Merge tag 'nfs-for-5.14-1' of git://git.linux-nfs.org/projects/trondmy/linux-nfs

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

  Features:

   - Multiple patches to add support for fcntl() leases over NFSv4.

   - A sysfs interface to display more information about the various
     transport connections used by the RPC client

   - A sysfs interface to allow a suitably privileged user to offline a
     transport that may no longer point to a valid server

   - A sysfs interface to allow a suitably privileged user to change the
     server IP address used by the RPC client

  Stable fixes:

   - Two sunrpc fixes for deadlocks involving privileged rpc_wait_queues

  Bugfixes:

   - SUNRPC: Avoid a KASAN slab-out-of-bounds bug in xdr_set_page_base()

   - SUNRPC: prevent port reuse on transports which don't request it.

   - NFSv3: Fix memory leak in posix_acl_create()

   - NFS: Various fixes to attribute revalidation timeouts

   - NFSv4: Fix handling of non-atomic change attribute updates

   - NFSv4: If a server is down, don't cause mounts to other servers to
     hang as well

   - pNFS: Fix an Oops in pnfs_mark_request_commit() when doing O_DIRECT

   - NFS: Fix mount failures due to incorrect setting of the
     has_sec_mnt_opts filesystem flag

   - NFS: Ensure nfs_readpage returns promptly when an internal error
     occurs

   - NFS: Fix fscache read from NFS after cache error

   - pNFS: Various bugfixes around the LAYOUTGET operation"

* tag 'nfs-for-5.14-1' of git://git.linux-nfs.org/projects/trondmy/linux-nfs: (46 commits)
  NFSv4/pNFS: Return an error if _nfs4_pnfs_v3_ds_connect can't load NFSv3
  NFSv4/pNFS: Don't call _nfs4_pnfs_v3_ds_connect multiple times
  NFSv4/pnfs: Clean up layout get on open
  NFSv4/pnfs: Fix layoutget behaviour after invalidation
  NFSv4/pnfs: Fix the layout barrier update
  NFS: Fix fscache read from NFS after cache error
  NFS: Ensure nfs_readpage returns promptly when internal error occurs
  sunrpc: remove an offlined xprt using sysfs
  sunrpc: provide showing transport's state info in the sysfs directory
  sunrpc: display xprt's queuelen of assigned tasks via sysfs
  sunrpc: provide multipath info in the sysfs directory
  NFSv4.1 identify and mark RPC tasks that can move between transports
  sunrpc: provide transport info in the sysfs directory
  SUNRPC: take a xprt offline using sysfs
  sunrpc: add dst_attr attributes to the sysfs xprt directory
  SUNRPC for TCP display xprt's source port in sysfs xprt_info
  SUNRPC query transport's source port
  SUNRPC display xprt's main value in sysfs's xprt_info
  SUNRPC mark the first transport
  sunrpc: add add sysfs directory per xprt under each xprt_switch
  ...

Author: torvalds

fs/nfs/delegation.c

@@ -75,6 +75,13 @@ void nfs_mark_delegation_referenced(struct nfs_delegation *delegation)
 	set_bit(NFS_DELEGATION_REFERENCED, &delegation->flags);
 }
 
+static void nfs_mark_return_delegation(struct nfs_server *server,
+				       struct nfs_delegation *delegation)
+{
+	set_bit(NFS_DELEGATION_RETURN, &delegation->flags);
+	set_bit(NFS4CLNT_DELEGRETURN, &server->nfs_client->cl_state);
+}
+
 static bool
 nfs4_is_valid_delegation(const struct nfs_delegation *delegation,
 		fmode_t flags)
@@ -293,6 +300,7 @@ nfs_start_delegation_return_locked(struct nfs_inode *nfsi)
 		goto out;
 	spin_lock(&delegation->lock);
 	if (!test_and_set_bit(NFS_DELEGATION_RETURNING, &delegation->flags)) {
+		clear_bit(NFS_DELEGATION_RETURN_DELAYED, &delegation->flags);
 		/* Refcount matched in nfs_end_delegation_return() */
 		ret = nfs_get_delegation(delegation);
 	}
@@ -314,16 +322,17 @@ nfs_start_delegation_return(struct nfs_inode *nfsi)
 	return delegation;
 }
 
-static void
-nfs_abort_delegation_return(struct nfs_delegation *delegation,
-		struct nfs_client *clp)
+static void nfs_abort_delegation_return(struct nfs_delegation *delegation,
+					struct nfs_client *clp, int err)
 {
 
 	spin_lock(&delegation->lock);
 	clear_bit(NFS_DELEGATION_RETURNING, &delegation->flags);
-	set_bit(NFS_DELEGATION_RETURN, &delegation->flags);
+	if (err == -EAGAIN) {
+		set_bit(NFS_DELEGATION_RETURN_DELAYED, &delegation->flags);
+		set_bit(NFS4CLNT_DELEGRETURN_DELAYED, &clp->cl_state);
+	}
 	spin_unlock(&delegation->lock);
-	set_bit(NFS4CLNT_DELEGRETURN, &clp->cl_state);
 }
 
 static struct nfs_delegation *
@@ -521,11 +530,18 @@ int nfs_inode_set_delegation(struct inode *inode, const struct cred *cred,
 static int nfs_end_delegation_return(struct inode *inode, struct nfs_delegation *delegation, int issync)
 {
 	struct nfs_client *clp = NFS_SERVER(inode)->nfs_client;
+	unsigned int mode = O_WRONLY | O_RDWR;
 	int err = 0;
 
 	if (delegation == NULL)
 		return 0;
-	do {
+
+	if (!issync)
+		mode |= O_NONBLOCK;
+	/* Recall of any remaining application leases */
+	err = break_lease(inode, mode);
+
+	while (err == 0) {
 		if (test_bit(NFS_DELEGATION_REVOKED, &delegation->flags))
 			break;
 		err = nfs_delegation_claim_opens(inode, &delegation->stateid,
@@ -536,10 +552,10 @@ static int nfs_end_delegation_return(struct inode *inode, struct nfs_delegation
 		 * Guard against state recovery
 		 */
 		err = nfs4_wait_clnt_recover(clp);
-	} while (err == 0);
+	}
 
 	if (err) {
-		nfs_abort_delegation_return(delegation, clp);
+		nfs_abort_delegation_return(delegation, clp, err);
 		goto out;
 	}
 
@@ -568,6 +584,7 @@ static bool nfs_delegation_need_return(struct nfs_delegation *delegation)
 	if (ret)
 		clear_bit(NFS_DELEGATION_RETURN_IF_CLOSED, &delegation->flags);
 	if (test_bit(NFS_DELEGATION_RETURNING, &delegation->flags) ||
+	    test_bit(NFS_DELEGATION_RETURN_DELAYED, &delegation->flags) ||
 	    test_bit(NFS_DELEGATION_REVOKED, &delegation->flags))
 		ret = false;
 
@@ -647,6 +664,38 @@ static int nfs_server_return_marked_delegations(struct nfs_server *server,
 	return err;
 }
 
+static bool nfs_server_clear_delayed_delegations(struct nfs_server *server)
+{
+	struct nfs_delegation *d;
+	bool ret = false;
+
+	list_for_each_entry_rcu (d, &server->delegations, super_list) {
+		if (!test_bit(NFS_DELEGATION_RETURN_DELAYED, &d->flags))
+			continue;
+		nfs_mark_return_delegation(server, d);
+		clear_bit(NFS_DELEGATION_RETURN_DELAYED, &d->flags);
+		ret = true;
+	}
+	return ret;
+}
+
+static bool nfs_client_clear_delayed_delegations(struct nfs_client *clp)
+{
+	struct nfs_server *server;
+	bool ret = false;
+
+	if (!test_and_clear_bit(NFS4CLNT_DELEGRETURN_DELAYED, &clp->cl_state))
+		goto out;
+	rcu_read_lock();
+	list_for_each_entry_rcu (server, &clp->cl_superblocks, client_link) {
+		if (nfs_server_clear_delayed_delegations(server))
+			ret = true;
+	}
+	rcu_read_unlock();
+out:
+	return ret;
+}
+
 /**
  * nfs_client_return_marked_delegations - return previously marked delegations
  * @clp: nfs_client to process
@@ -659,8 +708,14 @@ static int nfs_server_return_marked_delegations(struct nfs_server *server,
  */
 int nfs_client_return_marked_delegations(struct nfs_client *clp)
 {
-	return nfs_client_for_each_server(clp,
-			nfs_server_return_marked_delegations, NULL);
+	int err = nfs_client_for_each_server(
+		clp, nfs_server_return_marked_delegations, NULL);
+	if (err)
+		return err;
+	/* If a return was delayed, sleep to prevent hard looping */
+	if (nfs_client_clear_delayed_delegations(clp))
+		ssleep(1);
+	return 0;
 }
 
