Skip to contentSkip to conversation

drm/i915/gen9: implement WaConextSwitchWithConcurrentTLBInvalidate

PATCH Superseded [email protected] 3 participants 6 messages Jun 9, 2016 – Jun 10, 2016

Subsystems (2)

v2 of 3 Superseded Latest is v3 →

lore
Share & tools
lore
#0 lore raw

PATCH v2 1/1 drm/i915/gen9: implement WaConextSwitchWithConcurrentTLBInvalidate

From: Tim Gore <[email protected]>

This patch enables a workaround for a mid thread preemption issue where a hardware timing problem can prevent the context restore from happening, leading to a hang.

v2: move to gen9_init_workarounds (Arun)

---
 drivers/gpu/drm/i915/i915_reg.h         | 4 ++++
 drivers/gpu/drm/i915/intel_ringbuffer.c | 3 +++
 2 files changed, 7 insertions(+)
drivers/gpu/drm/i915/i915_reg.h +4 -0
index 81d1896..2a6fc62 100644
@@ -1810,6 +1810,10 @@ enum skl_disp_power_wells {
 #define   GEN9_IZ_HASHING_MASK(slice)			(0x3 << ((slice) * 2))
 #define   GEN9_IZ_HASHING(slice, val)			((val) << ((slice) * 2))
 
+/* chicken reg for WaConextSwitchWithConcurrentTLBInvalidate */
+#define GEN9_CSFE_CHICKEN1_RCS _MMIO(0x20D4)
+#define   GEN9_PREEMPT_GPGPU_SYNC_SWITCH_DISABLE (1 << 2)
+
 /* WaClearTdlStateAckDirtyBits */
 #define GEN8_STATE_ACK		_MMIO(0x20F0)
 #define GEN9_STATE_ACK_SLICE1	_MMIO(0x20F8)
drivers/gpu/drm/i915/intel_ringbuffer.c +3 -0
index cf8d0bf..7c756ac 100644
@@ -1022,6 +1022,9 @@ static int gen9_init_workarounds(struct intel_engine_cs *engine)
 	if (ret)
 		return ret;
 
+	/* WaConextSwitchWithConcurrentTLBInvalidate:skl,bxt,kbl */
+	I915_WRITE(GEN9_CSFE_CHICKEN1_RCS, _MASKED_BIT_ENABLE(GEN9_PREEMPT_GPGPU_SYNC_SWITCH_DISABLE));
+
 	return 0;
 }
 
Signature · 6 lines
1.9.1

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

✗ Ro.CI.BAT: warning for drm/i915/gen9: implement WaConextSwitchWithConcurrentTLBInvalidate (rev2)

== Series Details ==

== Series Details ==

Series: drm/i915/gen9: implement WaConextSwitchWithConcurrentTLBInvalidate (rev2)
URL : https://patchwork.freedesktop.org/series/8487/
State : warning

== Summary ==

Series 8487v2 drm/i915/gen9: implement WaConextSwitchWithConcurrentTLBInvalidate
http://patchwork.freedesktop.org/api/1.0/series/8487/revisions/2/mbox

Test gem_exec_flush:

        Subgroup basic-batch-kernel-default-cmd:
                fail       -> PASS       (ro-byt-n2820)
Test kms_pipe_crc_basic:
        Subgroup suspend-read-crc-pipe-a:
                pass       -> DMESG-WARN (fi-snb-i7-2600)
        Subgroup suspend-read-crc-pipe-c:
                skip       -> DMESG-WARN (ro-bdw-i5-5250u)
fi-bdw-i7-5557u  total:213  pass:201  dwarn:0   dfail:0   fail:0   skip:12 
fi-skl-i5-6260u  total:213  pass:202  dwarn:0   dfail:0   fail:0   skip:11 
fi-skl-i7-6700k  total:213  pass:188  dwarn:0   dfail:0   fail:0   skip:25 
fi-snb-i7-2600   total:213  pass:173  dwarn:1   dfail:0   fail:0   skip:39 
ro-bdw-i5-5250u  total:213  pass:197  dwarn:4   dfail:0   fail:0   skip:12 
ro-bdw-i7-5600u  total:213  pass:185  dwarn:0   dfail:0   fail:0   skip:28 
ro-bsw-n3050     total:213  pass:172  dwarn:0   dfail:0   fail:2   skip:39 
ro-byt-n2820     total:213  pass:174  dwarn:0   dfail:0   fail:2   skip:37 
ro-hsw-i7-4770r  total:213  pass:190  dwarn:0   dfail:0   fail:0   skip:23 
ro-ilk1-i5-650   total:208  pass:150  dwarn:0   dfail:0   fail:1   skip:57 
ro-ivb-i7-3770   total:213  pass:181  dwarn:0   dfail:0   fail:0   skip:32 
ro-ivb2-i7-3770  total:213  pass:185  dwarn:0   dfail:0   fail:0   skip:28 
ro-snb-i7-2620M  total:213  pass:174  dwarn:0   dfail:0   fail:1   skip:38 

fi-hsw-i7-4770k failed to connect after reboot

Results at /archive/results/CI_IGT_test/RO_Patchwork_1149/

487a126 drm-intel-nightly: 2016y-06m-09d-10h-05m-55s UTC integration manifest 6755d5e drm/i915/gen9: implement WaConextSwitchWithConcurrentTLBInvalidate

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

#2 lore raw
In reply to [email protected]
Quoted 34 lines · On 09/06/2016 20:19, [email protected] wrote:

From: Tim Gore <[email protected]>

This patch enables a workaround for a mid thread preemption issue where a hardware timing problem can prevent the context restore from happening, leading to a hang.

v2: move to gen9_init_workarounds (Arun)

---
  drivers/gpu/drm/i915/i915_reg.h         | 4 ++++
  drivers/gpu/drm/i915/intel_ringbuffer.c | 3 +++
  2 files changed, 7 insertions(+)
drivers/gpu/drm/i915/i915_reg.h +4 -0
index 81d1896..2a6fc62 100644
@@ -1810,6 +1810,10 @@ enum skl_disp_power_wells {
  #define   GEN9_IZ_HASHING_MASK(slice)			(0x3 << ((slice) * 2))
  #define   GEN9_IZ_HASHING(slice, val)			((val) << ((slice) * 2))
 
+/* chicken reg for WaConextSwitchWithConcurrentTLBInvalidate */
+#define GEN9_CSFE_CHICKEN1_RCS _MMIO(0x20D4)
+#define   GEN9_PREEMPT_GPGPU_SYNC_SWITCH_DISABLE (1 << 2)
+
  /* WaClearTdlStateAckDirtyBits */
  #define GEN8_STATE_ACK		_MMIO(0x20F0)
  #define GEN9_STATE_ACK_SLICE1	_MMIO(0x20F8)
drivers/gpu/drm/i915/intel_ringbuffer.c +0 -0
index cf8d0bf..7c756ac 100644

drivers/gpu/drm/i915/intel_ringbuffer.c static int gen9_init_workarounds(struct intel_engine_cs *engine) · line 1027

Quoted 6 lines
drivers/gpu/drm/i915/intel_ringbuffer.c +3 -0
@@ -1022,6 +1022,9 @@ static int gen9_init_workarounds(struct intel_engine_cs *engine)
  	if (ret)
  		return ret;
 
+	/* WaConextSwitchWithConcurrentTLBInvalidate:skl,bxt,kbl */
+	I915_WRITE(GEN9_CSFE_CHICKEN1_RCS, _MASKED_BIT_ENABLE(GEN9_PREEMPT_GPGPU_SYNC_SWITCH_DISABLE));
+

WA_SET_BIT_MASKED(GEN9_CSFE_CHICKEN1_RCS, GEN9_PREEMPT_GPGPU_SYNC_SWITCH_DISABLE);

Please correct the spelling.
We should try to keep WA regs in some order although it is not true for some of the existing ones but we should try to follow this rule for the new ones; HW whitelist registers are normally kept at the end. I think the correct place for this one is at the beginning of this function to maintain increasing order.

regards
Arun

drivers/gpu/drm/i915/intel_ringbuffer.c static int gen9_init_workarounds(struct intel_engine_cs *engine) · line 1030

Quoted 4 lines
drivers/gpu/drm/i915/intel_ringbuffer.c +0 -0
  	return 0;
  }
 

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

author
#3 lore raw

Tim Gore
Intel Corporation (UK) Ltd. - Co. Reg. #1134945 - Pipers Way, Swindon SN3 1RJ

Quoted 65 lines

-----Original Message-----
From: Arun Siluvery [mailto:[email protected]]
Sent: Friday, June 10, 2016 7:30 AM
To: Gore, Tim; [email protected]
Subject: Re: [PATCH v2] drm/i915/gen9: implement WaConextSwitchWithConcurrentTLBInvalidate

Quoted 19 lines · On 09/06/2016 20:19, [email protected] wrote:

From: Tim Gore <[email protected]>

This patch enables a workaround for a mid thread preemption issue where a hardware timing problem can prevent the context restore from happening, leading to a hang.

