Add handling for STATX_CHANGE_COOKIE

This commit adds handling for the STATX_CHANGE_COOKIE so that
we can properly surface the ZFS znode sequence to NFS clients via
knfsd.

If knfsd does not have STATX_CHANGE_COOKIE in statx result then
it will synthesize the NFS change_info4 structure and related
change4id values algorithmically based on the ctime value of the
file. Since internally ZFS is using ktime_get_coarse_real_ts64()
for the timestamp calculation here it introduces the possiblity
that the change will not increment the change4id of directories
/ files causing a failure in the client to invalidate its attr
cache (among other things). See RFC 8881 Section 10.8 for
discussion of how clients may implement name and directory
caching.

Notable in this commit is that we are not initializing the
inode->i_version to the znode->z_seq number. The reason for this
is that we're intentionally not setting `SB_I_VERSION`. This
indicates that the filesystem manages its own i_version and
so it is not populated in the generic_fillattr.

The following compares tight loop of setattr over NFSv4
protocol while traching nfsd4_change_attribute.

Before change:
inode, change_attribute
4723, 7590032215978780890
4723, 7590032215978780890
4723, 7590032215978780890
4723, 7590032215982780865
4723, 7590032215982780865

After change:
inode, change_attribute
7602, 7590032992517123951
7602, 7590032992517123952
7602, 7590032992517123953
7602, 7590032992517123954
7602, 7590032992517123955

Signed-off-by: Andrew Walker <andrew.walker@truenas.com>

Author: anodos325

module/os/linux/zfs/zfs_vnops_os.c

@@ -2581,8 +2581,19 @@ zfs_setattr(znode_t *zp, vattr_t *vap, int flags, cred_t *cr, zidmap_t *mnt_ns)
 	if (fuid_dirtied)
 		zfs_fuid_sync(zfsvfs, tx);
 
-	if (mask != 0)
+	if (mask != 0) {
 		zfs_log_setattr(zilog, tx, TX_SETATTR, zp, vap, mask, fuidp);
+		/*
+		 * Ensure that the z_seq is always incremented on setattr
+		 * operation. This is required for change accounting for
+		 * NFS clients.
+		 *
+		 * ATTR_MODE already increments via zfs_acl_chmod_setattr.
+		 * ATTR_SIZE already increments via zfs_freesp.
+		 */
+		if (!(mask & (ATTR_MODE | ATTR_SIZE)))
+			zp->z_seq++;
+	}
 
 	mutex_exit(&zp->z_lock);
 	if (mask & (ATTR_UID|ATTR_GID|ATTR_MODE))

module/os/linux/zfs/zpl_inode.c

@@ -506,6 +506,32 @@ zpl_getattr_impl(const struct path *path, struct kstat *stat, u32 request_mask,
 	}
 #endif
 
+#ifdef STATX_CHANGE_COOKIE
+	if (request_mask & STATX_CHANGE_COOKIE) {
+		/*
+		 * knfsd uses the STATX_CHANGE_COOKIE to surface to clients
+		 * change_info4 data, which is used to implement NFS client
+		 * name caching (see RFC 8881 Section 10.8). This number
+		 * should always increase with changes and should not be
+		 * reused. We cannot simply present ctime here because
+		 * ZFS uses a coarse timer to set them, which may cause
+		 * clients to fail to detect changes and invalidate cache.
+		 *
+		 * ZFS always increments znode z_seq number, but this is
+		 * uint_t and so we mask in ctime to upper bits.
+		 *
+		 * STATX_ATTR_CHANGE_MONOTONIC is advertised
+		 * to prevent knfsd from generating the change cookie
+		 * based on ctime. C.f. nfsd4_change_attribute in
+		 * fs/nfsd/nfsfh.c.
+		 */
+		stat->change_cookie =
+		    ((u64)stat->ctime.tv_sec << 32) | zp->z_seq;
+		stat->attributes |= STATX_ATTR_CHANGE_MONOTONIC;
+		stat->result_mask |= STATX_CHANGE_COOKIE;
+	}
+#endif
+
 #ifdef STATX_DIOALIGN
 	if (request_mask & STATX_DIOALIGN) {
 		uint64_t align;