Merge tag 'ceph-for-4.13-rc1' of git://github.com/ceph/ceph-client

Pull ceph updates from Ilya Dryomov:
 "The main item here is support for v12.y.z ("Luminous") clusters:
  RESEND_ON_SPLIT, RADOS_BACKOFF, OSDMAP_PG_UPMAP and CRUSH_CHOOSE_ARGS
  feature bits, and various other changes in the RADOS client protocol.

  On top of that we have a new fsc mount option to allow supplying
  fscache uniquifier (similar to NFS) and the usual pile of filesystem
  fixes from Zheng"

* tag 'ceph-for-4.13-rc1' of git://github.com/ceph/ceph-client: (44 commits)
  libceph: advertise support for NEW_OSDOP_ENCODING and SERVER_LUMINOUS
  libceph: osd_state is 32 bits wide in luminous
  crush: remove an obsolete comment
  crush: crush_init_workspace starts with struct crush_work
  libceph, crush: per-pool crush_choose_arg_map for crush_do_rule()
  crush: implement weight and id overrides for straw2
  libceph: apply_upmap()
  libceph: compute actual pgid in ceph_pg_to_up_acting_osds()
  libceph: pg_upmap[_items] infrastructure
  libceph: ceph_decode_skip_* helpers
  libceph: kill __{insert,lookup,remove}_pg_mapping()
  libceph: introduce and switch to decode_pg_mapping()
  libceph: don't pass pgid by value
  libceph: respect RADOS_BACKOFF backoffs
  libceph: make DEFINE_RB_* helpers more general
  libceph: avoid unnecessary pi lookups in calc_target()
  libceph: use target pi for calc_target() calculations
  libceph: always populate t->target_{oid,oloc} in calc_target()
  libceph: make sure need_resend targets reflect latest map
  libceph: delete from need_resend_linger before check_linger_pool_dne()
  ...

fs/ceph/addr.c

@@ -530,14 +530,10 @@ static int writepage_nounlock(struct page *page, struct writeback_control *wbc)
 	long writeback_stat;
 	u64 truncate_size;
 	u32 truncate_seq;
-	int err = 0, len = PAGE_SIZE;
+	int err, len = PAGE_SIZE;
 
 	dout("writepage %p idx %lu\n", page, page->index);
 
-	if (!page->mapping || !page->mapping->host) {
-		dout("writepage %p - no mapping\n", page);
-		return -EFAULT;
-	}
 	inode = page->mapping->host;
 	ci = ceph_inode(inode);
 	fsc = ceph_inode_to_client(inode);
@@ -547,17 +543,18 @@ static int writepage_nounlock(struct page *page, struct writeback_control *wbc)
 	snapc = page_snap_context(page);
 	if (snapc == NULL) {
 		dout("writepage %p page %p not dirty?\n", inode, page);
-		goto out;
+		return 0;
 	}
 	oldest = get_oldest_context(inode, &snap_size,
 				    &truncate_size, &truncate_seq);
 	if (snapc->seq > oldest->seq) {
 		dout("writepage %p page %p snapc %p not writeable - noop\n",
 		     inode, page, snapc);
 		/* we should only noop if called by kswapd */
-		WARN_ON((current->flags & PF_MEMALLOC) == 0);
+		WARN_ON(!(current->flags & PF_MEMALLOC));
 		ceph_put_snap_context(oldest);
-		goto out;
+		redirty_page_for_writepage(wbc, page);
+		return 0;
 	}
 	ceph_put_snap_context(oldest);
 
@@ -567,8 +564,9 @@ static int writepage_nounlock(struct page *page, struct writeback_control *wbc)
 	/* is this a partial page at end of file? */
 	if (page_off >= snap_size) {
 		dout("%p page eof %llu\n", page, snap_size);
-		goto out;
+		return 0;
 	}
+
 	if (snap_size < page_off + len)
 		len = snap_size - page_off;
 
@@ -595,7 +593,7 @@ static int writepage_nounlock(struct page *page, struct writeback_control *wbc)
 			dout("writepage interrupted page %p\n", page);
 			redirty_page_for_writepage(wbc, page);
 			end_page_writeback(page);
-			goto out;
+			return err;
 		}
 		dout("writepage setting page/mapping error %d %p\n",
 		     err, page);
@@ -611,7 +609,6 @@ static int writepage_nounlock(struct page *page, struct writeback_control *wbc)
 	end_page_writeback(page);
 	ceph_put_wrbuffer_cap_refs(ci, 1, snapc);
 	ceph_put_snap_context(snapc);  /* page's reference */
-out:
 	return err;
 }
 
