Merge tag 'nfsd-6.7' of git://git.kernel.org/pub/scm/linux/kernel/git/cel/linux

Pull nfsd updates from Chuck Lever:
 "This release completes the SunRPC thread scheduler work that was begun
  in v6.6. The scheduler can now find an svc thread to wake in constant
  time and without a list walk. Thanks again to Neil Brown for this
  overhaul.

  Lorenzo Bianconi contributed infrastructure for a netlink-based NFSD
  control plane. The long-term plan is to provide the same functionality
  as found in /proc/fs/nfsd, plus some interesting additions, and then
  migrate the NFSD user space utilities to netlink.

  A long series to overhaul NFSD's NFSv4 operation encoding was applied
  in this release. The goals are to bring this family of encoding
  functions in line with the matching NFSv4 decoding functions and with
  the NFSv2 and NFSv3 XDR functions, preparing the way for better memory
  safety and maintainability.

  A further improvement to NFSD's write delegation support was
  contributed by Dai Ngo. This adds a CB_GETATTR callback, enabling the
  server to retrieve cached size and mtime data from clients holding
  write delegations. If the server can retrieve this information, it
  does not have to recall the delegation in some cases.

  The usual panoply of bug fixes and minor improvements round out this
  release. As always I am grateful to all contributors, reviewers, and
  testers"

* tag 'nfsd-6.7' of git://git.kernel.org/pub/scm/linux/kernel/git/cel/linux: (127 commits)
  svcrdma: Fix tracepoint printk format
  svcrdma: Drop connection after an RDMA Read error
  NFSD: clean up alloc_init_deleg()
  NFSD: Fix frame size warning in svc_export_parse()
  NFSD: Rewrite synopsis of nfsd_percpu_counters_init()
  nfsd: Clean up errors in nfs3proc.c
  nfsd: Clean up errors in nfs4state.c
  NFSD: Clean up errors in stats.c
  NFSD: simplify error paths in nfsd_svc()
  NFSD: Clean up nfsd4_encode_seek()
  NFSD: Clean up nfsd4_encode_offset_status()
  NFSD: Clean up nfsd4_encode_copy_notify()
  NFSD: Clean up nfsd4_encode_copy()
  NFSD: Clean up nfsd4_encode_test_stateid()
  NFSD: Clean up nfsd4_encode_exchange_id()
  NFSD: Clean up nfsd4_do_encode_secinfo()
  NFSD: Clean up nfsd4_encode_access()
  NFSD: Clean up nfsd4_encode_readdir()
  NFSD: Clean up nfsd4_encode_entry4()
  NFSD: Add an nfsd4_encode_nfs_cookie4() helper
  ...

Author: torvalds

Documentation/filesystems/nfs/exporting.rst

@@ -241,3 +241,10 @@ following flags are defined:
     all of an inode's dirty data on last close. Exports that behave this
     way should set EXPORT_OP_FLUSH_ON_CLOSE so that NFSD knows to skip
     waiting for writeback when closing such files.
+
+  EXPORT_OP_ASYNC_LOCK - Indicates a capable filesystem to do async lock
+    requests from lockd. Only set EXPORT_OP_ASYNC_LOCK if the filesystem has
+    it's own ->lock() functionality as core posix_lock_file() implementation
+    has no async lock request handling yet. For more information about how to
+    indicate an async lock request from a ->lock() file_operations struct, see
+    fs/locks.c and comment for the function vfs_lock_file().

Documentation/netlink/specs/nfsd.yaml

@@ -0,0 +1,89 @@
+# SPDX-License-Identifier: ((GPL-2.0 WITH Linux-syscall-note) OR BSD-3-Clause)
+
+name: nfsd
+protocol: genetlink
+uapi-header: linux/nfsd_netlink.h
+
+doc: NFSD configuration over generic netlink.
+
+attribute-sets:
+  -
+    name: rpc-status
+    attributes:
+      -
+        name: xid
+        type: u32
+        byte-order: big-endian
+      -
+        name: flags
+        type: u32
+      -
+        name: prog
+        type: u32
+      -
+        name: version
+        type: u8
+      -
+        name: proc
+        type: u32
+      -
+        name: service_time
+        type: s64
+      -
+        name: pad
+        type: pad
+      -
+        name: saddr4
+        type: u32
+        byte-order: big-endian
+        display-hint: ipv4
+      -
+        name: daddr4
+        type: u32
+        byte-order: big-endian
+        display-hint: ipv4
+      -
+        name: saddr6
+        type: binary
+        display-hint: ipv6
+      -
+        name: daddr6
+        type: binary
+        display-hint: ipv6
+      -
+        name: sport
+        type: u16
+        byte-order: big-endian
+      -
+        name: dport
+        type: u16
+        byte-order: big-endian
+      -
+        name: compound-ops
+        type: u32
+        multi-attr: true
+
+operations:
+  list:
+    -
+      name: rpc-status-get
+      doc: dump pending nfsd rpc
+      attribute-set: rpc-status
+      dump:
+        pre: nfsd-nl-rpc-status-get-start
+        post: nfsd-nl-rpc-status-get-done
+        reply:
+          attributes:
+            - xid
+            - flags
+            - prog
+            - version
+            - proc
+            - service_time
+            - saddr4
+            - daddr4
+            - saddr6
+            - daddr6
+            - sport
+            - dport
+            - compound-ops

