Re: [Qemu-devel] Re: [PATCH v4 2/5] qed: Add QEMU Enhanced Disk image format

[Kevin Wolf <kwolf@redhat.com>]

Am 12.11.2010 18:24, schrieb Markus Armbruster:
> Kevin Wolf <kwolf@redhat.com> writes:
> 
>> Am 12.11.2010 17:34, schrieb Stefan Hajnoczi:
>>> On Fri, Nov 12, 2010 at 3:43 PM, Kevin Wolf <kwolf@redhat.com> wrote:
>>>> Am 28.10.2010 13:01, schrieb Stefan Hajnoczi:
>>>>> +/**
>>>>> + * Check whether an image format is raw
>>>>> + *
>>>>> + * @fmt:    Backing file format, may be NULL
>>>>> + */
>>>>> +static bool qed_fmt_is_raw(const char *fmt)
>>>>> +{
>>>>> +    return fmt && strcmp(fmt, "raw") == 0;
>>>>> +}
>>>>
>>>> People shouldn't use them directly, but should we also consider file,
>>>> host_device, etc.?
>>>
>>> Hrm..I will look into it for v5.  I thought we always have a "raw"
>>> format on top of "file", "host_device", etc protocols?
>>
>> That's how it's meant to be used. I think at the moment we still allow
>> to directly use file etc. and you know that if something is possible,
>> users will do it.
> 
> It's possible, it's perfectly well-defined, and it works.

Wow, I'm not used to be confirmed that quickly!

Yes, it's possible, it works, but it's only exposed because we didn't
pay attention when host_* was split out. There's no real reason to use
it (or do you have any example where format=raw doesn't work?) and it
should be considered an implementation detail.

Kevin