Skip to contentSkip to conversation

drm/i915/skl: Implement WaBarrierPerformanceFixDisable (again)

PATCH In review Ben Widawsky 5 participants 6 messages Jun 2, 2015 – Jun 3, 2015

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

PATCH 1/1 drm/i915/skl: Implement WaBarrierPerformanceFixDisable (again)

in
commit 65ca7514e21adbee25b8175fc909759c735d00ff

Author: Damien Lespiau <[email protected]>
Date:   Mon Feb 9 19:33:22 2015 +0000

drm/i915/skl: Implement WaBarrierPerformanceFixDisable

The workaround ended up in the chv workarounds. Not sure what the reason or history of that is, but it /seems/ wrong. Don't know if this fixes anything since I have many other problems with my platform.

Cc Damien Lespiau <[email protected]>
Cc Nick Hoath <[email protected]>
---
 drivers/gpu/drm/i915/intel_ringbuffer.c | 14 +++++++-------
 1 file changed, 7 insertions(+), 7 deletions(-)
drivers/gpu/drm/i915/intel_ringbuffer.c +7 -7
index d934f85..0fd6033d 100644
@@ -901,13 +901,6 @@ static int chv_init_workarounds(struct intel_engine_cs *ring)
 			    GEN6_WIZ_HASHING_MASK,
 			    GEN6_WIZ_HASHING_16x4);
 
-	if (INTEL_REVID(dev) == SKL_REVID_C0 ||
-	    INTEL_REVID(dev) == SKL_REVID_D0)
-		/* WaBarrierPerformanceFixDisable:skl */
-		WA_SET_BIT_MASKED(HDC_CHICKEN0,
-				  HDC_FENCE_DEST_SLM_DISABLE |
-				  HDC_BARRIER_PERFORMANCE_DISABLE);
-
 	return 0;
 }
 
@@ -972,6 +965,13 @@ static int gen9_init_workarounds(struct intel_engine_cs *ring)
 		tmp |= HDC_FORCE_CSR_NON_COHERENT_OVR_DISABLE;
 	WA_SET_BIT_MASKED(HDC_CHICKEN0, tmp);
 
+	if (INTEL_REVID(dev) == SKL_REVID_C0 ||
+	    INTEL_REVID(dev) == SKL_REVID_D0)
+		/* WaBarrierPerformanceFixDisable:skl */
+		WA_SET_BIT_MASKED(HDC_CHICKEN0,
+				  HDC_FENCE_DEST_SLM_DISABLE |
+				  HDC_BARRIER_PERFORMANCE_DISABLE);
+
 	return 0;
 }
 
Signature · 6 lines
2.4.2

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Tested-by on v1 1/1
#1 lore raw

Task id: 6527 -------------------------------------Summary-------------------------------------
Platform Delta drm-intel-nightly Series Applied

PNV                                  276/276              276/276
ILK                                  303/303              303/303
SNB                 -1              315/315              314/315
IVB                                  343/343              343/343
BYT                                  287/287              287/287
BDW                                  321/321              321/321

-------------------------------------Detailed-------------------------------------
Platform Test drm-intel-nightly Series Applied
*SNB igt@pm_rpm@dpms-mode-unset-non-lpsp PASS(1) DMESG_WARN(1)
(dmesg patch applied)WARNING:at_drivers/gpu/drm/i915/intel_uncore.c:#assert_device_not_suspended[i915]()@WARNING:.* at .* assert_device_not_suspended+0x
Note: You need to pay more attention to line start with '*'
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

maintainer
#2 lore raw
In reply to Ben Widawsky
Quoted 7 lines · Ben Widawsky

in
commit 65ca7514e21adbee25b8175fc909759c735d00ff

Author: Damien Lespiau <[email protected]>
Date:   Mon Feb 9 19:33:22 2015 +0000

drm/i915/skl: Implement WaBarrierPerformanceFixDisable

Quoted 7 lines

The workaround ended up in the chv workarounds. Not sure what the reason or history of that is, but it /seems/ wrong. Don't know if this fixes anything since I have many other problems with my platform.

Cc Damien Lespiau <[email protected]>
Cc Nick Hoath <[email protected]>

Ville beat you to it [1]. However he put it in skl workarounds, not gen9, so I'm presuming this does not apply to bxt.

BR,
Jani.

[1] http://mid.gmane.org/[email protected]

Quoted 43 lines
---
 drivers/gpu/drm/i915/intel_ringbuffer.c | 14 +++++++-------
 1 file changed, 7 insertions(+), 7 deletions(-)
drivers/gpu/drm/i915/intel_ringbuffer.c +7 -7
index d934f85..0fd6033d 100644
@@ -901,13 +901,6 @@ static int chv_init_workarounds(struct intel_engine_cs *ring)
 			    GEN6_WIZ_HASHING_MASK,
 			    GEN6_WIZ_HASHING_16x4);
 
-	if (INTEL_REVID(dev) == SKL_REVID_C0 ||
-	    INTEL_REVID(dev) == SKL_REVID_D0)
-		/* WaBarrierPerformanceFixDisable:skl */
-		WA_SET_BIT_MASKED(HDC_CHICKEN0,
-				  HDC_FENCE_DEST_SLM_DISABLE |
-				  HDC_BARRIER_PERFORMANCE_DISABLE);
-
 	return 0;
 }
 
