Merge tag 'ceph-for-6.8-rc1' of https://github.com/ceph/ceph-client

Pull ceph updates from Ilya Dryomov:
 "Assorted CephFS fixes and cleanups with nothing standing out"

* tag 'ceph-for-6.8-rc1' of https://github.com/ceph/ceph-client:
  ceph: get rid of passing callbacks in __dentry_leases_walk()
  ceph: d_obtain_{alias,root}(ERR_PTR(...)) will do the right thing
  ceph: fix invalid pointer access if get_quota_realm return ERR_PTR
  ceph: remove duplicated code in ceph_netfs_issue_read()
  ceph: send oldest_client_tid when renewing caps
  ceph: rename create_session_open_msg() to create_session_full_msg()
  ceph: select FS_ENCRYPTION_ALGS if FS_ENCRYPTION
  ceph: fix deadlock or deadcode of misusing dget()
  ceph: try to allocate a smaller extent map for sparse read
  libceph: remove MAX_EXTENTS check for sparse reads
  ceph: reinitialize mds feature bit even when session in open
  ceph: skip reconnecting if MDS is not ready

11 files changed, 98 insertions, 63 deletions

diff --git a/fs/ceph/Kconfig b/fs/ceph/Kconfig
index 94df854147d3..7249d70e1a43 100644
--- a/fs/ceph/Kconfig
+++ b/fs/ceph/Kconfig
@@ -7,6 +7,7 @@ config CEPH_FS
 	select CRYPTO_AES
 	select CRYPTO
 	select NETFS_SUPPORT
+	select FS_ENCRYPTION_ALGS if FS_ENCRYPTION
 	default n
 	help
 	  Choose Y or M here to include support for mounting the
diff --git a/fs/ceph/addr.c b/fs/ceph/addr.c
index 500a87b68a9a..1340d77124ae 100644
--- a/fs/ceph/addr.c
+++ b/fs/ceph/addr.c
@@ -337,6 +337,7 @@ static void ceph_netfs_issue_read(struct netfs_io_subrequest *subreq)
 	u64 len = subreq->len;
 	bool sparse = IS_ENCRYPTED(inode) || ceph_test_mount_opt(fsc, SPARSEREAD);
 	u64 off = subreq->start;
+	int extent_cnt;
 
 	if (ceph_inode_is_shutdown(inode)) {
 		err = -EIO;
@@ -350,8 +351,8 @@ static void ceph_netfs_issue_read(struct netfs_io_subrequest *subreq)
 
 	req = ceph_osdc_new_request(&fsc->client->osdc, &ci->i_layout, vino,
 			off, &len, 0, 1, sparse ? CEPH_OSD_OP_SPARSE_READ : CEPH_OSD_OP_READ,
-			CEPH_OSD_FLAG_READ | fsc->client->osdc.client->options->read_from_replica,
-			NULL, ci->i_truncate_seq, ci->i_truncate_size, false);
+			CEPH_OSD_FLAG_READ, NULL, ci->i_truncate_seq,
+			ci->i_truncate_size, false);
 	if (IS_ERR(req)) {
 		err = PTR_ERR(req);
 		req = NULL;
@@ -359,7 +360,8 @@ static void ceph_netfs_issue_read(struct netfs_io_subrequest *subreq)
 	}
 
 	if (sparse) {
-		err = ceph_alloc_sparse_ext_map(&req->r_ops[0]);
+		extent_cnt = __ceph_sparse_read_ext_count(inode, len);
+		err = ceph_alloc_sparse_ext_map(&req->r_ops[0], extent_cnt);
 		if (err)
 			goto out;
 	}
diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c
index 2c0b8dc3dd0d..9c02f328c966 100644
--- a/fs/ceph/caps.c
+++ b/fs/ceph/caps.c
@@ -4887,13 +4887,15 @@ int ceph_encode_dentry_release(void **p, struct dentry *dentry,
 			       struct inode *dir,
 			       int mds, int drop, int unless)
 {
-	struct dentry *parent = NULL;
 	struct ceph_mds_request_release *rel = *p;
 	struct ceph_dentry_info *di = ceph_dentry(dentry);
 	struct ceph_client *cl;
 	int force = 0;
 	int ret;
 
+	/* This shouldn't happen */
+	BUG_ON(!dir);
+
 	/*
 	 * force an record for the directory caps if we have a dentry lease.
 	 * this is racy (can't take i_ceph_lock and d_lock together), but it
@@ -4903,14 +4905,9 @@ int ceph_encode_dentry_release(void **p, struct dentry *dentry,
 	spin_lock(&dentry->d_lock);
 	if (di->lease_session && di->lease_session->s_mds == mds)
 		force = 1;
-	if (!dir) {
-		parent = dget(dentry->d_parent);
-		dir = d_inode(parent);
-	}
 	spin_unlock(&dentry->d_lock);
 
 	ret = ceph_encode_inode_release(p, dir, mds, drop, unless, force);
-	dput(parent);
 
 	cl = ceph_inode_to_client(dir);
 	spin_lock(&dentry->d_lock);
diff --git a/fs/ceph/dir.c b/fs/ceph/dir.c
index 678596684596..0e9f56eaba1e 100644
--- a/fs/ceph/dir.c
+++ b/fs/ceph/dir.c
@@ -1593,10 +1593,12 @@ struct ceph_lease_walk_control {
 	unsigned long dir_lease_ttl;
 };
 
+static int __dir_lease_check(const struct dentry *, struct ceph_lease_walk_control *);
+static int __dentry_lease_check(const struct dentry *);
+
 static unsigned long
 __dentry_leases_walk(struct ceph_mds_client *mdsc,
-		     struct ceph_lease_walk_control *lwc,
-		     int (*check)(struct dentry*, void*))
+		     struct ceph_lease_walk_control *lwc)
 {
 	struct ceph_dentry_info *di, *tmp;
 	struct dentry *dentry, *last = NULL;
@@ -1624,7 +1626,10 @@ __dentry_leases_walk(struct ceph_mds_client *mdsc,
 			goto next;
 		}
 
-		ret = check(dentry, lwc);
+		if (lwc->dir_lease)
+			ret = __dir_lease_check(dentry, lwc);
+		else
+			ret = __dentry_lease_check(dentry);
 		if (ret & TOUCH) {
 			/* move it into tail of dir lease list */
 			__dentry_dir_lease_touch(mdsc, di);
@@ -1681,7 +1686,7 @@ next:
 	return freed;
 }
 
-static int __dentry_lease_check(struct dentry *dentry, void *arg)
+static int __dentry_lease_check(const struct dentry *dentry)
 {
 	struct ceph_dentry_info *di = ceph_dentry(dentry);
 	int ret;
@@ -1696,9 +1701,9 @@ static int __dentry_lease_check(struct dentry *dentry, void *arg)
 	return DELETE;
 }
 
-static int __dir_lease_check(struct dentry *dentry, void *arg)
+static int __dir_lease_check(const struct dentry *dentry,
+			     struct ceph_lease_walk_control *lwc)
 {
-	struct ceph_lease_walk_control *lwc = arg;
 	struct ceph_dentry_info *di = ceph_dentry(dentry);
 
 	int ret = __dir_lease_try_check(dentry);
@@ -1737,7 +1742,7 @@ int ceph_trim_dentries(struct ceph_mds_client *mdsc)
 
 	lwc.dir_lease = false;
 	lwc.nr_to_scan  = CEPH_CAPS_PER_RELEASE * 2;
-	freed = __dentry_leases_walk(mdsc, &lwc, __dentry_lease_check);
+	freed = __dentry_leases_walk(mdsc, &lwc);
 	if (!lwc.nr_to_scan) /* more invalid leases */
 		return -EAGAIN;
 
@@ -1747,7 +1752,7 @@ int ceph_trim_dentries(struct ceph_mds_client *mdsc)
 	lwc.dir_lease = true;
 	lwc.expire_dir_lease = freed < count;
 	lwc.dir_lease_ttl = mdsc->fsc->mount_options->caps_wanted_delay_max * HZ;
-	freed +=__dentry_leases_walk(mdsc, &lwc, __dir_lease_check);
+	freed +=__dentry_leases_walk(mdsc, &lwc);
 	if (!lwc.nr_to_scan) /* more to check */
 		return -EAGAIN;
 
diff --git a/fs/ceph/export.c b/fs/ceph/export.c
index 726af69d4d62..a79f163ae4ed 100644
--- a/fs/ceph/export.c
+++ b/fs/ceph/export.c
@@ -286,8 +286,6 @@ static struct dentry *__snapfh_to_dentry(struct super_block *sb,
 		doutc(cl, "%llx.%llx parent %llx hash %x err=%d", vino.ino,
 		      vino.snap, sfh->parent_ino, sfh->hash, err);
 	}
-	if (IS_ERR(inode))
-		return ERR_CAST(inode);
 	/* see comments in ceph_get_parent() */
 	return unlinked ? d_obtain_root(inode) : d_obtain_alias(inode);
 }
diff --git a/fs/ceph/file.c b/fs/ceph/file.c
index d380d9dad0e0..abe8028d95bf 100644
--- a/fs/ceph/file.c
+++ b/fs/ceph/file.c
@@ -1029,6 +1029,7 @@ ssize_t __ceph_sync_read(struct inode *inode, loff_t *ki_pos,
 		struct ceph_osd_req_op *op;
 		u64 read_off = off;
 		u64 read_len = len;
+		int extent_cnt;
 
 		/* determine new offset/length if encrypted */
 		ceph_fscrypt_adjust_off_and_len(inode, &read_off, &read_len);
@@ -1068,7 +1069,8 @@ ssize_t __ceph_sync_read(struct inode *inode, loff_t *ki_pos,
 
 		op = &req->r_ops[0];
 		if (sparse) {
-			ret = ceph_alloc_sparse_ext_map(op);
+			extent_cnt = __ceph_sparse_read_ext_count(inode, read_len);
+			ret = ceph_alloc_sparse_ext_map(op, extent_cnt);
 			if (ret) {
 				ceph_osdc_put_request(req);
 				break;
@@ -1465,6 +1467,7 @@ ceph_direct_read_write(struct kiocb *iocb, struct iov_iter *iter,
 		ssize_t len;
 		struct ceph_osd_req_op *op;
 		int readop = sparse ? CEPH_OSD_OP_SPARSE_READ : CEPH_OSD_OP_READ;
+		int extent_cnt;
 
 		if (write)
 			size = min_t(u64, size, fsc->mount_options->wsize);
@@ -1528,7 +1531,8 @@ ceph_direct_read_write(struct kiocb *iocb, struct iov_iter *iter,
 		osd_req_op_extent_osd_data_bvecs(req, 0, bvecs, num_pages, len);
 		op = &req->r_ops[0];
 		if (sparse) {
-			ret = ceph_alloc_sparse_ext_map(op);
+			extent_cnt = __ceph_sparse_read_ext_count(inode, size);
+			ret = ceph_alloc_sparse_ext_map(op, extent_cnt);
 			if (ret) {
 				ceph_osdc_put_request(req);
 				break;
diff --git a/fs/ceph/mds_client.c b/fs/ceph/mds_client.c
index 02ebfabfc8ee..548d1de379f3 100644
--- a/fs/ceph/mds_client.c
+++ b/fs/ceph/mds_client.c
@@ -1534,7 +1534,8 @@ static int encode_metric_spec(void **p, void *end)
  * session message, specialization for CEPH_SESSION_REQUEST_OPEN
  * to include additional client metadata fields.
  */
-static struct ceph_msg *create_session_open_msg(struct ceph_mds_client *mdsc, u64 seq)
+static struct ceph_msg *
+create_session_full_msg(struct ceph_mds_client *mdsc, int op, u64 seq)
 {
 	struct ceph_msg *msg;
 	struct ceph_mds_session_head *h;
@@ -1578,6 +1579,9 @@ static struct ceph_msg *create_session_open_msg(struct ceph_mds_client *mdsc, u6
 		size = METRIC_BYTES(count);
 	extra_bytes += 2 + 4 + 4 + size;
 
+	/* flags, mds auth caps and oldest_client_tid */
+	extra_bytes += 4 + 4 + 8;
+
 	/* Allocate the message */
 	msg = ceph_msg_new(CEPH_MSG_CLIENT_SESSION, sizeof(*h) + extra_bytes,
 			   GFP_NOFS, false);
@@ -1589,16 +1593,16 @@ static struct ceph_msg *create_session_open_msg(struct ceph_mds_client *mdsc, u6
 	end = p + msg->front.iov_len;
 
 	h = p;
-	h->op = cpu_to_le32(CEPH_SESSION_REQUEST_OPEN);
+	h->op = cpu_to_le32(op);
 	h->seq = cpu_to_le64(seq);
 
 	/*
 	 * Serialize client metadata into waiting buffer space, using
 	 * the format that userspace expects for map<string, string>
 	 *
-	 * ClientSession messages with metadata are v4
+	 * ClientSession messages with metadata are v7
 	 */
-	msg->hdr.version = cpu_to_le16(4);
+	msg->hdr.version = cpu_to_le16(7);
 	msg->hdr.compat_version = cpu_to_le16(1);
 
 	/* The write pointer, following the session_head structure */
@@ -1634,6 +1638,15 @@ static struct ceph_msg *create_session_open_msg(struct ceph_mds_client *mdsc, u6
 		return ERR_PTR(ret);
 	}
 
+	/* version == 5, flags */
+	ceph_encode_32(&p, 0);
+
+	/* version == 6, mds auth caps */
+	ceph_encode_32(&p, 0);
+
+	/* version == 7, oldest_client_tid */
+	ceph_encode_64(&p, mdsc->oldest_tid);
+
 	msg->front.iov_len = p - msg->front.iov_base;
 	msg->hdr.front_len = cpu_to_le32(msg->front.iov_len);
 
@@ -1663,7 +1676,8 @@ static int __open_session(struct ceph_mds_client *mdsc,
 	session->s_renew_requested = jiffies;
 
 	/* send connect message */
-	msg = create_session_open_msg(mdsc, session->s_seq);
+	msg = create_session_full_msg(mdsc, CEPH_SESSION_REQUEST_OPEN,
+				      session->s_seq);
 	if (IS_ERR(msg))
 		return PTR_ERR(msg);
 	ceph_con_send(&session->s_con, msg);
@@ -2028,10 +2042,10 @@ static int send_renew_caps(struct ceph_mds_client *mdsc,
 
 	doutc(cl, "to mds%d (%s)\n", session->s_mds,
 	      ceph_mds_state_name(state));
-	msg = ceph_create_session_msg(CEPH_SESSION_REQUEST_RENEWCAPS,
+	msg = create_session_full_msg(mdsc, CEPH_SESSION_REQUEST_RENEWCAPS,
 				      ++session->s_renew_seq);
-	if (!msg)
-		return -ENOMEM;
+	if (IS_ERR(msg))
+		return PTR_ERR(msg);
 	ceph_con_send(&session->s_con, msg);
 	return 0;
 }
@@ -4128,12 +4142,12 @@ static void handle_session(struct ceph_mds_session *session,
 			pr_info_client(cl, "mds%d reconnect success\n",
 				       session->s_mds);
 
+		session->s_features = features;
 		if (session->s_state == CEPH_MDS_SESSION_OPEN) {
 			pr_notice_client(cl, "mds%d is already opened\n",
 					 session->s_mds);
 		} else {
 			session->s_state = CEPH_MDS_SESSION_OPEN;
-			session->s_features = features;
 			renewed_caps(mdsc, session, 0);
 			if (test_bit(CEPHFS_FEATURE_METRIC_COLLECT,
 				     &session->s_features))
@@ -5870,7 +5884,8 @@ static void mds_peer_reset(struct ceph_connection *con)
 
 	pr_warn_client(mdsc->fsc->client, "mds%d closed our session\n",
 		       s->s_mds);
-	if (READ_ONCE(mdsc->fsc->mount_state) != CEPH_MOUNT_FENCE_IO)
+	if (READ_ONCE(mdsc->fsc->mount_state) != CEPH_MOUNT_FENCE_IO &&
+	    ceph_mdsmap_get_state(mdsc->mdsmap, s->s_mds) >= CEPH_MDS_STATE_RECONNECT)
 		send_mds_reconnect(mdsc, s);
 }
 
diff --git a/fs/ceph/quota.c b/fs/ceph/quota.c
index 9d36c3532de1..06ee397e0c3a 100644
--- a/fs/ceph/quota.c
+++ b/fs/ceph/quota.c
@@ -197,10 +197,10 @@ void ceph_cleanup_quotarealms_inodes(struct ceph_mds_client *mdsc)
 }
 
 /*
- * This function walks through the snaprealm for an inode and returns the
- * ceph_snap_realm for the first snaprealm that has quotas set (max_files,
+ * This function walks through the snaprealm for an inode and set the
+ * realmp with the first snaprealm that has quotas set (max_files,
  * max_bytes, or any, depending on the 'which_quota' argument).  If the root is
- * reached, return the root ceph_snap_realm instead.
+ * reached, set the realmp with the root ceph_snap_realm instead.
  *
  * Note that the caller is responsible for calling ceph_put_snap_realm() on the
  * returned realm.
@@ -211,10 +211,9 @@ void ceph_cleanup_quotarealms_inodes(struct ceph_mds_client *mdsc)
  * this function will return -EAGAIN; otherwise, the snaprealms walk-through
  * will be restarted.
  */
-static struct ceph_snap_realm *get_quota_realm(struct ceph_mds_client *mdsc,
-					       struct inode *inode,
-					       enum quota_get_realm which_quota,
-					       bool retry)
+static int get_quota_realm(struct ceph_mds_client *mdsc, struct inode *inode,
+			   enum quota_get_realm which_quota,
+			   struct ceph_snap_realm **realmp, bool retry)
 {
 	struct ceph_client *cl = mdsc->fsc->client;
 	struct ceph_inode_info *ci = NULL;
@@ -222,8 +221,10 @@ static struct ceph_snap_realm *get_quota_realm(struct ceph_mds_client *mdsc,
 	struct inode *in;
 	bool has_quota;
 
+	if (realmp)
+		*realmp = NULL;
 	if (ceph_snap(inode) != CEPH_NOSNAP)
-		return NULL;
+		return 0;
 
 restart:
 	realm = ceph_inode(inode)->i_snap_realm;
@@ -250,7 +251,7 @@ restart:
 				break;
 			ceph_put_snap_realm(mdsc, realm);
 			if (!retry)
-				return ERR_PTR(-EAGAIN);
+				return -EAGAIN;
 			goto restart;
 		}
 
@@ -259,8 +260,11 @@ restart:
 		iput(in);
 
 		next = realm->parent;
-		if (has_quota || !next)
-		       return realm;
+		if (has_quota || !next) {
+			if (realmp)
+				*realmp = realm;
+			return 0;
+		}
 
 		ceph_get_snap_realm(mdsc, next);
 		ceph_put_snap_realm(mdsc, realm);
@@ -269,7 +273,7 @@ restart:
 	if (realm)
 		ceph_put_snap_realm(mdsc, realm);
 
-	return NULL;
+	return 0;
 }
 
 bool ceph_quota_is_same_realm(struct inode *old, struct inode *new)
@@ -277,6 +281,7 @@ bool ceph_quota_is_same_realm(struct inode *old, struct inode *new)
 	struct ceph_mds_client *mdsc = ceph_sb_to_mdsc(old->i_sb);
 	struct ceph_snap_realm *old_realm, *new_realm;
 	bool is_same;
+	int ret;
 
 restart:
 	/*
@@ -286,9 +291,9 @@ restart:
 	 * dropped and we can then restart the whole operation.
 	 */
 	down_read(&mdsc->snap_rwsem);
-	old_realm = get_quota_realm(mdsc, old, QUOTA_GET_ANY, true);
-	new_realm = get_quota_realm(mdsc, new, QUOTA_GET_ANY, false);
-	if (PTR_ERR(new_realm) == -EAGAIN) {
+	get_quota_realm(mdsc, old, QUOTA_GET_ANY, &old_realm, true);
+	ret = get_quota_realm(mdsc, new, QUOTA_GET_ANY, &new_realm, false);
+	if (ret == -EAGAIN) {
 		up_read(&mdsc->snap_rwsem);
 		if (old_realm)
 			ceph_put_snap_realm(mdsc, old_realm);
@@ -492,8 +497,8 @@ bool ceph_quota_update_statfs(struct ceph_fs_client *fsc, struct kstatfs *buf)
 	bool is_updated = false;
 
 	down_read(&mdsc->snap_rwsem);
-	realm = get_quota_realm(mdsc, d_inode(fsc->sb->s_root),
-				QUOTA_GET_MAX_BYTES, true);
+	get_quota_realm(mdsc, d_inode(fsc->sb->s_root), QUOTA_GET_MAX_BYTES,
+			&realm, true);
 	up_read(&mdsc->snap_rwsem);
 	if (!realm)
 		return false;
diff --git a/fs/ceph/super.h b/fs/ceph/super.h
index fe0f64a0acb2..b06e2bc86221 100644
--- a/fs/ceph/super.h
+++ b/fs/ceph/super.h
@@ -3,6 +3,7 @@
 #define _FS_CEPH_SUPER_H
 
 #include <linux/ceph/ceph_debug.h>
+#include <linux/ceph/osd_client.h>
 
 #include <asm/unaligned.h>
 #include <linux/backing-dev.h>
@@ -1407,6 +1408,19 @@ static inline void __ceph_update_quota(struct ceph_inode_info *ci,
 		ceph_adjust_quota_realms_count(&ci->netfs.inode, has_quota);
 }
 
+static inline int __ceph_sparse_read_ext_count(struct inode *inode, u64 len)
+{
+	int cnt = 0;
+
+	if (IS_ENCRYPTED(inode)) {
+		cnt = len >> CEPH_FSCRYPT_BLOCK_SHIFT;
+		if (cnt > CEPH_SPARSE_EXT_ARRAY_INITIAL)
+			cnt = 0;
+	}
+
+	return cnt;
+}
+
 extern void ceph_handle_quota(struct ceph_mds_client *mdsc,
 			      struct ceph_mds_session *session,
 			      struct ceph_msg *msg);
diff --git a/include/linux/ceph/osd_client.h b/include/linux/ceph/osd_client.h
index b8610e9d2471..fa018d5864e7 100644
--- a/include/linux/ceph/osd_client.h
+++ b/include/linux/ceph/osd_client.h
@@ -572,9 +572,12 @@ int __ceph_alloc_sparse_ext_map(struct ceph_osd_req_op *op, int cnt);
  */
 #define CEPH_SPARSE_EXT_ARRAY_INITIAL  16
 
-static inline int ceph_alloc_sparse_ext_map(struct ceph_osd_req_op *op)
+static inline int ceph_alloc_sparse_ext_map(struct ceph_osd_req_op *op, int cnt)
 {
-	return __ceph_alloc_sparse_ext_map(op, CEPH_SPARSE_EXT_ARRAY_INITIAL);
+	if (!cnt)
+		cnt = CEPH_SPARSE_EXT_ARRAY_INITIAL;
+
+	return __ceph_alloc_sparse_ext_map(op, cnt);
 }
 
 extern void ceph_osdc_get_request(struct ceph_osd_request *req);
diff --git a/net/ceph/osd_client.c b/net/ceph/osd_client.c
index d3a759e052c8..625622016f57 100644
--- a/net/ceph/osd_client.c
+++ b/net/ceph/osd_client.c
@@ -5850,8 +5850,6 @@ static inline void convert_extent_map(struct ceph_sparse_read *sr)
 }
 #endif
 
-#define MAX_EXTENTS 4096
-
 static int osd_sparse_read(struct ceph_connection *con,
 			   struct ceph_msg_data_cursor *cursor,
 			   char **pbuf)
@@ -5882,23 +5880,16 @@ next_op:
 
 		if (count > 0) {
 			if (!sr->sr_extent || count > sr->sr_ext_len) {
-				/*
-				 * Apply a hard cap to the number of extents.
-				 * If we have more, assume something is wrong.
-				 */
-				if (count > MAX_EXTENTS) {
-					dout("%s: OSD returned 0x%x extents in a single reply!\n",
-					     __func__, count);
-					return -EREMOTEIO;
-				}
-
 				/* no extent array provided, or too short */
 				kfree(sr->sr_extent);
 				sr->sr_extent = kmalloc_array(count,
 							      sizeof(*sr->sr_extent),
 							      GFP_NOIO);
-				if (!sr->sr_extent)
+				if (!sr->sr_extent) {
+					pr_err("%s: failed to allocate %u extents\n",
+					       __func__, count);
 					return -ENOMEM;
+				}
 				sr->sr_ext_len = count;
 			}
 			ret = count * sizeof(*sr->sr_extent);