Merge tag 'ceph-for-5.2-rc1' of git://github.com/ceph/ceph-client

Pull ceph updates from Ilya Dryomov:
 "On the filesystem side we have:

   - a fix to enforce quotas set above the mount point (Luis Henriques)

   - support for exporting snapshots through NFS (Zheng Yan)

   - proper statx implementation (Jeff Layton). statx flags are mapped
     to MDS caps, with AT_STATX_{DONT,FORCE}_SYNC taken into account.

   - some follow-up dentry name handling fixes, in particular
     elimination of our hand-rolled helper and the switch to __getname()
     as suggested by Al (Jeff Layton)

   - a set of MDS client cleanups in preparation for async MDS requests
     in the future (Jeff Layton)

   - a fix to sync the filesystem before remounting (Jeff Layton)

  On the rbd side, work is on-going on object-map and fast-diff image
  features"

* tag 'ceph-for-5.2-rc1' of git://github.com/ceph/ceph-client: (29 commits)
  ceph: flush dirty inodes before proceeding with remount
  ceph: fix unaligned access in ceph_send_cap_releases
  libceph: make ceph_pr_addr take an struct ceph_entity_addr pointer
  libceph: fix unaligned accesses in ceph_entity_addr handling
  rbd: don't assert on writes to snapshots
  rbd: client_mutex is never nested
  ceph: print inode number in __caps_issued_mask debugging messages
  ceph: just call get_session in __ceph_lookup_mds_session
  ceph: simplify arguments and return semantics of try_get_cap_refs
  ceph: fix comment over ceph_drop_caps_for_unlink
  ceph: move wait for mds request into helper function
  ceph: have ceph_mdsc_do_request call ceph_mdsc_submit_request
  ceph: after an MDS request, do callback and completions
  ceph: use pathlen values returned by set_request_path_attr
  ceph: use __getname/__putname in ceph_mdsc_build_path
  ceph: use ceph_mdsc_build_path instead of clone_dentry_name
  ceph: fix potential use-after-free in ceph_mdsc_build_path
  ceph: dump granular cap info in "caps" debugfs file
  ceph: make iterate_session_caps a public symbol
  ceph: fix NULL pointer deref when debugging is enabled
  ...

drivers/block/rbd.c

@@ -934,7 +934,7 @@ static struct rbd_client *rbd_get_client(struct ceph_options *ceph_opts)
 	struct rbd_client *rbdc;
 	int ret;
 
-	mutex_lock_nested(&client_mutex, SINGLE_DEPTH_NESTING);
+	mutex_lock(&client_mutex);
 	rbdc = rbd_client_find(ceph_opts);
 	if (rbdc) {
 		ceph_destroy_options(ceph_opts);
@@ -1326,7 +1326,7 @@ static void rbd_obj_zero_range(struct rbd_obj_request *obj_req, u32 off,
 		zero_bvecs(&obj_req->bvec_pos, off, bytes);
 		break;
 	default:
-		rbd_assert(0);
+		BUG();
 	}
 }
 
@@ -1581,7 +1581,7 @@ static void rbd_obj_request_destroy(struct kref *kref)
 		kfree(obj_request->bvec_pos.bvecs);
 		break;
 	default:
-		rbd_assert(0);
+		BUG();
 	}
 
 	kfree(obj_request->img_extents);
@@ -1781,7 +1781,7 @@ static void rbd_osd_req_setup_data(struct rbd_obj_request *obj_req, u32 which)
 						    &obj_req->bvec_pos);
 		break;
 	default:
-		rbd_assert(0);
+		BUG();
 	}
 }
 
@@ -2036,7 +2036,7 @@ static int __rbd_img_fill_request(struct rbd_img_request *img_req)
 			ret = rbd_obj_setup_zeroout(obj_req);
 			break;
 		default:
-			rbd_assert(0);
+			BUG();
 		}
 		if (ret < 0)
 			return ret;
@@ -2383,7 +2383,7 @@ static int rbd_obj_read_from_parent(struct rbd_obj_request *obj_req)
 						      &obj_req->bvec_pos);
 			break;
 		default:
-			rbd_assert(0);
+			BUG();
 		}
 	} else {
 		ret = rbd_img_fill_from_bvecs(child_img_req,
@@ -2515,7 +2515,7 @@ static int rbd_obj_issue_copyup_ops(struct rbd_obj_request *obj_req, u32 bytes)
 		num_osd_ops += count_zeroout_ops(obj_req);
 		break;
 	default:
-		rbd_assert(0);
+		BUG();
 	}
 
 	obj_req->osd_req = rbd_osd_req_create(obj_req, num_osd_ops);
@@ -2542,7 +2542,7 @@ static int rbd_obj_issue_copyup_ops(struct rbd_obj_request *obj_req, u32 bytes)
 		__rbd_obj_setup_zeroout(obj_req, which);
 		break;
 	default:
-		rbd_assert(0);
+		BUG();
 	}
 
 	ret = ceph_osdc_alloc_messages(obj_req->osd_req, GFP_NOIO);
@@ -3842,8 +3842,12 @@ static void rbd_queue_workfn(struct work_struct *work)
 		goto err_rq;
 	}
 
-	rbd_assert(op_type == OBJ_OP_READ ||
-		   rbd_dev->spec->snap_id == CEPH_NOSNAP);
+	if (op_type != OBJ_OP_READ && rbd_dev->spec->snap_id != CEPH_NOSNAP) {
+		rbd_warn(rbd_dev, "%s on read-only snapshot",
+			 obj_op_name(op_type));
+		result = -EIO;
+		goto err;
+	}
 
