[Intel-gfx] [PATCH (resend) 1/3] drm/i915: Ignore previous watermarks on ILK if inherited

[Daniel Vetter]

On Wed, Nov 08, 2017 at 10:29:20AM +0100, Maarten Lankhorst wrote:
> Fixes the following error when fastset is enabled, caught by CI:
> 
> [drm:ilk_validate_wm_level.part.8 [i915]] Sprite WM0 too large 56 (max 0)
> [drm:ilk_validate_pipe_wm [i915]] LP0 watermark invalid
> [drm:intel_crtc_atomic_check [i915]] No valid intermediate pipe watermarks are possible
> 
> Triggered on debugfs_test.read_all_entries, but could have been any igt
> test depending on ordering.
> 
> Signed-off-by: Maarten Lankhorst <maarten.lankhorst at linux.intel.com>

So I'm no expert on this, but why is this not needed for g4x and vlv
intermediate wm? I think the commit message should explain that. I think
it should also explain why simply shutting the above warnings up is safe.

I think once that's fixed I understand why we need this and why it works,
but probably better to get an Ack from Ville to make sure.
-Daniel

> ---
>  drivers/gpu/drm/i915/intel_pm.c | 4 +++-
>  1 file changed, 3 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c
> index 46440e2ecb33..9e8a0a9cac02 100644
> --- a/drivers/gpu/drm/i915/intel_pm.c
> +++ b/drivers/gpu/drm/i915/intel_pm.c
> @@ -3137,7 +3137,9 @@ static int ilk_compute_intermediate_wm(struct drm_device *dev,
>  	 * and after the vblank.
>  	 */
>  	*a = newstate->wm.ilk.optimal;
> -	if (!newstate->base.active || drm_atomic_crtc_needs_modeset(&newstate->base))
> +	if (!newstate->base.active ||
> +	    drm_atomic_crtc_needs_modeset(&newstate->base) ||
> +	    oldstate->base.mode.private_flags & I915_MODE_FLAG_INHERITED)
>  		return 0;
>  
>  	a->pipe_enabled |= b->pipe_enabled;
> -- 
> 2.15.0
> 
> _______________________________________________
> Intel-gfx mailing list
> Intel-gfx at lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/intel-gfx

-- 
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch