ceph: wait unsafe sync writes for evicting inode

Otherwise ceph_sync_write_unsafe() may access/modify freed inode.

Signed-off-by: Yan, Zheng <zyan@redhat.com>

5 files changed, 61 insertions, 48 deletions

diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c
index a08d245f16f5..1e48377f18a7 100644
--- a/fs/ceph/caps.c
+++ b/fs/ceph/caps.c
@@ -1928,53 +1928,6 @@ static int caps_are_flushed(struct inode *inode, u64 flush_tid)
 }
 
 /*
- * Wait on any unsafe replies for the given inode.  First wait on the
- * newest request, and make that the upper bound.  Then, if there are
- * more requests, keep waiting on the oldest as long as it is still older
- * than the original request.
- */
-static void sync_write_wait(struct inode *inode)
-{
-	struct ceph_inode_info *ci = ceph_inode(inode);
-	struct list_head *head = &ci->i_unsafe_writes;
-	struct ceph_osd_request *req;
-	u64 last_tid;
-
-	if (!S_ISREG(inode->i_mode))
-		return;
-
-	spin_lock(&ci->i_unsafe_lock);
-	if (list_empty(head))
-		goto out;
-
-	/* set upper bound as _last_ entry in chain */
-	req = list_last_entry(head, struct ceph_osd_request,
-			      r_unsafe_item);
-	last_tid = req->r_tid;
-
-	do {
-		ceph_osdc_get_request(req);
-		spin_unlock(&ci->i_unsafe_lock);
-		dout("sync_write_wait on tid %llu (until %llu)\n",
-		     req->r_tid, last_tid);
-		wait_for_completion(&req->r_safe_completion);
-		spin_lock(&ci->i_unsafe_lock);
-		ceph_osdc_put_request(req);
-
-		/*
-		 * from here on look at first entry in chain, since we
-		 * only want to wait for anything older than last_tid
-		 */
-		if (list_empty(head))
-			break;
-		req = list_first_entry(head, struct ceph_osd_request,
-				       r_unsafe_item);
-	} while (req->r_tid < last_tid);
-out:
-	spin_unlock(&ci->i_unsafe_lock);
-}
-
-/*
  * wait for any unsafe requests to complete.
  */
 static int unsafe_request_wait(struct inode *inode)
@@ -2026,7 +1979,8 @@ int ceph_fsync(struct file *file, loff_t start, loff_t end, int datasync)
 	int dirty;
 
 	dout("fsync %p%s\n", inode, datasync ? " datasync" : "");
-	sync_write_wait(inode);
+
+	ceph_sync_write_wait(inode);
 
 	ret = filemap_write_and_wait_range(inode->i_mapping, start, end);
 	if (ret < 0)
diff --git a/fs/ceph/file.c b/fs/ceph/file.c
index 033e88753875..7f2ef262cdf7 100644
--- a/fs/ceph/file.c
+++ b/fs/ceph/file.c
@@ -821,6 +821,54 @@ static void ceph_sync_write_unsafe(struct ceph_osd_request *req, bool unsafe)
 	}
 }
 
+/*
+ * Wait on any unsafe replies for the given inode.  First wait on the
+ * newest request, and make that the upper bound.  Then, if there are
+ * more requests, keep waiting on the oldest as long as it is still older
+ * than the original request.
+ */
+void ceph_sync_write_wait(struct inode *inode)
+{
+	struct ceph_inode_info *ci = ceph_inode(inode);
+	struct list_head *head = &ci->i_unsafe_writes;
+	struct ceph_osd_request *req;
+	u64 last_tid;
+
+	if (!S_ISREG(inode->i_mode))
+		return;
+
+	spin_lock(&ci->i_unsafe_lock);
+	if (list_empty(head))
+		goto out;
+
+	/* set upper bound as _last_ entry in chain */
+
+	req = list_last_entry(head, struct ceph_osd_request,
+			      r_unsafe_item);
+	last_tid = req->r_tid;
+
+	do {
+		ceph_osdc_get_request(req);
+		spin_unlock(&ci->i_unsafe_lock);
+
+		dout("sync_write_wait on tid %llu (until %llu)\n",
+		     req->r_tid, last_tid);
+		wait_for_completion(&req->r_safe_completion);
+		ceph_osdc_put_request(req);
+
+		spin_lock(&ci->i_unsafe_lock);
+		/*
+		 * from here on look at first entry in chain, since we
+		 * only want to wait for anything older than last_tid
+		 */
+		if (list_empty(head))
+			break;
+		req = list_first_entry(head, struct ceph_osd_request,
+				       r_unsafe_item);
+	} while (req->r_tid < last_tid);
+out:
+	spin_unlock(&ci->i_unsafe_lock);
+}
 
 static ssize_t
 ceph_direct_read_write(struct kiocb *iocb, struct iov_iter *iter,
diff --git a/fs/ceph/inode.c b/fs/ceph/inode.c
index 8ca843371d4b..6e16269277bd 100644
--- a/fs/ceph/inode.c
+++ b/fs/ceph/inode.c
@@ -585,6 +585,14 @@ int ceph_drop_inode(struct inode *inode)
 	return 1;
 }
 
+void ceph_evict_inode(struct inode *inode)
+{
+	/* wait unsafe sync writes */
+	ceph_sync_write_wait(inode);
+	truncate_inode_pages_final(&inode->i_data);
+	clear_inode(inode);
+}
+
 static inline blkcnt_t calc_inode_blocks(u64 size)
 {
 	return (size + (1<<9) - 1) >> 9;
diff --git a/fs/ceph/super.c b/fs/ceph/super.c
index 91e02481ce06..a5b2275e1573 100644
--- a/fs/ceph/super.c
+++ b/fs/ceph/super.c
@@ -731,6 +731,7 @@ static const struct super_operations ceph_super_ops = {
 	.destroy_inode	= ceph_destroy_inode,
 	.write_inode    = ceph_write_inode,
 	.drop_inode	= ceph_drop_inode,
+	.evict_inode	= ceph_evict_inode,
 	.sync_fs        = ceph_sync_fs,
 	.put_super	= ceph_put_super,
 	.show_options   = ceph_show_options,
diff --git a/fs/ceph/super.h b/fs/ceph/super.h
index 50846e6f6a8c..d5b9077467a4 100644
--- a/fs/ceph/super.h
+++ b/fs/ceph/super.h
@@ -749,6 +749,7 @@ extern const struct inode_operations ceph_file_iops;
 extern struct inode *ceph_alloc_inode(struct super_block *sb);
 extern void ceph_destroy_inode(struct inode *inode);
 extern int ceph_drop_inode(struct inode *inode);
+extern void ceph_evict_inode(struct inode *inode);
 
 extern struct inode *ceph_get_inode(struct super_block *sb,
 				    struct ceph_vino vino);
@@ -927,6 +928,7 @@ extern int ceph_atomic_open(struct inode *dir, struct dentry *dentry,
 extern int ceph_release(struct inode *inode, struct file *filp);
 extern void ceph_fill_inline_data(struct inode *inode, struct page *locked_page,
 				  char *data, size_t len);
+extern void ceph_sync_write_wait(struct inode *inode);
 /* dir.c */
 extern const struct file_operations ceph_dir_fops;
 extern const struct file_operations ceph_snapdir_fops;