 	/*
 	 * Quit early if the mapped snapshot no longer exists.  It's

fs/ceph/caps.c

@@ -892,8 +892,8 @@ int __ceph_caps_issued_mask(struct ceph_inode_info *ci, int mask, int touch)
 	int have = ci->i_snap_caps;
 
 	if ((have & mask) == mask) {
-		dout("__ceph_caps_issued_mask %p snap issued %s"
-		     " (mask %s)\n", &ci->vfs_inode,
+		dout("__ceph_caps_issued_mask ino 0x%lx snap issued %s"
+		     " (mask %s)\n", ci->vfs_inode.i_ino,
 		     ceph_cap_string(have),
 		     ceph_cap_string(mask));
 		return 1;
@@ -904,8 +904,8 @@ int __ceph_caps_issued_mask(struct ceph_inode_info *ci, int mask, int touch)
 		if (!__cap_is_valid(cap))
 			continue;
 		if ((cap->issued & mask) == mask) {
-			dout("__ceph_caps_issued_mask %p cap %p issued %s"
-			     " (mask %s)\n", &ci->vfs_inode, cap,
+			dout("__ceph_caps_issued_mask ino 0x%lx cap %p issued %s"
+			     " (mask %s)\n", ci->vfs_inode.i_ino, cap,
 			     ceph_cap_string(cap->issued),
 			     ceph_cap_string(mask));
 			if (touch)
@@ -916,8 +916,8 @@ int __ceph_caps_issued_mask(struct ceph_inode_info *ci, int mask, int touch)
 		/* does a combination of caps satisfy mask? */
 		have |= cap->issued;
 		if ((have & mask) == mask) {
-			dout("__ceph_caps_issued_mask %p combo issued %s"
-			     " (mask %s)\n", &ci->vfs_inode,
+			dout("__ceph_caps_issued_mask ino 0x%lx combo issued %s"
+			     " (mask %s)\n", ci->vfs_inode.i_ino,
 			     ceph_cap_string(cap->issued),
 			     ceph_cap_string(mask));
 			if (touch) {
@@ -2257,8 +2257,6 @@ int ceph_fsync(struct file *file, loff_t start, loff_t end, int datasync)
 	if (datasync)
 		goto out;
 
-	inode_lock(inode);
-
 	dirty = try_flush_caps(inode, &flush_tid);
 	dout("fsync dirty caps are %s\n", ceph_cap_string(dirty));
 
@@ -2273,7 +2271,6 @@ int ceph_fsync(struct file *file, loff_t start, loff_t end, int datasync)
 		ret = wait_event_interruptible(ci->i_cap_wq,
 					caps_are_flushed(inode, flush_tid));
 	}
-	inode_unlock(inode);
 out:
 	dout("fsync %p%s result=%d\n", inode, datasync ? " datasync" : "", ret);
 	return ret;
@@ -2528,9 +2525,14 @@ static void __take_cap_refs(struct ceph_inode_info *ci, int got,
  * to (when applicable), and check against max_size here as well.
  * Note that caller is responsible for ensuring max_size increases are
  * requested from the MDS.
+ *
+ * Returns 0 if caps were not able to be acquired (yet), a 1 if they were,
+ * or a negative error code.
+ *
+ * FIXME: how does a 0 return differ from -EAGAIN?
  */
 static int try_get_cap_refs(struct ceph_inode_info *ci, int need, int want,
-			    loff_t endoff, bool nonblock, int *got, int *err)
+			    loff_t endoff, bool nonblock, int *got)
 {
 	struct inode *inode = &ci->vfs_inode;
 	struct ceph_mds_client *mdsc = ceph_inode_to_client(inode)->mdsc;
@@ -2550,8 +2552,7 @@ static int try_get_cap_refs(struct ceph_inode_info *ci, int need, int want,
 	if ((file_wanted & need) != need) {
 		dout("try_get_cap_refs need %s file_wanted %s, EBADF\n",
 		     ceph_cap_string(need), ceph_cap_string(file_wanted));
-		*err = -EBADF;
-		ret = 1;
+		ret = -EBADF;
 		goto out_unlock;
 	}
 
@@ -2572,10 +2573,8 @@ static int try_get_cap_refs(struct ceph_inode_info *ci, int need, int want,
 		if (endoff >= 0 && endoff > (loff_t)ci->i_max_size) {
 			dout("get_cap_refs %p endoff %llu > maxsize %llu\n",
 			     inode, endoff, ci->i_max_size);
-			if (endoff > ci->i_requested_max_size) {
-				*err = -EAGAIN;
-				ret = 1;
-			}
+			if (endoff > ci->i_requested_max_size)
+				ret = -EAGAIN;
 			goto out_unlock;
 		}
 		/*
@@ -2610,8 +2609,7 @@ static int try_get_cap_refs(struct ceph_inode_info *ci, int need, int want,
 					 * task isn't in TASK_RUNNING state
 					 */
 					if (nonblock) {
-						*err = -EAGAIN;
-						ret = 1;
+						ret = -EAGAIN;
 						goto out_unlock;
 					}
 
@@ -2640,8 +2638,7 @@ static int try_get_cap_refs(struct ceph_inode_info *ci, int need, int want,
 		if (session_readonly) {
 			dout("get_cap_refs %p needed %s but mds%d readonly\n",
 			     inode, ceph_cap_string(need), ci->i_auth_cap->mds);
-			*err = -EROFS;
-			ret = 1;
+			ret = -EROFS;
 			goto out_unlock;
 		}
 
