Skip to contentSkip to conversation

drm/i915: Limit VF cache invalidate workaround usage to gen9

PATCH Superseded Ben Widawsky 5 participants 9 messages Dec 17, 2015 – Dec 21, 2015

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 drm/i915: Limit VF cache invalidate workaround usage to gen9

It is unclear if this is even required on BXT.

---
 drivers/gpu/drm/i915/intel_lrc.c | 14 +++++++-------
 1 file changed, 7 insertions(+), 7 deletions(-)
drivers/gpu/drm/i915/intel_lrc.c +7 -7
index 519cea32..0d35e94 100644
@@ -1716,14 +1716,14 @@ static int gen8_emit_flush_render(struct drm_i915_gem_request *request,
 		flags |= PIPE_CONTROL_STATE_CACHE_INVALIDATE;
 		flags |= PIPE_CONTROL_QW_WRITE;
 		flags |= PIPE_CONTROL_GLOBAL_GTT_IVB;
-	}
 
-	/*
-	 * On GEN9+ Before VF_CACHE_INVALIDATE we need to emit a NULL pipe
-	 * control.
-	 */
-	vf_flush_wa = INTEL_INFO(ring->dev)->gen >= 9 &&
-		      flags & PIPE_CONTROL_VF_CACHE_INVALIDATE;
+		/*
+		 * On GEN9: before VF_CACHE_INVALIDATE we need to emit a NULL
+		 * pipe control.
+		 */
+		if (IS_GEN9(ring->dev))
+			vf_flush_wa = true;
+	}
 
 	ret = intel_logical_ring_begin(request, vf_flush_wa ? 12 : 6);
 	if (ret)
Signature · 6 lines
2.6.4

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
#1 lore raw

PATCH v2 1/1 drm/i915: Limit VF cache invalidate workaround usage to gen9

It is unclear if this is even required on BXT.

v2: Make sure to set the default value to false. Uncertain how my compiler doesn't complain with v1.

---
 drivers/gpu/drm/i915/intel_lrc.c | 16 ++++++++--------
 1 file changed, 8 insertions(+), 8 deletions(-)
drivers/gpu/drm/i915/intel_lrc.c +8 -8
index 519cea32..af1e001 100644
@@ -1695,7 +1695,7 @@ static int gen8_emit_flush_render(struct drm_i915_gem_request *request,
 	struct intel_ringbuffer *ringbuf = request->ringbuf;
 	struct intel_engine_cs *ring = ringbuf->ring;
 	u32 scratch_addr = ring->scratch.gtt_offset + 2 * CACHELINE_BYTES;
-	bool vf_flush_wa;
+	bool vf_flush_wa = false;
 	u32 flags = 0;
 	int ret;
 
@@ -1716,14 +1716,14 @@ static int gen8_emit_flush_render(struct drm_i915_gem_request *request,
 		flags |= PIPE_CONTROL_STATE_CACHE_INVALIDATE;
 		flags |= PIPE_CONTROL_QW_WRITE;
 		flags |= PIPE_CONTROL_GLOBAL_GTT_IVB;
-	}
 
-	/*
-	 * On GEN9+ Before VF_CACHE_INVALIDATE we need to emit a NULL pipe
-	 * control.
-	 */
-	vf_flush_wa = INTEL_INFO(ring->dev)->gen >= 9 &&
-		      flags & PIPE_CONTROL_VF_CACHE_INVALIDATE;
+		/*
+		 * On GEN9: before VF_CACHE_INVALIDATE we need to emit a NULL
+		 * pipe control.
+		 */
+		if (IS_GEN9(ring->dev))
+			vf_flush_wa = true;
+	}
 
 	ret = intel_logical_ring_begin(request, vf_flush_wa ? 12 : 6);
 	if (ret)
Signature · 6 lines
2.6.4

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Reviewed-by on v2 1/1
#2 lore raw
Quoted 1 line · Ben Widawsky

It is unclear if this is even required on BXT.

I'm not sure either, I only added it on the premise that it was marked as SKL+ originally in BSpec. The revision log entry in BSpec has this much to say:
"""
The workaround that requires an empty PIPE_CONTROL before a PIPE_CONTROL with a VF Cache Invalidation Enable is only for SKL and not SKL+. The bug was fixed in CNL with the following HSD:... """
Which doesn't make this clear either imo.

Quoted 29 lines

v2: Make sure to set the default value to false. Uncertain how my compiler doesn't complain with v1.

