Re: [Qemu-block] [Qemu-devel] [PATCH for-2.5 1/1] blockdev: Mark {insert, remove}-medium experimental

[Eric Blake]

On 12/11/2015 08:23 AM, Max Reitz wrote:
> While in the long term we want throttling to be its own block filter
> BDS, in the short term we want it to be part of the BB instead of a BDS;
> even in the long term we may want legacy throttling to be automatically
> tied to the BB.
> 
> blockdev-insert-medium and blockdev-remove-medium do not retain
> throttling information in the BB (deliberately so). Therefore, using
> them means tying this information to a BDS, which would break the model
> described above. (The same applies to other flags such as
> detect_zeroes.) We probably want to move this information to the BB or
> its own filter BDS before blockdev-{insert,remove}-medium can be
> considered completely stable.
> 
> Therefore, mark these functions experimental for the time being.
> 
> Suggested-by: Markus Armbruster <address@hidden>
> Signed-off-by: Max Reitz <address@hidden>
> Acked-by: Markus Armbruster <address@hidden>
> Acked-by: Kevin Wolf <address@hidden>
> ---

Reviewed-by: Eric Blake <address@hidden>


> +++ b/qmp-commands.hx
> @@ -4203,13 +4203,13 @@ Example:
>  EQMP
>  
>      {
> -        .name       = "blockdev-remove-medium",
> +        .name       = "x-blockdev-remove-medium",
>          .args_type  = "device:s",
> -        .mhandler.cmd_new = qmp_marshal_blockdev_remove_medium,
> +        .mhandler.cmd_new = qmp_marshal_x_blockdev_remove_medium,
>      },
>  
>  SQMP
> -blockdev-remove-medium
> +x-blockdev-remove-medium
>  ----------------------

Formatting nit, but not worth holding this up (as it is really our last
chance to get it in 2.5 before baking in something we'd be stuck with).

>  
>  SQMP
> -blockdev-insert-medium
> +x-blockdev-insert-medium
>  ----------------------

Ditto.

-- 
Eric Blake   eblake redhat com    +1-919-301-3266
Libvirt virtualization library http://libvirt.org

signature.asc
Description: OpenPGP digital signature