Re: [Qemu-devel] [PATCH v2 00/49] audio: -audiodev option, multiple options

[Markus Armbruster]

"Kővágó Zoltán" <address@hidden> writes:

> 2015-09-03 17:07 keltezéssel, Eric Blake írta:
>> On 09/03/2015 04:15 AM, Gerd Hoffmann wrote:
>>> On Fr, 2015-08-21 at 17:36 +0200, Kővágó, Zoltán wrote:
>>>> This patch series adds support to multiple audio backends.  Afterwards I
>>>> add support to multiple backends.  Audio fronteds gain a new audiodev
>>>> option to specify the id of the audiodev to use.  The audiodev= option
>>>> is required, unless you use the old environment variable based
>>>> configuration, in that case it must not used (and you can't use multiple
>>>> backends).
>>>>
>>>> Finally I also make mixeng usage optional, it can save us some useless
>>>> format converting when not needed.  Also makes easier to support formats
>>>> currently not supported by qemu (as only the backend has to support it,
>>>> not the mixeng).
>>>>
>>>> For easier testing pull https://github.com/DirtYiCE/qemu.git tag
>>>> audio-multi-v2.
>>>>
>>>> Please review.
>>>
>>> Adding qapi & net folks to Cc.  Ping.
>>>
>>> How to go forward with the QAPI bits in this series (Patches 1-9)?
>>> Patch #1 got reviews from block folks.  Patch #3 has a comment from
>>> Eduardo pending.  What about the other ones?  Fine as-is?  I remember
>>> from the pre-2.4 freeze discussions that there at least was agreement
>>> that flattening the qapi structs is the way to go.
>>
>> Still on my list of patches to review (seems to be a rather large list,
>> sadly).  I should get to it before Monday.
>>
>>>
>>> I'd prefer if the qapi maintainers can pick up and merge these patches.
>>> But I can also merge them via audio queue if I get reviews from the qapi
>>> maintainers.
>>
>> I'll leave it up to Markus on which tree is better for the qapi patches
>> to go through.
>>
>
> Okay, I've rebased the patches to master and applied (some of) the
> changes suggested by Eric.  Since it looks like there are still some
> problems with the qapi parts, it may be better if I split out the qapi
> patches into a separate series, and go back to the audio patches when
> the qapi bits are merged.  Should I go ahead?

Given the size of this series, splitting it up into separately
reviewable parts sounds like an excellent idea.