Re: [PATCH] btrfs-progs: Skip device tree when we failed to read it

[Anand Jain <anand.jain@oracle.com>]

On 6/12/19 11:44 AM, Qu Wenruo wrote:
> Device tree is one of the least important tree, it only contains:
> - Device status
>    Like various error count
> - Device extents
>    Only makes sense for chunk allocation and physical->logical map, and
>    even for that only purpose, we can rebuild it easily from chunk tree.
> 
> So device tree even makes less sense compared to extent tree, while we
> still can't skip it at btrfs-progs.
> 
> This makes restore less useful. So this patch will make device tree to
> follow the same requirement for OPEN_CTREE_PARTIAL.
> 
> Reported-by: Christian Wimmer <telefonchris@icloud.com>
> Signed-off-by: Qu Wenruo <wqu@suse.com>
> ---
>   disk-io.c | 10 ++++------
>   1 file changed, 4 insertions(+), 6 deletions(-)
> 
> diff --git a/disk-io.c b/disk-io.c
> index 659f8b93..22aa2a61 100644
> --- a/disk-io.c
> +++ b/disk-io.c
> @@ -947,12 +947,10 @@ int btrfs_setup_all_roots(struct btrfs_fs_info *fs_info, u64 root_tree_bytenr,
>   		return ret;
>   	fs_info->extent_root->track_dirty = 1;
>   
> -	ret = find_and_setup_root(root, fs_info, BTRFS_DEV_TREE_OBJECTID,
> -				  fs_info->dev_root);
> -	if (ret) {
> -		printk("Couldn't setup device tree\n");
> -		return -EIO;
> -	}
> +	ret = setup_root_or_create_block(fs_info, flags, fs_info->dev_root,
> +					 BTRFS_DEV_TREE_OBJECTID, "device");
> +	if (ret)
> +		return ret;
>   	fs_info->dev_root->track_dirty = 1;
>   
>   	ret = setup_root_or_create_block(fs_info, flags, fs_info->csum_root,
> 

Cool.
Reviewed-by: Anand Jain <anand.jain@oracle.com>

Thanks Anand