v2: move to gen9_init_workarounds (Arun)

---
  drivers/gpu/drm/i915/i915_reg.h         | 4 ++++
  drivers/gpu/drm/i915/intel_ringbuffer.c | 3 +++
  2 files changed, 7 insertions(+)
(quoted diff) +0 -0

b/drivers/gpu/drm/i915/i915_reg.h index 81d1896..2a6fc62 100644

drivers/gpu/drm/i915/i915_reg.h enum skl_disp_power_wells { · line 1810

Quoted 1 line
drivers/gpu/drm/i915/i915_reg.h +0 -0
@@ -1810,6 +1810,10 @@ enum skl_disp_power_wells {
  #define   GEN9_IZ_HASHING_MASK(slice)			(0x3 <<

((slice) * 2))

drivers/gpu/drm/i915/i915_reg.h enum skl_disp_power_wells { · line 1811

Quoted 1 line
drivers/gpu/drm/i915/i915_reg.h +0 -0
  #define   GEN9_IZ_HASHING(slice, val)			((val) <<

((slice) * 2))

drivers/gpu/drm/i915/intel_ringbuffer.c static int gen9_init_workarounds(struct

Quoted 14 lines
drivers/gpu/drm/i915/i915_reg.h +4 -0
+/* chicken reg for WaConextSwitchWithConcurrentTLBInvalidate */
+#define GEN9_CSFE_CHICKEN1_RCS _MMIO(0x20D4)
+#define   GEN9_PREEMPT_GPGPU_SYNC_SWITCH_DISABLE (1 << 2)
+
  /* WaClearTdlStateAckDirtyBits */
  #define GEN8_STATE_ACK		_MMIO(0x20F0)
  #define GEN9_STATE_ACK_SLICE1	_MMIO(0x20F8)
drivers/gpu/drm/i915/intel_ringbuffer.c +0 -0
index cf8d0bf..7c756ac 100644
@@ -1022,6 +1022,9 @@ static int gen9_init_workarounds(struct

intel_engine_cs *engine)

drivers/gpu/drm/i915/intel_ringbuffer.c static int gen9_init_workarounds(struct intel_engine_cs *engine) · line 1030

Quoted 6 lines
drivers/gpu/drm/i915/intel_ringbuffer.c +2 -0
  	if (ret)
  		return ret;
 
+	/* WaConextSwitchWithConcurrentTLBInvalidate:skl,bxt,kbl */
+	I915_WRITE(GEN9_CSFE_CHICKEN1_RCS,
 

+_MASKED_BIT_ENABLE(GEN9_PREEMPT_GPGPU_SYNC_SWITCH_DISABLE) );

drivers/gpu/drm/i915/intel_ringbuffer.c static int gen9_init_workarounds(struct · line 1028

Quoted 1 line
drivers/gpu/drm/i915/intel_ringbuffer.c +1 -0
+

WA_SET_BIT_MASKED(GEN9_CSFE_CHICKEN1_RCS,
GEN9_PREEMPT_GPGPU_SYNC_SWITCH_DISABLE);

Please correct the spelling.
We should try to keep WA regs in some order although it is not true for some of the existing ones but we should try to follow this rule for the new ones; HW whitelist registers are normally kept at the end.
I think the correct place for this one is at the beginning of this function to maintain increasing order.

Quoted 4 lines

regards
Arun

Which spelling do you want corrected?

Tim

Quoted 4 lines

drivers/gpu/drm/i915/intel_ringbuffer.c static int gen9_init_workarounds(struct intel_engine_cs *engine) · line 1029

Quoted 4 lines
drivers/gpu/drm/i915/intel_ringbuffer.c +0 -0
  	return 0;
  }

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

#4 lore raw
Quoted 71 lines · Tim

Tim Gore
Intel Corporation (UK) Ltd. - Co. Reg. #1134945 - Pipers Way, Swindon SN3 1RJ

Quoted 65 lines

-----Original Message-----
From: Arun Siluvery [mailto:[email protected]]
Sent: Friday, June 10, 2016 7:30 AM
To: Gore, Tim; [email protected]
Subject: Re: [PATCH v2] drm/i915/gen9: implement WaConextSwitchWithConcurrentTLBInvalidate

Quoted 19 lines · On 09/06/2016 20:19, [email protected] wrote:

From: Tim Gore <[email protected]>

This patch enables a workaround for a mid thread preemption issue where a hardware timing problem can prevent the context restore from happening, leading to a hang.

v2: move to gen9_init_workarounds (Arun)

