Skip to contentSkip to conversation

btrfs: bail out of uuid tree scanning if we're closing

PATCH In review Josef Bacik 3 participants 3 messages Feb 14, 2020 – Feb 21, 2020 Btrfs file system

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 btrfs: bail out of uuid tree scanning if we're closing

In doing my fsstress+EIO stress testing I started running into issues where umount would get stuck forever because the uuid checker was chewing through the thousands of subvolumes I had created. We shouldn't block umount on this, simply bail if we're unmounting the fs. We need to make sure we don't mark the UUID tree as ok, so we only set that bit if we made it through the whole rescan operation, but otherwise this is completely safe.

---
 fs/btrfs/uuid-tree.c |  4 ++++
 fs/btrfs/volumes.c   | 11 +++++++++--
 2 files changed, 13 insertions(+), 2 deletions(-)
fs/btrfs/uuid-tree.c +4 -0
index 76b84f2397b1..30078b74220e 100644
@@ -278,6 +278,10 @@ int btrfs_uuid_tree_iterate(struct btrfs_fs_info *fs_info,
 	}
 
 	while (1) {
+		if (btrfs_fs_closing(fs_info)) {
+			ret = -EINTR;
+			goto out;
+		}
 		cond_resched();
 		leaf = path->nodes[0];
 		slot = path->slots[0];
fs/btrfs/volumes.c +9 -2
index 387f80656476..e40adf38ae5c 100644
@@ -4286,6 +4286,7 @@ static int btrfs_uuid_scan_kthread(void *data)
 	struct btrfs_root_item root_item;
 	u32 item_size;
 	struct btrfs_trans_handle *trans = NULL;
+	bool closing = false;
 
 	path = btrfs_alloc_path();
 	if (!path) {
@@ -4298,6 +4299,10 @@ static int btrfs_uuid_scan_kthread(void *data)
 	key.offset = 0;
 
 	while (1) {
+		if (btrfs_fs_closing(fs_info)) {
+			closing = true;
+			break;
+		}
 		ret = btrfs_search_forward(root, &key, path,
 				BTRFS_OLDEST_GENERATION);
 		if (ret) {
@@ -4397,7 +4402,7 @@ static int btrfs_uuid_scan_kthread(void *data)
 		btrfs_end_transaction(trans);
 	if (ret)
 		btrfs_warn(fs_info, "btrfs_uuid_scan_kthread failed %d", ret);
-	else
+	else if (!closing)
 		set_bit(BTRFS_FS_UPDATE_UUID_TREE_GEN, &fs_info->flags);
 	up(&fs_info->uuid_tree_rescan_sem);
 	return 0;
@@ -4460,7 +4465,9 @@ static int btrfs_uuid_rescan_kthread(void *data)
 	 */
 	ret = btrfs_uuid_tree_iterate(fs_info, btrfs_check_uuid_tree_entry);
 	if (ret < 0) {
-		btrfs_warn(fs_info, "iterating uuid_tree failed %d", ret);
+		if (ret != -EINTR)
+			btrfs_warn(fs_info, "iterating uuid_tree failed %d",
+				   ret);
 		up(&fs_info->uuid_tree_rescan_sem);
 		return ret;
 	}
Signature · 1 line
2.24.1
#2 lore raw
In reply to Josef Bacik
Quoted 9 lines · Josef Bacik

In doing my fsstress+EIO stress testing I started running into issues where umount would get stuck forever because the uuid checker was chewing through the thousands of subvolumes I had created. We shouldn't block umount on this, simply bail if we're unmounting the fs. We need to make sure we don't mark the UUID tree as ok, so we only set that bit if we made it through the whole rescan operation, but otherwise this is completely safe.

Added to misc-next, thanks.