Re: [Qemu-devel] [PATCH v2 1/1] block: Support Archipelago as a QEMU block backend

[Chrysostomos Nanakos <cnanakos@grnet.gr>]

On 06/02/2014 06:31 PM, Kevin Wolf wrote:
> Am 02.06.2014 um 17:04 hat Chrysostomos Nanakos geschrieben:
>> Hi,
>>
>> On 06/02/2014 05:30 PM, Eric Blake wrote:
>>> On 06/02/2014 08:15 AM, Chrysostomos Nanakos wrote:
>>>> VM Image on Archipelago volume is specified like this:
>>>>
>>>> file=archipelago:<volumename>[/mport=<mapperd_port>[:vport=<vlmcd_port>]]
>>>> or
>>> I still think that there is no need to burden the code with this
>>> shorthand...
>> If everyone agrees I can totally remove that part.
> I don't think it's a bad idea to provide a shortcut for human users. The
> important point is just that internally it's not the main interface, but
> mapped to the structured option (which I think your patch achieves now).
>
>>>> file.driver=archipelago,file.volume=<volumename>[,file.mport=<mapperd_port>[,
>>>> file.vport=<vlmcd_port>]]
>>> ...when this version is easier to support.  Older protocols have file=
>>> abuse due to back-compat reasons, but new protocols should not start out
>>> with baggage.
>>>
>>>
>>>> ---
>>>>   MAINTAINERS         |    5 +
>>>>   block/Makefile.objs |    1 +
>>>>   block/archipelago.c | 1174 +++++++++++++++++++++++++++++++++++++++++++++++++++
>>> This is a lot to review at once - is there any reasonable way to break
>>> it into a series of smaller patches?
>> It will be really difficult to break it in smaller patches, and I
>> believe it will more difficult for someone to review it.
> Not sure how much easier the review would become, but the obvious split
> that works almost always is to do a read-only driver in patch 1, and
> then add write and create support in patches 2 and 3.

I can do that. Eric do you think that I should do it now and resend the 
patches or in the next version in which I will add support for the QMP 
blockdev-add command?


> Kevin