---
   drivers/gpu/drm/i915/i915_reg.h         | 4 ++++
   drivers/gpu/drm/i915/intel_ringbuffer.c | 3 +++
   2 files changed, 7 insertions(+)
(quoted diff) +0 -0

b/drivers/gpu/drm/i915/i915_reg.h index 81d1896..2a6fc62 100644

drivers/gpu/drm/i915/i915_reg.h enum skl_disp_power_wells { · line 1810

Quoted 1 line
drivers/gpu/drm/i915/i915_reg.h +0 -0
@@ -1810,6 +1810,10 @@ enum skl_disp_power_wells {
   #define   GEN9_IZ_HASHING_MASK(slice)			(0x3 <<

((slice) * 2))

drivers/gpu/drm/i915/i915_reg.h enum skl_disp_power_wells { · line 1811

Quoted 1 line
drivers/gpu/drm/i915/i915_reg.h +0 -0
   #define   GEN9_IZ_HASHING(slice, val)			((val) <<

((slice) * 2))

drivers/gpu/drm/i915/intel_ringbuffer.c static int gen9_init_workarounds(struct

Quoted 14 lines
drivers/gpu/drm/i915/i915_reg.h +4 -0
+/* chicken reg for WaConextSwitchWithConcurrentTLBInvalidate */
+#define GEN9_CSFE_CHICKEN1_RCS _MMIO(0x20D4)
+#define   GEN9_PREEMPT_GPGPU_SYNC_SWITCH_DISABLE (1 << 2)
+
   /* WaClearTdlStateAckDirtyBits */
   #define GEN8_STATE_ACK		_MMIO(0x20F0)
   #define GEN9_STATE_ACK_SLICE1	_MMIO(0x20F8)
drivers/gpu/drm/i915/intel_ringbuffer.c +0 -0
index cf8d0bf..7c756ac 100644
@@ -1022,6 +1022,9 @@ static int gen9_init_workarounds(struct

intel_engine_cs *engine)

drivers/gpu/drm/i915/intel_ringbuffer.c static int gen9_init_workarounds(struct intel_engine_cs *engine) · line 1030

Quoted 6 lines
drivers/gpu/drm/i915/intel_ringbuffer.c +2 -0
   	if (ret)
   		return ret;
 
+	/* WaConextSwitchWithConcurrentTLBInvalidate:skl,bxt,kbl */
+	I915_WRITE(GEN9_CSFE_CHICKEN1_RCS,
 

+_MASKED_BIT_ENABLE(GEN9_PREEMPT_GPGPU_SYNC_SWITCH_DISABLE) );

drivers/gpu/drm/i915/intel_ringbuffer.c static int gen9_init_workarounds(struct · line 1028

Quoted 1 line
drivers/gpu/drm/i915/intel_ringbuffer.c +1 -0
+

WA_SET_BIT_MASKED(GEN9_CSFE_CHICKEN1_RCS,
GEN9_PREEMPT_GPGPU_SYNC_SWITCH_DISABLE);

Please correct the spelling.
We should try to keep WA regs in some order although it is not true for some of the existing ones but we should try to follow this rule for the new ones; HW whitelist registers are normally kept at the end.
I think the correct place for this one is at the beginning of this function to maintain increasing order.

Quoted 5 lines
Quoted 4 lines

regards
Arun

Which spelling do you want corrected?

WaConextSwitchWithConcurrentTLBInvalidate - two occurrences.

regards
Arun

Quoted 8 lines

Tim

Quoted 4 lines
Quoted 4 lines
drivers/gpu/drm/i915/intel_ringbuffer.c +0 -0
   	return 0;
   }

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

author
#5 lore raw

Tim Gore
Intel Corporation (UK) Ltd. - Co. Reg. #1134945 - Pipers Way, Swindon SN3 1RJ

Quoted 86 lines

-----Original Message-----
From: Arun Siluvery [mailto:[email protected]]
Sent: Friday, June 10, 2016 9:38 AM
To: Gore, Tim; [email protected]
Subject: Re: [PATCH v2] drm/i915/gen9: implement WaConextSwitchWithConcurrentTLBInvalidate

Quoted 29 lines · Tim

Tim Gore
Intel Corporation (UK) Ltd. - Co. Reg. #1134945 - Pipers Way, Swindon SN3 1RJ

Quoted 22 lines

-----Original Message-----
From: Arun Siluvery [mailto:[email protected]]
Sent: Friday, June 10, 2016 7:30 AM
To: Gore, Tim; [email protected]
Subject: Re: [PATCH v2] drm/i915/gen9: implement WaConextSwitchWithConcurrentTLBInvalidate

Quoted 14 lines · On 09/06/2016 20:19, [email protected] wrote:

From: Tim Gore <[email protected]>

This patch enables a workaround for a mid thread preemption issue where a hardware timing problem can prevent the context restore from happening, leading to a hang.

v2: move to gen9_init_workarounds (Arun)

---
   drivers/gpu/drm/i915/i915_reg.h         | 4 ++++
   drivers/gpu/drm/i915/intel_ringbuffer.c | 3 +++
   2 files changed, 7 insertions(+)
Quoted 6 lines
Quoted 6 lines

drivers/gpu/drm/i915/i915_reg.h enum skl_disp_power_wells { · line 1810

Quoted 6 lines
(quoted diff) +0 -0

b/drivers/gpu/drm/i915/i915_reg.h index 81d1896..2a6fc62 100644

drivers/gpu/drm/i915/i915_reg.h +0 -0
@@ -1810,6 +1810,10 @@ enum skl_disp_power_wells {
   #define   GEN9_IZ_HASHING_MASK(slice)			(0x3

<<

Quoted 18 lines
Quoted 18 lines

((slice) * 2))

drivers/gpu/drm/i915/i915_reg.h enum skl_disp_power_wells { · line 1811

Quoted 1 line
drivers/gpu/drm/i915/i915_reg.h +0 -0
   #define   GEN9_IZ_HASHING(slice, val)			((val) <<

((slice) * 2))

drivers/gpu/drm/i915/intel_ringbuffer.c static int gen9_init_workarounds(struct

Quoted 14 lines
drivers/gpu/drm/i915/i915_reg.h +4 -0
+/* chicken reg for WaConextSwitchWithConcurrentTLBInvalidate */
+#define GEN9_CSFE_CHICKEN1_RCS _MMIO(0x20D4)
+#define   GEN9_PREEMPT_GPGPU_SYNC_SWITCH_DISABLE (1 << 2)
+
   /* WaClearTdlStateAckDirtyBits */
   #define GEN8_STATE_ACK		_MMIO(0x20F0)
   #define GEN9_STATE_ACK_SLICE1	_MMIO(0x20F8)
drivers/gpu/drm/i915/intel_ringbuffer.c +0 -0
index cf8d0bf..7c756ac 100644
@@ -1022,6 +1022,9 @@ static int gen9_init_workarounds(struct

intel_engine_cs *engine)

Quoted 7 lines
Quoted 7 lines

drivers/gpu/drm/i915/intel_ringbuffer.c static int gen9_init_workarounds(struct intel_engine_cs *engine) · line 1030

Quoted 6 lines
drivers/gpu/drm/i915/intel_ringbuffer.c +2 -0
   	if (ret)
   		return ret;
 
+	/* WaConextSwitchWithConcurrentTLBInvalidate:skl,bxt,kbl */
+	I915_WRITE(GEN9_CSFE_CHICKEN1_RCS,
 

+_MASKED_BIT_ENABLE(GEN9_PREEMPT_GPGPU_SYNC_SWITCH_DISABLE)

Quoted 16 lines
Quoted 16 lines

);

drivers/gpu/drm/i915/intel_ringbuffer.c static int gen9_init_workarounds(struct · line 1028

Quoted 1 line
drivers/gpu/drm/i915/intel_ringbuffer.c +1 -0
+

WA_SET_BIT_MASKED(GEN9_CSFE_CHICKEN1_RCS,
GEN9_PREEMPT_GPGPU_SYNC_SWITCH_DISABLE);

Please correct the spelling.
We should try to keep WA regs in some order although it is not true for some of the existing ones but we should try to follow this rule for the new ones; HW whitelist registers are normally kept at the end. I think the correct place for this one is at the beginning of this function to maintain increasing order.

regards
Arun

Quoted 7 lines
Quoted 1 line

Which spelling do you want corrected?

WaConextSwitchWithConcurrentTLBInvalidate - two occurrences.

regards
Arun

As per my previous replies to Jani and you, WaConextSwitchWithConcurrentTLBInvalidate is copied directly from the documentation (bug report etc) and also from another implementation of this workaround. Actually, to be precise the spec has "WAConextSwitchWithConcurrentTLBInvalidate" (capitalised WA), but the bug report and other implementation use "WaConextSwitchWithConcurrentTLBInvalidate".
I am not really comfortable with using a different spelling here, even if it is "correct". Seems like it could lead to confusion or at least trip up some script.

Tim

Quoted 8 lines
Quoted 8 lines

Tim

Quoted 4 lines
Quoted 4 lines
drivers/gpu/drm/i915/intel_ringbuffer.c +0 -0
   	return 0;
   }

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