Re: [PATCH 04/17] block: Improve documentation of .bdrv_has_zero_init

[Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>]

31.01.2020 20:44, Eric Blake wrote:
> Several drivers supply .bdrv_has_zero_init that returns 1, but lack
> the .bdrv_has_zero_init_truncate callback (parallels and qed outright,
> vdi in some scenarios).  A literal reading of the existing
> documentation says such drivers are broken, because
> bdrv_has_zero_init_truncate() defaults to zero if the callback is
> missing; but in practice, the tie between the two functions is only
> relevant when truncate is supported.  Clarify the documentation to
> make it obvious that this is okay.
> 
> Fixes: 1dcaf527
> Signed-off-by: Eric Blake <eblake@redhat.com>
> ---
>   include/block/block_int.h | 3 ++-
>   1 file changed, 2 insertions(+), 1 deletion(-)
> 
> diff --git a/include/block/block_int.h b/include/block/block_int.h
> index 640fb82c789e..77ab45dc87cf 100644
> --- a/include/block/block_int.h
> +++ b/include/block/block_int.h
> @@ -444,7 +444,8 @@ struct BlockDriver {
>       /*
>        * Returns 1 if newly created images are guaranteed to contain only
>        * zeros, 0 otherwise.
> -     * Must return 0 if .bdrv_has_zero_init_truncate() returns 0.
> +     * Must return 0 if .bdrv_co_truncate is set and
> +     * .bdrv_has_zero_init_truncate() returns 0.
>        */
>       int (*bdrv_has_zero_init)(BlockDriverState *bs);
> 

I doubt, shouldn't 1dcaf527 be fixed by adding all needed bdrv_has_zero_init_truncate functions?

(sorry, I started to dig in the code and patches around bdrv_has_zero_init_truncate and tired :(.,
  hope Max will comment this).

-- 
Best regards,
Vladimir