Merge tag 'nfs-for-6.7-1' of git://git.linux-nfs.org/projects/trondmy/linux-nfs

Pull NFS client updates from Trond Myklebust:
 "Bugfixes:

   - SUNRPC:
       - re-probe the target RPC port after an ECONNRESET error
       - handle allocation errors from rpcb_call_async()
       - fix a use-after-free condition in rpc_pipefs
       - fix up various checks for timeouts

   - NFSv4.1:
       - Handle NFS4ERR_DELAY errors during session trunking
       - fix SP4_MACH_CRED protection for pnfs IO

   - NFSv4:
       - Ensure that we test all delegations when the server notifies
         us that it may have revoked some of them

  Features:

   - Allow knfsd processes to break out of NFS4ERR_DELAY loops when
     re-exporting NFSv4.x by setting appropriate values for the
     'delay_retrans' module parameter

   - nfs: Convert nfs_symlink() to use a folio"

* tag 'nfs-for-6.7-1' of git://git.linux-nfs.org/projects/trondmy/linux-nfs:
  nfs: Convert nfs_symlink() to use a folio
  SUNRPC: Fix RPC client cleaned up the freed pipefs dentries
  NFSv4.1: fix SP4_MACH_CRED protection for pnfs IO
  SUNRPC: Add an IS_ERR() check back to where it was
  NFSv4.1: fix handling NFS4ERR_DELAY when testing for session trunking
  nfs41: drop dependency between flexfiles layout driver and NFSv3 modules
  NFSv4: fairly test all delegations on a SEQ4_ revocation
  SUNRPC: SOFTCONN tasks should time out when on the sending list
  SUNRPC: Force close the socket when a hard error is reported
  SUNRPC: Don't skip timeout checks in call_connect_status()
  SUNRPC: ECONNRESET might require a rebind
  NFSv4/pnfs: Allow layoutget to return EAGAIN for softerr mounts
  NFSv4: Add a parameter to limit the number of retries after NFS4ERR_DELAY

Author: torvalds

Documentation/admin-guide/kernel-parameters.txt

@@ -3596,6 +3596,13 @@
 			[NFS] set the TCP port on which the NFSv4 callback
 			channel should listen.
 
+	nfs.delay_retrans=
+			[NFS] specifies the number of times the NFSv4 client
+			retries the request before returning an EAGAIN error,
+			after a reply of NFS4ERR_DELAY from the server.
+			Only applies if the softerr mount option is enabled,
+			and the specified value is >= 0.
+
 	nfs.enable_ino64=
 			[NFS] enable 64-bit inode numbers.
 			If zero, the NFS client will fake up a 32-bit inode

fs/nfs/Kconfig

@@ -125,7 +125,7 @@ config PNFS_BLOCK
 
 config PNFS_FLEXFILE_LAYOUT
 	tristate
-	depends on NFS_V4_1 && NFS_V3
+	depends on NFS_V4_1
 	default NFS_V4
 
 config NFS_V4_1_IMPLEMENTATION_ID_DOMAIN

fs/nfs/delegation.c

@@ -448,6 +448,7 @@ int nfs_inode_set_delegation(struct inode *inode, const struct cred *cred,
 	delegation->cred = get_cred(cred);
 	delegation->inode = inode;
 	delegation->flags = 1<<NFS_DELEGATION_REFERENCED;
+	delegation->test_gen = 0;
 	spin_lock_init(&delegation->lock);
 