@@ -1318,7 +1315,7 @@ static int ceph_write_end(struct file *file, struct address_space *mapping,
 			  struct page *page, void *fsdata)
 {
 	struct inode *inode = file_inode(file);
-	int check_cap = 0;
+	bool check_cap = false;
 
 	dout("write_end file %p inode %p page %p %d~%d (%d)\n", file,
 	     inode, page, (int)pos, (int)copied, (int)len);

fs/ceph/cache.c

@@ -35,18 +35,34 @@ struct fscache_netfs ceph_cache_netfs = {
 	.version	= 0,
 };
 
+static DEFINE_MUTEX(ceph_fscache_lock);
+static LIST_HEAD(ceph_fscache_list);
+
+struct ceph_fscache_entry {
+	struct list_head list;
+	struct fscache_cookie *fscache;
+	struct ceph_fsid fsid;
+	size_t uniq_len;
+	char uniquifier[0];
+};
+
 static uint16_t ceph_fscache_session_get_key(const void *cookie_netfs_data,
 					     void *buffer, uint16_t maxbuf)
 {
 	const struct ceph_fs_client* fsc = cookie_netfs_data;
-	uint16_t klen;
+	const char *fscache_uniq = fsc->mount_options->fscache_uniq;
+	uint16_t fsid_len, uniq_len;
 
-	klen = sizeof(fsc->client->fsid);
-	if (klen > maxbuf)
+	fsid_len = sizeof(fsc->client->fsid);
+	uniq_len = fscache_uniq ? strlen(fscache_uniq) : 0;
+	if (fsid_len + uniq_len > maxbuf)
 		return 0;
 
-	memcpy(buffer, &fsc->client->fsid, klen);
-	return klen;
+	memcpy(buffer, &fsc->client->fsid, fsid_len);
+	if (uniq_len)
+		memcpy(buffer + fsid_len, fscache_uniq, uniq_len);
+
+	return fsid_len + uniq_len;
 }
 
 static const struct fscache_cookie_def ceph_fscache_fsid_object_def = {
@@ -67,13 +83,54 @@ void ceph_fscache_unregister(void)
 
 int ceph_fscache_register_fs(struct ceph_fs_client* fsc)
 {
+	const struct ceph_fsid *fsid = &fsc->client->fsid;
+	const char *fscache_uniq = fsc->mount_options->fscache_uniq;
+	size_t uniq_len = fscache_uniq ? strlen(fscache_uniq) : 0;
+	struct ceph_fscache_entry *ent;
+	int err = 0;
+
+	mutex_lock(&ceph_fscache_lock);
+	list_for_each_entry(ent, &ceph_fscache_list, list) {
+		if (memcmp(&ent->fsid, fsid, sizeof(*fsid)))
+			continue;
+		if (ent->uniq_len != uniq_len)
+			continue;
+		if (uniq_len && memcmp(ent->uniquifier, fscache_uniq, uniq_len))
+			continue;
+
+		pr_err("fscache cookie already registered for fsid %pU\n", fsid);
+		pr_err("  use fsc=%%s mount option to specify a uniquifier\n");
+		err = -EBUSY;
+		goto out_unlock;
+	}
+
+	ent = kzalloc(sizeof(*ent) + uniq_len, GFP_KERNEL);
+	if (!ent) {
+		err = -ENOMEM;
+		goto out_unlock;
+	}
+
 	fsc->fscache = fscache_acquire_cookie(ceph_cache_netfs.primary_index,
 					      &ceph_fscache_fsid_object_def,
 					      fsc, true);
-	if (!fsc->fscache)
-		pr_err("Unable to register fsid: %p fscache cookie\n", fsc);
 
-	return 0;
+	if (fsc->fscache) {
+		memcpy(&ent->fsid, fsid, sizeof(*fsid));
+		if (uniq_len > 0) {
+			memcpy(&ent->uniquifier, fscache_uniq, uniq_len);
+			ent->uniq_len = uniq_len;
+		}
+		ent->fscache = fsc->fscache;
+		list_add_tail(&ent->list, &ceph_fscache_list);
+	} else {
+		kfree(ent);
+		pr_err("unable to register fscache cookie for fsid %pU\n",
+		       fsid);
+		/* all other fs ignore this error */
+	}
+out_unlock:
+	mutex_unlock(&ceph_fscache_lock);
+	return err;
 }
 
 static uint16_t ceph_fscache_inode_get_key(const void *cookie_netfs_data,
@@ -349,7 +406,24 @@ void ceph_invalidate_fscache_page(struct inode* inode, struct page *page)
 
