Re: [PATCH v2 05/33] block: Pass BdrvChildRole to bdrv_child_perm()

[Eric Blake]

On 2/4/20 11:08 AM, Max Reitz wrote:
> Signed-off-by: Max Reitz <address@hidden>

Sparse commit message - is the intent that this is no semantic change 
and just adding a parameter?

> ---
>   block.c                     | 22 ++++++++++++----------
>   block/backup-top.c          |  3 ++-
>   block/blkdebug.c            |  5 +++--
>   block/blklogwrites.c        |  9 +++++----
>   block/commit.c              |  1 +
>   block/copy-on-read.c        |  1 +
>   block/mirror.c              |  1 +
>   block/quorum.c              |  1 +
>   block/replication.c         |  1 +
>   block/vvfat.c               |  1 +
>   include/block/block_int.h   |  5 ++++-
>   tests/test-bdrv-drain.c     |  5 +++--
>   tests/test-bdrv-graph-mod.c |  1 +
>   13 files changed, 36 insertions(+), 20 deletions(-)
>
> diff --git a/block.c b/block.c
> index c576377650..7fa7830428 100644
> --- a/block.c
> +++ b/block.c
> @@ -1764,12 +1764,12 @@ bool bdrv_is_writable(BlockDriverState *bs)
>   
>   static void bdrv_child_perm(BlockDriverState *bs, BlockDriverState *child_bs,
>                               BdrvChild *c, const BdrvChildClass *child_class,
> -                            BlockReopenQueue *reopen_queue,
> +                            BdrvChildRole role, BlockReopenQueue *reopen_queue,

Again, using an enum name where a bitmask of non-enum values is 
self-documenting, but somewhat abusive of C's loose type system.  Is an 
unsigned int any better?

Looks mechanical enough. I'd like a better commit message, but:
Reviewed-by: Eric Blake <address@hidden>

--
Eric Blake, Principal Software Engineer
Red Hat, Inc.           +1-919-301-3226
Virtualization:  qemu.org | libvirt.org