@@ -2650,16 +2647,14 @@ static int try_get_cap_refs(struct ceph_inode_info *ci, int need, int want,
 			if (READ_ONCE(mdsc->fsc->mount_state) ==
 			    CEPH_MOUNT_SHUTDOWN) {
 				dout("get_cap_refs %p forced umount\n", inode);
-				*err = -EIO;
-				ret = 1;
+				ret = -EIO;
 				goto out_unlock;
 			}
 			mds_wanted = __ceph_caps_mds_wanted(ci, false);
 			if (need & ~(mds_wanted & need)) {
 				dout("get_cap_refs %p caps were dropped"
 				     " (session killed?)\n", inode);
-				*err = -ESTALE;
-				ret = 1;
+				ret = -ESTALE;
 				goto out_unlock;
 			}
 			if (!(file_wanted & ~mds_wanted))
@@ -2710,7 +2705,7 @@ static void check_max_size(struct inode *inode, loff_t endoff)
 int ceph_try_get_caps(struct ceph_inode_info *ci, int need, int want,
 		      bool nonblock, int *got)
 {
-	int ret, err = 0;
+	int ret;
 
 	BUG_ON(need & ~CEPH_CAP_FILE_RD);
 	BUG_ON(want & ~(CEPH_CAP_FILE_CACHE|CEPH_CAP_FILE_LAZYIO|CEPH_CAP_FILE_SHARED));
@@ -2718,15 +2713,8 @@ int ceph_try_get_caps(struct ceph_inode_info *ci, int need, int want,
 	if (ret < 0)
 		return ret;
 
-	ret = try_get_cap_refs(ci, need, want, 0, nonblock, got, &err);
-	if (ret) {
-		if (err == -EAGAIN) {
-			ret = 0;
-		} else if (err < 0) {
-			ret = err;
-		}
-	}
-	return ret;
+	ret = try_get_cap_refs(ci, need, want, 0, nonblock, got);
+	return ret == -EAGAIN ? 0 : ret;
 }
 
 /*
@@ -2737,7 +2725,7 @@ int ceph_try_get_caps(struct ceph_inode_info *ci, int need, int want,
 int ceph_get_caps(struct ceph_inode_info *ci, int need, int want,
 		  loff_t endoff, int *got, struct page **pinned_page)
 {
-	int _got, ret, err = 0;
+	int _got, ret;
 
 	ret = ceph_pool_perm_check(ci, need);
 	if (ret < 0)
@@ -2747,21 +2735,19 @@ int ceph_get_caps(struct ceph_inode_info *ci, int need, int want,
 		if (endoff > 0)
 			check_max_size(&ci->vfs_inode, endoff);
 
-		err = 0;
 		_got = 0;
 		ret = try_get_cap_refs(ci, need, want, endoff,
-				       false, &_got, &err);
-		if (ret) {
-			if (err == -EAGAIN)
-				continue;
-			if (err < 0)
-				ret = err;
-		} else {
+				       false, &_got);
+		if (ret == -EAGAIN) {
+			continue;
+		} else if (!ret) {
+			int err;
+
 			DEFINE_WAIT_FUNC(wait, woken_wake_function);
 			add_wait_queue(&ci->i_cap_wq, &wait);
 
-			while (!try_get_cap_refs(ci, need, want, endoff,
-						 true, &_got, &err)) {
+			while (!(err = try_get_cap_refs(ci, need, want, endoff,
+							true, &_got))) {
 				if (signal_pending(current)) {
 					ret = -ERESTARTSYS;
 					break;
@@ -2770,19 +2756,14 @@ int ceph_get_caps(struct ceph_inode_info *ci, int need, int want,
 			}
 
 			remove_wait_queue(&ci->i_cap_wq, &wait);
-
 			if (err == -EAGAIN)
 				continue;
-			if (err < 0)
-				ret = err;
 		}
-		if (ret < 0) {
-			if (err == -ESTALE) {
-				/* session was killed, try renew caps */
-				ret = ceph_renew_caps(&ci->vfs_inode);
-				if (ret == 0)
-					continue;
-			}
+		if (ret == -ESTALE) {
+			/* session was killed, try renew caps */
+			ret = ceph_renew_caps(&ci->vfs_inode);
+			if (ret == 0)
+				continue;
 			return ret;
 		}
 
@@ -4099,7 +4080,7 @@ void ceph_put_fmode(struct ceph_inode_info *ci, int fmode)
 }
 
 /*
- * For a soon-to-be unlinked file, drop the AUTH_RDCACHE caps. If it
+ * For a soon-to-be unlinked file, drop the LINK caps. If it
  * looks like the link count will hit 0, drop any other caps (other
  * than PIN) we don't specifically want (due to the file still being
  * open).

fs/ceph/debugfs.c

@@ -37,7 +37,7 @@ static int mdsmap_show(struct seq_file *s, void *p)
 		struct ceph_entity_addr *addr = &mdsmap->m_info[i].addr;
 		int state = mdsmap->m_info[i].state;
 		seq_printf(s, "\tmds%d\t%s\t(%s)\n", i,
-			       ceph_pr_addr(&addr->in_addr),
+			       ceph_pr_addr(addr),
 			       ceph_mds_state_name(state));
 	}
 	return 0;
@@ -88,7 +88,7 @@ static int mdsc_show(struct seq_file *s, void *p)
 				   req->r_dentry,
 				   path ? path : "");
 			spin_unlock(&req->r_dentry->d_lock);
-			kfree(path);
+			ceph_mdsc_free_path(path, pathlen);
 		} else if (req->r_path1) {
 			seq_printf(s, " #%llx/%s", req->r_ino1.ino,
 				   req->r_path1);
@@ -108,7 +108,7 @@ static int mdsc_show(struct seq_file *s, void *p)
 				   req->r_old_dentry,
 				   path ? path : "");
 			spin_unlock(&req->r_old_dentry->d_lock);
-			kfree(path);
+			ceph_mdsc_free_path(path, pathlen);
 		} else if (req->r_path2 && req->r_op != CEPH_MDS_OP_SYMLINK) {
 			if (req->r_ino2.ino)
 				seq_printf(s, " #%llx/%s", req->r_ino2.ino,
@@ -124,18 +124,48 @@ static int mdsc_show(struct seq_file *s, void *p)
 	return 0;
 }
 
+static int caps_show_cb(struct inode *inode, struct ceph_cap *cap, void *p)
+{
+	struct seq_file *s = p;
+
+	seq_printf(s, "0x%-17lx%-17s%-17s\n", inode->i_ino,
+		   ceph_cap_string(cap->issued),
+		   ceph_cap_string(cap->implemented));
+	return 0;
+}
+
 static int caps_show(struct seq_file *s, void *p)
 {
 	struct ceph_fs_client *fsc = s->private;
-	int total, avail, used, reserved, min;
+	struct ceph_mds_client *mdsc = fsc->mdsc;
+	int total, avail, used, reserved, min, i;
 
 	ceph_reservation_status(fsc, &total, &avail, &used, &reserved, &min);
 	seq_printf(s, "total\t\t%d\n"
 		   "avail\t\t%d\n"
 		   "used\t\t%d\n"
 		   "reserved\t%d\n"
-		   "min\t%d\n",
+		   "min\t\t%d\n\n",
 		   total, avail, used, reserved, min);
+	seq_printf(s, "ino                issued           implemented\n");
+	seq_printf(s, "-----------------------------------------------\n");
+
+	mutex_lock(&mdsc->mutex);
+	for (i = 0; i < mdsc->max_sessions; i++) {
+		struct ceph_mds_session *session;
+
+		session = __ceph_lookup_mds_session(mdsc, i);
+		if (!session)
+			continue;
+		mutex_unlock(&mdsc->mutex);
+		mutex_lock(&session->s_mutex);
+		ceph_iterate_session_caps(session, caps_show_cb, s);
+		mutex_unlock(&session->s_mutex);
+		ceph_put_mds_session(session);
+		mutex_lock(&mdsc->mutex);
+	}
+	mutex_unlock(&mdsc->mutex);
+
 	return 0;
 }
 

fs/ceph/file.c

@@ -929,7 +929,7 @@ ceph_direct_read_write(struct kiocb *iocb, struct iov_iter *iter,
 
 	dout("sync_direct_%s on file %p %lld~%u snapc %p seq %lld\n",
 	     (write ? "write" : "read"), file, pos, (unsigned)count,
-	     snapc, snapc->seq);
+	     snapc, snapc ? snapc->seq : 0);
 
 	ret = filemap_write_and_wait_range(inode->i_mapping,
 					   pos, pos + count - 1);

fs/ceph/inode.c

@@ -2266,43 +2266,72 @@ int ceph_permission(struct inode *inode, int mask)
 	return err;
 }
 
+/* Craft a mask of needed caps given a set of requested statx attrs. */
+static int statx_to_caps(u32 want)
+{
+	int mask = 0;
+
+	if (want & (STATX_MODE|STATX_UID|STATX_GID|STATX_CTIME))
+		mask |= CEPH_CAP_AUTH_SHARED;
+
+	if (want & (STATX_NLINK|STATX_CTIME))
+		mask |= CEPH_CAP_LINK_SHARED;
+
+	if (want & (STATX_ATIME|STATX_MTIME|STATX_CTIME|STATX_SIZE|
+		    STATX_BLOCKS))
+		mask |= CEPH_CAP_FILE_SHARED;
+
+	if (want & (STATX_CTIME))
+		mask |= CEPH_CAP_XATTR_SHARED;
+
+	return mask;
+}
+
 /*
- * Get all attributes.  Hopefully somedata we'll have a statlite()
- * and can limit the fields we require to be accurate.
+ * Get all the attributes. If we have sufficient caps for the requested attrs,
+ * then we can avoid talking to the MDS at all.
  */
 int ceph_getattr(const struct path *path, struct kstat *stat,
 		 u32 request_mask, unsigned int flags)
 {
 	struct inode *inode = d_inode(path->dentry);
 	struct ceph_inode_info *ci = ceph_inode(inode);
-	int err;
+	int err = 0;
 
-	err = ceph_do_getattr(inode, CEPH_STAT_CAP_INODE_ALL, false);
-	if (!err) {
-		generic_fillattr(inode, stat);
-		stat->ino = ceph_translate_ino(inode->i_sb, inode->i_ino);
-		if (ceph_snap(inode) == CEPH_NOSNAP)
-			stat->dev = inode->i_sb->s_dev;
+	/* Skip the getattr altogether if we're asked not to sync */
+	if (!(flags & AT_STATX_DONT_SYNC)) {
+		err = ceph_do_getattr(inode, statx_to_caps(request_mask),
+				      flags & AT_STATX_FORCE_SYNC);
+		if (err)
+			return err;
+	}
+
+	generic_fillattr(inode, stat);
+	stat->ino = ceph_translate_ino(inode->i_sb, inode->i_ino);
+	if (ceph_snap(inode) == CEPH_NOSNAP)
+		stat->dev = inode->i_sb->s_dev;
+	else
+		stat->dev = ci->i_snapid_map ? ci->i_snapid_map->dev : 0;
+
+	if (S_ISDIR(inode->i_mode)) {
+		if (ceph_test_mount_opt(ceph_sb_to_client(inode->i_sb),
+					RBYTES))
+			stat->size = ci->i_rbytes;
 		else
-			stat->dev = ci->i_snapid_map ? ci->i_snapid_map->dev : 0;
-
-		if (S_ISDIR(inode->i_mode)) {
-			if (ceph_test_mount_opt(ceph_sb_to_client(inode->i_sb),
-						RBYTES))
-				stat->size = ci->i_rbytes;
-			else
-				stat->size = ci->i_files + ci->i_subdirs;
-			stat->blocks = 0;
-			stat->blksize = 65536;
-			/*
-			 * Some applications rely on the number of st_nlink
-			 * value on directories to be either 0 (if unlinked)
-			 * or 2 + number of subdirectories.
-			 */
-			if (stat->nlink == 1)
-				/* '.' + '..' + subdirs */
-				stat->nlink = 1 + 1 + ci->i_subdirs;
-		}
+			stat->size = ci->i_files + ci->i_subdirs;
+		stat->blocks = 0;
+		stat->blksize = 65536;
+		/*
+		 * Some applications rely on the number of st_nlink
+		 * value on directories to be either 0 (if unlinked)
+		 * or 2 + number of subdirectories.
+		 */
+		if (stat->nlink == 1)
+			/* '.' + '..' + subdirs */
+			stat->nlink = 1 + 1 + ci->i_subdirs;
 	}
+
+	/* Mask off any higher bits (e.g. btime) until we have support */
+	stat->result_mask = request_mask & STATX_BASIC_STATS;
 	return err;
 }

