Merge tag 'nfs-for-4.2-2' of git://git.linux-nfs.org/projects/trondmy/linux-nfs

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

  Stable patches:
   - Fix a situation where the client uses the wrong (zero) stateid.
   - Fix a memory leak in nfs_do_recoalesce

  Bugfixes:
   - Plug a memory leak when ->prepare_layoutcommit fails
   - Fix an Oops in the NFSv4 open code
   - Fix a backchannel deadlock
   - Fix a livelock in sunrpc when sendmsg fails due to low memory
     availability
   - Don't revalidate the mapping if both size and change attr are up to
     date
   - Ensure we don't miss a file extension when doing pNFS
   - Several fixes to handle NFSv4.1 sequence operation status bits
     correctly
   - Several pNFS layout return bugfixes"

* tag 'nfs-for-4.2-2' of git://git.linux-nfs.org/projects/trondmy/linux-nfs: (28 commits)
  nfs: Fix an oops caused by using other thread's stack space in ASYNC mode
  nfs: plug memory leak when ->prepare_layoutcommit fails
  SUNRPC: Report TCP errors to the caller
  sunrpc: translate -EAGAIN to -ENOBUFS when socket is writable.
  NFSv4.2: handle NFS-specific llseek errors
  NFS: Don't clear desc->pg_moreio in nfs_do_recoalesce()
  NFS: Fix a memory leak in nfs_do_recoalesce
  NFS: nfs_mark_for_revalidate should always set NFS_INO_REVAL_PAGECACHE
  NFS: Remove the "NFS_CAP_CHANGE_ATTR" capability
  NFS: Set NFS_INO_REVAL_PAGECACHE if the change attribute is uninitialised
  NFS: Don't revalidate the mapping if both size and change attr are up to date
  NFSv4/pnfs: Ensure we don't miss a file extension
  NFSv4: We must set NFS_OPEN_STATE flag in nfs_resync_open_stateid_locked
  SUNRPC: xprt_complete_bc_request must also decrement the free slot count
  SUNRPC: Fix a backchannel deadlock
  pNFS: Don't throw out valid layout segments
  pNFS: pnfs_roc_drain() fix a race with open
  pNFS: Fix races between return-on-close and layoutreturn.
  pNFS: pnfs_roc_drain should return 'true' when sleeping
  pNFS: Layoutreturn must invalidate all existing layout segments.
  ...

fs/nfs/client.c

@@ -775,7 +775,7 @@ static int nfs_init_server(struct nfs_server *server,
 	server->options = data->options;
 	server->caps |= NFS_CAP_HARDLINKS|NFS_CAP_SYMLINKS|NFS_CAP_FILEID|
 		NFS_CAP_MODE|NFS_CAP_NLINK|NFS_CAP_OWNER|NFS_CAP_OWNER_GROUP|
-		NFS_CAP_ATIME|NFS_CAP_CTIME|NFS_CAP_MTIME|NFS_CAP_CHANGE_ATTR;
+		NFS_CAP_ATIME|NFS_CAP_CTIME|NFS_CAP_MTIME;
 
 	if (data->rsize)
 		server->rsize = nfs_block_size(data->rsize, NULL);

fs/nfs/flexfilelayout/flexfilelayout.c

@@ -1852,7 +1852,7 @@ ff_layout_mirror_prepare_stats(struct nfs42_layoutstat_args *args,
 	struct nfs42_layoutstat_devinfo *devinfo;
 	int i;
 
-	for (i = 0; i <= FF_LAYOUT_MIRROR_COUNT(pls); i++) {
+	for (i = 0; i < FF_LAYOUT_MIRROR_COUNT(pls); i++) {
 		if (*dev_count >= dev_limit)
 			break;
 		mirror = FF_LAYOUT_COMP(pls, i);

fs/nfs/inode.c

@@ -442,8 +442,9 @@ nfs_fhget(struct super_block *sb, struct nfs_fh *fh, struct nfs_fattr *fattr, st
 			nfs_set_cache_invalid(inode, NFS_INO_INVALID_ATTR);
 		if (fattr->valid & NFS_ATTR_FATTR_CHANGE)
 			inode->i_version = fattr->change_attr;
-		else if (nfs_server_capable(inode, NFS_CAP_CHANGE_ATTR))
-			nfs_set_cache_invalid(inode, NFS_INO_INVALID_ATTR);
+		else
+			nfs_set_cache_invalid(inode, NFS_INO_INVALID_ATTR
+				| NFS_INO_REVAL_PAGECACHE);
 		if (fattr->valid & NFS_ATTR_FATTR_SIZE)
 			inode->i_size = nfs_size_to_loff_t(fattr->size);
 		else
@@ -1244,9 +1245,11 @@ static int nfs_check_inode_attributes(struct inode *inode, struct nfs_fattr *fat
 	if (fattr->valid & NFS_ATTR_FATTR_SIZE) {
 		cur_size = i_size_read(inode);
 		new_isize = nfs_size_to_loff_t(fattr->size);
-		if (cur_size != new_isize && nfsi->nrequests == 0)
+		if (cur_size != new_isize)
 			invalid |= NFS_INO_INVALID_ATTR|NFS_INO_REVAL_PAGECACHE;
 	}
+	if (nfsi->nrequests != 0)
+		invalid &= ~NFS_INO_REVAL_PAGECACHE;
 
