[Intel-gfx] [PATCH v3 01/11] drm/i915: Update watermark state correctly in sanitize_watermarks

Maarten Lankhorst maarten.lankhorst at linux.intel.com
Sat Nov 11 14:14:16 UTC 2017


Op 10-11-17 om 16:28 schreef Ville Syrjälä:
> On Fri, Nov 10, 2017 at 12:34:53PM +0100, Maarten Lankhorst wrote:
>> We no longer use intel_crtc->wm.active for watermarks any more,
>> which was incorrect. But this uncovered a bug in sanitize_watermarks(),
>> which meant that we wrote the correct watermarks, but the next
>> update would still use the wrong hw watermarks for calculating.
>> This caused all further updates to fail with -EINVAL and the
>> log would reveal an error like the one below:
>>
>> [   10.043902] [drm:ilk_validate_wm_level.part.8 [i915]] Sprite WM0 too large 56 (max 0)
>> [   10.043960] [drm:ilk_validate_pipe_wm [i915]] LP0 watermark invalid
>> [   10.044030] [drm:intel_crtc_atomic_check [i915]] No valid intermediate pipe watermarks are possible
>>
>> Signed-off-by: Maarten Lankhorst <maarten.lankhorst at linux.intel.com>
>> Fixes: b6b178a77210 ("drm/i915: Calculate ironlake intermediate watermarks correctly, v2.")
>> Cc: stable at vger.kernel.org #v4.8+
>> ---
>>  drivers/gpu/drm/i915/intel_display.c | 2 ++
>>  1 file changed, 2 insertions(+)
>>
>> diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
>> index 84817ccc5305..17665ee06c9a 100644
>> --- a/drivers/gpu/drm/i915/intel_display.c
>> +++ b/drivers/gpu/drm/i915/intel_display.c
>> @@ -14442,6 +14442,8 @@ static void sanitize_watermarks(struct drm_device *dev)
>>  
>>  		cs->wm.need_postvbl_update = true;
>>  		dev_priv->display.optimize_watermarks(intel_state, cs);
>> +
>> +		to_intel_crtc_state(crtc->state)->wm = cs->wm;
> Hmm. Oh we don't swap the state here. Looks like we shouldn't be using
> anything else from the crtc state, so this should work AFAICS.
>
> Reviewed-by: Ville Syrjälä <ville.syrjala at linux.intel.com>
We should, but I don't trust that we keep the swapped state completely identical to the original.
For  this reason I felt it was safer to only copy the wm part of the state. Thanks, pushed. :)


More information about the Intel-gfx mailing list