Op 12-05-15 om 12:03 schreef Daniel Vetter:
> On Mon, May 11, 2015 at 4:25 PM, Maarten Lankhorst
> <maarten.lankho...@linux.intel.com> wrote:
>> @@ -11953,16 +11930,14 @@ check_shared_dpll_state(struct drm_device *dev)
>>
>>                 for_each_intel_crtc(dev, crtc) {
>>                         if (crtc->base.state->active && 
>> intel_crtc_to_shared_dpll(crtc) == pll)
>> -                               enabled_crtcs++;
>> -                       if (crtc->active && intel_crtc_to_shared_dpll(crtc) 
>> == pll)
>>                                 active_crtcs++;
>>                 }
>>                 I915_STATE_WARN(pll->active != active_crtcs,
>>                      "pll active crtcs mismatch (expected %i, found %i)\n",
>>                      pll->active, active_crtcs);
>> -               I915_STATE_WARN(hweight32(pll->config.crtc_mask) != 
>> enabled_crtcs,
>> +               I915_STATE_WARN(hweight32(pll->config.crtc_mask) != 
>> active_crtcs,
>>                      "pll enabled crtcs mismatch (expected %i, found %i)\n",
>> -                    hweight32(pll->config.crtc_mask), enabled_crtcs);
>> +                    hweight32(pll->config.crtc_mask), active_crtcs);
>
> Missed one: Why do you remove this? Imo that's a fairly crucial
> consistency check.
> -Daniel
It's not removed, but crtc->active is the same as crtc->base.state->active now. 
The check still works as intended. :-)
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

Reply via email to