Hi,
There are already patches floated to get the total_display_bandwidth
in-order to implement correct memory bandwidth related WA for GEN9. This
is the series floated earlier for this:
https://patchwork.freedesktop.org/series/18842/
above series is hitting the performance issue, because we were including
all CRTC in each flip even if crtc is not part of flip & we don't have
i915 level global ww_mutex lock.
There was another RFC doc to list down different approaches to implement
bandwidth related WA, which will be implemented eventually.
http://www.spinics.net/lists/intel-gfx/msg124872.html
unconditional implementation of these WA's restricts us from using more
planes per CRTC due to limited availability of DDB.
Do we still need this patch? because once above patches comes in, this
patch will require modification.
-Mahesh
On Saturday 27 May 2017 04:53 AM, Rodrigo Vivi wrote:
No detection will ever be needed. So let's simplify
and clean it up.
Cc: Mahesh Kumar <mahesh1.kumar@xxxxxxxxx>
Cc: Paulo Zanoni <paulo.r.zanoni@xxxxxxxxx>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@xxxxxxxxx>
---
drivers/gpu/drm/i915/i915_drv.h | 4 ++++
drivers/gpu/drm/i915/intel_pm.c | 21 ++-------------------
2 files changed, 6 insertions(+), 19 deletions(-)
diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index d2a5749..5abeb3f 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -2904,6 +2904,10 @@ static inline struct scatterlist *__sg_next(struct scatterlist *sg)
#define NEEDS_WaRsDisableCoarsePowerGating(dev_priv) \
(IS_SKL_GT3(dev_priv) || IS_SKL_GT4(dev_priv))
+/* WA#0893 */
+#define NEEDS_WaIncreaseMemoryBW(dev_priv) (IS_GEN9_BC(dev_priv) || \
+ IS_BROXTON(dev_priv))
+
/*
* dp aux and gmbus irq on gen4 seems to be able to generate legacy interrupts
* even when in MSI mode. This results in spurious interrupt warnings if the
diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c
index 32b3b3e..919235c 100644
--- a/drivers/gpu/drm/i915/intel_pm.c
+++ b/drivers/gpu/drm/i915/intel_pm.c
@@ -3532,21 +3532,6 @@ bool ilk_disable_lp_wm(struct drm_device *dev)
#define SKL_SAGV_BLOCK_TIME 30 /* µs */
-/*
- * FIXME: We still don't have the proper code detect if we need to apply the WA,
- * so assume we'll always need it in order to avoid underruns.
- */
-static bool skl_needs_memory_bw_wa(struct intel_atomic_state *state)
-{
- struct drm_i915_private *dev_priv = to_i915(state->base.dev);
-
- /* WA#0893: GEN9:ALL */
- if (IS_GEN9(dev_priv))
- return true;
-
- return false;
-}
-
static bool
intel_has_sagv(struct drm_i915_private *dev_priv)
{
@@ -3694,7 +3679,7 @@ bool intel_can_enable_sagv(struct drm_atomic_state *state)
latency = dev_priv->wm.skl_latency[level];
- if (skl_needs_memory_bw_wa(intel_state) &&
+ if (NEEDS_WaIncreaseMemoryBW(dev_priv) &&
plane->base.state->fb->modifier ==
I915_FORMAT_MOD_X_TILED)
latency += 15;
@@ -4265,9 +4250,7 @@ static int skl_compute_plane_wm(const struct drm_i915_private *dev_priv,
uint32_t plane_pixel_rate;
uint_fixed_16_16_t y_tile_minimum;
uint32_t y_min_scanlines;
- struct intel_atomic_state *state =
- to_intel_atomic_state(cstate->base.state);
- bool apply_memory_bw_wa = skl_needs_memory_bw_wa(state);
+ bool apply_memory_bw_wa = NEEDS_WaIncreaseMemoryBW(dev_priv);
bool y_tiled, x_tiled;
if (latency == 0 ||
_______________________________________________
Intel-gfx mailing list
Intel-gfx@xxxxxxxxxxxxxxxxxxxxx
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