 	spin_lock(&clp->cl_lock);
@@ -1294,6 +1295,8 @@ static int nfs_server_reap_expired_delegations(struct nfs_server *server,
 	struct inode *inode;
 	const struct cred *cred;
 	nfs4_stateid stateid;
+	unsigned long gen = ++server->delegation_gen;
+
 restart:
 	rcu_read_lock();
 restart_locked:
@@ -1303,7 +1306,8 @@ static int nfs_server_reap_expired_delegations(struct nfs_server *server,
 		    test_bit(NFS_DELEGATION_RETURNING,
 					&delegation->flags) ||
 		    test_bit(NFS_DELEGATION_TEST_EXPIRED,
-					&delegation->flags) == 0)
+					&delegation->flags) == 0 ||
+			delegation->test_gen == gen)
 			continue;
 		inode = nfs_delegation_grab_inode(delegation);
 		if (inode == NULL)
@@ -1312,6 +1316,7 @@ static int nfs_server_reap_expired_delegations(struct nfs_server *server,
 		cred = get_cred_rcu(delegation->cred);
 		nfs4_stateid_copy(&stateid, &delegation->stateid);
 		spin_unlock(&delegation->lock);
+		delegation->test_gen = gen;
 		clear_bit(NFS_DELEGATION_TEST_EXPIRED, &delegation->flags);
 		rcu_read_unlock();
 		nfs_delegation_test_free_expired(inode, &stateid, cred);

fs/nfs/delegation.h

@@ -21,6 +21,7 @@ struct nfs_delegation {
 	fmode_t type;
 	unsigned long pagemod_limit;
 	__u64 change_attr;
+	unsigned long test_gen;
 	unsigned long flags;
 	refcount_t refcount;
 	spinlock_t lock;

fs/nfs/dir.c

@@ -2532,7 +2532,7 @@ EXPORT_SYMBOL_GPL(nfs_unlink);
 int nfs_symlink(struct mnt_idmap *idmap, struct inode *dir,
 		struct dentry *dentry, const char *symname)
 {
-	struct page *page;
+	struct folio *folio;
 	char *kaddr;
 	struct iattr attr;
 	unsigned int pathlen = strlen(symname);
@@ -2547,24 +2547,24 @@ int nfs_symlink(struct mnt_idmap *idmap, struct inode *dir,
 	attr.ia_mode = S_IFLNK | S_IRWXUGO;
 	attr.ia_valid = ATTR_MODE;
 
-	page = alloc_page(GFP_USER);
-	if (!page)
+	folio = folio_alloc(GFP_USER, 0);
+	if (!folio)
 		return -ENOMEM;
 
-	kaddr = page_address(page);
+	kaddr = folio_address(folio);
 	memcpy(kaddr, symname, pathlen);
 	if (pathlen < PAGE_SIZE)
 		memset(kaddr + pathlen, 0, PAGE_SIZE - pathlen);
 
 	trace_nfs_symlink_enter(dir, dentry);
-	error = NFS_PROTO(dir)->symlink(dir, dentry, page, pathlen, &attr);
+	error = NFS_PROTO(dir)->symlink(dir, dentry, folio, pathlen, &attr);
 	trace_nfs_symlink_exit(dir, dentry, error);
 	if (error != 0) {
 		dfprintk(VFS, "NFS: symlink(%s/%lu, %pd, %s) error %d\n",
 			dir->i_sb->s_id, dir->i_ino,
 			dentry, symname, error);
 		d_drop(dentry);
-		__free_page(page);
+		folio_put(folio);
 		return error;
 	}
 
@@ -2574,18 +2574,13 @@ int nfs_symlink(struct mnt_idmap *idmap, struct inode *dir,
 	 * No big deal if we can't add this page to the page cache here.
 	 * READLINK will get the missing page from the server if needed.
 	 */
-	if (!add_to_page_cache_lru(page, d_inode(dentry)->i_mapping, 0,
-							GFP_KERNEL)) {
-		SetPageUptodate(page);
-		unlock_page(page);
-		/*
-		 * add_to_page_cache_lru() grabs an extra page refcount.
-		 * Drop it here to avoid leaking this page later.
-		 */
-		put_page(page);
-	} else
-		__free_page(page);
+	if (filemap_add_folio(d_inode(dentry)->i_mapping, folio, 0,
+							GFP_KERNEL) == 0) {
+		folio_mark_uptodate(folio);
+		folio_unlock(folio);
+	}
 
+	folio_put(folio);
 	return 0;
 }
 EXPORT_SYMBOL_GPL(nfs_symlink);

fs/nfs/nfs3proc.c

@@ -543,9 +543,10 @@ nfs3_proc_link(struct inode *inode, struct inode *dir, const struct qstr *name)
 }
 
 static int
-nfs3_proc_symlink(struct inode *dir, struct dentry *dentry, struct page *page,
+nfs3_proc_symlink(struct inode *dir, struct dentry *dentry, struct folio *folio,
 		  unsigned int len, struct iattr *sattr)
 {
+	struct page *page = &folio->page;
 	struct nfs3_createdata *data;
 	struct dentry *d_alias;
 	int status = -ENOMEM;

fs/nfs/nfs4_fs.h

@@ -209,6 +209,7 @@ struct nfs4_exception {
 	struct inode *inode;
 	nfs4_stateid *stateid;
 	long timeout;
+	unsigned short retrans;
 	unsigned char task_is_privileged : 1;
 	unsigned char delay : 1,
 		      recovering : 1,
@@ -546,6 +547,7 @@ extern unsigned short max_session_slots;
 extern unsigned short max_session_cb_slots;
 extern unsigned short send_implementation_id;
 extern bool recover_lost_locks;
+extern short nfs_delay_retrans;
 
 #define NFS4_CLIENT_ID_UNIQ_LEN		(64)
 extern char nfs4_client_id_uniquifier[NFS4_CLIENT_ID_UNIQ_LEN];

fs/nfs/nfs4proc.c

@@ -585,6 +585,21 @@ static int nfs4_do_handle_exception(struct nfs_server *server,
 	return 0;
 }
 
+/*
+ * Track the number of NFS4ERR_DELAY related retransmissions and return
+ * EAGAIN if the 'softerr' mount option is set, and we've exceeded the limit
+ * set by 'nfs_delay_retrans'.
+ */
+static int nfs4_exception_should_retrans(const struct nfs_server *server,
+					 struct nfs4_exception *exception)
+{
+	if (server->flags & NFS_MOUNT_SOFTERR && nfs_delay_retrans >= 0) {
+		if (exception->retrans++ >= (unsigned short)nfs_delay_retrans)
+			return -EAGAIN;
+	}
+	return 0;
+}
+
 /* This is the error handling routine for processes that are allowed
  * to sleep.
  */
@@ -595,6 +610,11 @@ int nfs4_handle_exception(struct nfs_server *server, int errorcode, struct nfs4_
 
 	ret = nfs4_do_handle_exception(server, errorcode, exception);
 	if (exception->delay) {
+		int ret2 = nfs4_exception_should_retrans(server, exception);
+		if (ret2 < 0) {
+			exception->retry = 0;
+			return ret2;
+		}
 		ret = nfs4_delay(&exception->timeout,
 				exception->interruptible);
 		goto out_retry;
@@ -623,6 +643,11 @@ nfs4_async_handle_exception(struct rpc_task *task, struct nfs_server *server,
 
 	ret = nfs4_do_handle_exception(server, errorcode, exception);
 	if (exception->delay) {
+		int ret2 = nfs4_exception_should_retrans(server, exception);
+		if (ret2 < 0) {
+			exception->retry = 0;
+			return ret2;
+		}
 		rpc_delay(task, nfs4_update_delay(&exception->timeout));
 		goto out_retry;
 	}
@@ -5011,9 +5036,10 @@ static void nfs4_free_createdata(struct nfs4_createdata *data)
 }
 
 static int _nfs4_proc_symlink(struct inode *dir, struct dentry *dentry,
-		struct page *page, unsigned int len, struct iattr *sattr,
+		struct folio *folio, unsigned int len, struct iattr *sattr,
 		struct nfs4_label *label)
 {
+	struct page *page = &folio->page;
 	struct nfs4_createdata *data;
 	int status = -ENAMETOOLONG;
 
@@ -5038,7 +5064,7 @@ static int _nfs4_proc_symlink(struct inode *dir, struct dentry *dentry,
 }
 
 static int nfs4_proc_symlink(struct inode *dir, struct dentry *dentry,
-		struct page *page, unsigned int len, struct iattr *sattr)
+		struct folio *folio, unsigned int len, struct iattr *sattr)
 {
 	struct nfs4_exception exception = {
 		.interruptible = true,
@@ -5049,7 +5075,7 @@ static int nfs4_proc_symlink(struct inode *dir, struct dentry *dentry,
 	label = nfs4_label_init_security(dir, dentry, sattr, &l);
 
 	do {
-		err = _nfs4_proc_symlink(dir, dentry, page, len, sattr, label);
+		err = _nfs4_proc_symlink(dir, dentry, folio, len, sattr, label);
 		trace_nfs4_symlink(dir, &dentry->d_name, err);
 		err = nfs4_handle_exception(NFS_SERVER(dir), err,
 				&exception);
@@ -5622,7 +5648,7 @@ static void nfs4_proc_write_setup(struct nfs_pgio_header *hdr,
 
 	msg->rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_WRITE];
 	nfs4_init_sequence(&hdr->args.seq_args, &hdr->res.seq_res, 0, 0);
-	nfs4_state_protect_write(server->nfs_client, clnt, msg, hdr);
+	nfs4_state_protect_write(hdr->ds_clp ? hdr->ds_clp : server->nfs_client, clnt, msg, hdr);
 }
 
 static void nfs4_proc_commit_rpc_prepare(struct rpc_task *task, struct nfs_commit_data *data)
@@ -5663,7 +5689,8 @@ static void nfs4_proc_commit_setup(struct nfs_commit_data *data, struct rpc_mess
 	data->res.server = server;
 	msg->rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_COMMIT];
 	nfs4_init_sequence(&data->args.seq_args, &data->res.seq_res, 1, 0);
-	nfs4_state_protect(server->nfs_client, NFS_SP4_MACH_CRED_COMMIT, clnt, msg);
+	nfs4_state_protect(data->ds_clp ? data->ds_clp : server->nfs_client,
+			NFS_SP4_MACH_CRED_COMMIT, clnt, msg);
 }
 
 static int _nfs4_proc_commit(struct file *dst, struct nfs_commitargs *args,
@@ -8934,6 +8961,7 @@ void nfs4_test_session_trunk(struct rpc_clnt *clnt, struct rpc_xprt *xprt,
 
 	sp4_how = (adata->clp->cl_sp4_flags == 0 ? SP4_NONE : SP4_MACH_CRED);
 
+try_again:
 	/* Test connection for session trunking. Async exchange_id call */
 	task = nfs4_run_exchange_id(adata->clp, adata->cred, sp4_how, xprt);
 	if (IS_ERR(task))
@@ -8946,11 +8974,15 @@ void nfs4_test_session_trunk(struct rpc_clnt *clnt, struct rpc_xprt *xprt,
 
 	if (status == 0)
 		rpc_clnt_xprt_switch_add_xprt(clnt, xprt);
-	else if (rpc_clnt_xprt_switch_has_addr(clnt,
+	else if (status != -NFS4ERR_DELAY && rpc_clnt_xprt_switch_has_addr(clnt,
 				(struct sockaddr *)&xprt->addr))
 		rpc_clnt_xprt_switch_remove_xprt(clnt, xprt);
 
 	rpc_put_task(task);
+	if (status == -NFS4ERR_DELAY) {
+		ssleep(1);
+		goto try_again;
+	}
 }
 EXPORT_SYMBOL_GPL(nfs4_test_session_trunk);
 
@@ -9621,6 +9653,9 @@ nfs4_layoutget_handle_exception(struct rpc_task *task,
 
 	nfs4_sequence_free_slot(&lgp->res.seq_res);
 
+	exception->state = NULL;
+	exception->stateid = NULL;
+
 	switch (nfs4err) {
 	case 0:
 		goto out;
@@ -9716,7 +9751,8 @@ static const struct rpc_call_ops nfs4_layoutget_call_ops = {
 };
 
 struct pnfs_layout_segment *
-nfs4_proc_layoutget(struct nfs4_layoutget *lgp, long *timeout)
+nfs4_proc_layoutget(struct nfs4_layoutget *lgp,
+		    struct nfs4_exception *exception)
 {
 	struct inode *inode = lgp->args.inode;
 	struct nfs_server *server = NFS_SERVER(inode);
@@ -9736,13 +9772,10 @@ nfs4_proc_layoutget(struct nfs4_layoutget *lgp, long *timeout)
 			 RPC_TASK_MOVEABLE,
 	};
 	struct pnfs_layout_segment *lseg = NULL;
-	struct nfs4_exception exception = {
-		.inode = inode,
-		.timeout = *timeout,
-	};
 	int status = 0;
 
 	nfs4_init_sequence(&lgp->args.seq_args, &lgp->res.seq_res, 0, 0);
+	exception->retry = 0;
 
 	task = rpc_run_task(&task_setup_data);
 	if (IS_ERR(task))
@@ -9753,11 +9786,12 @@ nfs4_proc_layoutget(struct nfs4_layoutget *lgp, long *timeout)
 		goto out;
 
 	if (task->tk_status < 0) {
-		status = nfs4_layoutget_handle_exception(task, lgp, &exception);
-		*timeout = exception.timeout;
+		exception->retry = 1;
+		status = nfs4_layoutget_handle_exception(task, lgp, exception);
 	} else if (lgp->res.layoutp->len == 0) {
+		exception->retry = 1;
 		status = -EAGAIN;
-		*timeout = nfs4_update_delay(&exception.timeout);
+		nfs4_update_delay(&exception->timeout);
 	} else
 		lseg = pnfs_layout_process(lgp);
 out:

fs/nfs/pnfs.c

@@ -1980,7 +1980,9 @@ pnfs_update_layout(struct inode *ino,
 	struct pnfs_layout_segment *lseg = NULL;
 	struct nfs4_layoutget *lgp;
 	nfs4_stateid stateid;
-	long timeout = 0;
+	struct nfs4_exception exception = {
+		.inode = ino,
+	};
 	unsigned long giveup = jiffies + (clp->cl_lease_time << 1);
 	bool first;
 
@@ -2144,7 +2146,7 @@ pnfs_update_layout(struct inode *ino,
 	lgp->lo = lo;
 	pnfs_get_layout_hdr(lo);
 
-	lseg = nfs4_proc_layoutget(lgp, &timeout);
+	lseg = nfs4_proc_layoutget(lgp, &exception);
 	trace_pnfs_update_layout(ino, pos, count, iomode, lo, lseg,
 				 PNFS_UPDATE_LAYOUT_SEND_LAYOUTGET);
 	nfs_layoutget_end(lo);
@@ -2171,6 +2173,8 @@ pnfs_update_layout(struct inode *ino,
 			goto out_put_layout_hdr;
 		}
 		if (lseg) {
+			if (!exception.retry)
+				goto out_put_layout_hdr;
 			if (first)
 				pnfs_clear_first_layoutget(lo);
 			trace_pnfs_update_layout(ino, pos, count,

fs/nfs/pnfs.h

@@ -35,6 +35,7 @@
 #include <linux/nfs_page.h>
 #include <linux/workqueue.h>
 
+struct nfs4_exception;
 struct nfs4_opendata;
 
 enum {
@@ -245,7 +246,9 @@ extern size_t max_response_pages(struct nfs_server *server);
 extern int nfs4_proc_getdeviceinfo(struct nfs_server *server,
 				   struct pnfs_device *dev,
 				   const struct cred *cred);
-extern struct pnfs_layout_segment* nfs4_proc_layoutget(struct nfs4_layoutget *lgp, long *timeout);
+extern struct pnfs_layout_segment *
+nfs4_proc_layoutget(struct nfs4_layoutget *lgp,
+		    struct nfs4_exception *exception);
 extern int nfs4_proc_layoutreturn(struct nfs4_layoutreturn *lrp, bool sync);
 
 /* pnfs.c */

fs/nfs/proc.c

@@ -396,9 +396,10 @@ nfs_proc_link(struct inode *inode, struct inode *dir, const struct qstr *name)
 }
 
 static int
-nfs_proc_symlink(struct inode *dir, struct dentry *dentry, struct page *page,
+nfs_proc_symlink(struct inode *dir, struct dentry *dentry, struct folio *folio,
 		 unsigned int len, struct iattr *sattr)
 {
+	struct page *page = &folio->page;
 	struct nfs_fh *fh;
 	struct nfs_fattr *fattr;
 	struct nfs_symlinkargs	arg = {