Merge tag 'for-5.19-rc6-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux

Pull btrfs fixes from David Sterba:
 "A more fixes that seem to me to be important enough to get merged
  before release:

   - in zoned mode, fix leak of a structure when reading zone info, this
     happens on normal path so this can be significant

   - in zoned mode, revert an optimization added in 5.19-rc1 to finish a
     zone when the capacity is full, but this is not reliable in all
     cases

   - try to avoid short reads for compressed data or inline files when
     it's a NOWAIT read, applications should handle that but there are
     two, qemu and mariadb, that are affected"

* tag 'for-5.19-rc6-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux:
  btrfs: zoned: drop optimization of zone finish
  btrfs: zoned: fix a leaked bioc in read_zone_info
  btrfs: return -EAGAIN for NOWAIT dio reads/writes on compressed and inline extents

Author: torvalds

fs/btrfs/inode.c

@@ -7681,7 +7681,19 @@ static int btrfs_dio_iomap_begin(struct inode *inode, loff_t start,
 	if (test_bit(EXTENT_FLAG_COMPRESSED, &em->flags) ||
 	    em->block_start == EXTENT_MAP_INLINE) {
 		free_extent_map(em);
-		ret = -ENOTBLK;
+		/*
+		 * If we are in a NOWAIT context, return -EAGAIN in order to
+		 * fallback to buffered IO. This is not only because we can
+		 * block with buffered IO (no support for NOWAIT semantics at
+		 * the moment) but also to avoid returning short reads to user
+		 * space - this happens if we were able to read some data from
+		 * previous non-compressed extents and then when we fallback to
+		 * buffered IO, at btrfs_file_read_iter() by calling
+		 * filemap_read(), we fail to fault in pages for the read buffer,
+		 * in which case filemap_read() returns a short read (the number
+		 * of bytes previously read is > 0, so it does not return -EFAULT).
+		 */
+		ret = (flags & IOMAP_NOWAIT) ? -EAGAIN : -ENOTBLK;
 		goto unlock_err;
 	}
 

fs/btrfs/zoned.c

@@ -1735,12 +1735,14 @@ static int read_zone_info(struct btrfs_fs_info *fs_info, u64 logical,
 	ret = btrfs_map_sblock(fs_info, BTRFS_MAP_GET_READ_MIRRORS, logical,
 			       &mapped_length, &bioc);
 	if (ret || !bioc || mapped_length < PAGE_SIZE) {
-		btrfs_put_bioc(bioc);
-		return -EIO;
+		ret = -EIO;
+		goto out_put_bioc;
 	}
 
-	if (bioc->map_type & BTRFS_BLOCK_GROUP_RAID56_MASK)
-		return -EINVAL;
+	if (bioc->map_type & BTRFS_BLOCK_GROUP_RAID56_MASK) {
+		ret = -EINVAL;
+		goto out_put_bioc;
+	}
 
 	nofs_flag = memalloc_nofs_save();
 	nmirrors = (int)bioc->num_stripes;
@@ -1759,7 +1761,8 @@ static int read_zone_info(struct btrfs_fs_info *fs_info, u64 logical,
 		break;
 	}
 	memalloc_nofs_restore(nofs_flag);
-
+out_put_bioc:
+	btrfs_put_bioc(bioc);
 	return ret;
 }
 
@@ -1885,7 +1888,6 @@ static int do_zone_finish(struct btrfs_block_group *block_group, bool fully_writ
 {
 	struct btrfs_fs_info *fs_info = block_group->fs_info;
 	struct map_lookup *map;
-	bool need_zone_finish;
 	int ret = 0;
 	int i;
 
@@ -1942,12 +1944,6 @@ static int do_zone_finish(struct btrfs_block_group *block_group, bool fully_writ
 		}
 	}
 
-	/*
-	 * The block group is not fully allocated, so not fully written yet. We
-	 * need to send ZONE_FINISH command to free up an active zone.
-	 */
-	need_zone_finish = !btrfs_zoned_bg_is_full(block_group);
-
 	block_group->zone_is_active = 0;
 	block_group->alloc_offset = block_group->zone_capacity;
 	block_group->free_space_ctl->free_space = 0;
@@ -1963,15 +1959,13 @@ static int do_zone_finish(struct btrfs_block_group *block_group, bool fully_writ
 		if (device->zone_info->max_active_zones == 0)
 			continue;
 
-		if (need_zone_finish) {
-			ret = blkdev_zone_mgmt(device->bdev, REQ_OP_ZONE_FINISH,
-					       physical >> SECTOR_SHIFT,
-					       device->zone_info->zone_size >> SECTOR_SHIFT,
-					       GFP_NOFS);
+		ret = blkdev_zone_mgmt(device->bdev, REQ_OP_ZONE_FINISH,
+				       physical >> SECTOR_SHIFT,
+				       device->zone_info->zone_size >> SECTOR_SHIFT,
+				       GFP_NOFS);
 
-			if (ret)
-				return ret;
-		}
+		if (ret)
+			return ret;
 
 		btrfs_dev_clear_active_zone(device, physical);
 	}