +354,20 @@ void __exit unregister_nfs_fs(void)
 	unregister_filesystem(&nfs_fs_type);
 }
 
-void nfs_sb_active(struct nfs_server *server)
+void nfs_sb_active(struct super_block *sb)
 {
-	atomic_inc(&server->active);
-}
+	struct nfs_server *server = NFS_SB(sb);
 
-void nfs_sb_deactive(struct nfs_server *server)
-{
-	if (atomic_dec_and_test(&server->active))
-		wake_up(&server->active_wq);
+	if (atomic_inc_return(&server->active) == 1)
+		atomic_inc(&sb->s_active);
 }
 
-static void nfs_put_super(struct super_block *sb)
+void nfs_sb_deactive(struct super_block *sb)
 {
 	struct nfs_server *server = NFS_SB(sb);
-	/*
-	 * Make sure there are no outstanding ops to this server.
-	 * If so, wait for them to finish before allowing the
-	 * unmount to continue.
-	 */
-	wait_event(server->active_wq, atomic_read(&server->active) == 0);
+
+	if (atomic_dec_and_test(&server->active))
+		deactivate_super(sb);
 }
 
 /*
@@ -663,25 +674,6 @@ static void nfs_umount_begin(struct super_block *sb)
 		rpc_killall_tasks(rpc);
 }
 
-/*
- * Set the port number in an address.  Be agnostic about the address family.
- */
-static void nfs_set_port(struct sockaddr *sap, unsigned short port)
-{
-	switch (sap->sa_family) {
-	case AF_INET: {
-		struct sockaddr_in *ap = (struct sockaddr_in *)sap;
-		ap->sin_port = htons(port);
-		break;
-	}
-	case AF_INET6: {
-		struct sockaddr_in6 *ap = (struct sockaddr_in6 *)sap;
-		ap->sin6_port = htons(port);
-		break;
-	}
-	}
-}
-
 /*
  * Sanity-check a server address provided by the mount command.
  *
@@ -724,20 +716,22 @@ static void nfs_parse_ipv4_address(char *string, size_t str_len,
 	*addr_len = 0;
 }
 
-#define IPV6_SCOPE_DELIMITER	'%'
-
 #if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE)
-static void nfs_parse_ipv6_scope_id(const char *string, const size_t str_len,
+static int nfs_parse_ipv6_scope_id(const char *string, const size_t str_len,
 				   const char *delim,
 				   struct sockaddr_in6 *sin6)
 {
 	char *p;
 	size_t len;
 
-	if (!(ipv6_addr_type(&sin6->sin6_addr) & IPV6_ADDR_LINKLOCAL))
-		return ;
+	if ((string + str_len) == delim)
+		return 1;
+
 	if (*delim != IPV6_SCOPE_DELIMITER)
-		return;
+		return 0;
+
+	if (!(ipv6_addr_type(&sin6->sin6_addr) & IPV6_ADDR_LINKLOCAL))
+		return 0;
 
 	len = (string + str_len) - delim - 1;
 	p = kstrndup(delim + 1, len, GFP_KERNEL);
@@ -750,14 +744,20 @@ static void nfs_parse_ipv6_scope_id(const char *string, const size_t str_len,
 			scope_id = dev->ifindex;
 			dev_put(dev);
 		} else {
-			/* scope_id is set to zero on error */
-			strict_strtoul(p, 10, &scope_id);
+			if (strict_strtoul(p, 10, &scope_id) == 0) {
+				kfree(p);
+				return 0;
+			}
 		}
 
 		kfree(p);
+
 		sin6->sin6_scope_id = scope_id;
 		dfprintk(MOUNT, "NFS: IPv6 scope ID = %lu\n", scope_id);
+		return 1;
 	}
+
+	return 0;
 }
 
 static void nfs_parse_ipv6_address(char *string, size_t str_len,
@@ -773,8 +773,10 @@ static void nfs_parse_ipv6_address(char *string, size_t str_len,
 
 		sin6->sin6_family = AF_INET6;
 		*addr_len = sizeof(*sin6);
-		if (in6_pton(string, str_len, addr, IPV6_SCOPE_DELIMITER, &delim)) {
-			nfs_parse_ipv6_scope_id(string, str_len, delim, sin6);
+		if (in6_pton(string, str_len, addr,
+					IPV6_SCOPE_DELIMITER, &delim) != 0) {
+			if (nfs_parse_ipv6_scope_id(string, str_len,
+							delim, sin6) != 0)
 				return;
 		}
 	}
@@ -798,7 +800,7 @@ static void nfs_parse_ipv6_address(char *string, size_t str_len,
  * If there is a problem constructing the new sockaddr, set the address
  * family to AF_UNSPEC.
  */
-static void nfs_parse_ip_address(char *string, size_t str_len,
+void nfs_parse_ip_address(char *string, size_t str_len,
 				 struct sockaddr *sap, size_t *addr_len)
 {
 	unsigned int i, colons;
@@ -1258,6 +1260,30 @@ static int nfs_parse_mount_options(char *raw,
 					     &mnt->mount_server.addrlen);
 			kfree(string);
 			break;
+		case Opt_lookupcache:
+			string = match_strdup(args);
+			if (string == NULL)
+				goto out_nomem;
+			token = match_token(string,
+					nfs_lookupcache_tokens, args);
+			kfree(string);
+			switch (token) {
+				case Opt_lookupcache_all:
+					mnt->flags &= ~(NFS_MOUNT_LOOKUP_CACHE_NONEG|NFS_MOUNT_LOOKUP_CACHE_NONE);
+					break;
+				case Opt_lookupcache_positive:
+					mnt->flags &= ~NFS_MOUNT_LOOKUP_CACHE_NONE;
+					mnt->flags |= NFS_MOUNT_LOOKUP_CACHE_NONEG;
+					break;
+				case Opt_lookupcache_none:
+					mnt->flags |= NFS_MOUNT_LOOKUP_CACHE_NONEG|NFS_MOUNT_LOOKUP_CACHE_NONE;
+					break;
+				default:
+					errors++;
+					dfprintk(MOUNT, "NFS:   invalid "
+							"lookupcache argument\n");
+			};
+			break;
 
 		/*
 		 * Special options
@@ -1558,7 +1584,7 @@ static int nfs_validate_mount_data(void *options,
 		 * Translate to nfs_parsed_mount_data, which nfs_fill_super
 		 * can deal with.
 		 */
-		args->flags		= data->flags;
+		args->flags		= data->flags & NFS_MOUNT_FLAGMASK;
 		args->rsize		= data->rsize;
 		args->wsize		= data->wsize;
 		args->timeo		= data->timeo;

fs/nfs/unlink.c

@@ -99,7 +99,7 @@ static void nfs_async_unlink_release(void *calldata)
 
 	nfs_dec_sillycount(data->dir);
 	nfs_free_unlinkdata(data);
-	nfs_sb_deactive(NFS_SB(sb));
+	nfs_sb_deactive(sb);
 }
 
 static const struct rpc_call_ops nfs_unlink_ops = {
@@ -118,6 +118,7 @@ static int nfs_do_call_unlink(struct dentry *parent, struct inode *dir, struct n
 		.rpc_message = &msg,
 		.callback_ops = &nfs_unlink_ops,
 		.callback_data = data,
+		.workqueue = nfsiod_workqueue,
 		.flags = RPC_TASK_ASYNC,
 	};
 	struct rpc_task *task;
@@ -149,7 +150,7 @@ static int nfs_do_call_unlink(struct dentry *parent, struct inode *dir, struct n
 		nfs_dec_sillycount(dir);
 		return 0;
 	}
-	nfs_sb_active(NFS_SERVER(dir));
+	nfs_sb_active(dir->i_sb);
 	data->args.fh = NFS_FH(dir);
 	nfs_fattr_init(&data->res.dir_attr);
 

fs/nfs/write.c

@@ -1427,8 +1427,9 @@ static int nfs_write_mapping(struct address_space *mapping, int how)
 		.bdi = mapping->backing_dev_info,
 		.sync_mode = WB_SYNC_NONE,
 		.nr_to_write = LONG_MAX,
+		.range_start = 0,
+		.range_end = LLONG_MAX,
 		.for_writepages = 1,
-		.range_cyclic = 1,
 	};
 	int ret;
 

include/linux/nfs_fs.h

@@ -137,7 +137,7 @@ struct nfs_inode {
 	unsigned long		attrtimeo_timestamp;
 	__u64			change_attr;		/* v4 only */
 
-	unsigned long		last_updated;
+	unsigned long		attr_gencount;
 	/* "Generation counter" for the attribute cache. This is
 	 * bumped whenever we update the metadata on the
 	 * server.
@@ -200,11 +200,10 @@ struct nfs_inode {
 /*
  * Bit offsets in flags field
  */
-#define NFS_INO_REVALIDATING	(0)		/* revalidating attrs */
-#define NFS_INO_ADVISE_RDPLUS	(1)		/* advise readdirplus */
-#define NFS_INO_STALE		(2)		/* possible stale inode */
-#define NFS_INO_ACL_LRU_SET	(3)		/* Inode is on the LRU list */
-#define NFS_INO_MOUNTPOINT	(4)		/* inode is remote mountpoint */
+#define NFS_INO_ADVISE_RDPLUS	(0)		/* advise readdirplus */
+#define NFS_INO_STALE		(1)		/* possible stale inode */
+#define NFS_INO_ACL_LRU_SET	(2)		/* Inode is on the LRU list */
+#define NFS_INO_MOUNTPOINT	(3)		/* inode is remote mountpoint */
 
 static inline struct nfs_inode *NFS_I(const struct inode *inode)
 {
@@ -345,15 +344,11 @@ extern struct nfs_open_context *get_nfs_open_context(struct nfs_open_context *ct
 extern void put_nfs_open_context(struct nfs_open_context *ctx);
 extern struct nfs_open_context *nfs_find_open_context(struct inode *inode, struct rpc_cred *cred, int mode);
 extern u64 nfs_compat_user_ino64(u64 fileid);
+extern void nfs_fattr_init(struct nfs_fattr *fattr);
 
 /* linux/net/ipv4/ipconfig.c: trims ip addr off front of name, too. */
 extern __be32 root_nfs_parse_addr(char *name); /*__init*/
-
-static inline void nfs_fattr_init(struct nfs_fattr *fattr)
-{
-	fattr->valid = 0;
-	fattr->time_start = jiffies;
-}
+extern unsigned long nfs_inc_attr_generation_counter(void);
 
 /*
  * linux/fs/nfs/file.c

include/linux/nfs_fs_sb.h

@@ -119,7 +119,6 @@ struct nfs_server {
 	void (*destroy)(struct nfs_server *);
 
 	atomic_t active; /* Keep trace of any activity to this server */
-	wait_queue_head_t active_wq;  /* Wait for any activity to stop  */
 
 	/* mountd-related mount options */
 	struct sockaddr_storage	mountd_address;

include/linux/nfs_mount.h

@@ -65,4 +65,8 @@ struct nfs_mount_data {
 #define NFS_MOUNT_UNSHARED	0x8000	/* 5 */
 #define NFS_MOUNT_FLAGMASK	0xFFFF
 
+/* The following are for internal use only */
+#define NFS_MOUNT_LOOKUP_CACHE_NONEG	0x10000
+#define NFS_MOUNT_LOOKUP_CACHE_NONE	0x20000
+
 #endif

include/linux/nfs_xdr.h

@@ -36,6 +36,7 @@ struct nfs_fattr {
 	__u32			nlink;
 	__u32			uid;
 	__u32			gid;
+	dev_t			rdev;
 	__u64			size;
 	union {
 		struct {
@@ -46,7 +47,6 @@ struct nfs_fattr {
 			__u64	used;
 		} nfs3;
 	} du;
-	dev_t			rdev;
 	struct nfs_fsid		fsid;
 	__u64			fileid;
 	struct timespec		atime;
@@ -56,6 +56,7 @@ struct nfs_fattr {
 	__u64			change_attr;	/* NFSv4 change attribute */
 	__u64			pre_change_attr;/* pre-op NFSv4 change attribute */
 	unsigned long		time_start;
+	unsigned long		gencount;
 };
 
 #define NFS_ATTR_WCC		0x0001		/* pre-op WCC data    */
@@ -672,16 +673,16 @@ struct nfs4_rename_res {
 	struct nfs_fattr *		new_fattr;
 };
 
-#define NFS4_SETCLIENTID_NAMELEN	(56)
+#define NFS4_SETCLIENTID_NAMELEN	(127)
 struct nfs4_setclientid {
 	const nfs4_verifier *		sc_verifier;
 	unsigned int			sc_name_len;
-	char				sc_name[NFS4_SETCLIENTID_NAMELEN];
+	char				sc_name[NFS4_SETCLIENTID_NAMELEN + 1];
 	u32				sc_prog;
 	unsigned int			sc_netid_len;
-	char				sc_netid[RPCBIND_MAXNETIDLEN];
+	char				sc_netid[RPCBIND_MAXNETIDLEN + 1];
 	unsigned int			sc_uaddr_len;
-	char				sc_uaddr[RPCBIND_MAXUADDRLEN];
+	char				sc_uaddr[RPCBIND_MAXUADDRLEN + 1];
 	u32				sc_cb_ident;
 };
 

include/linux/sunrpc/xprtrdma.h

@@ -66,9 +66,6 @@
 
 #define RPCRDMA_INLINE_PAD_THRESH  (512)/* payload threshold to pad (bytes) */
 
-#define RDMA_RESOLVE_TIMEOUT	(5*HZ)	/* TBD 5 seconds */
-#define RDMA_CONNECT_RETRY_MAX	(2)	/* retries if no listener backlog */
-
 /* memory registration strategies */
 #define RPCRDMA_PERSISTENT_REGISTRATION (1)
 
@@ -78,6 +75,7 @@ enum rpcrdma_memreg {
 	RPCRDMA_MEMWINDOWS,
 	RPCRDMA_MEMWINDOWS_ASYNC,
 	RPCRDMA_MTHCAFMR,
+	RPCRDMA_FRMR,
 	RPCRDMA_ALLPHYSICAL,
 	RPCRDMA_LAST
 };

net/sunrpc/clnt.c

@@ -213,10 +213,10 @@ static struct rpc_clnt * rpc_new_client(const struct rpc_create_args *args, stru
 	}
 
 	/* save the nodename */
-	clnt->cl_nodelen = strlen(utsname()->nodename);
+	clnt->cl_nodelen = strlen(init_utsname()->nodename);
 	if (clnt->cl_nodelen > UNX_MAXNODENAME)
 		clnt->cl_nodelen = UNX_MAXNODENAME;
-	memcpy(clnt->cl_nodename, utsname()->nodename, clnt->cl_nodelen);
+	memcpy(clnt->cl_nodename, init_utsname()->nodename, clnt->cl_nodelen);
 	rpc_register_client(clnt);
 	return clnt;
 

net/sunrpc/rpcb_clnt.c

@@ -460,6 +460,28 @@ static struct rpc_task *rpcb_call_async(struct rpc_clnt *rpcb_clnt, struct rpcbi
 	return rpc_run_task(&task_setup_data);
 }
 
+/*
+ * In the case where rpc clients have been cloned, we want to make
+ * sure that we use the program number/version etc of the actual
+ * owner of the xprt. To do so, we walk back up the tree of parents
+ * to find whoever created the transport and/or whoever has the
+ * autobind flag set.
+ */
+static struct rpc_clnt *rpcb_find_transport_owner(struct rpc_clnt *clnt)
+{
+	struct rpc_clnt *parent = clnt->cl_parent;
+
+	while (parent != clnt) {
+		if (parent->cl_xprt != clnt->cl_xprt)
+			break;
+		if (clnt->cl_autobind)
+			break;
+		clnt = parent;
+		parent = parent->cl_parent;
+	}
+	return clnt;
+}
+
 /**
  * rpcb_getport_async - obtain the port for a given RPC service on a given host
  * @task: task that is waiting for portmapper request
@@ -469,10 +491,10 @@ static struct rpc_task *rpcb_call_async(struct rpc_clnt *rpcb_clnt, struct rpcbi
  */
 void rpcb_getport_async(struct rpc_task *task)
 {
-	struct rpc_clnt *clnt = task->tk_client;
+	struct rpc_clnt *clnt;
 	struct rpc_procinfo *proc;
 	u32 bind_version;
-	struct rpc_xprt *xprt = task->tk_xprt;
+	struct rpc_xprt *xprt;
 	struct rpc_clnt	*rpcb_clnt;
 	static struct rpcbind_args *map;
 	struct rpc_task	*child;
@@ -481,13 +503,13 @@ void rpcb_getport_async(struct rpc_task *task)
 	size_t salen;
 	int status;
 
+	clnt = rpcb_find_transport_owner(task->tk_client);
+	xprt = clnt->cl_xprt;
+
 	dprintk("RPC: %5u %s(%s, %u, %u, %d)\n",
 		task->tk_pid, __func__,
 		clnt->cl_server, clnt->cl_prog, clnt->cl_vers, xprt->prot);
 
-	/* Autobind on cloned rpc clients is discouraged */
-	BUG_ON(clnt->cl_parent != clnt);
-
 	/* Put self on the wait queue to ensure we get notified if
 	 * some other task is already attempting to bind the port */
 	rpc_sleep_on(&xprt->binding, task, NULL);
@@ -549,7 +571,7 @@ void rpcb_getport_async(struct rpc_task *task)
 		status = -ENOMEM;
 		dprintk("RPC: %5u %s: no memory available\n",
 			task->tk_pid, __func__);
-		goto bailout_nofree;
+		goto bailout_release_client;
 	}
 	map->r_prog = clnt->cl_prog;
 	map->r_vers = clnt->cl_vers;
@@ -569,11 +591,13 @@ void rpcb_getport_async(struct rpc_task *task)
 			task->tk_pid, __func__);
 		return;
 	}
-	rpc_put_task(child);
 
-	task->tk_xprt->stat.bind_count++;
+	xprt->stat.bind_count++;
+	rpc_put_task(child);
 	return;
 
+bailout_release_client:
+	rpc_release_client(rpcb_clnt);
 bailout_nofree:
 	rpcb_wake_rpcbind_waiters(xprt, status);
 	task->tk_status = status;

net/sunrpc/xprt.c

@@ -108,13 +108,10 @@ int xprt_register_transport(struct xprt_class *transport)
 			goto out;
 	}
 
-	result = -EINVAL;
-	if (try_module_get(THIS_MODULE)) {
 	list_add_tail(&transport->list, &xprt_list);
 	printk(KERN_INFO "RPC: Registered %s transport module.\n",
 	       transport->name);
 	result = 0;
-	}
 
 out:
 	spin_unlock(&xprt_list_lock);
@@ -143,7 +140,6 @@ int xprt_unregister_transport(struct xprt_class *transport)
 				"RPC: Unregistered %s transport module.\n",
 				transport->name);
 			list_del_init(&transport->list);
-			module_put(THIS_MODULE);
 			goto out;
 		}
 	}

net/sunrpc/xprtrdma/rpc_rdma.c

@@ -118,6 +118,10 @@ rpcrdma_convert_iovs(struct xdr_buf *xdrbuf, unsigned int pos,
 	}
 
 	if (xdrbuf->tail[0].iov_len) {
+		/* the rpcrdma protocol allows us to omit any trailing
+		 * xdr pad bytes, saving the server an RDMA operation. */
+		if (xdrbuf->tail[0].iov_len < 4 && xprt_rdma_pad_optimize)
+			return n;
 		if (n == nsegs)
 			return 0;
 		seg[n].mr_page = NULL;
@@ -508,7 +512,7 @@ rpcrdma_marshal_req(struct rpc_rqst *rqst)
 	if (hdrlen == 0)
 		return -1;
 
-	dprintk("RPC:       %s: %s: hdrlen %zd rpclen %zd padlen %zd\n"
+	dprintk("RPC:       %s: %s: hdrlen %zd rpclen %zd padlen %zd"
 		" headerp 0x%p base 0x%p lkey 0x%x\n",
 		__func__, transfertypes[wtype], hdrlen, rpclen, padlen,
 		headerp, base, req->rl_iov.lkey);
@@ -594,7 +598,7 @@ rpcrdma_count_chunks(struct rpcrdma_rep *rep, unsigned int max, int wrchunk, __b
  * Scatter inline received data back into provided iov's.
  */
 static void
-rpcrdma_inline_fixup(struct rpc_rqst *rqst, char *srcp, int copy_len)
+rpcrdma_inline_fixup(struct rpc_rqst *rqst, char *srcp, int copy_len, int pad)
 {
 	int i, npages, curlen, olen;
 	char *destp;
@@ -660,6 +664,13 @@ rpcrdma_inline_fixup(struct rpc_rqst *rqst, char *srcp, int copy_len)
 	} else
 		rqst->rq_rcv_buf.tail[0].iov_len = 0;
 
+	if (pad) {
+		/* implicit padding on terminal chunk */
+		unsigned char *p = rqst->rq_rcv_buf.tail[0].iov_base;
+		while (pad--)
+			p[rqst->rq_rcv_buf.tail[0].iov_len++] = 0;
+	}
+
 	if (copy_len)
 		dprintk("RPC:       %s: %d bytes in"
 			" %d extra segments (%d lost)\n",
@@ -681,12 +692,14 @@ rpcrdma_conn_func(struct rpcrdma_ep *ep)
 	struct rpc_xprt *xprt = ep->rep_xprt;
 
 	spin_lock_bh(&xprt->transport_lock);
+	if (++xprt->connect_cookie == 0)	/* maintain a reserved value */
+		++xprt->connect_cookie;
 	if (ep->rep_connected > 0) {
 		if (!xprt_test_and_set_connected(xprt))
 			xprt_wake_pending_tasks(xprt, 0);
 	} else {
 		if (xprt_test_and_clear_connected(xprt))
-			xprt_wake_pending_tasks(xprt, ep->rep_connected);
+			xprt_wake_pending_tasks(xprt, -ENOTCONN);
 	}
 	spin_unlock_bh(&xprt->transport_lock);
 }
@@ -792,14 +805,20 @@ rpcrdma_reply_handler(struct rpcrdma_rep *rep)
 			    ((unsigned char *)iptr - (unsigned char *)headerp);
 			status = rep->rr_len + rdmalen;
 			r_xprt->rx_stats.total_rdma_reply += rdmalen;
+			/* special case - last chunk may omit padding */
+			if (rdmalen &= 3) {
+				rdmalen = 4 - rdmalen;
+				status += rdmalen;
+			}
 		} else {
 			/* else ordinary inline */
+			rdmalen = 0;
 			iptr = (__be32 *)((unsigned char *)headerp + 28);
 			rep->rr_len -= 28; /*sizeof *headerp;*/
 			status = rep->rr_len;
 		}
 		/* Fix up the rpc results for upper layer */
-		rpcrdma_inline_fixup(rqst, (char *)iptr, rep->rr_len);
+		rpcrdma_inline_fixup(rqst, (char *)iptr, rep->rr_len, rdmalen);
 		break;
 
 	case htonl(RDMA_NOMSG):

net/sunrpc/xprtrdma/transport.c

@@ -70,11 +70,8 @@ static unsigned int xprt_rdma_slot_table_entries = RPCRDMA_DEF_SLOT_TABLE;
 static 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;
-#if !RPCRDMA_PERSISTENT_REGISTRATION
-static unsigned int xprt_rdma_memreg_strategy = RPCRDMA_REGISTER; /* FMR? */
-#else
-static unsigned int xprt_rdma_memreg_strategy = RPCRDMA_ALLPHYSICAL;
-#endif
+static unsigned int xprt_rdma_memreg_strategy = RPCRDMA_FRMR;
+                int xprt_rdma_pad_optimize = 0;
 
 #ifdef RPC_DEBUG
 
@@ -139,6 +136,14 @@ static ctl_table xr_tunables_table[] = {
 		.extra1		= &min_memreg,
 		.extra2		= &max_memreg,
 	},
+	{
+		.ctl_name       = CTL_UNNUMBERED,
+		.procname	= "rdma_pad_optimize",
+		.data		= &xprt_rdma_pad_optimize,
+		.maxlen		= sizeof(unsigned int),
+		.mode		= 0644,
+		.proc_handler	= &proc_dointvec,
+	},
 	{
 		.ctl_name = 0,
 	},
@@ -458,6 +463,8 @@ xprt_rdma_close(struct rpc_xprt *xprt)
 	struct rpcrdma_xprt *r_xprt = rpcx_to_rdmax(xprt);
 
 	dprintk("RPC:       %s: closing\n", __func__);
+	if (r_xprt->rx_ep.rep_connected > 0)
+		xprt->reestablish_timeout = 0;
 	xprt_disconnect_done(xprt);
 	(void) rpcrdma_ep_disconnect(&r_xprt->rx_ep, &r_xprt->rx_ia);
 }
@@ -485,6 +492,11 @@ xprt_rdma_connect(struct rpc_task *task)
 			/* Reconnect */
 			schedule_delayed_work(&r_xprt->rdma_connect,
 				xprt->reestablish_timeout);
+			xprt->reestablish_timeout <<= 1;
+			if (xprt->reestablish_timeout > (30 * HZ))
+				xprt->reestablish_timeout = (30 * HZ);
+			else if (xprt->reestablish_timeout < (5 * HZ))
+				xprt->reestablish_timeout = (5 * HZ);
 		} else {
 			schedule_delayed_work(&r_xprt->rdma_connect, 0);
 			if (!RPC_IS_ASYNC(task))
@@ -591,6 +603,7 @@ xprt_rdma_allocate(struct rpc_task *task, size_t size)
 	}
 	dprintk("RPC:       %s: size %zd, request 0x%p\n", __func__, size, req);
 out:
+	req->rl_connect_cookie = 0;	/* our reserved value */
 	return req->rl_xdr_buf;
 
 outfail:
@@ -694,13 +707,21 @@ xprt_rdma_send_request(struct rpc_task *task)
 		req->rl_reply->rr_xprt = xprt;
 	}
 
-	if (rpcrdma_ep_post(&r_xprt->rx_ia, &r_xprt->rx_ep, req)) {
-		xprt_disconnect_done(xprt);
-		return -ENOTCONN;	/* implies disconnect */
-	}
+	/* Must suppress retransmit to maintain credits */
+	if (req->rl_connect_cookie == xprt->connect_cookie)
+		goto drop_connection;
+	req->rl_connect_cookie = xprt->connect_cookie;
+
+	if (rpcrdma_ep_post(&r_xprt->rx_ia, &r_xprt->rx_ep, req))
+		goto drop_connection;
 
+	task->tk_bytes_sent += rqst->rq_snd_buf.len;
 	rqst->rq_bytes_sent = 0;
 	return 0;
+
+drop_connection:
+	xprt_disconnect_done(xprt);
+	return -ENOTCONN;	/* implies disconnect */
 }
 
 static void xprt_rdma_print_stats(struct rpc_xprt *xprt, struct seq_file *seq)
@@ -770,7 +791,7 @@ static void __exit xprt_rdma_cleanup(void)
 {
 	int rc;
 
-	dprintk("RPCRDMA Module Removed, deregister RPC RDMA transport\n");
+	dprintk(KERN_INFO "RPCRDMA Module Removed, deregister RPC RDMA transport\n");
 #ifdef RPC_DEBUG
 	if (sunrpc_table_header) {
 		unregister_sysctl_table(sunrpc_table_header);

net/sunrpc/xprtrdma/verbs.c

@@ -284,6 +284,7 @@ rpcrdma_conn_upcall(struct rdma_cm_id *id, struct rdma_cm_event *event)
 	switch (event->event) {
 	case RDMA_CM_EVENT_ADDR_RESOLVED:
 	case RDMA_CM_EVENT_ROUTE_RESOLVED:
+		ia->ri_async_rc = 0;
 		complete(&ia->ri_done);
 		break;
 	case RDMA_CM_EVENT_ADDR_ERROR:
@@ -338,13 +339,32 @@ rpcrdma_conn_upcall(struct rdma_cm_id *id, struct rdma_cm_event *event)
 		wake_up_all(&ep->rep_connect_wait);
 		break;
 	default:
-		ia->ri_async_rc = -EINVAL;
-		dprintk("RPC:       %s: unexpected CM event %X\n",
+		dprintk("RPC:       %s: unexpected CM event %d\n",
 			__func__, event->event);
-		complete(&ia->ri_done);
 		break;
 	}
 
+#ifdef RPC_DEBUG
+	if (connstate == 1) {
+		int ird = attr.max_dest_rd_atomic;
+		int tird = ep->rep_remote_cma.responder_resources;
+		printk(KERN_INFO "rpcrdma: connection to %u.%u.%u.%u:%u "
+			"on %s, memreg %d slots %d ird %d%s\n",
+			NIPQUAD(addr->sin_addr.s_addr),
+			ntohs(addr->sin_port),
+			ia->ri_id->device->name,
+			ia->ri_memreg_strategy,
+			xprt->rx_buf.rb_max_requests,
+			ird, ird < 4 && ird < tird / 2 ? " (low!)" : "");
+	} else if (connstate < 0) {
+		printk(KERN_INFO "rpcrdma: connection to %u.%u.%u.%u:%u "
+			"closed (%d)\n",
+			NIPQUAD(addr->sin_addr.s_addr),
+			ntohs(addr->sin_port),
+			connstate);
+	}
+#endif
+
 	return 0;
 }
 
@@ -355,6 +375,8 @@ rpcrdma_create_id(struct rpcrdma_xprt *xprt,
 	struct rdma_cm_id *id;
 	int rc;
 
+	init_completion(&ia->ri_done);
+
 	id = rdma_create_id(rpcrdma_conn_upcall, xprt, RDMA_PS_TCP);
 	if (IS_ERR(id)) {
 		rc = PTR_ERR(id);
@@ -363,26 +385,28 @@ rpcrdma_create_id(struct rpcrdma_xprt *xprt,
 		return id;
 	}
 
-	ia->ri_async_rc = 0;
+	ia->ri_async_rc = -ETIMEDOUT;
 	rc = rdma_resolve_addr(id, NULL, addr, RDMA_RESOLVE_TIMEOUT);
 	if (rc) {
 		dprintk("RPC:       %s: rdma_resolve_addr() failed %i\n",
 			__func__, rc);
 		goto out;
 	}
-	wait_for_completion(&ia->ri_done);
+	wait_for_completion_interruptible_timeout(&ia->ri_done,
+				msecs_to_jiffies(RDMA_RESOLVE_TIMEOUT) + 1);
 	rc = ia->ri_async_rc;
 	if (rc)
 		goto out;
 
-	ia->ri_async_rc = 0;
+	ia->ri_async_rc = -ETIMEDOUT;
 	rc = rdma_resolve_route(id, RDMA_RESOLVE_TIMEOUT);
 	if (rc) {
 		dprintk("RPC:       %s: rdma_resolve_route() failed %i\n",
 			__func__, rc);
 		goto out;
 	}
-	wait_for_completion(&ia->ri_done);
+	wait_for_completion_interruptible_timeout(&ia->ri_done,
+				msecs_to_jiffies(RDMA_RESOLVE_TIMEOUT) + 1);
 	rc = ia->ri_async_rc;
 	if (rc)
 		goto out;
@@ -423,11 +447,10 @@ rpcrdma_clean_cq(struct ib_cq *cq)
 int
 rpcrdma_ia_open(struct rpcrdma_xprt *xprt, struct sockaddr *addr, int memreg)
 {
-	int rc;
+	int rc, mem_priv;
+	struct ib_device_attr devattr;
 	struct rpcrdma_ia *ia = &xprt->rx_ia;
 
-	init_completion(&ia->ri_done);
-
 	ia->ri_id = rpcrdma_create_id(xprt, ia, addr);
 	if (IS_ERR(ia->ri_id)) {
 		rc = PTR_ERR(ia->ri_id);
@@ -442,6 +465,73 @@ rpcrdma_ia_open(struct rpcrdma_xprt *xprt, struct sockaddr *addr, int memreg)
 		goto out2;
 	}
 
+	/*
+	 * Query the device to determine if the requested memory
+	 * registration strategy is supported. If it isn't, set the
+	 * strategy to a globally supported model.
+	 */
+	rc = ib_query_device(ia->ri_id->device, &devattr);
+	if (rc) {
+		dprintk("RPC:       %s: ib_query_device failed %d\n",
+			__func__, rc);
+		goto out2;
+	}
+
+	if (devattr.device_cap_flags & IB_DEVICE_LOCAL_DMA_LKEY) {
+		ia->ri_have_dma_lkey = 1;
+		ia->ri_dma_lkey = ia->ri_id->device->local_dma_lkey;
+	}
+
+	switch (memreg) {
+	case RPCRDMA_MEMWINDOWS:
+	case RPCRDMA_MEMWINDOWS_ASYNC:
+		if (!(devattr.device_cap_flags & IB_DEVICE_MEM_WINDOW)) {
+			dprintk("RPC:       %s: MEMWINDOWS registration "
+				"specified but not supported by adapter, "
+				"using slower RPCRDMA_REGISTER\n",
+				__func__);
+			memreg = RPCRDMA_REGISTER;
+		}
+		break;
+	case RPCRDMA_MTHCAFMR:
+		if (!ia->ri_id->device->alloc_fmr) {
+#if RPCRDMA_PERSISTENT_REGISTRATION
+			dprintk("RPC:       %s: MTHCAFMR registration "
+				"specified but not supported by adapter, "
+				"using riskier RPCRDMA_ALLPHYSICAL\n",
+				__func__);
+			memreg = RPCRDMA_ALLPHYSICAL;
+#else
+			dprintk("RPC:       %s: MTHCAFMR registration "
+				"specified but not supported by adapter, "
+				"using slower RPCRDMA_REGISTER\n",
+				__func__);
+			memreg = RPCRDMA_REGISTER;
+#endif
+		}
+		break;
+	case RPCRDMA_FRMR:
+		/* Requires both frmr reg and local dma lkey */
+		if ((devattr.device_cap_flags &
+		     (IB_DEVICE_MEM_MGT_EXTENSIONS|IB_DEVICE_LOCAL_DMA_LKEY)) !=
+		    (IB_DEVICE_MEM_MGT_EXTENSIONS|IB_DEVICE_LOCAL_DMA_LKEY)) {
+#if RPCRDMA_PERSISTENT_REGISTRATION
+			dprintk("RPC:       %s: FRMR registration "
+				"specified but not supported by adapter, "
+				"using riskier RPCRDMA_ALLPHYSICAL\n",
+				__func__);
+			memreg = RPCRDMA_ALLPHYSICAL;
+#else
+			dprintk("RPC:       %s: FRMR registration "
+				"specified but not supported by adapter, "
+				"using slower RPCRDMA_REGISTER\n",
+				__func__);
+			memreg = RPCRDMA_REGISTER;
+#endif
+		}
+		break;
+	}
+
 	/*
 	 * Optionally obtain an underlying physical identity mapping in
 	 * order to do a memory window-based bind. This base registration
@@ -450,22 +540,28 @@ rpcrdma_ia_open(struct rpcrdma_xprt *xprt, struct sockaddr *addr, int memreg)
 	 * revoked after the corresponding completion similar to a storage
 	 * adapter.
 	 */
-	if (memreg > RPCRDMA_REGISTER) {
-		int mem_priv = IB_ACCESS_LOCAL_WRITE;
 	switch (memreg) {
+	case RPCRDMA_BOUNCEBUFFERS:
+	case RPCRDMA_REGISTER:
+	case RPCRDMA_FRMR:
+		break;
 #if RPCRDMA_PERSISTENT_REGISTRATION
 	case RPCRDMA_ALLPHYSICAL:
-			mem_priv |= IB_ACCESS_REMOTE_WRITE;
-			mem_priv |= IB_ACCESS_REMOTE_READ;
-			break;
+		mem_priv = IB_ACCESS_LOCAL_WRITE |
+				IB_ACCESS_REMOTE_WRITE |
+				IB_ACCESS_REMOTE_READ;
+		goto register_setup;
 #endif
 	case RPCRDMA_MEMWINDOWS_ASYNC:
 	case RPCRDMA_MEMWINDOWS:
-			mem_priv |= IB_ACCESS_MW_BIND;
-			break;
-		default:
+		mem_priv = IB_ACCESS_LOCAL_WRITE |
+				IB_ACCESS_MW_BIND;
+		goto register_setup;
+	case RPCRDMA_MTHCAFMR:
+		if (ia->ri_have_dma_lkey)
 			break;
-		}
+		mem_priv = IB_ACCESS_LOCAL_WRITE;
+	register_setup:
 		ia->ri_bind_mem = ib_get_dma_mr(ia->ri_pd, mem_priv);
 		if (IS_ERR(ia->ri_bind_mem)) {
 			printk(KERN_ALERT "%s: ib_get_dma_mr for "
@@ -475,7 +571,15 @@ rpcrdma_ia_open(struct rpcrdma_xprt *xprt, struct sockaddr *addr, int memreg)
 			memreg = RPCRDMA_REGISTER;
 			ia->ri_bind_mem = NULL;
 		}
+		break;
+	default:
+		printk(KERN_ERR "%s: invalid memory registration mode %d\n",
+				__func__, memreg);
+		rc = -EINVAL;
+		goto out2;
 	}
+	dprintk("RPC:       %s: memory registration strategy is %d\n",
+		__func__, memreg);
 
 	/* Else will do memory reg/dereg for each chunk */
 	ia->ri_memreg_strategy = memreg;
@@ -483,6 +587,7 @@ rpcrdma_ia_open(struct rpcrdma_xprt *xprt, struct sockaddr *addr, int memreg)
 	return 0;
 out2:
 	rdma_destroy_id(ia->ri_id);
+	ia->ri_id = NULL;
 out1:
 	return rc;
 }
@@ -503,15 +608,17 @@ rpcrdma_ia_close(struct rpcrdma_ia *ia)
 		dprintk("RPC:       %s: ib_dereg_mr returned %i\n",
 			__func__, rc);
 	}
-	if (ia->ri_id != NULL && !IS_ERR(ia->ri_id) && ia->ri_id->qp)
+	if (ia->ri_id != NULL && !IS_ERR(ia->ri_id)) {
+		if (ia->ri_id->qp)
 			rdma_destroy_qp(ia->ri_id);
+		rdma_destroy_id(ia->ri_id);
+		ia->ri_id = NULL;
+	}
 	if (ia->ri_pd != NULL && !IS_ERR(ia->ri_pd)) {
 		rc = ib_dealloc_pd(ia->ri_pd);
 		dprintk("RPC:       %s: ib_dealloc_pd returned %i\n",
 			__func__, rc);
 	}
-	if (ia->ri_id != NULL && !IS_ERR(ia->ri_id))
-		rdma_destroy_id(ia->ri_id);
 }
 
 /*
@@ -541,6 +648,12 @@ rpcrdma_ep_create(struct rpcrdma_ep *ep, struct rpcrdma_ia *ia,
 	ep->rep_attr.srq = NULL;
 	ep->rep_attr.cap.max_send_wr = cdata->max_requests;
 	switch (ia->ri_memreg_strategy) {
+	case RPCRDMA_FRMR:
+		/* Add room for frmr register and invalidate WRs */
+		ep->rep_attr.cap.max_send_wr *= 3;
+		if (ep->rep_attr.cap.max_send_wr > devattr.max_qp_wr)
+			return -EINVAL;
+		break;
 	case RPCRDMA_MEMWINDOWS_ASYNC:
 	case RPCRDMA_MEMWINDOWS:
 		/* Add room for mw_binds+unbinds - overkill! */
@@ -617,29 +730,13 @@ rpcrdma_ep_create(struct rpcrdma_ep *ep, struct rpcrdma_ia *ia,
 	ep->rep_remote_cma.private_data_len = 0;
 
 	/* Client offers RDMA Read but does not initiate */
-	switch (ia->ri_memreg_strategy) {
-	case RPCRDMA_BOUNCEBUFFERS:
+	ep->rep_remote_cma.initiator_depth = 0;
+	if (ia->ri_memreg_strategy == RPCRDMA_BOUNCEBUFFERS)
 		ep->rep_remote_cma.responder_resources = 0;
-		break;
-	case RPCRDMA_MTHCAFMR:
-	case RPCRDMA_REGISTER:
-		ep->rep_remote_cma.responder_resources = cdata->max_requests *
-				(RPCRDMA_MAX_DATA_SEGS / 8);
-		break;
-	case RPCRDMA_MEMWINDOWS:
-	case RPCRDMA_MEMWINDOWS_ASYNC:
-#if RPCRDMA_PERSISTENT_REGISTRATION
-	case RPCRDMA_ALLPHYSICAL:
-#endif
-		ep->rep_remote_cma.responder_resources = cdata->max_requests *
-				(RPCRDMA_MAX_DATA_SEGS / 2);
-		break;
-	default:
-		break;
-	}
-	if (ep->rep_remote_cma.responder_resources > devattr.max_qp_rd_atom)
+	else if (devattr.max_qp_rd_atom > 32)	/* arbitrary but <= 255 */
+		ep->rep_remote_cma.responder_resources = 32;
+	else
 		ep->rep_remote_cma.responder_resources = devattr.max_qp_rd_atom;
-	ep->rep_remote_cma.initiator_depth = 0;
 
 	ep->rep_remote_cma.retry_count = 7;
 	ep->rep_remote_cma.flow_control = 0;
@@ -679,21 +776,16 @@ rpcrdma_ep_destroy(struct rpcrdma_ep *ep, struct rpcrdma_ia *ia)
 		if (rc)
 			dprintk("RPC:       %s: rpcrdma_ep_disconnect"
 				" returned %i\n", __func__, rc);
+		rdma_destroy_qp(ia->ri_id);
+		ia->ri_id->qp = NULL;
 	}
 
-	ep->rep_func = NULL;
-
 	/* padding - could be done in rpcrdma_buffer_destroy... */
 	if (ep->rep_pad_mr) {
 		rpcrdma_deregister_internal(ia, ep->rep_pad_mr, &ep->rep_pad);
 		ep->rep_pad_mr = NULL;
 	}
 
-	if (ia->ri_id->qp) {
-		rdma_destroy_qp(ia->ri_id);
-		ia->ri_id->qp = NULL;
-	}
-
 	rpcrdma_clean_cq(ep->rep_cq);
 	rc = ib_destroy_cq(ep->rep_cq);
 	if (rc)
@@ -712,9 +804,8 @@ rpcrdma_ep_connect(struct rpcrdma_ep *ep, struct rpcrdma_ia *ia)
 	struct rdma_cm_id *id;
 	int rc = 0;
 	int retry_count = 0;
-	int reconnect = (ep->rep_connected != 0);
 
-	if (reconnect) {
+	if (ep->rep_connected != 0) {
 		struct rpcrdma_xprt *xprt;
 retry:
 		rc = rpcrdma_ep_disconnect(ep, ia);
@@ -745,6 +836,7 @@ rpcrdma_ep_connect(struct rpcrdma_ep *ep, struct rpcrdma_ia *ia)
 			goto out;
 		}
 		/* END TEMP */
+		rdma_destroy_qp(ia->ri_id);
 		rdma_destroy_id(ia->ri_id);
 		ia->ri_id = id;
 	}
@@ -769,14 +861,6 @@ if (strnicmp(ia->ri_id->device->dma_device->bus->name, "pci", 3) == 0) {
 	}
 }
 
-	/* Theoretically a client initiator_depth > 0 is not needed,
-	 * but many peers fail to complete the connection unless they
-	 * == responder_resources! */
-	if (ep->rep_remote_cma.initiator_depth !=
-				ep->rep_remote_cma.responder_resources)
-		ep->rep_remote_cma.initiator_depth =
-			ep->rep_remote_cma.responder_resources;
-
 	ep->rep_connected = 0;
 
 	rc = rdma_connect(ia->ri_id, &ep->rep_remote_cma);
@@ -786,9 +870,6 @@ if (strnicmp(ia->ri_id->device->dma_device->bus->name, "pci", 3) == 0) {
 		goto out;
 	}
 
-	if (reconnect)
-		return 0;
-
 	wait_event_interruptible(ep->rep_connect_wait, ep->rep_connected != 0);
 
 	/*
@@ -805,14 +886,16 @@ if (strnicmp(ia->ri_id->device->dma_device->bus->name, "pci", 3) == 0) {
 	if (ep->rep_connected <= 0) {
 		/* Sometimes, the only way to reliably connect to remote
 		 * CMs is to use same nonzero values for ORD and IRD. */
+		if (retry_count++ <= RDMA_CONNECT_RETRY_MAX + 1 &&
+		    (ep->rep_remote_cma.responder_resources == 0 ||
+		     ep->rep_remote_cma.initiator_depth !=
+				ep->rep_remote_cma.responder_resources)) {
+			if (ep->rep_remote_cma.responder_resources == 0)
+				ep->rep_remote_cma.responder_resources = 1;
 			ep->rep_remote_cma.initiator_depth =
 				ep->rep_remote_cma.responder_resources;
-		if (ep->rep_remote_cma.initiator_depth == 0)
-			++ep->rep_remote_cma.initiator_depth;
-		if (ep->rep_remote_cma.responder_resources == 0)
-			++ep->rep_remote_cma.responder_resources;
-		if (retry_count++ == 0)
 			goto retry;
+		}
 		rc = ep->rep_connected;
 	} else {
 		dprintk("RPC:       %s: connected\n", __func__);
@@ -863,6 +946,7 @@ rpcrdma_buffer_create(struct rpcrdma_buffer *buf, struct rpcrdma_ep *ep,
 	char *p;
 	size_t len;
 	int i, rc;
+	struct rpcrdma_mw *r;
 
 	buf->rb_max_requests = cdata->max_requests;
 	spin_lock_init(&buf->rb_lock);
@@ -873,7 +957,7 @@ rpcrdma_buffer_create(struct rpcrdma_buffer *buf, struct rpcrdma_ep *ep,
 	 *   2.  arrays of struct rpcrdma_req to fill in pointers
 	 *   3.  array of struct rpcrdma_rep for replies
 	 *   4.  padding, if any
-	 *   5.  mw's, if any
+	 *   5.  mw's, fmr's or frmr's, if any
 	 * Send/recv buffers in req/rep need to be registered
 	 */
 
@@ -881,6 +965,10 @@ rpcrdma_buffer_create(struct rpcrdma_buffer *buf, struct rpcrdma_ep *ep,
 		(sizeof(struct rpcrdma_req *) + sizeof(struct rpcrdma_rep *));
 	len += cdata->padding;
 	switch (ia->ri_memreg_strategy) {
+	case RPCRDMA_FRMR:
+		len += buf->rb_max_requests * RPCRDMA_MAX_SEGS *
+				sizeof(struct rpcrdma_mw);
+		break;
 	case RPCRDMA_MTHCAFMR:
 		/* TBD we are perhaps overallocating here */
 		len += (buf->rb_max_requests + 1) * RPCRDMA_MAX_SEGS *
@@ -927,15 +1015,37 @@ rpcrdma_buffer_create(struct rpcrdma_buffer *buf, struct rpcrdma_ep *ep,
 	 * and also reduce unbind-to-bind collision.
 	 */
 	INIT_LIST_HEAD(&buf->rb_mws);
+	r = (struct rpcrdma_mw *)p;
 	switch (ia->ri_memreg_strategy) {
+	case RPCRDMA_FRMR:
+		for (i = buf->rb_max_requests * RPCRDMA_MAX_SEGS; i; i--) {
+			r->r.frmr.fr_mr = ib_alloc_fast_reg_mr(ia->ri_pd,
+							 RPCRDMA_MAX_SEGS);
+			if (IS_ERR(r->r.frmr.fr_mr)) {
+				rc = PTR_ERR(r->r.frmr.fr_mr);
+				dprintk("RPC:       %s: ib_alloc_fast_reg_mr"
+					" failed %i\n", __func__, rc);
+				goto out;
+			}
+			r->r.frmr.fr_pgl =
+				ib_alloc_fast_reg_page_list(ia->ri_id->device,
+							    RPCRDMA_MAX_SEGS);
+			if (IS_ERR(r->r.frmr.fr_pgl)) {
+				rc = PTR_ERR(r->r.frmr.fr_pgl);
+				dprintk("RPC:       %s: "
+					"ib_alloc_fast_reg_page_list "
+					"failed %i\n", __func__, rc);
+				goto out;
+			}
+			list_add(&r->mw_list, &buf->rb_mws);
+			++r;
+		}
+		break;
 	case RPCRDMA_MTHCAFMR:
-		{
-		struct rpcrdma_mw *r = (struct rpcrdma_mw *)p;
-		struct ib_fmr_attr fa = {
-			RPCRDMA_MAX_DATA_SEGS, 1, PAGE_SHIFT
-		};
 		/* TBD we are perhaps overallocating here */
 		for (i = (buf->rb_max_requests+1) * RPCRDMA_MAX_SEGS; i; i--) {
+			static struct ib_fmr_attr fa =
+				{ RPCRDMA_MAX_DATA_SEGS, 1, PAGE_SHIFT };
 			r->r.fmr = ib_alloc_fmr(ia->ri_pd,
 				IB_ACCESS_REMOTE_WRITE | IB_ACCESS_REMOTE_READ,
 				&fa);
@@ -948,12 +1058,9 @@ rpcrdma_buffer_create(struct rpcrdma_buffer *buf, struct rpcrdma_ep *ep,
 			list_add(&r->mw_list, &buf->rb_mws);
 			++r;
 		}
-		}
 		break;
 	case RPCRDMA_MEMWINDOWS_ASYNC:
 	case RPCRDMA_MEMWINDOWS:
-		{
-		struct rpcrdma_mw *r = (struct rpcrdma_mw *)p;
 		/* Allocate one extra request's worth, for full cycling */
 		for (i = (buf->rb_max_requests+1) * RPCRDMA_MAX_SEGS; i; i--) {
 			r->r.mw = ib_alloc_mw(ia->ri_pd);
@@ -966,7 +1073,6 @@ rpcrdma_buffer_create(struct rpcrdma_buffer *buf, struct rpcrdma_ep *ep,
 			list_add(&r->mw_list, &buf->rb_mws);
 			++r;
 		}
-		}
 		break;
 	default:
 		break;
@@ -1046,6 +1152,7 @@ rpcrdma_buffer_destroy(struct rpcrdma_buffer *buf)
 {
 	int rc, i;
 	struct rpcrdma_ia *ia = rdmab_to_ia(buf);
+	struct rpcrdma_mw *r;
 
 	/* clean up in reverse order from create
 	 *   1.  recv mr memory (mr free, then kfree)
@@ -1065,11 +1172,19 @@ rpcrdma_buffer_destroy(struct rpcrdma_buffer *buf)
 		}
 		if (buf->rb_send_bufs && buf->rb_send_bufs[i]) {
 			while (!list_empty(&buf->rb_mws)) {
-				struct rpcrdma_mw *r;
 				r = list_entry(buf->rb_mws.next,
 					struct rpcrdma_mw, mw_list);
 				list_del(&r->mw_list);
 				switch (ia->ri_memreg_strategy) {
+				case RPCRDMA_FRMR:
+					rc = ib_dereg_mr(r->r.frmr.fr_mr);
+					if (rc)
+						dprintk("RPC:       %s:"
+							" ib_dereg_mr"
+							" failed %i\n",
+							__func__, rc);
+					ib_free_fast_reg_page_list(r->r.frmr.fr_pgl);
+					break;
 				case RPCRDMA_MTHCAFMR:
 					rc = ib_dealloc_fmr(r->r.fmr);
 					if (rc)
@@ -1115,6 +1230,8 @@ rpcrdma_buffer_get(struct rpcrdma_buffer *buffers)
 {
 	struct rpcrdma_req *req;
 	unsigned long flags;
+	int i;
+	struct rpcrdma_mw *r;
 
 	spin_lock_irqsave(&buffers->rb_lock, flags);
 	if (buffers->rb_send_index == buffers->rb_max_requests) {
@@ -1135,9 +1252,8 @@ rpcrdma_buffer_get(struct rpcrdma_buffer *buffers)
 	}
 	buffers->rb_send_bufs[buffers->rb_send_index++] = NULL;
 	if (!list_empty(&buffers->rb_mws)) {
-		int i = RPCRDMA_MAX_SEGS - 1;
+		i = RPCRDMA_MAX_SEGS - 1;
 		do {
-			struct rpcrdma_mw *r;
 			r = list_entry(buffers->rb_mws.next,
 					struct rpcrdma_mw, mw_list);
 			list_del(&r->mw_list);
@@ -1171,6 +1287,7 @@ rpcrdma_buffer_put(struct rpcrdma_req *req)
 		req->rl_reply = NULL;
 	}
 	switch (ia->ri_memreg_strategy) {
+	case RPCRDMA_FRMR:
 	case RPCRDMA_MTHCAFMR:
 	case RPCRDMA_MEMWINDOWS_ASYNC:
 	case RPCRDMA_MEMWINDOWS:
@@ -1252,7 +1369,11 @@ rpcrdma_register_internal(struct rpcrdma_ia *ia, void *va, int len,
 			va, len, DMA_BIDIRECTIONAL);
 	iov->length = len;
 
-	if (ia->ri_bind_mem != NULL) {
+	if (ia->ri_have_dma_lkey) {
+		*mrp = NULL;
+		iov->lkey = ia->ri_dma_lkey;
+		return 0;
+	} else if (ia->ri_bind_mem != NULL) {
 		*mrp = NULL;
 		iov->lkey = ia->ri_bind_mem->lkey;
 		return 0;
@@ -1329,74 +1450,169 @@ rpcrdma_unmap_one(struct rpcrdma_ia *ia, struct rpcrdma_mr_seg *seg)
 				seg->mr_dma, seg->mr_dmalen, seg->mr_dir);
 }
 
-int
-rpcrdma_register_external(struct rpcrdma_mr_seg *seg,
-			int nsegs, int writing, struct rpcrdma_xprt *r_xprt)
+static int
+rpcrdma_register_frmr_external(struct rpcrdma_mr_seg *seg,
+			int *nsegs, int writing, struct rpcrdma_ia *ia,
+			struct rpcrdma_xprt *r_xprt)
 {
-	struct rpcrdma_ia *ia = &r_xprt->rx_ia;
-	int mem_priv = (writing ? IB_ACCESS_REMOTE_WRITE :
-				  IB_ACCESS_REMOTE_READ);
 	struct rpcrdma_mr_seg *seg1 = seg;
-	int i;
-	int rc = 0;
-
-	switch (ia->ri_memreg_strategy) {
+	struct ib_send_wr frmr_wr, *bad_wr;
+	u8 key;
+	int len, pageoff;
+	int i, rc;
 
-#if RPCRDMA_PERSISTENT_REGISTRATION
-	case RPCRDMA_ALLPHYSICAL:
+	pageoff = offset_in_page(seg1->mr_offset);
+	seg1->mr_offset -= pageoff;	/* start of page */
+	seg1->mr_len += pageoff;
+	len = -pageoff;
+	if (*nsegs > RPCRDMA_MAX_DATA_SEGS)
+		*nsegs = RPCRDMA_MAX_DATA_SEGS;
+	for (i = 0; i < *nsegs;) {
 		rpcrdma_map_one(ia, seg, writing);
-		seg->mr_rkey = ia->ri_bind_mem->rkey;
-		seg->mr_base = seg->mr_dma;
-		seg->mr_nsegs = 1;
-		nsegs = 1;
+		seg1->mr_chunk.rl_mw->r.frmr.fr_pgl->page_list[i] = seg->mr_dma;
+		len += seg->mr_len;
+		++seg;
+		++i;
+		/* Check for holes */
+		if ((i < *nsegs && offset_in_page(seg->mr_offset)) ||
+		    offset_in_page((seg-1)->mr_offset + (seg-1)->mr_len))
 			break;
-#endif
+	}
+	dprintk("RPC:       %s: Using frmr %p to map %d segments\n",
+		__func__, seg1->mr_chunk.rl_mw, i);
+
+	/* Bump the key */
+	key = (u8)(seg1->mr_chunk.rl_mw->r.frmr.fr_mr->rkey & 0x000000FF);
+	ib_update_fast_reg_key(seg1->mr_chunk.rl_mw->r.frmr.fr_mr, ++key);
+
+	/* Prepare FRMR WR */
+	memset(&frmr_wr, 0, sizeof frmr_wr);
+	frmr_wr.opcode = IB_WR_FAST_REG_MR;
+	frmr_wr.send_flags = 0;			/* unsignaled */
+	frmr_wr.wr.fast_reg.iova_start = (unsigned long)seg1->mr_dma;
+	frmr_wr.wr.fast_reg.page_list = seg1->mr_chunk.rl_mw->r.frmr.fr_pgl;
+	frmr_wr.wr.fast_reg.page_list_len = i;
+	frmr_wr.wr.fast_reg.page_shift = PAGE_SHIFT;
+	frmr_wr.wr.fast_reg.length = i << PAGE_SHIFT;
+	frmr_wr.wr.fast_reg.access_flags = (writing ?
+				IB_ACCESS_REMOTE_WRITE : IB_ACCESS_REMOTE_READ);
+	frmr_wr.wr.fast_reg.rkey = seg1->mr_chunk.rl_mw->r.frmr.fr_mr->rkey;
+	DECR_CQCOUNT(&r_xprt->rx_ep);
 
-	/* Registration using fast memory registration */
-	case RPCRDMA_MTHCAFMR:
-		{
+	rc = ib_post_send(ia->ri_id->qp, &frmr_wr, &bad_wr);
+
+	if (rc) {
+		dprintk("RPC:       %s: failed ib_post_send for register,"
+			" status %i\n", __func__, rc);
+		while (i--)
+			rpcrdma_unmap_one(ia, --seg);
+	} else {
+		seg1->mr_rkey = seg1->mr_chunk.rl_mw->r.frmr.fr_mr->rkey;
+		seg1->mr_base = seg1->mr_dma + pageoff;
+		seg1->mr_nsegs = i;
+		seg1->mr_len = len;
+	}
+	*nsegs = i;
+	return rc;
+}
+
+static int
+rpcrdma_deregister_frmr_external(struct rpcrdma_mr_seg *seg,
+			struct rpcrdma_ia *ia, struct rpcrdma_xprt *r_xprt)
+{
+	struct rpcrdma_mr_seg *seg1 = seg;
+	struct ib_send_wr invalidate_wr, *bad_wr;
+	int rc;
+
+	while (seg1->mr_nsegs--)
+		rpcrdma_unmap_one(ia, seg++);
+
+	memset(&invalidate_wr, 0, sizeof invalidate_wr);
+	invalidate_wr.opcode = IB_WR_LOCAL_INV;
+	invalidate_wr.send_flags = 0;			/* unsignaled */
+	invalidate_wr.ex.invalidate_rkey = seg1->mr_chunk.rl_mw->r.frmr.fr_mr->rkey;
+	DECR_CQCOUNT(&r_xprt->rx_ep);
+
+	rc = ib_post_send(ia->ri_id->qp, &invalidate_wr, &bad_wr);
+	if (rc)
+		dprintk("RPC:       %s: failed ib_post_send for invalidate,"
+			" status %i\n", __func__, rc);
+	return rc;
+}
+
+static int
+rpcrdma_register_fmr_external(struct rpcrdma_mr_seg *seg,
+			int *nsegs, int writing, struct rpcrdma_ia *ia)
+{
+	struct rpcrdma_mr_seg *seg1 = seg;
 	u64 physaddrs[RPCRDMA_MAX_DATA_SEGS];
-		int len, pageoff = offset_in_page(seg->mr_offset);
+	int len, pageoff, i, rc;
+
+	pageoff = offset_in_page(seg1->mr_offset);
 	seg1->mr_offset -= pageoff;	/* start of page */
 	seg1->mr_len += pageoff;
 	len = -pageoff;
-		if (nsegs > RPCRDMA_MAX_DATA_SEGS)
-			nsegs = RPCRDMA_MAX_DATA_SEGS;
-		for (i = 0; i < nsegs;) {
+	if (*nsegs > RPCRDMA_MAX_DATA_SEGS)
+		*nsegs = RPCRDMA_MAX_DATA_SEGS;
+	for (i = 0; i < *nsegs;) {
 		rpcrdma_map_one(ia, seg, writing);
 		physaddrs[i] = seg->mr_dma;
 		len += seg->mr_len;
 		++seg;
 		++i;
 		/* Check for holes */
-			if ((i < nsegs && offset_in_page(seg->mr_offset)) ||
-			    offset_in_page((seg-1)->mr_offset+(seg-1)->mr_len))
+		if ((i < *nsegs && offset_in_page(seg->mr_offset)) ||
+		    offset_in_page((seg-1)->mr_offset + (seg-1)->mr_len))
 			break;
 	}
-		nsegs = i;
 	rc = ib_map_phys_fmr(seg1->mr_chunk.rl_mw->r.fmr,
-					physaddrs, nsegs, seg1->mr_dma);
+				physaddrs, i, seg1->mr_dma);
 	if (rc) {
 		dprintk("RPC:       %s: failed ib_map_phys_fmr "
 			"%u@0x%llx+%i (%d)... status %i\n", __func__,
 			len, (unsigned long long)seg1->mr_dma,
-				pageoff, nsegs, rc);
-			while (nsegs--)
+			pageoff, i, rc);
+		while (i--)
 			rpcrdma_unmap_one(ia, --seg);
 	} else {
 		seg1->mr_rkey = seg1->mr_chunk.rl_mw->r.fmr->rkey;
 		seg1->mr_base = seg1->mr_dma + pageoff;
-			seg1->mr_nsegs = nsegs;
+		seg1->mr_nsegs = i;
 		seg1->mr_len = len;
 	}
-		}
-		break;
+	*nsegs = i;
+	return rc;
+}
 
-	/* Registration using memory windows */
-	case RPCRDMA_MEMWINDOWS_ASYNC:
-	case RPCRDMA_MEMWINDOWS:
-		{
+static int
+rpcrdma_deregister_fmr_external(struct rpcrdma_mr_seg *seg,
+			struct rpcrdma_ia *ia)
+{
+	struct rpcrdma_mr_seg *seg1 = seg;
+	LIST_HEAD(l);
+	int rc;
+
+	list_add(&seg1->mr_chunk.rl_mw->r.fmr->list, &l);
+	rc = ib_unmap_fmr(&l);
+	while (seg1->mr_nsegs--)
+		rpcrdma_unmap_one(ia, seg++);
+	if (rc)
+		dprintk("RPC:       %s: failed ib_unmap_fmr,"
+			" status %i\n", __func__, rc);
+	return rc;
+}
+
+static int
+rpcrdma_register_memwin_external(struct rpcrdma_mr_seg *seg,
+			int *nsegs, int writing, struct rpcrdma_ia *ia,
+			struct rpcrdma_xprt *r_xprt)
+{
+	int mem_priv = (writing ? IB_ACCESS_REMOTE_WRITE :
+				  IB_ACCESS_REMOTE_READ);
 	struct ib_mw_bind param;
+	int rc;
+
+	*nsegs = 1;
 	rpcrdma_map_one(ia, seg, writing);
 	param.mr = ia->ri_bind_mem;
 	param.wr_id = 0ULL;	/* no send cookie */
@@ -1406,8 +1622,7 @@ rpcrdma_register_external(struct rpcrdma_mr_seg *seg,
 	param.mw_access_flags = mem_priv;
 
 	DECR_CQCOUNT(&r_xprt->rx_ep);
-		rc = ib_bind_mw(ia->ri_id->qp,
-					seg->mr_chunk.rl_mw->r.mw, &param);
+	rc = ib_bind_mw(ia->ri_id->qp, seg->mr_chunk.rl_mw->r.mw, &param);
 	if (rc) {
 		dprintk("RPC:       %s: failed ib_bind_mw "
 			"%u@0x%llx status %i\n",
@@ -1418,19 +1633,56 @@ rpcrdma_register_external(struct rpcrdma_mr_seg *seg,
 		seg->mr_rkey = seg->mr_chunk.rl_mw->r.mw->rkey;
 		seg->mr_base = param.addr;
 		seg->mr_nsegs = 1;
-			nsegs = 1;
 	}
+	return rc;
+}
+
+static int
+rpcrdma_deregister_memwin_external(struct rpcrdma_mr_seg *seg,
+			struct rpcrdma_ia *ia,
+			struct rpcrdma_xprt *r_xprt, void **r)
+{
+	struct ib_mw_bind param;
+	LIST_HEAD(l);
+	int rc;
+
+	BUG_ON(seg->mr_nsegs != 1);
+	param.mr = ia->ri_bind_mem;
+	param.addr = 0ULL;	/* unbind */
+	param.length = 0;
+	param.mw_access_flags = 0;
+	if (*r) {
+		param.wr_id = (u64) (unsigned long) *r;
+		param.send_flags = IB_SEND_SIGNALED;
+		INIT_CQCOUNT(&r_xprt->rx_ep);
+	} else {
+		param.wr_id = 0ULL;
+		param.send_flags = 0;
+		DECR_CQCOUNT(&r_xprt->rx_ep);
 	}
-		break;
+	rc = ib_bind_mw(ia->ri_id->qp, seg->mr_chunk.rl_mw->r.mw, &param);
+	rpcrdma_unmap_one(ia, seg);
+	if (rc)
+		dprintk("RPC:       %s: failed ib_(un)bind_mw,"
+			" status %i\n", __func__, rc);
+	else
+		*r = NULL;	/* will upcall on completion */
+	return rc;
+}
 
-	/* Default registration each time */
-	default:
-		{
+static int
+rpcrdma_register_default_external(struct rpcrdma_mr_seg *seg,
+			int *nsegs, int writing, struct rpcrdma_ia *ia)
+{
+	int mem_priv = (writing ? IB_ACCESS_REMOTE_WRITE :
+				  IB_ACCESS_REMOTE_READ);
+	struct rpcrdma_mr_seg *seg1 = seg;
 	struct ib_phys_buf ipb[RPCRDMA_MAX_DATA_SEGS];
-		int len = 0;
-		if (nsegs > RPCRDMA_MAX_DATA_SEGS)
-			nsegs = RPCRDMA_MAX_DATA_SEGS;
-		for (i = 0; i < nsegs;) {
+	int len, i, rc = 0;
+
+	if (*nsegs > RPCRDMA_MAX_DATA_SEGS)
+		*nsegs = RPCRDMA_MAX_DATA_SEGS;
+	for (len = 0, i = 0; i < *nsegs;) {
 		rpcrdma_map_one(ia, seg, writing);
 		ipb[i].addr = seg->mr_dma;
 		ipb[i].size = seg->mr_len;
@@ -1438,28 +1690,85 @@ rpcrdma_register_external(struct rpcrdma_mr_seg *seg,
 		++seg;
 		++i;
 		/* Check for holes */
-			if ((i < nsegs && offset_in_page(seg->mr_offset)) ||
+		if ((i < *nsegs && offset_in_page(seg->mr_offset)) ||
 		    offset_in_page((seg-1)->mr_offset+(seg-1)->mr_len))
 			break;
 	}
-		nsegs = i;
 	seg1->mr_base = seg1->mr_dma;
 	seg1->mr_chunk.rl_mr = ib_reg_phys_mr(ia->ri_pd,
-					ipb, nsegs, mem_priv, &seg1->mr_base);
+				ipb, i, mem_priv, &seg1->mr_base);
 	if (IS_ERR(seg1->mr_chunk.rl_mr)) {
 		rc = PTR_ERR(seg1->mr_chunk.rl_mr);
 		dprintk("RPC:       %s: failed ib_reg_phys_mr "
 			"%u@0x%llx (%d)... status %i\n",
 			__func__, len,
-				(unsigned long long)seg1->mr_dma, nsegs, rc);
-			while (nsegs--)
+			(unsigned long long)seg1->mr_dma, i, rc);
+		while (i--)
 			rpcrdma_unmap_one(ia, --seg);
 	} else {
 		seg1->mr_rkey = seg1->mr_chunk.rl_mr->rkey;
-			seg1->mr_nsegs = nsegs;
+		seg1->mr_nsegs = i;
 		seg1->mr_len = len;
 	}
-		}
+	*nsegs = i;
+	return rc;
+}
+
+static int
+rpcrdma_deregister_default_external(struct rpcrdma_mr_seg *seg,
+			struct rpcrdma_ia *ia)
+{
+	struct rpcrdma_mr_seg *seg1 = seg;
+	int rc;
+
+	rc = ib_dereg_mr(seg1->mr_chunk.rl_mr);
+	seg1->mr_chunk.rl_mr = NULL;
+	while (seg1->mr_nsegs--)
+		rpcrdma_unmap_one(ia, seg++);
+	if (rc)
+		dprintk("RPC:       %s: failed ib_dereg_mr,"
+			" status %i\n", __func__, rc);
+	return rc;
+}
+
+int
+rpcrdma_register_external(struct rpcrdma_mr_seg *seg,
+			int nsegs, int writing, struct rpcrdma_xprt *r_xprt)
+{
+	struct rpcrdma_ia *ia = &r_xprt->rx_ia;
+	int rc = 0;
+
+	switch (ia->ri_memreg_strategy) {
+
+#if RPCRDMA_PERSISTENT_REGISTRATION
+	case RPCRDMA_ALLPHYSICAL:
+		rpcrdma_map_one(ia, seg, writing);
+		seg->mr_rkey = ia->ri_bind_mem->rkey;
+		seg->mr_base = seg->mr_dma;
+		seg->mr_nsegs = 1;
+		nsegs = 1;
+		break;
+#endif
+
+	/* Registration using frmr registration */
+	case RPCRDMA_FRMR:
+		rc = rpcrdma_register_frmr_external(seg, &nsegs, writing, ia, r_xprt);
+		break;
+
+	/* Registration using fmr memory registration */
+	case RPCRDMA_MTHCAFMR:
+		rc = rpcrdma_register_fmr_external(seg, &nsegs, writing, ia);
+		break;
+
+	/* Registration using memory windows */
+	case RPCRDMA_MEMWINDOWS_ASYNC:
+	case RPCRDMA_MEMWINDOWS:
+		rc = rpcrdma_register_memwin_external(seg, &nsegs, writing, ia, r_xprt);
+		break;
+
+	/* Default registration each time */
+	default:
+		rc = rpcrdma_register_default_external(seg, &nsegs, writing, ia);
 		break;
 	}
 	if (rc)
@@ -1473,7 +1782,6 @@ rpcrdma_deregister_external(struct rpcrdma_mr_seg *seg,
 		struct rpcrdma_xprt *r_xprt, void *r)
 {
 	struct rpcrdma_ia *ia = &r_xprt->rx_ia;
-	struct rpcrdma_mr_seg *seg1 = seg;
 	int nsegs = seg->mr_nsegs, rc;
 
 	switch (ia->ri_memreg_strategy) {
@@ -1486,56 +1794,21 @@ rpcrdma_deregister_external(struct rpcrdma_mr_seg *seg,
 		break;
 #endif
 
+	case RPCRDMA_FRMR:
+		rc = rpcrdma_deregister_frmr_external(seg, ia, r_xprt);
+		break;
+
 	case RPCRDMA_MTHCAFMR:
-		{
-		LIST_HEAD(l);
-		list_add(&seg->mr_chunk.rl_mw->r.fmr->list, &l);
-		rc = ib_unmap_fmr(&l);
-		while (seg1->mr_nsegs--)
-			rpcrdma_unmap_one(ia, seg++);
-		}
-		if (rc)
-			dprintk("RPC:       %s: failed ib_unmap_fmr,"
-				" status %i\n", __func__, rc);
+		rc = rpcrdma_deregister_fmr_external(seg, ia);
 		break;
 
 	case RPCRDMA_MEMWINDOWS_ASYNC:
 	case RPCRDMA_MEMWINDOWS:
-		{
-		struct ib_mw_bind param;
-		BUG_ON(nsegs != 1);
-		param.mr = ia->ri_bind_mem;
-		param.addr = 0ULL;	/* unbind */
-		param.length = 0;
-		param.mw_access_flags = 0;
-		if (r) {
-			param.wr_id = (u64) (unsigned long) r;
-			param.send_flags = IB_SEND_SIGNALED;
-			INIT_CQCOUNT(&r_xprt->rx_ep);
-		} else {
-			param.wr_id = 0ULL;
-			param.send_flags = 0;
-			DECR_CQCOUNT(&r_xprt->rx_ep);
-		}
-		rc = ib_bind_mw(ia->ri_id->qp,
-				seg->mr_chunk.rl_mw->r.mw, &param);
-		rpcrdma_unmap_one(ia, seg);
-		}
-		if (rc)
-			dprintk("RPC:       %s: failed ib_(un)bind_mw,"
-				" status %i\n", __func__, rc);
-		else
-			r = NULL;	/* will upcall on completion */
+		rc = rpcrdma_deregister_memwin_external(seg, ia, r_xprt, &r);
 		break;
 
 	default:
-		rc = ib_dereg_mr(seg1->mr_chunk.rl_mr);
-		seg1->mr_chunk.rl_mr = NULL;
-		while (seg1->mr_nsegs--)
-			rpcrdma_unmap_one(ia, seg++);
-		if (rc)
-			dprintk("RPC:       %s: failed ib_dereg_mr,"
-				" status %i\n", __func__, rc);
+		rc = rpcrdma_deregister_default_external(seg, ia);
 		break;
 	}
 	if (r) {

net/sunrpc/xprtrdma/xprt_rdma.h

@@ -51,6 +51,9 @@
 #include <linux/sunrpc/rpc_rdma.h> 	/* RPC/RDMA protocol */
 #include <linux/sunrpc/xprtrdma.h> 	/* xprt parameters */
 
+#define RDMA_RESOLVE_TIMEOUT	(5000)	/* 5 seconds */
+#define RDMA_CONNECT_RETRY_MAX	(2)	/* retries if no listener backlog */
+
 /*
  * Interface Adapter -- one per transport instance
  */
@@ -58,6 +61,8 @@ struct rpcrdma_ia {
 	struct rdma_cm_id 	*ri_id;
 	struct ib_pd		*ri_pd;
 	struct ib_mr		*ri_bind_mem;
+	u32			ri_dma_lkey;
+	int			ri_have_dma_lkey;
 	struct completion	ri_done;
 	int			ri_async_rc;
 	enum rpcrdma_memreg	ri_memreg_strategy;
@@ -156,6 +161,10 @@ struct rpcrdma_mr_seg {		/* chunk descriptors */
 			union {
 				struct ib_mw	*mw;
 				struct ib_fmr	*fmr;
+				struct {
+					struct ib_fast_reg_page_list *fr_pgl;
+					struct ib_mr *fr_mr;
+				} frmr;
 			} r;
 			struct list_head mw_list;
 		} *rl_mw;
@@ -175,6 +184,7 @@ struct rpcrdma_req {
 	size_t 		rl_size;	/* actual length of buffer */
 	unsigned int	rl_niovs;	/* 0, 2 or 4 */
 	unsigned int	rl_nchunks;	/* non-zero if chunks */
+	unsigned int	rl_connect_cookie;	/* retry detection */
 	struct rpcrdma_buffer *rl_buffer; /* home base for this structure */
 	struct rpcrdma_rep	*rl_reply;/* holder for reply buffer */
 	struct rpcrdma_mr_seg rl_segments[RPCRDMA_MAX_SEGS];/* chunk segments */
@@ -198,7 +208,7 @@ struct rpcrdma_buffer {
 	atomic_t	rb_credits;	/* most recent server credits */
 	unsigned long	rb_cwndscale;	/* cached framework rpc_cwndscale */
 	int		rb_max_requests;/* client max requests */
-	struct list_head rb_mws;	/* optional memory windows/fmrs */
+	struct list_head rb_mws;	/* optional memory windows/fmrs/frmrs */
 	int		rb_send_index;
 	struct rpcrdma_req	**rb_send_bufs;
 	int		rb_recv_index;
@@ -273,6 +283,11 @@ struct rpcrdma_xprt {
 #define rpcx_to_rdmax(x) container_of(x, struct rpcrdma_xprt, xprt)
 #define rpcx_to_rdmad(x) (rpcx_to_rdmax(x)->rx_data)
 
+/* Setting this to 0 ensures interoperability with early servers.
+ * Setting this to 1 enhances certain unaligned read/write performance.
+ * Default is 0, see sysctl entry and rpc_rdma.c rpcrdma_convert_iovs() */
+extern int xprt_rdma_pad_optimize;
+
 /*
  * Interface Adapter calls - xprtrdma/verbs.c
  */