---
 drivers/gpu/drm/i915/intel_lrc.c | 16 ++++++++--------

 1 file changed, 8 insertions(+), 8 deletions(-)
drivers/gpu/drm/i915/intel_lrc.c +1 -1
index 519cea32..af1e001 100644
@@ -1695,7 +1695,7 @@ static int gen8_emit_flush_render(struct drm_i915_gem_request *request,
  	struct intel_ringbuffer *ringbuf = request->ringbuf;
  	struct intel_engine_cs *ring = ringbuf->ring;
  	u32 scratch_addr = ring->scratch.gtt_offset + 2 * CACHELINE_BYTES;
-	bool vf_flush_wa;
+	bool vf_flush_wa = false;
  	u32 flags = 0;
  	int ret;

drivers/gpu/drm/i915/intel_lrc.c +0 -1
@@ -1716,14 +1716,14 @@ static int gen8_emit_flush_render(struct drm_i915_gem_request *request,
  		flags |= PIPE_CONTROL_STATE_CACHE_INVALIDATE;
  		flags |= PIPE_CONTROL_QW_WRITE;
  		flags |= PIPE_CONTROL_GLOBAL_GTT_IVB;
-	}

drivers/gpu/drm/i915/intel_lrc.c +0 -1
-	/*

drivers/gpu/drm/i915/intel_lrc.c static int gen8_emit_flush_render(struct drm_i915_gem_request *request,

Quoted 10 lines
drivers/gpu/drm/i915/intel_lrc.c +5 -5
-	 * On GEN9+ Before VF_CACHE_INVALIDATE we need to emit a NULL pipe
-	 * control.
-	 */
-	vf_flush_wa = INTEL_INFO(ring->dev)->gen >= 9 &&
-		      flags & PIPE_CONTROL_VF_CACHE_INVALIDATE;
+		/*
+		 * On GEN9: before VF_CACHE_INVALIDATE we need to emit a NULL
+		 * pipe control.
+		 */
+> 		if (IS_GEN9(ring->dev))

Nitpick: INTEL_INFO()->gen == 9 is the preferred way. Either way:

Quoted 5 lines
drivers/gpu/drm/i915/intel_lrc.c +2 -0
+			vf_flush_wa = true;
+	}


  ret = intel_logical_ring_begin(request, vf_flush_wa ? 12 : 6); if (ret)

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

#3 lore raw
Quoted 54 lines · Imre Deak
Quoted 1 line · Ben Widawsky

It is unclear if this is even required on BXT.

I'm not sure either, I only added it on the premise that it was marked as SKL+ originally in BSpec. The revision log entry in BSpec has this much to say:
"""
The workaround that requires an empty PIPE_CONTROL before a PIPE_CONTROL with a VF Cache Invalidation Enable is only for SKL and not SKL+. The bug was fixed in CNL with the following HSD:... """
Which doesn't make this clear either imo.

Quoted 29 lines

v2: Make sure to set the default value to false. Uncertain how my compiler doesn't complain with v1.

---
 drivers/gpu/drm/i915/intel_lrc.c | 16 ++++++++--------

 1 file changed, 8 insertions(+), 8 deletions(-)
drivers/gpu/drm/i915/intel_lrc.c +1 -1
index 519cea32..af1e001 100644
@@ -1695,7 +1695,7 @@ static int gen8_emit_flush_render(struct drm_i915_gem_request *request,
  	struct intel_ringbuffer *ringbuf = request->ringbuf;
  	struct intel_engine_cs *ring = ringbuf->ring;
  	u32 scratch_addr = ring->scratch.gtt_offset + 2 * CACHELINE_BYTES;
-	bool vf_flush_wa;
+	bool vf_flush_wa = false;
  	u32 flags = 0;
  	int ret;

drivers/gpu/drm/i915/intel_lrc.c +0 -1
@@ -1716,14 +1716,14 @@ static int gen8_emit_flush_render(struct drm_i915_gem_request *request,
  		flags |= PIPE_CONTROL_STATE_CACHE_INVALIDATE;
  		flags |= PIPE_CONTROL_QW_WRITE;
  		flags |= PIPE_CONTROL_GLOBAL_GTT_IVB;
-	}

drivers/gpu/drm/i915/intel_lrc.c +0 -1
-	/*

drivers/gpu/drm/i915/intel_lrc.c static int gen8_emit_flush_render(struct drm_i915_gem_request *request,