 /**
@@ -698,13 +753,14 @@ int nfs4_inode_return_delegation(struct inode *inode)
 {
 	struct nfs_inode *nfsi = NFS_I(inode);
 	struct nfs_delegation *delegation;
-	int err = 0;
 
-	nfs_wb_all(inode);
 	delegation = nfs_start_delegation_return(nfsi);
+	/* Synchronous recall of any application leases */
+	break_lease(inode, O_WRONLY | O_RDWR);
+	nfs_wb_all(inode);
 	if (delegation != NULL)
-		err = nfs_end_delegation_return(inode, delegation, 1);
-	return err;
+		return nfs_end_delegation_return(inode, delegation, 1);
+	return 0;
 }
 
 /**
@@ -775,13 +831,6 @@ static void nfs_mark_return_if_closed_delegation(struct nfs_server *server,
 	set_bit(NFS4CLNT_DELEGRETURN, &server->nfs_client->cl_state);
 }
 
-static void nfs_mark_return_delegation(struct nfs_server *server,
-		struct nfs_delegation *delegation)
-{
-	set_bit(NFS_DELEGATION_RETURN, &delegation->flags);
-	set_bit(NFS4CLNT_DELEGRETURN, &server->nfs_client->cl_state);
-}
-
 static bool nfs_server_mark_return_all_delegations(struct nfs_server *server)
 {
 	struct nfs_delegation *delegation;
@@ -1010,6 +1059,9 @@ int nfs_async_inode_return_delegation(struct inode *inode,
 	nfs_mark_return_delegation(server, delegation);
 	rcu_read_unlock();
 
+	/* If there are any application leases or delegations, recall them */
+	break_lease(inode, O_WRONLY | O_RDWR | O_NONBLOCK);
+
 	nfs_delegation_run_state_manager(clp);
 	return 0;
 out_enoent:

fs/nfs/delegation.h

@@ -36,6 +36,7 @@ enum {
 	NFS_DELEGATION_REVOKED,
 	NFS_DELEGATION_TEST_EXPIRED,
 	NFS_DELEGATION_INODE_FREEING,
+	NFS_DELEGATION_RETURN_DELAYED,
 };
 
 int nfs_inode_set_delegation(struct inode *inode, const struct cred *cred,

fs/nfs/direct.c

@@ -700,8 +700,8 @@ static void nfs_direct_write_completion(struct nfs_pgio_header *hdr)
 {
 	struct nfs_direct_req *dreq = hdr->dreq;
 	struct nfs_commit_info cinfo;
-	bool request_commit = false;
 	struct nfs_page *req = nfs_list_entry(hdr->pages.next);
+	int flags = NFS_ODIRECT_DONE;
 
 	nfs_init_cinfo_from_dreq(&cinfo, dreq);
 
@@ -713,28 +713,25 @@ static void nfs_direct_write_completion(struct nfs_pgio_header *hdr)
 
 	nfs_direct_count_bytes(dreq, hdr);
 	if (hdr->good_bytes != 0 && nfs_write_need_commit(hdr)) {
-		switch (dreq->flags) {
-		case 0:
+		if (!dreq->flags)
 			dreq->flags = NFS_ODIRECT_DO_COMMIT;
-			request_commit = true;
-			break;
-		case NFS_ODIRECT_RESCHED_WRITES:
-		case NFS_ODIRECT_DO_COMMIT:
-			request_commit = true;
-		}
+		flags = dreq->flags;
 	}
 	spin_unlock(&dreq->lock);
 
 	while (!list_empty(&hdr->pages)) {
 
 		req = nfs_list_entry(hdr->pages.next);
 		nfs_list_remove_request(req);
-		if (request_commit) {
+		if (flags == NFS_ODIRECT_DO_COMMIT) {
 			kref_get(&req->wb_kref);
 			memcpy(&req->wb_verf, &hdr->verf.verifier,
 			       sizeof(req->wb_verf));
 			nfs_mark_request_commit(req, hdr->lseg, &cinfo,
 				hdr->ds_commit_idx);
+		} else if (flags == NFS_ODIRECT_RESCHED_WRITES) {
+			kref_get(&req->wb_kref);
+			nfs_mark_request_commit(req, NULL, &cinfo, 0);
 		}
 		nfs_unlock_and_release_request(req);
 	}

fs/nfs/fscache.c

@@ -385,12 +385,15 @@ static void nfs_readpage_from_fscache_complete(struct page *page,
 		 "NFS: readpage_from_fscache_complete (0x%p/0x%p/%d)\n",
 		 page, context, error);
 
-	/* if the read completes with an error, we just unlock the page and let
-	 * the VM reissue the readpage */
-	if (!error) {
+	/*
+	 * If the read completes with an error, mark the page with PG_checked,
+	 * unlock the page, and let the VM reissue the readpage.
+	 */
+	if (!error)
 		SetPageUptodate(page);
-		unlock_page(page);
-	}
+	else
+		SetPageChecked(page);
+	unlock_page(page);
 }
 
 /*
@@ -405,6 +408,11 @@ int __nfs_readpage_from_fscache(struct nfs_open_context *ctx,
 		 "NFS: readpage_from_fscache(fsc:%p/p:%p(i:%lx f:%lx)/0x%p)\n",
 		 nfs_i_fscache(inode), page, page->index, page->flags, inode);
 
+	if (PageChecked(page)) {
+		ClearPageChecked(page);
+		return 1;
+	}
+
 	ret = fscache_read_or_alloc_page(nfs_i_fscache(inode),
 					 page,
 					 nfs_readpage_from_fscache_complete,

fs/nfs/getroot.c

@@ -67,7 +67,7 @@ static int nfs_superblock_set_dummy_root(struct super_block *sb, struct inode *i
 int nfs_get_root(struct super_block *s, struct fs_context *fc)
 {
 	struct nfs_fs_context *ctx = nfs_fc2context(fc);
-	struct nfs_server *server = NFS_SB(s);
+	struct nfs_server *server = NFS_SB(s), *clone_server;
 	struct nfs_fsinfo fsinfo;
 	struct dentry *root;
 	struct inode *inode;
@@ -127,7 +127,7 @@ int nfs_get_root(struct super_block *s, struct fs_context *fc)
 	}
 	spin_unlock(&root->d_lock);
 	fc->root = root;
-	if (NFS_SB(s)->caps & NFS_CAP_SECURITY_LABEL)
+	if (server->caps & NFS_CAP_SECURITY_LABEL)
 		kflags |= SECURITY_LSM_NATIVE_LABELS;
 	if (ctx->clone_data.sb) {
 		if (d_inode(fc->root)->i_fop != &nfs_dir_operations) {
@@ -137,15 +137,19 @@ int nfs_get_root(struct super_block *s, struct fs_context *fc)
 		/* clone lsm security options from the parent to the new sb */
 		error = security_sb_clone_mnt_opts(ctx->clone_data.sb,
 						   s, kflags, &kflags_out);
+		if (error)
+			goto error_splat_root;
+		clone_server = NFS_SB(ctx->clone_data.sb);
+		server->has_sec_mnt_opts = clone_server->has_sec_mnt_opts;
 	} else {
 		error = security_sb_set_mnt_opts(s, fc->security,
 							kflags, &kflags_out);
 	}
 	if (error)
 		goto error_splat_root;
-	if (NFS_SB(s)->caps & NFS_CAP_SECURITY_LABEL &&
+	if (server->caps & NFS_CAP_SECURITY_LABEL &&
 		!(kflags_out & SECURITY_LSM_NATIVE_LABELS))
-		NFS_SB(s)->caps &= ~NFS_CAP_SECURITY_LABEL;
+		server->caps &= ~NFS_CAP_SECURITY_LABEL;
 
 	nfs_setsecurity(inode, fsinfo.fattr, fsinfo.fattr->label);
 	error = 0;