 void ceph_fscache_unregister_fs(struct ceph_fs_client* fsc)
 {
-	fscache_relinquish_cookie(fsc->fscache, 0);
+	if (fscache_cookie_valid(fsc->fscache)) {
+		struct ceph_fscache_entry *ent;
+		bool found = false;
+
+		mutex_lock(&ceph_fscache_lock);
+		list_for_each_entry(ent, &ceph_fscache_list, list) {
+			if (ent->fscache == fsc->fscache) {
+				list_del(&ent->list);
+				kfree(ent);
+				found = true;
+				break;
+			}
+		}
+		WARN_ON_ONCE(!found);
+		mutex_unlock(&ceph_fscache_lock);
+
+		__fscache_relinquish_cookie(fsc->fscache, 0);
+	}
 	fsc->fscache = NULL;
 }
 

fs/ceph/caps.c

@@ -1653,6 +1653,21 @@ static int try_nonblocking_invalidate(struct inode *inode)
 	return -1;
 }
 
+bool __ceph_should_report_size(struct ceph_inode_info *ci)
+{
+	loff_t size = ci->vfs_inode.i_size;
+	/* mds will adjust max size according to the reported size */
+	if (ci->i_flushing_caps & CEPH_CAP_FILE_WR)
+		return false;
+	if (size >= ci->i_max_size)
+		return true;
+	/* half of previous max_size increment has been used */
+	if (ci->i_max_size > ci->i_reported_size &&
+	    (size << 1) >= ci->i_max_size + ci->i_reported_size)
+		return true;
+	return false;
+}
+
 /*
  * Swiss army knife function to examine currently used and wanted
  * versus held caps.  Release, flush, ack revoked caps to mds as
@@ -1806,8 +1821,7 @@ void ceph_check_caps(struct ceph_inode_info *ci, int flags,
 			}
 
 			/* approaching file_max? */
-			if ((inode->i_size << 1) >= ci->i_max_size &&
-			    (ci->i_reported_size << 1) < ci->i_max_size) {
+			if (__ceph_should_report_size(ci)) {
 				dout("i_size approaching max_size\n");
 				goto ack;
 			}
@@ -3027,8 +3041,10 @@ static void handle_cap_grant(struct ceph_mds_client *mdsc,
 					le32_to_cpu(grant->truncate_seq),
 					le64_to_cpu(grant->truncate_size),
 					size);
-		/* max size increase? */
-		if (ci->i_auth_cap == cap && max_size != ci->i_max_size) {
+	}
+
+	if (ci->i_auth_cap == cap && (newcaps & CEPH_CAP_ANY_FILE_WR)) {
+		if (max_size != ci->i_max_size) {
 			dout("max_size %lld -> %llu\n",
 			     ci->i_max_size, max_size);
 			ci->i_max_size = max_size;
@@ -3037,6 +3053,10 @@ static void handle_cap_grant(struct ceph_mds_client *mdsc,
 				ci->i_requested_max_size = 0;
 			}
 			wake = true;
+		} else if (ci->i_wanted_max_size > ci->i_max_size &&
+			   ci->i_wanted_max_size > ci->i_requested_max_size) {
+			/* CEPH_CAP_OP_IMPORT */
+			wake = true;
 		}
 	}
 
@@ -3554,7 +3574,6 @@ static void handle_cap_import(struct ceph_mds_client *mdsc,
 	}
 
 	/* make sure we re-request max_size, if necessary */
-	ci->i_wanted_max_size = 0;
 	ci->i_requested_max_size = 0;
 
 	*old_issued = issued;
@@ -3790,6 +3809,7 @@ void ceph_handle_caps(struct ceph_mds_session *session,
  */
 void ceph_check_delayed_caps(struct ceph_mds_client *mdsc)
 {
+	struct inode *inode;
 	struct ceph_inode_info *ci;
 	int flags = CHECK_CAPS_NODELAY;
 
@@ -3805,9 +3825,15 @@ void ceph_check_delayed_caps(struct ceph_mds_client *mdsc)
 		    time_before(jiffies, ci->i_hold_caps_max))
 			break;
 		list_del_init(&ci->i_cap_delay_list);
+
+		inode = igrab(&ci->vfs_inode);
 		spin_unlock(&mdsc->cap_delay_lock);
-		dout("check_delayed_caps on %p\n", &ci->vfs_inode);
-		ceph_check_caps(ci, flags, NULL);
+
+		if (inode) {
+			dout("check_delayed_caps on %p\n", inode);
+			ceph_check_caps(ci, flags, NULL);
+			iput(inode);
+		}
 	}
 	spin_unlock(&mdsc->cap_delay_lock);
 }

fs/ceph/file.c

@@ -1040,8 +1040,8 @@ ceph_sync_write(struct kiocb *iocb, struct iov_iter *from, loff_t pos,
 	int num_pages;
 	int written = 0;
 	int flags;
-	int check_caps = 0;
 	int ret;
+	bool check_caps = false;
 	struct timespec mtime = current_time(inode);
 	size_t count = iov_iter_count(from);
 

fs/ceph/inode.c

@@ -1016,6 +1016,7 @@ static void update_dentry_lease(struct dentry *dentry,
 	long unsigned ttl = from_time + (duration * HZ) / 1000;
 	long unsigned half_ttl = from_time + (duration * HZ / 2) / 1000;
 	struct inode *dir;
+	struct ceph_mds_session *old_lease_session = NULL;
 
 	/*
 	 * Make sure dentry's inode matches tgt_vino. NULL tgt_vino means that
@@ -1051,8 +1052,10 @@ static void update_dentry_lease(struct dentry *dentry,
 	    time_before(ttl, di->time))
 		goto out_unlock;  /* we already have a newer lease. */
 
-	if (di->lease_session && di->lease_session != session)
-		goto out_unlock;
+	if (di->lease_session && di->lease_session != session) {
+		old_lease_session = di->lease_session;
+		di->lease_session = NULL;
+	}
 
 	ceph_dentry_lru_touch(dentry);
 
@@ -1065,6 +1068,8 @@ static void update_dentry_lease(struct dentry *dentry,
 	di->time = ttl;
 out_unlock:
 	spin_unlock(&dentry->d_lock);
+	if (old_lease_session)
+		ceph_put_mds_session(old_lease_session);
 	return;
 }
 
@@ -1653,20 +1658,17 @@ int ceph_readdir_prepopulate(struct ceph_mds_request *req,
 	return err;
 }
 
-int ceph_inode_set_size(struct inode *inode, loff_t size)
+bool ceph_inode_set_size(struct inode *inode, loff_t size)
 {
 	struct ceph_inode_info *ci = ceph_inode(inode);
-	int ret = 0;
+	bool ret;
 
 	spin_lock(&ci->i_ceph_lock);
 	dout("set_size %p %llu -> %llu\n", inode, inode->i_size, size);
 	i_size_write(inode, size);
 	inode->i_blocks = calc_inode_blocks(size);
 
-	/* tell the MDS if we are approaching max_size */
-	if ((size << 1) >= ci->i_max_size &&
-	    (ci->i_reported_size << 1) < ci->i_max_size)
-		ret = 1;
+	ret = __ceph_should_report_size(ci);
 
 	spin_unlock(&ci->i_ceph_lock);
 	return ret;

fs/ceph/locks.c

@@ -127,6 +127,29 @@ static int ceph_lock_wait_for_completion(struct ceph_mds_client *mdsc,
 	dout("ceph_lock_wait_for_completion: request %llu was interrupted\n",
 	     req->r_tid);
 
+	mutex_lock(&mdsc->mutex);
+	if (test_bit(CEPH_MDS_R_GOT_RESULT, &req->r_req_flags)) {
+		err = 0;
+	} else {
+		/*
+		 * ensure we aren't running concurrently with
+		 * ceph_fill_trace or ceph_readdir_prepopulate, which
+		 * rely on locks (dir mutex) held by our caller.
+		 */
+		mutex_lock(&req->r_fill_mutex);
+		req->r_err = err;
+		set_bit(CEPH_MDS_R_ABORTED, &req->r_req_flags);
+		mutex_unlock(&req->r_fill_mutex);
+
+		if (!req->r_session) {
+			// haven't sent the request
+			err = 0;
+		}
+	}
+	mutex_unlock(&mdsc->mutex);
+	if (!err)
+		return 0;
+
 	intr_req = ceph_mdsc_create_request(mdsc, CEPH_MDS_OP_SETFILELOCK,
 					    USE_AUTH_MDS);
 	if (IS_ERR(intr_req))
@@ -146,7 +169,7 @@ static int ceph_lock_wait_for_completion(struct ceph_mds_client *mdsc,
 	if (err && err != -ERESTARTSYS)
 		return err;
 
-	wait_for_completion(&req->r_completion);
+	wait_for_completion_killable(&req->r_safe_completion);
 	return 0;
 }