[Intel-gfx] [PATCH v3] drm/i915: Fix FBC cfb stride programming for non X-tiled FB

[Paulo Zanoni]

Em Qui, 2017-03-16 às 17:38 +0530, Praveen Paneri escreveu:
> When FBC is enabled for linear, legacy Y-tiled and Yf-tiled
> surfaces on gen9, the cfb stride must be programmed by SW as
> 
> cfb_stride = ceiling[(at least plane width in pixels)/
> 		     (32 * compression limit factor)] * 8
> 
> v2: Minor fix for a build error
> 
> v3: Fixed subject, register name and platform check (Ville)
> 
> Signed-off-by: Praveen Paneri <praveen.paneri at intel.com>
> ---
>  drivers/gpu/drm/i915/i915_reg.h  | 3 +++
>  drivers/gpu/drm/i915/intel_fbc.c | 8 ++++++++
>  2 files changed, 11 insertions(+)
> 
> diff --git a/drivers/gpu/drm/i915/i915_reg.h
> b/drivers/gpu/drm/i915/i915_reg.h
> index 5d88c35..f4f0cb5 100644
> --- a/drivers/gpu/drm/i915/i915_reg.h
> +++ b/drivers/gpu/drm/i915/i915_reg.h
> @@ -6506,6 +6506,9 @@ enum {
>  #define  GLK_CL1_PWR_DOWN	(1 << 11)
>  #define  GLK_CL2_PWR_DOWN	(1 << 12)
>  
> +#define CHICKEN_MISC_4		_MMIO(0x4208c)
> +#define   FBC_STRIDE_OVERRIDE	(1<<13)
> +
>  #define _CHICKEN_PIPESL_1_A	0x420b0
>  #define _CHICKEN_PIPESL_1_B	0x420b4
>  #define  HSW_FBCQ_DIS			(1 << 22)
> diff --git a/drivers/gpu/drm/i915/intel_fbc.c
> b/drivers/gpu/drm/i915/intel_fbc.c
> index ded2add..e7f259f 100644
> --- a/drivers/gpu/drm/i915/intel_fbc.c
> +++ b/drivers/gpu/drm/i915/intel_fbc.c
> @@ -298,9 +298,17 @@ static bool ilk_fbc_is_active(struct
> drm_i915_private *dev_priv)
>  static void gen7_fbc_activate(struct drm_i915_private *dev_priv)
>  {
>  	struct intel_fbc_reg_params *params = &dev_priv->fbc.params;
> +	struct intel_fbc_state_cache *cache = &dev_priv-
> >fbc.state_cache;
>  	u32 dpfc_ctl;
>  	int threshold = dev_priv->fbc.threshold;
>  

Please add:

/* Display WA #0529: skl, kbl, bxt, glk. */

(and if you find the real name of the WA, please also add it)

I wanted to give this patch a try with kms_frontbuffer_tracking, but I
just couldn't find your IGT patch where Y tiling support is added to
it. I can see your patches that touch lib/igt_draw and kms_draw_crc,
but no patches testing kms_frontbuffer_tracking. Can you please send me
the patchwork link?

The corollary of the paragraph above is the question: do we maintain
the same pass rate as X tiling when we test it against Y tiling now?

My original goal was to block FBC Y tiling support until we added the
IGT stuff, but I see it was already enabled by someone even though we
didn't have the missing WA nor the IGT stuff... Oh, well....

> +	if (IS_GEN9(dev_priv) &&
> +	    i915_gem_object_get_tiling(cache->vma->obj) !=
> I915_TILING_X) {
> +		int cfb_stride = DIV_ROUND_UP(cache->plane.src_w,
> +					      (32 * threshold)) * 8;
> +		I915_WRITE(CHICKEN_MISC_4, FBC_STRIDE_OVERRIDE |
> cfb_stride);
> +	}
> +
>  	dpfc_ctl = 0;
>  	if (IS_IVYBRIDGE(dev_priv))
>  		dpfc_ctl |= IVB_DPFC_CTL_PLANE(params->crtc.plane);