Merge tag 'ceph-for-5.6-rc8' of git://github.com/ceph/ceph-client

Pull ceph fixes from Ilya Dryomov:
 "A patch for a rather old regression in fullness handling and two
  memory leak fixes, marked for stable"

* tag 'ceph-for-5.6-rc8' of git://github.com/ceph/ceph-client:
  ceph: fix memory leak in ceph_cleanup_snapid_map()
  libceph: fix alloc_msg_with_page_vector() memory leaks
  ceph: check POOL_FLAG_FULL/NEARFULL in addition to OSDMAP_FULL/NEARFULL

Author: torvalds

fs/ceph/file.c

@@ -1415,10 +1415,13 @@ static ssize_t ceph_write_iter(struct kiocb *iocb, struct iov_iter *from)
 	struct inode *inode = file_inode(file);
 	struct ceph_inode_info *ci = ceph_inode(inode);
 	struct ceph_fs_client *fsc = ceph_inode_to_client(inode);
+	struct ceph_osd_client *osdc = &fsc->client->osdc;
 	struct ceph_cap_flush *prealloc_cf;
 	ssize_t count, written = 0;
 	int err, want, got;
 	bool direct_lock = false;
+	u32 map_flags;
+	u64 pool_flags;
 	loff_t pos;
 	loff_t limit = max(i_size_read(inode), fsc->max_file_size);
 
@@ -1481,8 +1484,12 @@ static ssize_t ceph_write_iter(struct kiocb *iocb, struct iov_iter *from)
 			goto out;
 	}
 
-	/* FIXME: not complete since it doesn't account for being at quota */
-	if (ceph_osdmap_flag(&fsc->client->osdc, CEPH_OSDMAP_FULL)) {
+	down_read(&osdc->lock);
+	map_flags = osdc->osdmap->flags;
+	pool_flags = ceph_pg_pool_flags(osdc->osdmap, ci->i_layout.pool_id);
+	up_read(&osdc->lock);
+	if ((map_flags & CEPH_OSDMAP_FULL) ||
+	    (pool_flags & CEPH_POOL_FLAG_FULL)) {
 		err = -ENOSPC;
 		goto out;
 	}
@@ -1575,7 +1582,8 @@ static ssize_t ceph_write_iter(struct kiocb *iocb, struct iov_iter *from)
 	}
 
 	if (written >= 0) {
-		if (ceph_osdmap_flag(&fsc->client->osdc, CEPH_OSDMAP_NEARFULL))
+		if ((map_flags & CEPH_OSDMAP_NEARFULL) ||
+		    (pool_flags & CEPH_POOL_FLAG_NEARFULL))
 			iocb->ki_flags |= IOCB_DSYNC;
 		written = generic_write_sync(iocb, written);
 	}

fs/ceph/snap.c

@@ -1155,5 +1155,6 @@ void ceph_cleanup_snapid_map(struct ceph_mds_client *mdsc)
 			pr_err("snapid map %llx -> %x still in use\n",
 			       sm->snap, sm->dev);
 		}
+		kfree(sm);
 	}
 }

include/linux/ceph/messenger.h

@@ -175,9 +175,10 @@ struct ceph_msg_data {
 #endif /* CONFIG_BLOCK */
 		struct ceph_bvec_iter	bvec_pos;
 		struct {
-			struct page	**pages;	/* NOT OWNER. */
+			struct page	**pages;
 			size_t		length;		/* total # bytes */
 			unsigned int	alignment;	/* first page */
+			bool		own_pages;
 		};
 		struct ceph_pagelist	*pagelist;
 	};
@@ -356,8 +357,8 @@ extern void ceph_con_keepalive(struct ceph_connection *con);
 extern bool ceph_con_keepalive_expired(struct ceph_connection *con,
 				       unsigned long interval);
 
-extern void ceph_msg_data_add_pages(struct ceph_msg *msg, struct page **pages,
-				size_t length, size_t alignment);
+void ceph_msg_data_add_pages(struct ceph_msg *msg, struct page **pages,
+			     size_t length, size_t alignment, bool own_pages);
 extern void ceph_msg_data_add_pagelist(struct ceph_msg *msg,
 				struct ceph_pagelist *pagelist);
 #ifdef CONFIG_BLOCK

include/linux/ceph/osdmap.h

@@ -37,6 +37,9 @@ int ceph_spg_compare(const struct ceph_spg *lhs, const struct ceph_spg *rhs);
 #define CEPH_POOL_FLAG_HASHPSPOOL	(1ULL << 0) /* hash pg seed and pool id
 						       together */
 #define CEPH_POOL_FLAG_FULL		(1ULL << 1) /* pool is full */
+#define CEPH_POOL_FLAG_FULL_QUOTA	(1ULL << 10) /* pool ran out of quota,
+							will set FULL too */
+#define CEPH_POOL_FLAG_NEARFULL		(1ULL << 11) /* pool is nearfull */
 
 struct ceph_pg_pool_info {
 	struct rb_node node;
@@ -304,5 +307,6 @@ extern struct ceph_pg_pool_info *ceph_pg_pool_by_id(struct ceph_osdmap *map,
 
 extern const char *ceph_pg_pool_name_by_id(struct ceph_osdmap *map, u64 id);
 extern int ceph_pg_poolid_by_name(struct ceph_osdmap *map, const char *name);
+u64 ceph_pg_pool_flags(struct ceph_osdmap *map, u64 id);
 
 #endif

include/linux/ceph/rados.h

@@ -143,8 +143,10 @@ extern const char *ceph_osd_state_name(int s);
 /*
  * osd map flag bits
  */
-#define CEPH_OSDMAP_NEARFULL (1<<0)  /* sync writes (near ENOSPC) */
-#define CEPH_OSDMAP_FULL     (1<<1)  /* no data writes (ENOSPC) */
+#define CEPH_OSDMAP_NEARFULL (1<<0)  /* sync writes (near ENOSPC),
+					not set since ~luminous */
+#define CEPH_OSDMAP_FULL     (1<<1)  /* no data writes (ENOSPC),
+					not set since ~luminous */
 #define CEPH_OSDMAP_PAUSERD  (1<<2)  /* pause all reads */
 #define CEPH_OSDMAP_PAUSEWR  (1<<3)  /* pause all writes */
 #define CEPH_OSDMAP_PAUSEREC (1<<4)  /* pause recovery */

net/ceph/messenger.c

@@ -3248,12 +3248,16 @@ static struct ceph_msg_data *ceph_msg_data_add(struct ceph_msg *msg)
 
 static void ceph_msg_data_destroy(struct ceph_msg_data *data)
 {
-	if (data->type == CEPH_MSG_DATA_PAGELIST)
+	if (data->type == CEPH_MSG_DATA_PAGES && data->own_pages) {
+		int num_pages = calc_pages_for(data->alignment, data->length);
+		ceph_release_page_vector(data->pages, num_pages);
+	} else if (data->type == CEPH_MSG_DATA_PAGELIST) {
 		ceph_pagelist_release(data->pagelist);
+	}
 }
 
 void ceph_msg_data_add_pages(struct ceph_msg *msg, struct page **pages,
-		size_t length, size_t alignment)
+			     size_t length, size_t alignment, bool own_pages)
 {
 	struct ceph_msg_data *data;
 
@@ -3265,6 +3269,7 @@ void ceph_msg_data_add_pages(struct ceph_msg *msg, struct page **pages,
 	data->pages = pages;
 	data->length = length;
 	data->alignment = alignment & ~PAGE_MASK;
+	data->own_pages = own_pages;
 
 	msg->data_length += length;
 }

net/ceph/osd_client.c

@@ -962,7 +962,7 @@ static void ceph_osdc_msg_data_add(struct ceph_msg *msg,
 		BUG_ON(length > (u64) SIZE_MAX);
 		if (length)
 			ceph_msg_data_add_pages(msg, osd_data->pages,
-					length, osd_data->alignment);
+					length, osd_data->alignment, false);
 	} else if (osd_data->type == CEPH_OSD_DATA_TYPE_PAGELIST) {
 		BUG_ON(!length);
 		ceph_msg_data_add_pagelist(msg, osd_data->pagelist);
@@ -4436,9 +4436,7 @@ static void handle_watch_notify(struct ceph_osd_client *osdc,
 							CEPH_MSG_DATA_PAGES);
 					*lreq->preply_pages = data->pages;
 					*lreq->preply_len = data->length;
-				} else {
-					ceph_release_page_vector(data->pages,
-					       calc_pages_for(0, data->length));
+					data->own_pages = false;
 				}
 			}
 			lreq->notify_finish_error = return_code;
@@ -5506,9 +5504,6 @@ static struct ceph_msg *get_reply(struct ceph_connection *con,
 	return m;
 }
 
-/*
- * TODO: switch to a msg-owned pagelist
- */
 static struct ceph_msg *alloc_msg_with_page_vector(struct ceph_msg_header *hdr)
 {
 	struct ceph_msg *m;
@@ -5522,7 +5517,6 @@ static struct ceph_msg *alloc_msg_with_page_vector(struct ceph_msg_header *hdr)
 
 	if (data_len) {
 		struct page **pages;
-		struct ceph_osd_data osd_data;
 
 		pages = ceph_alloc_page_vector(calc_pages_for(0, data_len),
 					       GFP_NOIO);
@@ -5531,9 +5525,7 @@ static struct ceph_msg *alloc_msg_with_page_vector(struct ceph_msg_header *hdr)
 			return NULL;
 		}
 
-		ceph_osd_data_pages_init(&osd_data, pages, data_len, 0, false,
-					 false);
-		ceph_osdc_msg_data_add(m, &osd_data);
+		ceph_msg_data_add_pages(m, pages, data_len, 0, true);
 	}
 
 	return m;

net/ceph/osdmap.c

@@ -710,6 +710,15 @@ int ceph_pg_poolid_by_name(struct ceph_osdmap *map, const char *name)
 }
 EXPORT_SYMBOL(ceph_pg_poolid_by_name);
 
+u64 ceph_pg_pool_flags(struct ceph_osdmap *map, u64 id)
+{
+	struct ceph_pg_pool_info *pi;
+
+	pi = __lookup_pg_pool(&map->pg_pools, id);
+	return pi ? pi->flags : 0;
+}
+EXPORT_SYMBOL(ceph_pg_pool_flags);
+
 static void __remove_pg_pool(struct rb_root *root, struct ceph_pg_pool_info *pi)
 {
 	rb_erase(&pi->node, root);