@@ -972,6 +965,13 @@ static int gen9_init_workarounds(struct intel_engine_cs *ring)
 		tmp |= HDC_FORCE_CSR_NON_COHERENT_OVR_DISABLE;
 	WA_SET_BIT_MASKED(HDC_CHICKEN0, tmp);
 
+	if (INTEL_REVID(dev) == SKL_REVID_C0 ||
+	    INTEL_REVID(dev) == SKL_REVID_D0)
+		/* WaBarrierPerformanceFixDisable:skl */
+		WA_SET_BIT_MASKED(HDC_CHICKEN0,
+				  HDC_FENCE_DEST_SLM_DISABLE |
+				  HDC_BARRIER_PERFORMANCE_DISABLE);
+
 	return 0;
 }
 
Signature · 6 lines
2.4.2

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Signature · 5 lines
Jani Nikula, Intel Open Source Technology Center
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Reviewed-by
#3 lore raw
Quoted 17 lines · On Wed, 03 Jun 2015 09:49:43 +0300 Jani Nikula <[email protected]>…
Quoted 7 lines · Ben Widawsky

in
commit 65ca7514e21adbee25b8175fc909759c735d00ff

Author: Damien Lespiau <[email protected]>
Date:   Mon Feb 9 19:33:22 2015 +0000

drm/i915/skl: Implement WaBarrierPerformanceFixDisable

Quoted 8 lines

The workaround ended up in the chv workarounds. Not sure what the reason or history of that is, but it /seems/ wrong. Don't know if this fixes anything since I have many other problems with my platform.

Cc Damien Lespiau <[email protected]>
Cc Nick Hoath <[email protected]>
Quoted 7 lines

Ville beat you to it [1]. However he put it in skl workarounds, not gen9, so I'm presuming this does not apply to bxt.

BR,
Jani.

Anyone who will talk to us and tells you they know the answer to that definitively is lying.

My patch is wrong though. Checking only the REVID without checking IS_SKYLAKE is not the right thing to do.

Consider Ville's patch

(I no longer subscribe to intel-gfx, so actually adding my review is kind of a pain).

Quoted 40 lines

[1]
http://mid.gmane.org/[email protected]

Quoted 36 lines
---
 drivers/gpu/drm/i915/intel_ringbuffer.c | 14 +++++++-------
 1 file changed, 7 insertions(+), 7 deletions(-)
(quoted diff) +0 -0

b/drivers/gpu/drm/i915/intel_ringbuffer.c index d934f85..0fd6033d 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.c +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c

drivers/gpu/drm/i915/intel_ringbuffer.c +0 -7
@@ -901,13 +901,6 @@ static int chv_init_workarounds(struct
 intel_engine_cs *ring) GEN6_WIZ_HASHING_MASK,
 			    GEN6_WIZ_HASHING_16x4);
 
-	if (INTEL_REVID(dev) == SKL_REVID_C0 ||
-	    INTEL_REVID(dev) == SKL_REVID_D0)
-		/* WaBarrierPerformanceFixDisable:skl */
-		WA_SET_BIT_MASKED(HDC_CHICKEN0,
-				  HDC_FENCE_DEST_SLM_DISABLE |
-				  HDC_BARRIER_PERFORMANCE_DISABLE);
-
 	return 0;
 }
 
@@ -972,6 +965,13 @@ static int gen9_init_workarounds(struct

intel_engine_cs *ring) tmp |=
HDC_FORCE_CSR_NON_COHERENT_OVR_DISABLE;
WA_SET_BIT_MASKED(HDC_CHICKEN0, tmp);

drivers/gpu/drm/i915/intel_ringbuffer.c +7 -0
+	if (INTEL_REVID(dev) == SKL_REVID_C0 ||
+	    INTEL_REVID(dev) == SKL_REVID_D0)
+		/* WaBarrierPerformanceFixDisable:skl */
+		WA_SET_BIT_MASKED(HDC_CHICKEN0,
+				  HDC_FENCE_DEST_SLM_DISABLE |
+				  HDC_BARRIER_PERFORMANCE_DISABLE);
+
 	return 0;
 }
Quoted 8 lines
Quoted 7 lines
Signature · 6 lines
2.4.2

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

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

Reviewed-by on v1 1/1
#4 lore raw
In reply to Ben Widawsky

On 03/06/2015 00:02, Widawsky, Benjamin wrote:

Probably should have a line like:
Problem introduced in:
instead of just 'in'

Quoted 9 lines

in
commit 65ca7514e21adbee25b8175fc909759c735d00ff

Author: Damien Lespiau <[email protected]>
Date:   Mon Feb 9 19:33:22 2015 +0000

drm/i915/skl: Implement WaBarrierPerformanceFixDisable

The workaround ended up in the chv workarounds. Not sure what the reason or history of that is, but it /seems/ wrong. Don't know if this fixes anything

patch doesn't always get it right...

Quoted 32 lines

since I have many other problems with my platform.

Cc Damien Lespiau <[email protected]>
Cc Nick Hoath <[email protected]>
---
  drivers/gpu/drm/i915/intel_ringbuffer.c | 14 +++++++-------
  1 file changed, 7 insertions(+), 7 deletions(-)
drivers/gpu/drm/i915/intel_ringbuffer.c +0 -7
index d934f85..0fd6033d 100644
@@ -901,13 +901,6 @@ static int chv_init_workarounds(struct intel_engine_cs *ring)
  			    GEN6_WIZ_HASHING_MASK,
  			    GEN6_WIZ_HASHING_16x4);
 
-	if (INTEL_REVID(dev) == SKL_REVID_C0 ||
-	    INTEL_REVID(dev) == SKL_REVID_D0)
-		/* WaBarrierPerformanceFixDisable:skl */
-		WA_SET_BIT_MASKED(HDC_CHICKEN0,
-				  HDC_FENCE_DEST_SLM_DISABLE |
-				  HDC_BARRIER_PERFORMANCE_DISABLE);
-
  	return 0;
  }
 
@@ -972,6 +965,13 @@ static int gen9_init_workarounds(struct intel_engine_cs *ring)
  		tmp |= HDC_FORCE_CSR_NON_COHERENT_OVR_DISABLE;
  	WA_SET_BIT_MASKED(HDC_CHICKEN0, tmp);
 

drivers/gpu/drm/i915/intel_ringbuffer.c static int gen9_init_workarounds(struct intel_engine_cs *ring) · line 977

Quoted 10 lines
drivers/gpu/drm/i915/intel_ringbuffer.c +7 -0
+	if (INTEL_REVID(dev) == SKL_REVID_C0 ||
+	    INTEL_REVID(dev) == SKL_REVID_D0)
+		/* WaBarrierPerformanceFixDisable:skl */
+		WA_SET_BIT_MASKED(HDC_CHICKEN0,
+				  HDC_FENCE_DEST_SLM_DISABLE |
+				  HDC_BARRIER_PERFORMANCE_DISABLE);
+
  	return 0;
  }
 

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

maintainer
#5 lore raw
Quoted 11 lines · Nick Hoath

On 03/06/2015 00:02, Widawsky, Benjamin wrote:

Probably should have a line like:
Problem introduced in:
instead of just 'in'

Quoted 5 lines

in
commit 65ca7514e21adbee25b8175fc909759c735d00ff

Author: Damien Lespiau <[email protected]>
Date:   Mon Feb 9 19:33:22 2015 +0000
Quoted 8 lines
Quoted 4 lines

drm/i915/skl: Implement WaBarrierPerformanceFixDisable

The workaround ended up in the chv workarounds. Not sure what the reason or history of that is, but it /seems/ wrong. Don't know if this fixes anything

patch doesn't always get it right...

Already pushed Ville's patch
http://mid.gmane.org/[email protected]

BR,
Jani.

Quoted 49 lines
Quoted 33 lines

since I have many other problems with my platform.

Cc Damien Lespiau <[email protected]>
Cc Nick Hoath <[email protected]>
---
  drivers/gpu/drm/i915/intel_ringbuffer.c | 14 +++++++-------
  1 file changed, 7 insertions(+), 7 deletions(-)
drivers/gpu/drm/i915/intel_ringbuffer.c +0 -7
index d934f85..0fd6033d 100644
@@ -901,13 +901,6 @@ static int chv_init_workarounds(struct intel_engine_cs *ring)
  			    GEN6_WIZ_HASHING_MASK,
  			    GEN6_WIZ_HASHING_16x4);
 
-	if (INTEL_REVID(dev) == SKL_REVID_C0 ||
-	    INTEL_REVID(dev) == SKL_REVID_D0)
-		/* WaBarrierPerformanceFixDisable:skl */
-		WA_SET_BIT_MASKED(HDC_CHICKEN0,
-				  HDC_FENCE_DEST_SLM_DISABLE |
-				  HDC_BARRIER_PERFORMANCE_DISABLE);
-
  	return 0;
  }
 
@@ -972,6 +965,13 @@ static int gen9_init_workarounds(struct intel_engine_cs *ring)
  		tmp |= HDC_FORCE_CSR_NON_COHERENT_OVR_DISABLE;
  	WA_SET_BIT_MASKED(HDC_CHICKEN0, tmp);
 

drivers/gpu/drm/i915/intel_ringbuffer.c static int gen9_init_workarounds(struct intel_engine_cs *ring) · line 977

Quoted 10 lines
drivers/gpu/drm/i915/intel_ringbuffer.c +7 -0
+	if (INTEL_REVID(dev) == SKL_REVID_C0 ||
+	    INTEL_REVID(dev) == SKL_REVID_D0)
+		/* WaBarrierPerformanceFixDisable:skl */
+		WA_SET_BIT_MASKED(HDC_CHICKEN0,
+				  HDC_FENCE_DEST_SLM_DISABLE |
+				  HDC_BARRIER_PERFORMANCE_DISABLE);
+
  	return 0;
  }
 

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

Signature · 5 lines
Jani Nikula, Intel Open Source Technology Center
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx