Commit 7d396cac authored by Ville Syrjälä's avatar Ville Syrjälä
Browse files

drm/i195: Make the async flip VT-d workaround dynamic



Since the VT-d vs. async flip issues are plaguing a wider range
of supported hw let's try to minimize the impact on normal
operation by flipping the relevant chicken bits on and off
as needed. I presume there is some power/perf impact on since
this is reducing some prefetching I think.

Cc: Karthik B S <karthik.b.s@intel.com>
Signed-off-by: default avatarVille Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210930190943.17547-2-ville.syrjala@linux.intel.com


Reviewed-by: default avatarMatt Roper <matthew.d.roper@intel.com>
parent d08df3b0
Loading
Loading
Loading
Loading
+35 −0
Original line number Diff line number Diff line
@@ -2311,6 +2311,33 @@ static bool needs_cursorclk_wa(const struct intel_crtc_state *crtc_state)
	return false;
}

static void intel_async_flip_vtd_wa(struct drm_i915_private *i915,
				    enum pipe pipe, bool enable)
{
	if (DISPLAY_VER(i915) == 9) {
		/*
		 * "Plane N strech max must be programmed to 11b (x1)
		 *  when Async flips are enabled on that plane."
		 */
		intel_de_rmw(i915, CHICKEN_PIPESL_1(pipe),
			     SKL_PLANE1_STRETCH_MAX_MASK,
			     enable ? SKL_PLANE1_STRETCH_MAX_X1 : SKL_PLANE1_STRETCH_MAX_X8);
	} else {
		/* Also needed on HSW/BDW albeit undocumented */
		intel_de_rmw(i915, CHICKEN_PIPESL_1(pipe),
			     HSW_PRI_STRETCH_MAX_MASK,
			     enable ? HSW_PRI_STRETCH_MAX_X1 : HSW_PRI_STRETCH_MAX_X8);
	}
}

static bool needs_async_flip_vtd_wa(const struct intel_crtc_state *crtc_state)
{
	struct drm_i915_private *i915 = to_i915(crtc_state->uapi.crtc->dev);

	return crtc_state->uapi.async_flip && intel_vtd_active() &&
		(DISPLAY_VER(i915) == 9 || IS_BROADWELL(i915) || IS_HASWELL(i915));
}

static bool planes_enabling(const struct intel_crtc_state *old_crtc_state,
			    const struct intel_crtc_state *new_crtc_state)
{
@@ -2346,6 +2373,10 @@ static void intel_post_plane_update(struct intel_atomic_state *state,
	intel_fbc_post_update(state, crtc);
	intel_drrs_page_flip(state, crtc);

	if (needs_async_flip_vtd_wa(old_crtc_state) &&
	    !needs_async_flip_vtd_wa(new_crtc_state))
		intel_async_flip_vtd_wa(dev_priv, pipe, false);

	if (needs_nv12_wa(old_crtc_state) &&
	    !needs_nv12_wa(new_crtc_state))
		skl_wa_827(dev_priv, pipe, false);
@@ -2444,6 +2475,10 @@ static void intel_pre_plane_update(struct intel_atomic_state *state,
	if (intel_fbc_pre_update(state, crtc))
		intel_wait_for_vblank(dev_priv, pipe);

	if (!needs_async_flip_vtd_wa(old_crtc_state) &&
	    needs_async_flip_vtd_wa(new_crtc_state))
		intel_async_flip_vtd_wa(dev_priv, pipe, true);

	/* Display WA 827 */
	if (!needs_nv12_wa(old_crtc_state) &&
	    needs_nv12_wa(new_crtc_state))
+0 −26
Original line number Diff line number Diff line
@@ -76,8 +76,6 @@ struct intel_wm_config {

static void gen9_init_clock_gating(struct drm_i915_private *dev_priv)
{
	enum pipe pipe;

	if (HAS_LLC(dev_priv)) {
		/*
		 * WaCompressedResourceDisplayNewHashMode:skl,kbl
@@ -91,16 +89,6 @@ static void gen9_init_clock_gating(struct drm_i915_private *dev_priv)
			   SKL_DE_COMPRESSED_HASH_MODE);
	}

	for_each_pipe(dev_priv, pipe) {
		/*
		 * "Plane N strech max must be programmed to 11b (x1)
		 *  when Async flips are enabled on that plane."
		 */
		if (!IS_GEMINILAKE(dev_priv) && intel_vtd_active())
			intel_uncore_rmw(&dev_priv->uncore, CHICKEN_PIPESL_1(pipe),
					 SKL_PLANE1_STRETCH_MAX_MASK, SKL_PLANE1_STRETCH_MAX_X1);
	}

	/* See Bspec note for PSR2_CTL bit 31, Wa#828:skl,bxt,kbl,cfl */
	intel_uncore_write(&dev_priv->uncore, CHICKEN_PAR1_1,
		   intel_uncore_read(&dev_priv->uncore, CHICKEN_PAR1_1) | SKL_EDP_PSR_FIX_RDWRAP);
@@ -7599,11 +7587,6 @@ static void bdw_init_clock_gating(struct drm_i915_private *dev_priv)
		intel_uncore_write(&dev_priv->uncore, CHICKEN_PIPESL_1(pipe),
			   intel_uncore_read(&dev_priv->uncore, CHICKEN_PIPESL_1(pipe)) |
			   BDW_DPRS_MASK_VBLANK_SRD);

		/* Undocumented but fixes async flip + VT-d corruption */
		if (intel_vtd_active())
			intel_uncore_rmw(&dev_priv->uncore, CHICKEN_PIPESL_1(pipe),
					 HSW_PRI_STRETCH_MAX_MASK, HSW_PRI_STRETCH_MAX_X1);
	}

	/* WaVSRefCountFullforceMissDisable:bdw */
@@ -7639,20 +7622,11 @@ static void bdw_init_clock_gating(struct drm_i915_private *dev_priv)

static void hsw_init_clock_gating(struct drm_i915_private *dev_priv)
{
	enum pipe pipe;

	/* WaFbcAsynchFlipDisableFbcQueue:hsw,bdw */
	intel_uncore_write(&dev_priv->uncore, CHICKEN_PIPESL_1(PIPE_A),
		   intel_uncore_read(&dev_priv->uncore, CHICKEN_PIPESL_1(PIPE_A)) |
		   HSW_FBCQ_DIS);

	for_each_pipe(dev_priv, pipe) {
		/* Undocumented but fixes async flip + VT-d corruption */
		if (intel_vtd_active())
			intel_uncore_rmw(&dev_priv->uncore, CHICKEN_PIPESL_1(pipe),
					 HSW_PRI_STRETCH_MAX_MASK, HSW_PRI_STRETCH_MAX_X1);
	}

	/* This is required by WaCatErrorRejectionIssue:hsw */
	intel_uncore_write(&dev_priv->uncore, GEN7_SQ_CHICKEN_MBCUNIT_CONFIG,
		   intel_uncore_read(&dev_priv->uncore, GEN7_SQ_CHICKEN_MBCUNIT_CONFIG) |