 	/* Have any file permissions changed? */
 	if ((fattr->valid & NFS_ATTR_FATTR_MODE) && (inode->i_mode & S_IALLUGO) != (fattr->mode & S_IALLUGO))
@@ -1684,13 +1687,12 @@ static int nfs_update_inode(struct inode *inode, struct nfs_fattr *fattr)
 			invalid |= NFS_INO_INVALID_ATTR
 				| NFS_INO_INVALID_DATA
 				| NFS_INO_INVALID_ACCESS
-				| NFS_INO_INVALID_ACL
-				| NFS_INO_REVAL_PAGECACHE;
+				| NFS_INO_INVALID_ACL;
 			if (S_ISDIR(inode->i_mode))
 				nfs_force_lookup_revalidate(inode);
 			inode->i_version = fattr->change_attr;
 		}
-	} else if (server->caps & NFS_CAP_CHANGE_ATTR)
+	} else
 		nfsi->cache_validity |= save_cache_validity;
 
 	if (fattr->valid & NFS_ATTR_FATTR_MTIME) {
@@ -1717,7 +1719,6 @@ static int nfs_update_inode(struct inode *inode, struct nfs_fattr *fattr)
 			if ((nfsi->nrequests == 0) || new_isize > cur_isize) {
 				i_size_write(inode, new_isize);
 				invalid |= NFS_INO_INVALID_ATTR|NFS_INO_INVALID_DATA;
-				invalid &= ~NFS_INO_REVAL_PAGECACHE;
 			}
 			dprintk("NFS: isize change on server for file %s/%ld "
 					"(%Ld to %Ld)\n",

fs/nfs/internal.h

@@ -296,6 +296,22 @@ extern struct rpc_procinfo nfs4_procedures[];
 
 #ifdef CONFIG_NFS_V4_SECURITY_LABEL
 extern struct nfs4_label *nfs4_label_alloc(struct nfs_server *server, gfp_t flags);
+static inline struct nfs4_label *
+nfs4_label_copy(struct nfs4_label *dst, struct nfs4_label *src)
+{
+	if (!dst || !src)
+		return NULL;
+
+	if (src->len > NFS4_MAXLABELLEN)
+		return NULL;
+
+	dst->lfs = src->lfs;
+	dst->pi = src->pi;
+	dst->len = src->len;
+	memcpy(dst->label, src->label, src->len);
+
+	return dst;
+}
 static inline void nfs4_label_free(struct nfs4_label *label)
 {
 	if (label) {
@@ -316,6 +332,11 @@ static inline void nfs4_label_free(void *label) {}
 static inline void nfs_zap_label_cache_locked(struct nfs_inode *nfsi)
 {
 }
+static inline struct nfs4_label *
+nfs4_label_copy(struct nfs4_label *dst, struct nfs4_label *src)
+{
+	return NULL;
+}
 #endif /* CONFIG_NFS_V4_SECURITY_LABEL */
 
 /* proc.c */

fs/nfs/nfs42proc.c

@@ -135,7 +135,7 @@ int nfs42_proc_deallocate(struct file *filep, loff_t offset, loff_t len)
 	return err;
 }
 
-loff_t nfs42_proc_llseek(struct file *filep, loff_t offset, int whence)
+static loff_t _nfs42_proc_llseek(struct file *filep, loff_t offset, int whence)
 {
 	struct inode *inode = file_inode(filep);
 	struct nfs42_seek_args args = {
@@ -171,6 +171,23 @@ loff_t nfs42_proc_llseek(struct file *filep, loff_t offset, int whence)
 	return vfs_setpos(filep, res.sr_offset, inode->i_sb->s_maxbytes);
 }
 
+loff_t nfs42_proc_llseek(struct file *filep, loff_t offset, int whence)
+{
+	struct nfs_server *server = NFS_SERVER(file_inode(filep));
+	struct nfs4_exception exception = { };
+	int err;
+
+	do {
+		err = _nfs42_proc_llseek(filep, offset, whence);
+		if (err == -ENOTSUPP)
+			return -EOPNOTSUPP;
+		err = nfs4_handle_exception(server, err, &exception);
+	} while (exception.retry);
+
+	return err;
+}
+
+
 static void
 nfs42_layoutstat_prepare(struct rpc_task *task, void *calldata)
 {

fs/nfs/nfs4proc.c

@@ -467,7 +467,10 @@ static void do_renew_lease(struct nfs_client *clp, unsigned long timestamp)
 
 static void renew_lease(const struct nfs_server *server, unsigned long timestamp)
 {
-	do_renew_lease(server->nfs_client, timestamp);
+	struct nfs_client *clp = server->nfs_client;
+
+	if (!nfs4_has_session(clp))
+		do_renew_lease(clp, timestamp);
 }
 
 struct nfs4_call_sync_data {
@@ -616,8 +619,7 @@ int nfs41_sequence_done(struct rpc_task *task, struct nfs4_sequence_res *res)
 		clp = session->clp;
 		do_renew_lease(clp, res->sr_timestamp);
 		/* Check sequence flags */
-		if (res->sr_status_flags != 0)
-			nfs4_schedule_lease_recovery(clp);
+		nfs41_handle_sequence_flag_errors(clp, res->sr_status_flags);
 		nfs41_update_target_slotid(slot->table, slot, res);
 		break;
 	case 1:
@@ -910,6 +912,7 @@ struct nfs4_opendata {
 	struct nfs_open_confirmres c_res;
 	struct nfs4_string owner_name;
 	struct nfs4_string group_name;
+	struct nfs4_label *a_label;
 	struct nfs_fattr f_attr;
 	struct nfs4_label *f_label;
 	struct dentry *dir;
@@ -1013,6 +1016,10 @@ static struct nfs4_opendata *nfs4_opendata_alloc(struct dentry *dentry,
 	if (IS_ERR(p->f_label))
 		goto err_free_p;
 
+	p->a_label = nfs4_label_alloc(server, gfp_mask);
+	if (IS_ERR(p->a_label))
+		goto err_free_f;
+
 	alloc_seqid = server->nfs_client->cl_mvops->alloc_seqid;
 	p->o_arg.seqid = alloc_seqid(&sp->so_seqid, gfp_mask);
 	if (IS_ERR(p->o_arg.seqid))
@@ -1041,7 +1048,7 @@ static struct nfs4_opendata *nfs4_opendata_alloc(struct dentry *dentry,
 	p->o_arg.server = server;
 	p->o_arg.bitmask = nfs4_bitmask(server, label);
 	p->o_arg.open_bitmap = &nfs4_fattr_bitmap[0];
-	p->o_arg.label = label;
+	p->o_arg.label = nfs4_label_copy(p->a_label, label);
 	p->o_arg.claim = nfs4_map_atomic_open_claim(server, claim);
 	switch (p->o_arg.claim) {
 	case NFS4_OPEN_CLAIM_NULL:
@@ -1074,6 +1081,8 @@ static struct nfs4_opendata *nfs4_opendata_alloc(struct dentry *dentry,
 	return p;
 
 err_free_label:
+	nfs4_label_free(p->a_label);
+err_free_f:
 	nfs4_label_free(p->f_label);
 err_free_p:
 	kfree(p);
@@ -1093,6 +1102,7 @@ static void nfs4_opendata_free(struct kref *kref)
 		nfs4_put_open_state(p->state);
 	nfs4_put_state_owner(p->owner);
 
+	nfs4_label_free(p->a_label);
 	nfs4_label_free(p->f_label);
 
 	dput(p->dir);
@@ -1198,12 +1208,15 @@ static bool nfs_need_update_open_stateid(struct nfs4_state *state,
 
 static void nfs_resync_open_stateid_locked(struct nfs4_state *state)
 {
+	if (!(state->n_wronly || state->n_rdonly || state->n_rdwr))
+		return;
 	if (state->n_wronly)
 		set_bit(NFS_O_WRONLY_STATE, &state->flags);
 	if (state->n_rdonly)
 		set_bit(NFS_O_RDONLY_STATE, &state->flags);
 	if (state->n_rdwr)
 		set_bit(NFS_O_RDWR_STATE, &state->flags);
+	set_bit(NFS_OPEN_STATE, &state->flags);
 }
 
 static void nfs_clear_open_stateid_locked(struct nfs4_state *state,
@@ -7571,13 +7584,8 @@ static int nfs4_proc_sequence(struct nfs_client *clp, struct rpc_cred *cred)
 		goto out;
 	}
 	ret = rpc_wait_for_completion_task(task);
-	if (!ret) {
-		struct nfs4_sequence_res *res = task->tk_msg.rpc_resp;
-
-		if (task->tk_status == 0)
-			nfs41_handle_sequence_flag_errors(clp, res->sr_status_flags);
+	if (!ret)
 		ret = task->tk_status;
-	}
 	rpc_put_task(task);
 out:
 	dprintk("<-- %s status=%d\n", __func__, ret);
@@ -7965,16 +7973,17 @@ static void nfs4_layoutreturn_release(void *calldata)
 {
 	struct nfs4_layoutreturn *lrp = calldata;
 	struct pnfs_layout_hdr *lo = lrp->args.layout;
+	LIST_HEAD(freeme);
 
 	dprintk("--> %s\n", __func__);
 	spin_lock(&lo->plh_inode->i_lock);
 	if (lrp->res.lrs_present)
 		pnfs_set_layout_stateid(lo, &lrp->res.stateid, true);
+	pnfs_mark_matching_lsegs_invalid(lo, &freeme, &lrp->args.range);
 	pnfs_clear_layoutreturn_waitbit(lo);
-	clear_bit(NFS_LAYOUT_RETURN_BEFORE_CLOSE, &lo->plh_flags);
-	rpc_wake_up(&NFS_SERVER(lo->plh_inode)->roc_rpcwaitq);
 	lo->plh_block_lgets--;
 	spin_unlock(&lo->plh_inode->i_lock);
+	pnfs_free_lseg_list(&freeme);
 	pnfs_put_layout_hdr(lrp->args.layout);
 	nfs_iput_and_deactive(lrp->inode);
 	kfree(calldata);
@@ -8588,7 +8597,6 @@ static const struct nfs4_minor_version_ops nfs_v4_0_minor_ops = {
 	.minor_version = 0,
 	.init_caps = NFS_CAP_READDIRPLUS
 		| NFS_CAP_ATOMIC_OPEN
-		| NFS_CAP_CHANGE_ATTR
 		| NFS_CAP_POSIX_LOCK,
 	.init_client = nfs40_init_client,
 	.shutdown_client = nfs40_shutdown_client,
@@ -8614,7 +8622,6 @@ static const struct nfs4_minor_version_ops nfs_v4_1_minor_ops = {
 	.minor_version = 1,
 	.init_caps = NFS_CAP_READDIRPLUS
 		| NFS_CAP_ATOMIC_OPEN
-		| NFS_CAP_CHANGE_ATTR
 		| NFS_CAP_POSIX_LOCK
 		| NFS_CAP_STATEID_NFSV41
 		| NFS_CAP_ATOMIC_OPEN_V1,
@@ -8637,7 +8644,6 @@ static const struct nfs4_minor_version_ops nfs_v4_2_minor_ops = {
 	.minor_version = 2,
 	.init_caps = NFS_CAP_READDIRPLUS
 		| NFS_CAP_ATOMIC_OPEN
-		| NFS_CAP_CHANGE_ATTR
 		| NFS_CAP_POSIX_LOCK
 		| NFS_CAP_STATEID_NFSV41
 		| NFS_CAP_ATOMIC_OPEN_V1

fs/nfs/nfs4state.c

@@ -2191,25 +2191,35 @@ static void nfs41_handle_server_reboot(struct nfs_client *clp)
 	}
 }
 
-static void nfs41_handle_state_revoked(struct nfs_client *clp)
+static void nfs41_handle_all_state_revoked(struct nfs_client *clp)
 {
 	nfs4_reset_all_state(clp);
 	dprintk("%s: state revoked on server %s\n", __func__, clp->cl_hostname);
 }
 
+static void nfs41_handle_some_state_revoked(struct nfs_client *clp)
+{
+	nfs4_state_mark_reclaim_helper(clp, nfs4_state_mark_reclaim_nograce);
+	nfs4_schedule_state_manager(clp);
+
+	dprintk("%s: state revoked on server %s\n", __func__, clp->cl_hostname);
+}
+
 static void nfs41_handle_recallable_state_revoked(struct nfs_client *clp)
 {
-	/* This will need to handle layouts too */
-	nfs_expire_all_delegations(clp);
+	/* FIXME: For now, we destroy all layouts. */
+	pnfs_destroy_all_layouts(clp);
+	/* FIXME: For now, we test all delegations+open state+locks. */
+	nfs41_handle_some_state_revoked(clp);
 	dprintk("%s: Recallable state revoked on server %s!\n", __func__,
 			clp->cl_hostname);
 }
 
 static void nfs41_handle_backchannel_fault(struct nfs_client *clp)
 {
-	nfs_expire_all_delegations(clp);
-	if (test_and_set_bit(NFS4CLNT_SESSION_RESET, &clp->cl_state) == 0)
-		nfs4_schedule_state_manager(clp);
+	set_bit(NFS4CLNT_SESSION_RESET, &clp->cl_state);
+	nfs4_schedule_state_manager(clp);
+
 	dprintk("%s: server %s declared a backchannel fault\n", __func__,
 			clp->cl_hostname);
 }
@@ -2231,10 +2241,11 @@ void nfs41_handle_sequence_flag_errors(struct nfs_client *clp, u32 flags)
 
 	if (flags & SEQ4_STATUS_RESTART_RECLAIM_NEEDED)
 		nfs41_handle_server_reboot(clp);
-	if (flags & (SEQ4_STATUS_EXPIRED_ALL_STATE_REVOKED |
-			    SEQ4_STATUS_EXPIRED_SOME_STATE_REVOKED |
+	if (flags & (SEQ4_STATUS_EXPIRED_ALL_STATE_REVOKED))
+		nfs41_handle_all_state_revoked(clp);
+	if (flags & (SEQ4_STATUS_EXPIRED_SOME_STATE_REVOKED |
 			    SEQ4_STATUS_ADMIN_STATE_REVOKED))
-		nfs41_handle_state_revoked(clp);
+		nfs41_handle_some_state_revoked(clp);
 	if (flags & SEQ4_STATUS_LEASE_MOVED)
 		nfs4_schedule_lease_moved_recovery(clp);
 	if (flags & SEQ4_STATUS_RECALLABLE_STATE_REVOKED)

fs/nfs/pagelist.c

@@ -1100,8 +1100,6 @@ static int nfs_do_recoalesce(struct nfs_pageio_descriptor *desc)
 		mirror->pg_base = 0;
 		mirror->pg_recoalesce = 0;
 
-		desc->pg_moreio = 0;
-
 		while (!list_empty(&head)) {
 			struct nfs_page *req;
 
@@ -1109,8 +1107,11 @@ static int nfs_do_recoalesce(struct nfs_pageio_descriptor *desc)
 			nfs_list_remove_request(req);
 			if (__nfs_pageio_add_request(desc, req))
 				continue;
-			if (desc->pg_error < 0)
+			if (desc->pg_error < 0) {
+				list_splice_tail(&head, &mirror->pg_list);
+				mirror->pg_recoalesce = 1;
 				return 0;
+			}
 			break;
 		}
 	} while (mirror->pg_recoalesce);

fs/nfs/pnfs.c

@@ -352,7 +352,7 @@ pnfs_layout_need_return(struct pnfs_layout_hdr *lo,
 {
 	struct pnfs_layout_segment *s;
 
-	if (!test_bit(NFS_LSEG_LAYOUTRETURN, &lseg->pls_flags))
+	if (!test_and_clear_bit(NFS_LSEG_LAYOUTRETURN, &lseg->pls_flags))
 		return false;
 
 	list_for_each_entry(s, &lo->plh_segs, pls_list)
@@ -362,6 +362,18 @@ pnfs_layout_need_return(struct pnfs_layout_hdr *lo,
 	return true;
 }
 
+static bool
+pnfs_prepare_layoutreturn(struct pnfs_layout_hdr *lo)
+{
+	if (test_and_set_bit(NFS_LAYOUT_RETURN, &lo->plh_flags))
+		return false;
+	lo->plh_return_iomode = 0;
+	lo->plh_block_lgets++;
+	pnfs_get_layout_hdr(lo);
+	clear_bit(NFS_LAYOUT_RETURN_BEFORE_CLOSE, &lo->plh_flags);
+	return true;
+}
+
 static void pnfs_layoutreturn_before_put_lseg(struct pnfs_layout_segment *lseg,
 		struct pnfs_layout_hdr *lo, struct inode *inode)
 {
@@ -372,17 +384,16 @@ static void pnfs_layoutreturn_before_put_lseg(struct pnfs_layout_segment *lseg,
 	if (pnfs_layout_need_return(lo, lseg)) {
 		nfs4_stateid stateid;
 		enum pnfs_iomode iomode;
+		bool send;
 
 		stateid = lo->plh_stateid;
 		iomode = lo->plh_return_iomode;
-		/* decreased in pnfs_send_layoutreturn() */
-		lo->plh_block_lgets++;
-		lo->plh_return_iomode = 0;
+		send = pnfs_prepare_layoutreturn(lo);
 		spin_unlock(&inode->i_lock);
-		pnfs_get_layout_hdr(lo);
-
-		/* Send an async layoutreturn so we dont deadlock */
-		pnfs_send_layoutreturn(lo, stateid, iomode, false);
+		if (send) {
+			/* Send an async layoutreturn so we dont deadlock */
+			pnfs_send_layoutreturn(lo, stateid, iomode, false);
+		}
 	} else
 		spin_unlock(&inode->i_lock);
 }
@@ -411,6 +422,10 @@ pnfs_put_lseg(struct pnfs_layout_segment *lseg)
 		pnfs_layoutreturn_before_put_lseg(lseg, lo, inode);
 
 	if (atomic_dec_and_lock(&lseg->pls_refcount, &inode->i_lock)) {
+		if (test_bit(NFS_LSEG_VALID, &lseg->pls_flags)) {
+			spin_unlock(&inode->i_lock);
+			return;
+		}
 		pnfs_get_layout_hdr(lo);
 		pnfs_layout_remove_lseg(lo, lseg);
 		spin_unlock(&inode->i_lock);
@@ -451,6 +466,8 @@ pnfs_put_lseg_locked(struct pnfs_layout_segment *lseg)
 		test_bit(NFS_LSEG_VALID, &lseg->pls_flags));
 	if (atomic_dec_and_test(&lseg->pls_refcount)) {
 		struct pnfs_layout_hdr *lo = lseg->pls_layout;
+		if (test_bit(NFS_LSEG_VALID, &lseg->pls_flags))
+			return;
 		pnfs_get_layout_hdr(lo);
 		pnfs_layout_remove_lseg(lo, lseg);
 		pnfs_free_lseg_async(lseg);
@@ -924,6 +941,7 @@ void pnfs_clear_layoutreturn_waitbit(struct pnfs_layout_hdr *lo)
 	clear_bit_unlock(NFS_LAYOUT_RETURN, &lo->plh_flags);
 	smp_mb__after_atomic();
 	wake_up_bit(&lo->plh_flags, NFS_LAYOUT_RETURN);
+	rpc_wake_up(&NFS_SERVER(lo->plh_inode)->roc_rpcwaitq);
 }
 
 static int
@@ -978,6 +996,7 @@ _pnfs_return_layout(struct inode *ino)
 	LIST_HEAD(tmp_list);
 	nfs4_stateid stateid;
 	int status = 0, empty;
+	bool send;
 
 	dprintk("NFS: %s for inode %lu\n", __func__, ino->i_ino);
 
@@ -1007,17 +1026,18 @@ _pnfs_return_layout(struct inode *ino)
 	/* Don't send a LAYOUTRETURN if list was initially empty */
 	if (empty) {
 		spin_unlock(&ino->i_lock);
-		pnfs_put_layout_hdr(lo);
 		dprintk("NFS: %s no layout segments to return\n", __func__);
-		goto out;
+		goto out_put_layout_hdr;
 	}
 
 	set_bit(NFS_LAYOUT_INVALID_STID, &lo->plh_flags);
-	lo->plh_block_lgets++;
+	send = pnfs_prepare_layoutreturn(lo);
 	spin_unlock(&ino->i_lock);
 	pnfs_free_lseg_list(&tmp_list);
-
-	status = pnfs_send_layoutreturn(lo, stateid, IOMODE_ANY, true);
+	if (send)
+		status = pnfs_send_layoutreturn(lo, stateid, IOMODE_ANY, true);
+out_put_layout_hdr:
+	pnfs_put_layout_hdr(lo);
 out:
 	dprintk("<-- %s status: %d\n", __func__, status);
 	return status;
@@ -1097,13 +1117,9 @@ bool pnfs_roc(struct inode *ino)
 out_noroc:
 	if (lo) {
 		stateid = lo->plh_stateid;
-		layoutreturn =
-			test_and_clear_bit(NFS_LAYOUT_RETURN_BEFORE_CLOSE,
-					   &lo->plh_flags);
-		if (layoutreturn) {
-			lo->plh_block_lgets++;
-			pnfs_get_layout_hdr(lo);
-		}
+		if (test_and_clear_bit(NFS_LAYOUT_RETURN_BEFORE_CLOSE,
+					   &lo->plh_flags))
+			layoutreturn = pnfs_prepare_layoutreturn(lo);
 	}
 	spin_unlock(&ino->i_lock);
 	if (layoutreturn) {
@@ -1146,15 +1162,18 @@ bool pnfs_roc_drain(struct inode *ino, u32 *barrier, struct rpc_task *task)
 	struct pnfs_layout_segment *lseg;
 	nfs4_stateid stateid;
 	u32 current_seqid;
-	bool found = false, layoutreturn = false;
+	bool layoutreturn = false;
 
 	spin_lock(&ino->i_lock);
-	list_for_each_entry(lseg, &nfsi->layout->plh_segs, pls_list)
-		if (test_bit(NFS_LSEG_ROC, &lseg->pls_flags)) {
-			rpc_sleep_on(&NFS_SERVER(ino)->roc_rpcwaitq, task, NULL);
-			found = true;
-			goto out;
-		}
+	list_for_each_entry(lseg, &nfsi->layout->plh_segs, pls_list) {
+		if (!test_bit(NFS_LSEG_ROC, &lseg->pls_flags))
+			continue;
+		if (test_bit(NFS_LSEG_VALID, &lseg->pls_flags))
+			continue;
+		rpc_sleep_on(&NFS_SERVER(ino)->roc_rpcwaitq, task, NULL);
+		spin_unlock(&ino->i_lock);
+		return true;
+	}
 	lo = nfsi->layout;
 	current_seqid = be32_to_cpu(lo->plh_stateid.seqid);
 
@@ -1162,23 +1181,19 @@ bool pnfs_roc_drain(struct inode *ino, u32 *barrier, struct rpc_task *task)
 	 * a barrier, we choose the worst-case barrier.
 	 */
 	*barrier = current_seqid + atomic_read(&lo->plh_outstanding);
-out:
-	if (!found) {
-		stateid = lo->plh_stateid;
-		layoutreturn =
-			test_and_clear_bit(NFS_LAYOUT_RETURN_BEFORE_CLOSE,
-					   &lo->plh_flags);
-		if (layoutreturn) {
-			lo->plh_block_lgets++;
-			pnfs_get_layout_hdr(lo);
-		}
-	}
+	stateid = lo->plh_stateid;
+	if (test_and_clear_bit(NFS_LAYOUT_RETURN_BEFORE_CLOSE,
+					   &lo->plh_flags))
+		layoutreturn = pnfs_prepare_layoutreturn(lo);
+	if (test_bit(NFS_LAYOUT_RETURN, &lo->plh_flags))
+		rpc_sleep_on(&NFS_SERVER(ino)->roc_rpcwaitq, task, NULL);
+
 	spin_unlock(&ino->i_lock);
 	if (layoutreturn) {
-		rpc_sleep_on(&NFS_SERVER(ino)->roc_rpcwaitq, task, NULL);
 		pnfs_send_layoutreturn(lo, stateid, IOMODE_ANY, false);
+		return true;
 	}
-	return found;
+	return false;
 }
 
 /*
@@ -1695,7 +1710,6 @@ void pnfs_error_mark_layout_for_return(struct inode *inode,
 	spin_lock(&inode->i_lock);
 	/* set failure bit so that pnfs path will be retried later */
 	pnfs_layout_set_fail_bit(lo, iomode);
-	set_bit(NFS_LAYOUT_RETURN, &lo->plh_flags);
 	if (lo->plh_return_iomode == 0)
 		lo->plh_return_iomode = range.iomode;
 	else if (lo->plh_return_iomode != range.iomode)
@@ -2207,13 +2221,12 @@ pnfs_layoutcommit_inode(struct inode *inode, bool sync)
 	if (ld->prepare_layoutcommit) {
 		status = ld->prepare_layoutcommit(&data->args);
 		if (status) {
+			put_rpccred(data->cred);
 			spin_lock(&inode->i_lock);
 			set_bit(NFS_INO_LAYOUTCOMMIT, &nfsi->flags);
 			if (end_pos > nfsi->layout->plh_lwb)
 				nfsi->layout->plh_lwb = end_pos;
-			spin_unlock(&inode->i_lock);
-			put_rpccred(data->cred);
-			goto clear_layoutcommitting;
+			goto out_unlock;
 		}
 	}
 

fs/nfs/write.c

@@ -1379,24 +1379,27 @@ static void nfs_writeback_check_extend(struct nfs_pgio_header *hdr,
 {
 	struct nfs_pgio_args *argp = &hdr->args;
 	struct nfs_pgio_res *resp = &hdr->res;
+	u64 size = argp->offset + resp->count;
 
 	if (!(fattr->valid & NFS_ATTR_FATTR_SIZE))
+		fattr->size = size;
+	if (nfs_size_to_loff_t(fattr->size) < i_size_read(hdr->inode)) {
+		fattr->valid &= ~NFS_ATTR_FATTR_SIZE;
 		return;
-	if (argp->offset + resp->count != fattr->size)
-		return;
-	if (nfs_size_to_loff_t(fattr->size) < i_size_read(hdr->inode))
+	}
+	if (size != fattr->size)
 		return;
 	/* Set attribute barrier */
 	nfs_fattr_set_barrier(fattr);
+	/* ...and update size */
+	fattr->valid |= NFS_ATTR_FATTR_SIZE;
 }
 
 void nfs_writeback_update_inode(struct nfs_pgio_header *hdr)
 {
-	struct nfs_fattr *fattr = hdr->res.fattr;
+	struct nfs_fattr *fattr = &hdr->fattr;
 	struct inode *inode = hdr->inode;
 
-	if (fattr == NULL)
-		return;
 	spin_lock(&inode->i_lock);
 	nfs_writeback_check_extend(hdr, fattr);
 	nfs_post_op_update_inode_force_wcc_locked(inode, fattr);

include/linux/nfs_fs.h

@@ -292,9 +292,12 @@ static inline void nfs_mark_for_revalidate(struct inode *inode)
 	struct nfs_inode *nfsi = NFS_I(inode);
 
 	spin_lock(&inode->i_lock);
-	nfsi->cache_validity |= NFS_INO_INVALID_ATTR|NFS_INO_INVALID_ACCESS;
+	nfsi->cache_validity |= NFS_INO_INVALID_ATTR |
+				NFS_INO_REVAL_PAGECACHE |
+				NFS_INO_INVALID_ACCESS |
+				NFS_INO_INVALID_ACL;
 	if (S_ISDIR(inode->i_mode))
-		nfsi->cache_validity |= NFS_INO_REVAL_PAGECACHE|NFS_INO_INVALID_DATA;
+		nfsi->cache_validity |= NFS_INO_INVALID_DATA;
 	spin_unlock(&inode->i_lock);
 }
 

include/linux/nfs_fs_sb.h

@@ -220,7 +220,7 @@ struct nfs_server {
 #define NFS_CAP_SYMLINKS	(1U << 2)
 #define NFS_CAP_ACLS		(1U << 3)
 #define NFS_CAP_ATOMIC_OPEN	(1U << 4)
-#define NFS_CAP_CHANGE_ATTR	(1U << 5)
+/* #define NFS_CAP_CHANGE_ATTR	(1U << 5) */
 #define NFS_CAP_FILEID		(1U << 6)
 #define NFS_CAP_MODE		(1U << 7)
 #define NFS_CAP_NLINK		(1U << 8)

net/sunrpc/backchannel_rqst.c

@@ -240,8 +240,8 @@ static struct rpc_rqst *xprt_alloc_bc_request(struct rpc_xprt *xprt, __be32 xid)
 		req = xprt_alloc_bc_req(xprt, GFP_ATOMIC);
 		if (!req)
 			goto not_found;
-		/* Note: this 'free' request adds it to xprt->bc_pa_list */
-		xprt_free_bc_request(req);
+		list_add_tail(&req->rq_bc_pa_list, &xprt->bc_pa_list);
+		xprt->bc_alloc_count++;
 	}
 	req = list_first_entry(&xprt->bc_pa_list, struct rpc_rqst,
 				rq_bc_pa_list);
@@ -336,7 +336,7 @@ void xprt_complete_bc_request(struct rpc_rqst *req, uint32_t copied)
 
 	spin_lock(&xprt->bc_pa_lock);
 	list_del(&req->rq_bc_pa_list);
-	xprt->bc_alloc_count--;
+	xprt_dec_alloc_count(xprt, 1);
 	spin_unlock(&xprt->bc_pa_lock);
 
 	req->rq_private_buf.len = copied;

net/sunrpc/clnt.c

@@ -1902,6 +1902,7 @@ call_transmit_status(struct rpc_task *task)
 
 	switch (task->tk_status) {
 	case -EAGAIN:
+	case -ENOBUFS:
 		break;
 	default:
 		dprint_status(task);
@@ -1928,7 +1929,6 @@ call_transmit_status(struct rpc_task *task)
 	case -ECONNABORTED:
 	case -EADDRINUSE:
 	case -ENOTCONN:
-	case -ENOBUFS:
 	case -EPIPE:
 		rpc_task_force_reencode(task);
 	}
@@ -2057,12 +2057,13 @@ call_status(struct rpc_task *task)
 	case -ECONNABORTED:
 		rpc_force_rebind(clnt);
 	case -EADDRINUSE:
-	case -ENOBUFS:
 		rpc_delay(task, 3*HZ);
 	case -EPIPE:
 	case -ENOTCONN:
 		task->tk_action = call_bind;
 		break;
+	case -ENOBUFS:
+		rpc_delay(task, HZ>>2);
 	case -EAGAIN:
 		task->tk_action = call_transmit;
 		break;

net/sunrpc/xprtsock.c

@@ -527,6 +527,10 @@ static int xs_local_send_request(struct rpc_task *task)
 			      true, &sent);
 	dprintk("RPC:       %s(%u) = %d\n",
 			__func__, xdr->len - req->rq_bytes_sent, status);
+
+	if (status == -EAGAIN && sock_writeable(transport->inet))
+		status = -ENOBUFS;
+
 	if (likely(sent > 0) || status == 0) {
 		req->rq_bytes_sent += sent;
 		req->rq_xmit_bytes_sent += sent;
@@ -539,6 +543,7 @@ static int xs_local_send_request(struct rpc_task *task)
 
 	switch (status) {
 	case -ENOBUFS:
+		break;
 	case -EAGAIN:
 		status = xs_nospace(task);
 		break;
@@ -589,6 +594,9 @@ static int xs_udp_send_request(struct rpc_task *task)
 	if (status == -EPERM)
 		goto process_status;
 
+	if (status == -EAGAIN && sock_writeable(transport->inet))
+		status = -ENOBUFS;
+
 	if (sent > 0 || status == 0) {
 		req->rq_xmit_bytes_sent += sent;
 		if (sent >= req->rq_slen)
@@ -669,9 +677,6 @@ static int xs_tcp_send_request(struct rpc_task *task)
 		dprintk("RPC:       xs_tcp_send_request(%u) = %d\n",
 				xdr->len - req->rq_bytes_sent, status);
 
-		if (unlikely(sent == 0 && status < 0))
-			break;
-
 		/* If we've sent the entire packet, immediately
 		 * reset the count of bytes sent. */
 		req->rq_bytes_sent += sent;
@@ -681,18 +686,21 @@ static int xs_tcp_send_request(struct rpc_task *task)
 			return 0;
 		}
 
-		if (sent != 0)
-			continue;
-		status = -EAGAIN;
-		break;
+		if (status < 0)
+			break;
+		if (sent == 0) {
+			status = -EAGAIN;
+			break;
+		}
 	}
+	if (status == -EAGAIN && sk_stream_is_writeable(transport->inet))
+		status = -ENOBUFS;
 
 	switch (status) {
 	case -ENOTSOCK:
 		status = -ENOTCONN;
 		/* Should we call xs_close() here? */
 		break;
-	case -ENOBUFS:
 	case -EAGAIN:
 		status = xs_nospace(task);
 		break;
@@ -703,6 +711,7 @@ static int xs_tcp_send_request(struct rpc_task *task)
 	case -ECONNREFUSED:
 	case -ENOTCONN:
 	case -EADDRINUSE:
+	case -ENOBUFS:
 	case -EPIPE:
 		clear_bit(SOCK_ASYNC_NOSPACE, &transport->sock->flags);
 	}