[Intel-gfx] [PATCH v4 1/8] drm/i915: Pass crtc state to modeset_get_crtc_power_domains.

[Maarten Lankhorst]

Op 18-02-16 om 14:08 schreef Zanoni, Paulo R:
> Em Qui, 2016-02-18 às 10:51 +0100, Maarten Lankhorst escreveu:
>> Op 17-02-16 om 18:54 schreef Zanoni, Paulo R:
>>> Em Qua, 2016-02-10 às 13:49 +0100, Maarten Lankhorst escreveu:
>>>> Use our newly created encoder_mask to iterate over the encoders.
>>> As someone who was not paying attention to the discussion of the
>>> previous patches related to this, I think it would be really good
>>> if
>>> your commit message could tell me why we should use the newly
>>> created
>>> encoder_mask instead of the current patch. What's bad about the
>>> current
>>> version? Please sell me your patch. If you think the answer is
>>> trivial,
>>> remember that it's not trivial to many people, and that random
>>> people
>>> may find this patch through git-bisect and have to judge its
>>> importance. Also, an explanation really helps the reviewers :)
>>>
>>> The patch looks correct, so if you improve the commit message:
>>> Reviewed-by: Paulo Zanoni <paulo.r.zanoni at intel.com>
>> Does this sound good?
>>
>> Use our newly created encoder_mask to iterate over the encoders.
>> This makes it possible to get the crtc power domains from the
>> crtc_state at any time, without any locks or having to look at
>> the legacy state.
> But we were already not grabbing any locks. Is this a bug fix? It would
> be good to point it if it's an actual fix.
>
> Anyway, it's definitely better now :)
>
But we're holding locks, crtc.mutex and mode_config.connection_mutex.

The changes allows us to remove the need of the locks, just a crtc_state is enough.