fs/ceph/locks.c

@@ -237,15 +237,6 @@ int ceph_lock(struct file *file, int cmd, struct file_lock *fl)
 	spin_lock(&ci->i_ceph_lock);
 	if (ci->i_ceph_flags & CEPH_I_ERROR_FILELOCK) {
 		err = -EIO;
-	} else if (op == CEPH_MDS_OP_SETFILELOCK) {
-		/*
-		 * increasing i_filelock_ref closes race window between
-		 * handling request reply and adding file_lock struct to
-		 * inode. Otherwise, i_auth_cap may get trimmed in the
-		 * window. Caller function will decrease the counter.
-		 */
-		fl->fl_ops = &ceph_fl_lock_ops;
-		atomic_inc(&ci->i_filelock_ref);
 	}
 	spin_unlock(&ci->i_ceph_lock);
 	if (err < 0) {
@@ -299,10 +290,6 @@ int ceph_flock(struct file *file, int cmd, struct file_lock *fl)
 	spin_lock(&ci->i_ceph_lock);
 	if (ci->i_ceph_flags & CEPH_I_ERROR_FILELOCK) {
 		err = -EIO;
-	} else {
-		/* see comment in ceph_lock */
-		fl->fl_ops = &ceph_fl_lock_ops;
-		atomic_inc(&ci->i_filelock_ref);
 	}
 	spin_unlock(&ci->i_ceph_lock);
 	if (err < 0) {

fs/ceph/mds_client.h

@@ -325,6 +325,18 @@ struct ceph_snapid_map {
 	unsigned long last_used;
 };
 
+/*
+ * node for list of quotarealm inodes that are not visible from the filesystem
+ * mountpoint, but required to handle, e.g. quotas.
+ */
+struct ceph_quotarealm_inode {
+	struct rb_node node;
+	u64 ino;
+	unsigned long timeout; /* last time a lookup failed for this inode */
+	struct mutex mutex;
+	struct inode *inode;
+};
+
 /*
  * mds client state
  */
@@ -344,6 +356,12 @@ struct ceph_mds_client {
 	int                     stopping;      /* true if shutting down */
 
 	atomic64_t		quotarealms_count; /* # realms with quota */
+	/*
+	 * We keep a list of inodes we don't see in the mountpoint but that we
+	 * need to track quota realms.
+	 */
+	struct rb_root		quotarealms_inodes;
+	struct mutex		quotarealms_inodes_mutex;
 
 	/*
 	 * snap_rwsem will cover cap linkage into snaprealms, and
@@ -447,8 +465,9 @@ extern int ceph_alloc_readdir_reply_buffer(struct ceph_mds_request *req,
 					   struct inode *dir);
 extern struct ceph_mds_request *
 ceph_mdsc_create_request(struct ceph_mds_client *mdsc, int op, int mode);
-extern void ceph_mdsc_submit_request(struct ceph_mds_client *mdsc,
-				     struct ceph_mds_request *req);
+extern int ceph_mdsc_submit_request(struct ceph_mds_client *mdsc,
+				    struct inode *dir,
+				    struct ceph_mds_request *req);
 extern int ceph_mdsc_do_request(struct ceph_mds_client *mdsc,
 				struct inode *dir,
 				struct ceph_mds_request *req);
@@ -468,8 +487,18 @@ extern void ceph_flush_cap_releases(struct ceph_mds_client *mdsc,
 				    struct ceph_mds_session *session);
 extern void ceph_queue_cap_reclaim_work(struct ceph_mds_client *mdsc);
 extern void ceph_reclaim_caps_nr(struct ceph_mds_client *mdsc, int nr);
+extern int ceph_iterate_session_caps(struct ceph_mds_session *session,
+				     int (*cb)(struct inode *,
+					       struct ceph_cap *, void *),
+				     void *arg);
 extern void ceph_mdsc_pre_umount(struct ceph_mds_client *mdsc);
 
+static inline void ceph_mdsc_free_path(char *path, int len)
+{
+	if (path)
+		__putname(path - (PATH_MAX - 1 - len));
+}
+
 extern char *ceph_mdsc_build_path(struct dentry *dentry, int *plen, u64 *base,
 				  int stop_on_nosnap);
 

fs/ceph/mdsmap.c

@@ -205,7 +205,7 @@ struct ceph_mdsmap *ceph_mdsmap_decode(void **p, void *end)
 
 		dout("mdsmap_decode %d/%d %lld mds%d.%d %s %s\n",
 		     i+1, n, global_id, mds, inc,
-		     ceph_pr_addr(&addr.in_addr),
+		     ceph_pr_addr(&addr),
 		     ceph_mds_state_name(state));
 
 		if (mds < 0 || state <= 0)

fs/ceph/quota.c

@@ -22,7 +22,16 @@ void ceph_adjust_quota_realms_count(struct inode *inode, bool inc)
 static inline bool ceph_has_realms_with_quotas(struct inode *inode)
 {
 	struct ceph_mds_client *mdsc = ceph_inode_to_client(inode)->mdsc;
-	return atomic64_read(&mdsc->quotarealms_count) > 0;
+	struct super_block *sb = mdsc->fsc->sb;
+
+	if (atomic64_read(&mdsc->quotarealms_count) > 0)
+		return true;
+	/* if root is the real CephFS root, we don't have quota realms */
+	if (sb->s_root->d_inode &&
+	    (sb->s_root->d_inode->i_ino == CEPH_INO_ROOT))
+		return false;
+	/* otherwise, we can't know for sure */
+	return true;
 }
 
 void ceph_handle_quota(struct ceph_mds_client *mdsc,
@@ -68,6 +77,108 @@ void ceph_handle_quota(struct ceph_mds_client *mdsc,
 	iput(inode);
 }
 
+static struct ceph_quotarealm_inode *
+find_quotarealm_inode(struct ceph_mds_client *mdsc, u64 ino)
+{
+	struct ceph_quotarealm_inode *qri = NULL;
+	struct rb_node **node, *parent = NULL;
+
+	mutex_lock(&mdsc->quotarealms_inodes_mutex);
+	node = &(mdsc->quotarealms_inodes.rb_node);
+	while (*node) {
+		parent = *node;
+		qri = container_of(*node, struct ceph_quotarealm_inode, node);
+
+		if (ino < qri->ino)
+			node = &((*node)->rb_left);
+		else if (ino > qri->ino)
+			node = &((*node)->rb_right);
+		else
+			break;
+	}
+	if (!qri || (qri->ino != ino)) {
+		/* Not found, create a new one and insert it */
+		qri = kmalloc(sizeof(*qri), GFP_KERNEL);
+		if (qri) {
+			qri->ino = ino;
+			qri->inode = NULL;
+			qri->timeout = 0;
+			mutex_init(&qri->mutex);
+			rb_link_node(&qri->node, parent, node);
+			rb_insert_color(&qri->node, &mdsc->quotarealms_inodes);
+		} else
+			pr_warn("Failed to alloc quotarealms_inode\n");
+	}
+	mutex_unlock(&mdsc->quotarealms_inodes_mutex);
+
+	return qri;
+}
+
+/*
+ * This function will try to lookup a realm inode which isn't visible in the
+ * filesystem mountpoint.  A list of these kind of inodes (not visible) is
+ * maintained in the mdsc and freed only when the filesystem is umounted.
+ *
+ * Note that these inodes are kept in this list even if the lookup fails, which
+ * allows to prevent useless lookup requests.
+ */
+static struct inode *lookup_quotarealm_inode(struct ceph_mds_client *mdsc,
+					     struct super_block *sb,
+					     struct ceph_snap_realm *realm)
+{
+	struct ceph_quotarealm_inode *qri;
+	struct inode *in;
+
+	qri = find_quotarealm_inode(mdsc, realm->ino);
+	if (!qri)
+		return NULL;
+
+	mutex_lock(&qri->mutex);
+	if (qri->inode) {
+		/* A request has already returned the inode */
+		mutex_unlock(&qri->mutex);
+		return qri->inode;
+	}
+	/* Check if this inode lookup has failed recently */
+	if (qri->timeout &&
+	    time_before_eq(jiffies, qri->timeout)) {
+		mutex_unlock(&qri->mutex);
+		return NULL;
+	}
+	in = ceph_lookup_inode(sb, realm->ino);
+	if (IS_ERR(in)) {
+		pr_warn("Can't lookup inode %llx (err: %ld)\n",
+			realm->ino, PTR_ERR(in));
+		qri->timeout = jiffies + msecs_to_jiffies(60 * 1000); /* XXX */
+	} else {
+		qri->timeout = 0;
+		qri->inode = in;
+	}
+	mutex_unlock(&qri->mutex);
+
+	return in;
+}
+
+void ceph_cleanup_quotarealms_inodes(struct ceph_mds_client *mdsc)
+{
+	struct ceph_quotarealm_inode *qri;
+	struct rb_node *node;
+
+	/*
+	 * It should now be safe to clean quotarealms_inode tree without holding
+	 * mdsc->quotarealms_inodes_mutex...
+	 */
+	mutex_lock(&mdsc->quotarealms_inodes_mutex);
+	while (!RB_EMPTY_ROOT(&mdsc->quotarealms_inodes)) {
+		node = rb_first(&mdsc->quotarealms_inodes);
+		qri = rb_entry(node, struct ceph_quotarealm_inode, node);
+		rb_erase(node, &mdsc->quotarealms_inodes);
+		iput(qri->inode);
+		kfree(qri);
+	}
+	mutex_unlock(&mdsc->quotarealms_inodes_mutex);
+}
+
 /*
  * This function walks through the snaprealm for an inode and returns the
  * ceph_snap_realm for the first snaprealm that has quotas set (either max_files
@@ -76,9 +187,15 @@ void ceph_handle_quota(struct ceph_mds_client *mdsc,
  *
  * Note that the caller is responsible for calling ceph_put_snap_realm() on the
  * returned realm.
+ *
+ * Callers of this function need to hold mdsc->snap_rwsem.  However, if there's
+ * a need to do an inode lookup, this rwsem will be temporarily dropped.  Hence
+ * the 'retry' argument: if rwsem needs to be dropped and 'retry' is 'false'
+ * 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)
+					       struct inode *inode, bool retry)
 {
 	struct ceph_inode_info *ci = NULL;
 	struct ceph_snap_realm *realm, *next;
@@ -88,6 +205,7 @@ static struct ceph_snap_realm *get_quota_realm(struct ceph_mds_client *mdsc,
 	if (ceph_snap(inode) != CEPH_NOSNAP)
 		return NULL;
 
+restart:
 	realm = ceph_inode(inode)->i_snap_realm;
 	if (realm)
 		ceph_get_snap_realm(mdsc, realm);
@@ -95,11 +213,25 @@ static struct ceph_snap_realm *get_quota_realm(struct ceph_mds_client *mdsc,
 		pr_err_ratelimited("get_quota_realm: ino (%llx.%llx) "
 				   "null i_snap_realm\n", ceph_vinop(inode));
 	while (realm) {
+		bool has_inode;
+
 		spin_lock(&realm->inodes_with_caps_lock);
-		in = realm->inode ? igrab(realm->inode) : NULL;
+		has_inode = realm->inode;
+		in = has_inode ? igrab(realm->inode) : NULL;
 		spin_unlock(&realm->inodes_with_caps_lock);
-		if (!in)
+		if (has_inode && !in)
 			break;
+		if (!in) {
+			up_read(&mdsc->snap_rwsem);
+			in = lookup_quotarealm_inode(mdsc, inode->i_sb, realm);
+			down_read(&mdsc->snap_rwsem);
+			if (IS_ERR_OR_NULL(in))
+				break;
+			ceph_put_snap_realm(mdsc, realm);
+			if (!retry)
+				return ERR_PTR(-EAGAIN);
+			goto restart;
+		}
 
 		ci = ceph_inode(in);
 		has_quota = __ceph_has_any_quota(ci);
@@ -125,9 +257,22 @@ bool ceph_quota_is_same_realm(struct inode *old, struct inode *new)
 	struct ceph_snap_realm *old_realm, *new_realm;
 	bool is_same;
 
+restart:
+	/*
+	 * We need to lookup 2 quota realms atomically, i.e. with snap_rwsem.
+	 * However, get_quota_realm may drop it temporarily.  By setting the
+	 * 'retry' parameter to 'false', we'll get -EAGAIN if the rwsem was
+	 * dropped and we can then restart the whole operation.
+	 */
 	down_read(&mdsc->snap_rwsem);
-	old_realm = get_quota_realm(mdsc, old);
-	new_realm = get_quota_realm(mdsc, new);
+	old_realm = get_quota_realm(mdsc, old, true);
+	new_realm = get_quota_realm(mdsc, new, false);
+	if (PTR_ERR(new_realm) == -EAGAIN) {
+		up_read(&mdsc->snap_rwsem);
+		if (old_realm)
+			ceph_put_snap_realm(mdsc, old_realm);
+		goto restart;
+	}
 	is_same = (old_realm == new_realm);
 	up_read(&mdsc->snap_rwsem);
 
@@ -166,6 +311,7 @@ static bool check_quota_exceeded(struct inode *inode, enum quota_check_op op,
 		return false;
 
 	down_read(&mdsc->snap_rwsem);
+restart:
 	realm = ceph_inode(inode)->i_snap_realm;
 	if (realm)
 		ceph_get_snap_realm(mdsc, realm);
@@ -173,12 +319,23 @@ static bool check_quota_exceeded(struct inode *inode, enum quota_check_op op,
 		pr_err_ratelimited("check_quota_exceeded: ino (%llx.%llx) "
 				   "null i_snap_realm\n", ceph_vinop(inode));
 	while (realm) {
+		bool has_inode;
+
 		spin_lock(&realm->inodes_with_caps_lock);
-		in = realm->inode ? igrab(realm->inode) : NULL;
+		has_inode = realm->inode;
+		in = has_inode ? igrab(realm->inode) : NULL;
 		spin_unlock(&realm->inodes_with_caps_lock);
-		if (!in)
+		if (has_inode && !in)
 			break;
-
+		if (!in) {
+			up_read(&mdsc->snap_rwsem);
+			in = lookup_quotarealm_inode(mdsc, inode->i_sb, realm);
+			down_read(&mdsc->snap_rwsem);
+			if (IS_ERR_OR_NULL(in))
+				break;
+			ceph_put_snap_realm(mdsc, realm);
+			goto restart;
+		}
 		ci = ceph_inode(in);
 		spin_lock(&ci->i_ceph_lock);
 		if (op == QUOTA_CHECK_MAX_FILES_OP) {
@@ -314,7 +471,7 @@ 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));
+	realm = get_quota_realm(mdsc, d_inode(fsc->sb->s_root), true);
 	up_read(&mdsc->snap_rwsem);
 	if (!realm)
 		return false;

fs/ceph/super.c

@@ -845,6 +845,12 @@ static void ceph_umount_begin(struct super_block *sb)
 	return;
 }
 
+static int ceph_remount(struct super_block *sb, int *flags, char *data)
+{
+	sync_filesystem(sb);
+	return 0;
+}
+
 static const struct super_operations ceph_super_ops = {
 	.alloc_inode	= ceph_alloc_inode,
 	.destroy_inode	= ceph_destroy_inode,
@@ -853,6 +859,7 @@ static const struct super_operations ceph_super_ops = {
 	.drop_inode	= ceph_drop_inode,
 	.sync_fs        = ceph_sync_fs,
 	.put_super	= ceph_put_super,
+	.remount_fs	= ceph_remount,
 	.show_options   = ceph_show_options,
 	.statfs		= ceph_statfs,
 	.umount_begin   = ceph_umount_begin,

fs/ceph/super.h

@@ -1083,6 +1083,7 @@ extern long ceph_ioctl(struct file *file, unsigned int cmd, unsigned long arg);
 
 /* export.c */
 extern const struct export_operations ceph_export_ops;
+struct inode *ceph_lookup_inode(struct super_block *sb, u64 ino);
 
 /* locks.c */
 extern __init void ceph_flock_init(void);
@@ -1133,5 +1134,6 @@ extern bool ceph_quota_is_max_bytes_approaching(struct inode *inode,
 						loff_t newlen);
 extern bool ceph_quota_update_statfs(struct ceph_fs_client *fsc,
 				     struct kstatfs *buf);
+extern void ceph_cleanup_quotarealms_inodes(struct ceph_mds_client *mdsc);
 
 #endif /* _FS_CEPH_SUPER_H */

include/linux/ceph/ceph_fs.h

@@ -436,6 +436,12 @@ union ceph_mds_request_args {
 		__le64 length; /* num bytes to lock from start */
 		__u8 wait; /* will caller wait for lock to become available? */
 	} __attribute__ ((packed)) filelock_change;
+	struct {
+		__le32 mask;                 /* CEPH_CAP_* */
+		__le64 snapid;
+		__le64 parent;
+		__le32 hash;
+	} __attribute__ ((packed)) lookupino;
 } __attribute__ ((packed));
 
 #define CEPH_MDS_FLAG_REPLAY        1  /* this is a replayed op */

include/linux/ceph/messenger.h

@@ -323,7 +323,8 @@ struct ceph_connection {
 };
 
 
-extern const char *ceph_pr_addr(const struct sockaddr_storage *ss);
+extern const char *ceph_pr_addr(const struct ceph_entity_addr *addr);
+
 extern int ceph_parse_ips(const char *c, const char *end,
 			  struct ceph_entity_addr *addr,
 			  int max_count, int *count);

include/linux/ceph/osdmap.h

@@ -110,17 +110,16 @@ struct ceph_object_id {
 	int name_len;
 };
 
+#define __CEPH_OID_INITIALIZER(oid) { .name = (oid).inline_name }
+
+#define CEPH_DEFINE_OID_ONSTACK(oid)				\
+	struct ceph_object_id oid = __CEPH_OID_INITIALIZER(oid)
+
 static inline void ceph_oid_init(struct ceph_object_id *oid)
 {
-	oid->name = oid->inline_name;
-	oid->name_len = 0;
+	*oid = (struct ceph_object_id) __CEPH_OID_INITIALIZER(*oid);
 }
 
-#define CEPH_OID_INIT_ONSTACK(oid)					\
-    ({ ceph_oid_init(&oid); oid; })
-#define CEPH_DEFINE_OID_ONSTACK(oid)					\
-	struct ceph_object_id oid = CEPH_OID_INIT_ONSTACK(oid)
-
 static inline bool ceph_oid_empty(const struct ceph_object_id *oid)
 {
 	return oid->name == oid->inline_name && !oid->name_len;

net/ceph/cls_lock_client.c

@@ -271,7 +271,7 @@ static int decode_locker(void **p, void *end, struct ceph_locker *locker)
 
 	dout("%s %s%llu cookie %s addr %s\n", __func__,
 	     ENTITY_NAME(locker->id.name), locker->id.cookie,
-	     ceph_pr_addr(&locker->info.addr.in_addr));
+	     ceph_pr_addr(&locker->info.addr));
 	return 0;
 }
 

net/ceph/debugfs.c

@@ -46,7 +46,7 @@ static int monmap_show(struct seq_file *s, void *p)
 
 		seq_printf(s, "\t%s%lld\t%s\n",
 			   ENTITY_NAME(inst->name),
-			   ceph_pr_addr(&inst->addr.in_addr));
+			   ceph_pr_addr(&inst->addr));
 	}
 	return 0;
 }
@@ -82,7 +82,7 @@ static int osdmap_show(struct seq_file *s, void *p)
 		char sb[64];
 
 		seq_printf(s, "osd%d\t%s\t%3d%%\t(%s)\t%3d%%\n",
-			   i, ceph_pr_addr(&addr->in_addr),
+			   i, ceph_pr_addr(addr),
 			   ((map->osd_weight[i]*100) >> 16),
 			   ceph_osdmap_state_str(sb, sizeof(sb), state),
 			   ((ceph_get_primary_affinity(map, i)*100) >> 16));

net/ceph/mon_client.c

@@ -76,7 +76,7 @@ struct ceph_monmap *ceph_monmap_decode(void *p, void *end)
 	     m->num_mon);
 	for (i = 0; i < m->num_mon; i++)
 		dout("monmap_decode  mon%d is %s\n", i,
-		     ceph_pr_addr(&m->mon_inst[i].addr.in_addr));
+		     ceph_pr_addr(&m->mon_inst[i].addr));
 	return m;
 
 bad:
@@ -203,7 +203,7 @@ static void reopen_session(struct ceph_mon_client *monc)
 {
 	if (!monc->hunting)
 		pr_info("mon%d %s session lost, hunting for new mon\n",
-		    monc->cur_mon, ceph_pr_addr(&monc->con.peer_addr.in_addr));
+		    monc->cur_mon, ceph_pr_addr(&monc->con.peer_addr));
 
 	__close_session(monc);
 	__open_session(monc);
@@ -1178,7 +1178,7 @@ static void handle_auth_reply(struct ceph_mon_client *monc,
 		__resend_generic_request(monc);
 
 		pr_info("mon%d %s session established\n", monc->cur_mon,
-			ceph_pr_addr(&monc->con.peer_addr.in_addr));
+			ceph_pr_addr(&monc->con.peer_addr));
 	}
 
 out:

net/ceph/osd_client.c

@@ -4926,7 +4926,7 @@ static int decode_watcher(void **p, void *end, struct ceph_watch_item *item)
 
 	dout("%s %s%llu cookie %llu addr %s\n", __func__,
 	     ENTITY_NAME(item->name), item->cookie,
-	     ceph_pr_addr(&item->addr.in_addr));
+	     ceph_pr_addr(&item->addr));
 	return 0;
 }