fs/lockd/svc.c

@@ -24,7 +24,6 @@
 #include <linux/uio.h>
 #include <linux/smp.h>
 #include <linux/mutex.h>
-#include <linux/kthread.h>
 #include <linux/freezer.h>
 #include <linux/inetdevice.h>
 
@@ -135,11 +134,11 @@ lockd(void *vrqstp)
 	 * The main request loop. We don't terminate until the last
 	 * NFS mount or NFS daemon has gone away.
 	 */
-	while (!kthread_should_stop()) {
+	while (!svc_thread_should_stop(rqstp)) {
 		/* update sv_maxconn if it has changed */
 		rqstp->rq_server->sv_maxconn = nlm_max_connections;
 
-		nlmsvc_retry_blocked();
+		nlmsvc_retry_blocked(rqstp);
 		svc_recv(rqstp);
 	}
 	if (nlmsvc_ops)
@@ -373,7 +372,9 @@ static void lockd_put(void)
 	unregister_inet6addr_notifier(&lockd_inet6addr_notifier);
 #endif
 
+	svc_get(nlmsvc_serv);
 	svc_set_num_threads(nlmsvc_serv, NULL, 0);
+	svc_put(nlmsvc_serv);
 	timer_delete_sync(&nlmsvc_retry);
 	nlmsvc_serv = NULL;
 	dprintk("lockd_down: service destroyed\n");

fs/lockd/svclock.c

@@ -30,7 +30,6 @@
 #include <linux/sunrpc/svc_xprt.h>
 #include <linux/lockd/nlm.h>
 #include <linux/lockd/lockd.h>
-#include <linux/kthread.h>
 #include <linux/exportfs.h>
 
 #define NLMDBG_FACILITY		NLMDBG_SVCLOCK
@@ -481,9 +480,7 @@ nlmsvc_lock(struct svc_rqst *rqstp, struct nlm_file *file,
 	    struct nlm_host *host, struct nlm_lock *lock, int wait,
 	    struct nlm_cookie *cookie, int reclaim)
 {
-#if IS_ENABLED(CONFIG_SUNRPC_DEBUG)
 	struct inode		*inode = nlmsvc_file_inode(file);
-#endif
 	struct nlm_block	*block = NULL;
 	int			error;
 	int			mode;
@@ -497,7 +494,7 @@ nlmsvc_lock(struct svc_rqst *rqstp, struct nlm_file *file,
 				(long long)lock->fl.fl_end,
 				wait);
 
-	if (nlmsvc_file_file(file)->f_op->lock) {
+	if (!exportfs_lock_op_is_async(inode->i_sb->s_export_op)) {
 		async_block = wait;
 		wait = 0;
 	}
@@ -543,6 +540,25 @@ nlmsvc_lock(struct svc_rqst *rqstp, struct nlm_file *file,
 		goto out;
 	}
 
+	spin_lock(&nlm_blocked_lock);
+	/*
+	 * If this is a lock request for an already pending
+	 * lock request we return nlm_lck_blocked without calling
+	 * vfs_lock_file() again. Otherwise we have two pending
+	 * requests on the underlaying ->lock() implementation but
+	 * only one nlm_block to being granted by lm_grant().
+	 */
+	if (exportfs_lock_op_is_async(inode->i_sb->s_export_op) &&
+	    !list_empty(&block->b_list)) {
+		spin_unlock(&nlm_blocked_lock);
+		ret = nlm_lck_blocked;
+		goto out;
+	}
+
+	/* Append to list of blocked */
+	nlmsvc_insert_block_locked(block, NLM_NEVER);
+	spin_unlock(&nlm_blocked_lock);
+
 	if (!wait)
 		lock->fl.fl_flags &= ~FL_SLEEP;
 	mode = lock_to_openmode(&lock->fl);
@@ -552,16 +568,12 @@ nlmsvc_lock(struct svc_rqst *rqstp, struct nlm_file *file,
 	dprintk("lockd: vfs_lock_file returned %d\n", error);
 	switch (error) {
 		case 0:
+			nlmsvc_remove_block(block);
 			ret = nlm_granted;
 			goto out;
 		case -EAGAIN:
-			/*
-			 * If this is a blocking request for an
-			 * already pending lock request then we need
-			 * to put it back on lockd's block list
-			 */
-			if (wait)
-				break;
+			if (!wait)
+				nlmsvc_remove_block(block);
 			ret = async_block ? nlm_lck_blocked : nlm_lck_denied;
 			goto out;
 		case FILE_LOCK_DEFERRED:
@@ -572,17 +584,16 @@ nlmsvc_lock(struct svc_rqst *rqstp, struct nlm_file *file,
 			ret = nlmsvc_defer_lock_rqst(rqstp, block);
 			goto out;
 		case -EDEADLK:
+			nlmsvc_remove_block(block);
 			ret = nlm_deadlock;
 			goto out;
 		default:			/* includes ENOLCK */
+			nlmsvc_remove_block(block);
 			ret = nlm_lck_denied_nolocks;
 			goto out;
 	}
 
 	ret = nlm_lck_blocked;
-
-	/* Append to list of blocked */
-	nlmsvc_insert_block(block, NLM_NEVER);
 out:
 	mutex_unlock(&file->f_mutex);
 	nlmsvc_release_block(block);
@@ -1020,13 +1031,13 @@ retry_deferred_block(struct nlm_block *block)
  * be retransmitted.
  */
 void
-nlmsvc_retry_blocked(void)
+nlmsvc_retry_blocked(struct svc_rqst *rqstp)
 {
 	unsigned long	timeout = MAX_SCHEDULE_TIMEOUT;
 	struct nlm_block *block;
 
 	spin_lock(&nlm_blocked_lock);
-	while (!list_empty(&nlm_blocked) && !kthread_should_stop()) {
+	while (!list_empty(&nlm_blocked) && !svc_thread_should_stop(rqstp)) {
 		block = list_entry(nlm_blocked.next, struct nlm_block, b_list);
 
 		if (block->b_when == NLM_NEVER)

fs/locks.c

@@ -2264,11 +2264,13 @@ int fcntl_getlk(struct file *filp, unsigned int cmd, struct flock *flock)
  * To avoid blocking kernel daemons, such as lockd, that need to acquire POSIX
  * locks, the ->lock() interface may return asynchronously, before the lock has
  * been granted or denied by the underlying filesystem, if (and only if)
- * lm_grant is set. Callers expecting ->lock() to return asynchronously
- * will only use F_SETLK, not F_SETLKW; they will set FL_SLEEP if (and only if)
- * the request is for a blocking lock. When ->lock() does return asynchronously,
- * it must return FILE_LOCK_DEFERRED, and call ->lm_grant() when the lock
- * request completes.
+ * lm_grant is set. Additionally EXPORT_OP_ASYNC_LOCK in export_operations
+ * flags need to be set.
+ *
+ * Callers expecting ->lock() to return asynchronously will only use F_SETLK,
+ * not F_SETLKW; they will set FL_SLEEP if (and only if) the request is for a
+ * blocking lock. When ->lock() does return asynchronously, it must return
+ * FILE_LOCK_DEFERRED, and call ->lm_grant() when the lock request completes.
  * If the request is for non-blocking lock the file system should return
  * FILE_LOCK_DEFERRED then try to get the lock and call the callback routine
  * with the result. If the request timed out the callback routine will return a

fs/nfs/callback.c

@@ -78,53 +78,14 @@ nfs4_callback_svc(void *vrqstp)
 
 	set_freezable();
 
-	while (!kthread_freezable_should_stop(NULL))
+	while (!svc_thread_should_stop(rqstp))
 		svc_recv(rqstp);
 
 	svc_exit_thread(rqstp);
 	return 0;
 }
 
 #if defined(CONFIG_NFS_V4_1)
-/*
- * The callback service for NFSv4.1 callbacks
- */
-static int
-nfs41_callback_svc(void *vrqstp)
-{
-	struct svc_rqst *rqstp = vrqstp;
-	struct svc_serv *serv = rqstp->rq_server;
-	struct rpc_rqst *req;
-	int error;
-	DEFINE_WAIT(wq);
-
-	set_freezable();
-
-	while (!kthread_freezable_should_stop(NULL)) {
-		prepare_to_wait(&serv->sv_cb_waitq, &wq, TASK_IDLE);
-		spin_lock_bh(&serv->sv_cb_lock);
-		if (!list_empty(&serv->sv_cb_list)) {
-			req = list_first_entry(&serv->sv_cb_list,
-					struct rpc_rqst, rq_bc_list);
-			list_del(&req->rq_bc_list);
-			spin_unlock_bh(&serv->sv_cb_lock);
-			finish_wait(&serv->sv_cb_waitq, &wq);
-			dprintk("Invoking bc_svc_process()\n");
-			error = bc_svc_process(serv, req, rqstp);
-			dprintk("bc_svc_process() returned w/ error code= %d\n",
-				error);
-		} else {
-			spin_unlock_bh(&serv->sv_cb_lock);
-			if (!kthread_should_stop())
-				schedule();
-			finish_wait(&serv->sv_cb_waitq, &wq);
-		}
-	}
-
-	svc_exit_thread(rqstp);
-	return 0;
-}
-
 static inline void nfs_callback_bc_serv(u32 minorversion, struct rpc_xprt *xprt,
 		struct svc_serv *serv)
 {
@@ -237,10 +198,7 @@ static struct svc_serv *nfs_callback_create_svc(int minorversion)
 			cb_info->users);
 
 	threadfn = nfs4_callback_svc;
-#if defined(CONFIG_NFS_V4_1)
-	if (minorversion)
-		threadfn = nfs41_callback_svc;
-#else
+#if !defined(CONFIG_NFS_V4_1)
 	if (minorversion)
 		return ERR_PTR(-ENOTSUPP);
 #endif

fs/nfsd/Makefile

@@ -12,7 +12,8 @@ nfsd-y			+= trace.o
 
 nfsd-y 			+= nfssvc.o nfsctl.o nfsfh.o vfs.o \
 			   export.o auth.o lockd.o nfscache.o \
-			   stats.o filecache.o nfs3proc.o nfs3xdr.o
+			   stats.o filecache.o nfs3proc.o nfs3xdr.o \
+			   netlink.o
 nfsd-$(CONFIG_NFSD_V2) += nfsproc.o nfsxdr.o
 nfsd-$(CONFIG_NFSD_V2_ACL) += nfs2acl.o
 nfsd-$(CONFIG_NFSD_V3_ACL) += nfs3acl.o

fs/nfsd/blocklayoutxdr.c

@@ -16,9 +16,9 @@
 
 __be32
 nfsd4_block_encode_layoutget(struct xdr_stream *xdr,
-		struct nfsd4_layoutget *lgp)
+		const struct nfsd4_layoutget *lgp)
 {
-	struct pnfs_block_extent *b = lgp->lg_content;
+	const struct pnfs_block_extent *b = lgp->lg_content;
 	int len = sizeof(__be32) + 5 * sizeof(__be64) + sizeof(__be32);
 	__be32 *p;
 
@@ -77,7 +77,7 @@ nfsd4_block_encode_volume(struct xdr_stream *xdr, struct pnfs_block_volume *b)
 
 __be32
 nfsd4_block_encode_getdeviceinfo(struct xdr_stream *xdr,
-		struct nfsd4_getdeviceinfo *gdp)
+		const struct nfsd4_getdeviceinfo *gdp)
 {
 	struct pnfs_block_deviceaddr *dev = gdp->gd_device;
 	int len = sizeof(__be32), ret, i;

fs/nfsd/blocklayoutxdr.h

@@ -51,9 +51,9 @@ struct pnfs_block_deviceaddr {
 };
 
 __be32 nfsd4_block_encode_getdeviceinfo(struct xdr_stream *xdr,
-		struct nfsd4_getdeviceinfo *gdp);
+		const struct nfsd4_getdeviceinfo *gdp);
 __be32 nfsd4_block_encode_layoutget(struct xdr_stream *xdr,
-		struct nfsd4_layoutget *lgp);
+		const struct nfsd4_layoutget *lgp);
 int nfsd4_block_decode_layoutupdate(__be32 *p, u32 len, struct iomap **iomapp,
 		u32 block_size);
 int nfsd4_scsi_decode_layoutupdate(__be32 *p, u32 len, struct iomap **iomapp,