Quoted 10 lines
drivers/gpu/drm/i915/intel_lrc.c +5 -5
-	 * On GEN9+ Before VF_CACHE_INVALIDATE we need to emit a NULL pipe
-	 * control.
-	 */
-	vf_flush_wa = INTEL_INFO(ring->dev)->gen >= 9 &&
-		      flags & PIPE_CONTROL_VF_CACHE_INVALIDATE;
+		/*
+		 * On GEN9: before VF_CACHE_INVALIDATE we need to emit a NULL
+		 * pipe control.
+		 */
+> 		if (IS_GEN9(ring->dev))

Nitpick: INTEL_INFO()->gen == 9 is the preferred way. Either way:

Quoted 2 lines

Sounds good to me. Does someone mind fixing this up when they push? This patch is a pre-requisite to another internal patch I have at the moment.

Quoted 5 lines
Quoted 5 lines
drivers/gpu/drm/i915/intel_lrc.c +2 -0
+			vf_flush_wa = true;
+	}


  ret = intel_logical_ring_begin(request, vf_flush_wa ? 12 : 6); if (ret)

Signature · 5 lines
Ben Widawsky, Intel Open Source Technology Center
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
maintainer
#4 lore raw
In reply to Imre Deak
Quoted 4 lines · Imre Deak
Quoted 1 line · Ben Widawsky

+> if (IS_GEN9(ring->dev))

Nitpick: INTEL_INFO()->gen == 9 is the preferred way.

Oh? Since when and says who?

BR,
Jani.

Signature · 5 lines
Jani Nikula, Intel Open Source Technology Center
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
#5 lore raw
Quoted 7 lines · Jani Nikula
Quoted 4 lines · Imre Deak
Quoted 1 line · Ben Widawsky

+>  if (IS_GEN9(ring->dev))

Nitpick: INTEL_INFO()->gen == 9 is the preferred way.

Oh? Since when and says who?

Haven't found  it after some digging, but I do remember someone having this comment earlier.

My opinion is that we should only use one form, and since INTEL_INFO()- gen can be used in a more generic manner I would prefer that. Currently we have it both ways in the code.

--Imre
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

#6 lore raw
Quoted 7 lines · Imre Deak
Quoted 6 lines · Jani Nikula
Quoted 4 lines · Imre Deak
Quoted 1 line · Ben Widawsky

+>  if (IS_GEN9(ring->dev))

Nitpick: INTEL_INFO()->gen == 9 is the preferred way.

Quoted 8 lines
Quoted 1 line

Oh? Since when and says who?

Haven't found  it after some digging, but I do remember someone having this comment earlier.

My opinion is that we should only use one form, and since INTEL_INFO()- gen can be used in a more generic manner I would prefer that. Currently we have it both ways in the code.

We use IS_GENx for individual gen checks and INTEL_INFO()->gen when checking for ranges. At least that's the usage I inferred ... -Daniel

Signature · 7 lines
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
#7 lore raw
In reply to Imre Deak
Quoted 54 lines · Imre Deak
Quoted 1 line · Ben Widawsky

It is unclear if this is even required on BXT.

I'm not sure either, I only added it on the premise that it was marked as SKL+ originally in BSpec. The revision log entry in BSpec has this much to say:
"""
The workaround that requires an empty PIPE_CONTROL before a PIPE_CONTROL with a VF Cache Invalidation Enable is only for SKL and not SKL+. The bug was fixed in CNL with the following HSD:... """
Which doesn't make this clear either imo.

Quoted 29 lines

v2: Make sure to set the default value to false. Uncertain how my compiler doesn't complain with v1.

---
 drivers/gpu/drm/i915/intel_lrc.c | 16 ++++++++--------

 1 file changed, 8 insertions(+), 8 deletions(-)
drivers/gpu/drm/i915/intel_lrc.c +1 -1
index 519cea32..af1e001 100644
@@ -1695,7 +1695,7 @@ static int gen8_emit_flush_render(struct drm_i915_gem_request *request,
  	struct intel_ringbuffer *ringbuf = request->ringbuf;
  	struct intel_engine_cs *ring = ringbuf->ring;
  	u32 scratch_addr = ring->scratch.gtt_offset + 2 * CACHELINE_BYTES;
-	bool vf_flush_wa;
+	bool vf_flush_wa = false;
  	u32 flags = 0;
  	int ret;

drivers/gpu/drm/i915/intel_lrc.c +0 -1
@@ -1716,14 +1716,14 @@ static int gen8_emit_flush_render(struct drm_i915_gem_request *request,
  		flags |= PIPE_CONTROL_STATE_CACHE_INVALIDATE;
  		flags |= PIPE_CONTROL_QW_WRITE;
  		flags |= PIPE_CONTROL_GLOBAL_GTT_IVB;
-	}

drivers/gpu/drm/i915/intel_lrc.c +0 -1
-	/*

drivers/gpu/drm/i915/intel_lrc.c static int gen8_emit_flush_render(struct drm_i915_gem_request *request,

Quoted 10 lines
drivers/gpu/drm/i915/intel_lrc.c +5 -5
-	 * On GEN9+ Before VF_CACHE_INVALIDATE we need to emit a NULL pipe
-	 * control.
-	 */
-	vf_flush_wa = INTEL_INFO(ring->dev)->gen >= 9 &&
-		      flags & PIPE_CONTROL_VF_CACHE_INVALIDATE;
+		/*
+		 * On GEN9: before VF_CACHE_INVALIDATE we need to emit a NULL
+		 * pipe control.
+		 */
+> 		if (IS_GEN9(ring->dev))

Nitpick: INTEL_INFO()->gen == 9 is the preferred way. Either way:

Quoted 1 line

Queued for -next, thanks for the patch.
-Daniel

Quoted 10 lines
Quoted 6 lines
drivers/gpu/drm/i915/intel_lrc.c +2 -0
+			vf_flush_wa = true;
+	}


  ret = intel_logical_ring_begin(request, vf_flush_wa ? 12 : 6); if (ret)

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

Signature · 7 lines
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
#8 lore raw

✗ warning: Fi.CI.BAT

In reply to Ben Widawsky
== Summary ==

== Summary ==

Built on ac2305b6c91b9a84cc12566016ece257c3ebcba3 drm-intel-nightly: 2015y-12m-17d-16h-19m-23s UTC integration manifest

Test igt@kms_flip@basic-flip-vs-wf_vblank on snb-x220t dmesg-warn -> pass Test igt@kms_flip@basic-flip-vs-wf_vblank on bsw-nuc-2 pass -> dmesg-warn Test igt@kms_flip@basic-flip-vs-wf_vblank on skl-i7k-2 dmesg-fail -> dmesg-warn Test igt@kms_pipe_crc_basic@read-crc-pipe-c-frame-sequence on bsw-nuc-2 dmesg-warn -> pass Test igt@kms_pipe_crc_basic@hang-read-crc-pipe-a on snb-x220t pass -> dmesg-warn Test igt@kms_flip@basic-flip-vs-dpms on ilk-hp8440p pass -> dmesg-warn
Test igt@kms_pipe_crc_basic@suspend-read-crc-pipe-b on snb-x220t dmesg-warn -> pass Test igt@kms_pipe_crc_basic@read-crc-pipe-a on snb-x220t pass -> dmesg-warn Test igt@kms_pipe_crc_basic@read-crc-pipe-c on skl-i7k-2 dmesg-warn -> pass Test igt@kms_pipe_crc_basic@read-crc-pipe-b on snb-dellxps dmesg-warn -> pass Test igt@kms_flip@basic-flip-vs-modeset on bsw-nuc-2 dmesg-warn -> pass Test igt@kms_flip@basic-flip-vs-modeset on skl-i5k-2 dmesg-warn -> pass

bsw-nuc-2        total:135  pass:114  dwarn:1   dfail:0   fail:0   skip:20 
hsw-brixbox      total:135  pass:127  dwarn:1   dfail:0   fail:0   skip:7  
hsw-gt2          total:135  pass:130  dwarn:1   dfail:0   fail:0   skip:4  
ilk-hp8440p      total:135  pass:99   dwarn:1   dfail:0   fail:2   skip:33 
ivb-t430s        total:135  pass:128  dwarn:1   dfail:1   fail:1   skip:4  
skl-i5k-2        total:135  pass:123  dwarn:4   dfail:0   fail:0   skip:8  
skl-i7k-2        total:135  pass:123  dwarn:4   dfail:0   fail:0   skip:8  
snb-dellxps      total:135  pass:122  dwarn:1   dfail:0   fail:0   skip:12 
snb-x220t        total:135  pass:121  dwarn:2   dfail:0   fail:1   skip:11 

Results at /archive/results/CI_IGT_test/Patchwork_706/

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx