Skip to contentSkip to conversation

Gen9 HW whitelist and Preemption WA patches

PATCH Superseded Arun Siluvery 5 participants 20 messages Jan 21, 2016 – Jan 25, 2016

Subsystems (2)

v2 of 4 Superseded Latest is v4 →

lore
Share & tools
lore
#0 lore raw

v2 cover Gen9 HW whitelist and Preemption WA patches

Resending all patches as told by Daniel because I didn't use correct message-id while replying updated version of Patch1 before which means patchwork won't pickup and we won't have CI results. Previous updates regarding Patch1 are available at https://patchwork.freedesktop.org/patch/70527/

Some of the WA patches are now reviewed so I added r-b tags for them.

Reviewed: 1, 2, 3, 5, 6
Updated: 4 (to address review comment)
To be reviewed: 4, 7, 8

Arun Siluvery (8):
drm/i915/gen9: Add framework to whitelist specific GPU registers drm/i915/gen9: Add GEN8_CS_CHICKEN1 to HW whitelist drm/i915/gen9: Add HDC_CHICKEN1 to HW whitelist drm/i915/bxt: Add GEN9_CS_DEBUG_MODE1 to HW whitelist drm/i915/bxt: Add GEN8_L3SQCREG4 to HW whitelist drm/i915/skl: Add GEN8_L3SQCREG4 to HW whitelist drm/i915/skl: Enable Per context Preemption granularity control drm/i915/gen9: Add WaOCLCoherentLineFlush

 drivers/gpu/drm/i915/i915_debugfs.c     | 15 +++++---
 drivers/gpu/drm/i915/i915_drv.h         |  9 ++++-
 drivers/gpu/drm/i915/i915_reg.h         | 11 ++++++
 drivers/gpu/drm/i915/intel_ringbuffer.c | 61 +++++++++++++++++++++++++++++++++
 4 files changed, 90 insertions(+), 6 deletions(-)
Signature · 6 lines
1.9.1

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

PATCH v2 1/8 drm/i915/gen9: Add framework to whitelist specific GPU registers

Some of the HW registers are privileged and cannot be written to from non-privileged batch buffers coming from userspace unless they are added to the HW whitelist. This whitelist is maintained by HW and it is different from SW whitelist. Userspace need write access to them to implement preemption related WA.

The reason for using this approach is, the register bits that control preemption granularity at the HW level are not context save/restored; so even if we set these bits always in kernel they are going to change once the context is switched out. We can consider making them non-privileged by default but these registers also contain other chicken bits which should not be allowed to be modified.

In the later revisions controlling bits are save/restored at context level but in the existing revisions these are exported via other debug registers and should be on the whitelist. This patch adds changes to provide HW with a list of registers to be whitelisted. HW checks this list during execution and provides access accordingly.

HW imposes a limit on the number of registers on whitelist and it is per-engine. At this point we are only enabling whitelist for RCS and we don't foresee any requirement for other engines.

The registers to be whitelisted are added using generic workaround list mechanism, even these are only enablers for userspace workarounds. But by sharing this mechanism we get some test assets without additional cost (Mika).

v2: rebase

v3: parameterize RING_FORCE_TO_NONPRIV() as _MMIO() should be limited to i915_reg.h (Ville), drop inline for wa_ring_whitelist_reg (Mika).

v4: improvements suggested by Chris Wilson.
Clarify that this is HW whitelist and different from the one maintained in driver. This list is engine specific but it gets initialized along with other WA which is RCS specific thing, so make it clear that we are not doing any cross engine setup during initialization.
Make HW whitelist count of each engine available in debugfs.

---
 drivers/gpu/drm/i915/i915_debugfs.c     | 15 ++++++++++-----
 drivers/gpu/drm/i915/i915_drv.h         |  9 ++++++++-
 drivers/gpu/drm/i915/i915_reg.h         |  3 +++
 drivers/gpu/drm/i915/intel_ringbuffer.c | 17 +++++++++++++++++
 4 files changed, 38 insertions(+), 6 deletions(-)
drivers/gpu/drm/i915/i915_debugfs.c +10 -5
index e3377ab..7eb002c 100644
@@ -3229,9 +3229,11 @@ static int i915_wa_registers(struct seq_file *m, void *unused)
 {
 	int i;
 	int ret;
+	struct intel_engine_cs *ring;
 	struct drm_info_node *node = (struct drm_info_node *) m->private;
 	struct drm_device *dev = node->minor->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct i915_workarounds *workarounds = &dev_priv->workarounds;
 
 	ret = mutex_lock_interruptible(&dev->struct_mutex);
 	if (ret)
@@ -3239,15 +3241,18 @@ static int i915_wa_registers(struct seq_file *m, void *unused)
 
 	intel_runtime_pm_get(dev_priv);
 
-	seq_printf(m, "Workarounds applied: %d\n", dev_priv->workarounds.count);
-	for (i = 0; i < dev_priv->workarounds.count; ++i) {
+	seq_printf(m, "Workarounds applied: %d\n", workarounds->count);
+	for_each_ring(ring, dev_priv, i)
+		seq_printf(m, "HW whitelist count for %s: %d\n",
+			   ring->name, workarounds->hw_whitelist_count[i]);
+	for (i = 0; i < workarounds->count; ++i) {
 		i915_reg_t addr;
 		u32 mask, value, read;
 		bool ok;
 
-		addr = dev_priv->workarounds.reg[i].addr;
-		mask = dev_priv->workarounds.reg[i].mask;
-		value = dev_priv->workarounds.reg[i].value;
+		addr = workarounds->reg[i].addr;
+		mask = workarounds->reg[i].mask;
+		value = workarounds->reg[i].value;
 		read = I915_READ(addr);
 		ok = (value & mask) == (read & mask);
 		seq_printf(m, "0x%X: 0x%08X, mask: 0x%08X, read: 0x%08x, status: %s\n",
drivers/gpu/drm/i915/i915_drv.h +8 -1
index 104bd18..83fccc0 100644
@@ -1653,11 +1653,18 @@ struct i915_wa_reg {
 	u32 mask;
 };
 
-#define I915_MAX_WA_REGS 16
+/*
+ * RING_MAX_NONPRIV_SLOTS is per-engine but at this point we are only
+ * allowing it for RCS as we don't foresee any requirement of having
+ * a whitelist for other engines. When it is really required for
+ * other engines then the limit need to be increased.
+ */
+#define I915_MAX_WA_REGS (16 + RING_MAX_NONPRIV_SLOTS)
 
 struct i915_workarounds {
 	struct i915_wa_reg reg[I915_MAX_WA_REGS];
 	u32 count;
+	u32 hw_whitelist_count[I915_NUM_RINGS];
 };
 
 struct i915_virtual_gpu {
drivers/gpu/drm/i915/i915_reg.h +3 -0
index 0a98889..7938814 100644
@@ -1635,6 +1635,9 @@ enum skl_disp_power_wells {
 #define   RING_WAIT		(1<<11) /* gen3+, PRBx_CTL */
 #define   RING_WAIT_SEMAPHORE	(1<<10) /* gen6+ */
 
+#define RING_FORCE_TO_NONPRIV(base, i) _MMIO(((base)+0x4D0) + (i)*4)
+#define   RING_MAX_NONPRIV_SLOTS  12
+
 #define GEN7_TLB_RD_ADDR	_MMIO(0x4700)
 
 #if 0
drivers/gpu/drm/i915/intel_ringbuffer.c +17 -0
index 4060acf..56af736 100644
@@ -787,6 +787,22 @@ static int wa_add(struct drm_i915_private *dev_priv,
 
 #define WA_WRITE(addr, val) WA_REG(addr, 0xffffffff, val)
 
+static int wa_ring_whitelist_reg(struct intel_engine_cs *ring, i915_reg_t reg)
+{
+	struct drm_i915_private *dev_priv = ring->dev->dev_private;
+	struct i915_workarounds *wa = &dev_priv->workarounds;
+	const uint32_t index = wa->hw_whitelist_count[ring->id];
+
+	if (WARN_ON(index >= RING_MAX_NONPRIV_SLOTS))
+		return -EINVAL;
+
+	WA_WRITE(RING_FORCE_TO_NONPRIV(ring->mmio_base, index),
+		 i915_mmio_reg_offset(reg));
+	wa->hw_whitelist_count[ring->id]++;
+
+	return 0;
+}
+
 static int gen8_init_workarounds(struct intel_engine_cs *ring)
 {
 	struct drm_device *dev = ring->dev;
@@ -1115,6 +1131,7 @@ int init_workarounds_ring(struct intel_engine_cs *ring)
 	WARN_ON(ring->id != RCS);
 
 	dev_priv->workarounds.count = 0;
+	dev_priv->workarounds.hw_whitelist_count[RCS] = 0;
 
 	if (IS_BROADWELL(dev))
 		return bdw_init_workarounds(ring);
Signature · 6 lines
1.9.1

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

PATCH v2 2/8 drm/i915/gen9: Add GEN8_CS_CHICKEN1 to HW whitelist

In reply to Arun Siluvery

Required for WaEnablePreemptionGranularityControlByUMD:skl,bxt

---
 drivers/gpu/drm/i915/i915_reg.h         | 2 ++
 drivers/gpu/drm/i915/intel_ringbuffer.c | 6 ++++++
 2 files changed, 8 insertions(+)
drivers/gpu/drm/i915/i915_reg.h +2 -0
index 7938814..511732e 100644
@@ -5998,6 +5998,8 @@ enum skl_disp_power_wells {
 #define FF_SLICE_CS_CHICKEN2			_MMIO(0x20e4)
 #define  GEN9_TSG_BARRIER_ACK_DISABLE		(1<<8)
 
+#define GEN8_CS_CHICKEN1		_MMIO(0x2580)
+
 /* GEN7 chicken */
 #define GEN7_COMMON_SLICE_CHICKEN1		_MMIO(0x7010)
 # define GEN7_CSC1_RHWO_OPT_DISABLE_IN_RCC	((1<<10) | (1<<26))
drivers/gpu/drm/i915/intel_ringbuffer.c +6 -0
index 56af736..47d8767 100644
@@ -908,6 +908,7 @@ static int gen9_init_workarounds(struct intel_engine_cs *ring)
 	struct drm_device *dev = ring->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	uint32_t tmp;
+	int ret;
 
 	/* WaEnableLbsSlaRetryTimerDecrement:skl */
 	I915_WRITE(BDW_SCRATCH1, I915_READ(BDW_SCRATCH1) |
@@ -978,6 +979,11 @@ static int gen9_init_workarounds(struct intel_engine_cs *ring)
 	/* WaDisableSTUnitPowerOptimization:skl,bxt */
 	WA_SET_BIT_MASKED(HALF_SLICE_CHICKEN2, GEN8_ST_PO_DISABLE);
 
+	/* WaEnablePreemptionGranularityControlByUMD:skl,bxt */
+	ret= wa_ring_whitelist_reg(ring, GEN8_CS_CHICKEN1);
+	if (ret)
+		return ret;
+
 	return 0;
 }
 
Signature · 6 lines
1.9.1

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

PATCH v2 3/8 drm/i915/gen9: Add HDC_CHICKEN1 to HW whitelist

In reply to Arun Siluvery

Required for WaAllowUMDToModifyHDCChicken1:skl,bxt

---
 drivers/gpu/drm/i915/i915_reg.h         | 2 ++
 drivers/gpu/drm/i915/intel_ringbuffer.c | 5 +++++
 2 files changed, 7 insertions(+)
drivers/gpu/drm/i915/i915_reg.h +2 -0
index 511732e..ed887cf 100644
@@ -6045,6 +6045,8 @@ enum skl_disp_power_wells {
 #define  HDC_FORCE_NON_COHERENT			(1<<4)
 #define  HDC_BARRIER_PERFORMANCE_DISABLE	(1<<10)
 
+#define GEN8_HDC_CHICKEN1			_MMIO(0x7304)
+
 /* GEN9 chicken */
 #define SLICE_ECO_CHICKEN0			_MMIO(0x7308)
 #define   PIXEL_MASK_CAMMING_DISABLE		(1 << 14)
drivers/gpu/drm/i915/intel_ringbuffer.c +5 -0
index 47d8767..fea632f 100644
@@ -984,6 +984,11 @@ static int gen9_init_workarounds(struct intel_engine_cs *ring)
 	if (ret)
 		return ret;
 
+	/* WaAllowUMDToModifyHDCChicken1:skl,bxt */
+	ret = wa_ring_whitelist_reg(ring, GEN8_HDC_CHICKEN1);
+	if (ret)
+		return ret;
+
 	return 0;
 }
 
Signature · 6 lines
1.9.1

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

PATCH v2 4/8 drm/i915/bxt: Add GEN9_CS_DEBUG_MODE1 to HW whitelist

In reply to Arun Siluvery

Required for, WaDisableObjectLevelPreemptionForTrifanOrPolygon:bxt WaDisableObjectLevelPreemptionForInstancedDraw:bxt WaDisableObjectLevelPreemtionForInstanceId:bxt

According to WA database these are only applicable for BXT:A0 but since A0 and A1 shares the same GT these are extended for A1 as well.

These are also required for SKL until B0 but not adding them because they are pre-production steppings.

v2: use lower case in register defines (Nick)

---
 drivers/gpu/drm/i915/i915_reg.h         | 1 +
 drivers/gpu/drm/i915/intel_ringbuffer.c | 9 +++++++++
 2 files changed, 10 insertions(+)
drivers/gpu/drm/i915/i915_reg.h +1 -0
index ed887cf..c51e7e9 100644
@@ -5998,6 +5998,7 @@ enum skl_disp_power_wells {
 #define FF_SLICE_CS_CHICKEN2			_MMIO(0x20e4)
 #define  GEN9_TSG_BARRIER_ACK_DISABLE		(1<<8)
 
+#define GEN9_CS_DEBUG_MODE1		_MMIO(0x20ec)
 #define GEN8_CS_CHICKEN1		_MMIO(0x2580)
 
 /* GEN7 chicken */
drivers/gpu/drm/i915/intel_ringbuffer.c +9 -0
index fea632f..72e89b6 100644
@@ -1131,6 +1131,15 @@ static int bxt_init_workarounds(struct intel_engine_cs *ring)
 			GEN7_SBE_SS_CACHE_DISPATCH_PORT_SHARING_DISABLE);
 	}
 
+	/* WaDisableObjectLevelPreemptionForTrifanOrPolygon:bxt */
+	/* WaDisableObjectLevelPreemptionForInstancedDraw:bxt */
+	/* WaDisableObjectLevelPreemtionForInstanceId:bxt */
+	if (IS_BXT_REVID(dev, 0, BXT_REVID_A1)) {
+		ret = wa_ring_whitelist_reg(ring, GEN9_CS_DEBUG_MODE1);
+		if (ret)
+			return ret;
+	}
+
 	return 0;
 }
 
Signature · 6 lines
1.9.1

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Reviewed-by on v2 4/8
#5 lore raw

Re: [PATCH v2 4/8] drm/i915/bxt: Add GEN9_CS_DEBUG_MODE1 to HW whitelist

Quoted 8 lines · Arun Siluvery

Required for, WaDisableObjectLevelPreemptionForTrifanOrPolygon:bxt WaDisableObjectLevelPreemptionForInstancedDraw:bxt WaDisableObjectLevelPreemtionForInstanceId:bxt

According to WA database these are only applicable for BXT:A0 but since A0 and A1 shares the same GT these are extended for A1 as well.

Quoted 6 lines

These are also required for SKL until B0 but not adding them because they are pre-production steppings.

v2: use lower case in register defines (Nick)

Quoted 28 lines
---
  drivers/gpu/drm/i915/i915_reg.h         | 1 +
  drivers/gpu/drm/i915/intel_ringbuffer.c | 9 +++++++++
  2 files changed, 10 insertions(+)
drivers/gpu/drm/i915/i915_reg.h +1 -0
index ed887cf..c51e7e9 100644
@@ -5998,6 +5998,7 @@ enum skl_disp_power_wells {
  #define FF_SLICE_CS_CHICKEN2			_MMIO(0x20e4)
  #define  GEN9_TSG_BARRIER_ACK_DISABLE		(1<<8)
 
+#define GEN9_CS_DEBUG_MODE1		_MMIO(0x20ec)
  #define GEN8_CS_CHICKEN1		_MMIO(0x2580)
 
  /* GEN7 chicken */
drivers/gpu/drm/i915/intel_ringbuffer.c +2 -0
index fea632f..72e89b6 100644
@@ -1131,6 +1131,15 @@ static int bxt_init_workarounds(struct intel_engine_cs *ring)
  			GEN7_SBE_SS_CACHE_DISPATCH_PORT_SHARING_DISABLE);
  	}
 
+	/* WaDisableObjectLevelPreemptionForTrifanOrPolygon:bxt */
+	/* WaDisableObjectLevelPreemptionForInstancedDraw:bxt */

drivers/gpu/drm/i915/intel_ringbuffer.c static int bxt_init_workarounds(struct intel_engine_cs *ring) · line 1145

Quoted 10 lines
drivers/gpu/drm/i915/intel_ringbuffer.c +7 -0
+	/* WaDisableObjectLevelPreemtionForInstanceId:bxt */
+	if (IS_BXT_REVID(dev, 0, BXT_REVID_A1)) {
+		ret = wa_ring_whitelist_reg(ring, GEN9_CS_DEBUG_MODE1);
+		if (ret)
+			return ret;
+	}
+
  	return 0;
  }
 

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

#6 lore raw

PATCH v2 5/8 drm/i915/bxt: Add GEN8_L3SQCREG4 to HW whitelist

In reply to Arun Siluvery

Required for WaDisableLSQCROPERFforOCL:bxt

According to WA database these are only applicable for BXT:A0 but since A0 and A1 shares the same GT these are extended for A1 as well.

---
 drivers/gpu/drm/i915/intel_ringbuffer.c | 5 +++++
 1 file changed, 5 insertions(+)
drivers/gpu/drm/i915/intel_ringbuffer.c +5 -0
index 72e89b6..1decaf1 100644
@@ -1134,10 +1134,15 @@ static int bxt_init_workarounds(struct intel_engine_cs *ring)
 	/* WaDisableObjectLevelPreemptionForTrifanOrPolygon:bxt */
 	/* WaDisableObjectLevelPreemptionForInstancedDraw:bxt */
 	/* WaDisableObjectLevelPreemtionForInstanceId:bxt */
+	/* WaDisableLSQCROPERFforOCL:bxt */
 	if (IS_BXT_REVID(dev, 0, BXT_REVID_A1)) {
 		ret = wa_ring_whitelist_reg(ring, GEN9_CS_DEBUG_MODE1);
 		if (ret)
 			return ret;
+
+		ret = wa_ring_whitelist_reg(ring, GEN8_L3SQCREG4);
+		if (ret)
+			return ret;
 	}
 
 	return 0;
Signature · 6 lines
1.9.1

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

PATCH v2 6/8 drm/i915/skl: Add GEN8_L3SQCREG4 to HW whitelist

In reply to Arun Siluvery

Required for WaDisableLSQCROPERFforOCL:skl

---
 drivers/gpu/drm/i915/intel_ringbuffer.c | 5 +++++
 1 file changed, 5 insertions(+)
drivers/gpu/drm/i915/intel_ringbuffer.c +5 -0
index 1decaf1..ce64519 100644
@@ -1096,6 +1096,11 @@ static int skl_init_workarounds(struct intel_engine_cs *ring)
 			GEN7_HALF_SLICE_CHICKEN1,
 			GEN7_SBE_SS_CACHE_DISPATCH_PORT_SHARING_DISABLE);
 
+	/* WaDisableLSQCROPERFforOCL:skl */
+	ret = wa_ring_whitelist_reg(ring, GEN8_L3SQCREG4);
+	if (ret)
+		return ret;
+
 	return skl_tune_iz_hashing(ring);
 }
 
Signature · 6 lines
1.9.1

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

PATCH v2 7/8 drm/i915/skl: Enable Per context Preemption granularity control

In reply to Arun Siluvery

Per context preemption granularity control is only available from SKL:E0+

---
 drivers/gpu/drm/i915/i915_reg.h         |  3 +++
 drivers/gpu/drm/i915/intel_ringbuffer.c | 10 ++++++++++
 2 files changed, 13 insertions(+)
drivers/gpu/drm/i915/i915_reg.h +3 -0
index c51e7e9..65e32a3 100644
@@ -5995,6 +5995,9 @@ enum skl_disp_power_wells {
 #define SKL_DFSM_CDCLK_LIMIT_450	(2 << 23)
 #define SKL_DFSM_CDCLK_LIMIT_337_5	(3 << 23)
 
+#define GEN7_FF_SLICE_CS_CHICKEN1	_MMIO(0x20e0)
+#define   GEN9_FFSC_PERCTX_PREEMPT_CTRL	(1<<14)
+
 #define FF_SLICE_CS_CHICKEN2			_MMIO(0x20e4)
 #define  GEN9_TSG_BARRIER_ACK_DISABLE		(1<<8)
 
drivers/gpu/drm/i915/intel_ringbuffer.c +10 -0
index ce64519..e91fb70 100644
@@ -1044,6 +1044,16 @@ static int skl_init_workarounds(struct intel_engine_cs *ring)
 	if (ret)
 		return ret;
 
+	/*
+	 * Actual WA is to disable percontext preemption granularity control
+	 * until D0 which is the default case so this is equivalent to
+	 * !WaDisablePerCtxtPreemptionGranularityControl:skl
+	 */
+	if (IS_SKL_REVID(dev, SKL_REVID_E0, REVID_FOREVER)) {
+		I915_WRITE(GEN7_FF_SLICE_CS_CHICKEN1,
+			   _MASKED_BIT_ENABLE(GEN9_FFSC_PERCTX_PREEMPT_CTRL));
+	}
+
 	if (IS_SKL_REVID(dev, 0, SKL_REVID_D0)) {
 		/* WaDisableChickenBitTSGBarrierAckForFFSliceCS:skl */
 		I915_WRITE(FF_SLICE_CS_CHICKEN2,
Signature · 6 lines
1.9.1

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Reviewed-by on v2 7/8
#9 lore raw

Re: [PATCH v2 7/8] drm/i915/skl: Enable Per context Preemption granularity control

Quoted 4 lines · Arun Siluvery

Per context preemption granularity control is only available from SKL:E0+

Quoted 31 lines
---
  drivers/gpu/drm/i915/i915_reg.h         |  3 +++
  drivers/gpu/drm/i915/intel_ringbuffer.c | 10 ++++++++++
  2 files changed, 13 insertions(+)
drivers/gpu/drm/i915/i915_reg.h +3 -0
index c51e7e9..65e32a3 100644
@@ -5995,6 +5995,9 @@ enum skl_disp_power_wells {
  #define SKL_DFSM_CDCLK_LIMIT_450	(2 << 23)
  #define SKL_DFSM_CDCLK_LIMIT_337_5	(3 << 23)
 
+#define GEN7_FF_SLICE_CS_CHICKEN1	_MMIO(0x20e0)
+#define   GEN9_FFSC_PERCTX_PREEMPT_CTRL	(1<<14)
+
  #define FF_SLICE_CS_CHICKEN2			_MMIO(0x20e4)
  #define  GEN9_TSG_BARRIER_ACK_DISABLE		(1<<8)
 
drivers/gpu/drm/i915/intel_ringbuffer.c +3 -0
index ce64519..e91fb70 100644
@@ -1044,6 +1044,16 @@ static int skl_init_workarounds(struct intel_engine_cs *ring)
  	if (ret)
  		return ret;
 
+	/*
+	 * Actual WA is to disable percontext preemption granularity control
+	 * until D0 which is the default case so this is equivalent to

drivers/gpu/drm/i915/intel_ringbuffer.c static int skl_init_workarounds(struct intel_engine_cs *ring) · line 1059

Quoted 10 lines
drivers/gpu/drm/i915/intel_ringbuffer.c +7 -0
+	 * !WaDisablePerCtxtPreemptionGranularityControl:skl
+	 */
+	if (IS_SKL_REVID(dev, SKL_REVID_E0, REVID_FOREVER)) {
+		I915_WRITE(GEN7_FF_SLICE_CS_CHICKEN1,
+			   _MASKED_BIT_ENABLE(GEN9_FFSC_PERCTX_PREEMPT_CTRL));
+	}
+
  	if (IS_SKL_REVID(dev, 0, SKL_REVID_D0)) {
  		/* WaDisableChickenBitTSGBarrierAckForFFSliceCS:skl */
  		I915_WRITE(FF_SLICE_CS_CHICKEN2,

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

#10 lore raw

PATCH v2 8/8 drm/i915/gen9: Add WaOCLCoherentLineFlush

In reply to Arun Siluvery

This is mainly required for preemption.

---
 drivers/gpu/drm/i915/intel_ringbuffer.c | 4 ++++
 1 file changed, 4 insertions(+)
drivers/gpu/drm/i915/intel_ringbuffer.c +4 -0
index e91fb70..f26f274 100644
@@ -979,6 +979,10 @@ static int gen9_init_workarounds(struct intel_engine_cs *ring)
 	/* WaDisableSTUnitPowerOptimization:skl,bxt */
 	WA_SET_BIT_MASKED(HALF_SLICE_CHICKEN2, GEN8_ST_PO_DISABLE);
 
+	/* WaOCLCoherentLineFlush:skl,bxt */
+	I915_WRITE(GEN8_L3SQCREG4, (I915_READ(GEN8_L3SQCREG4) |
+				    GEN8_LQSC_FLUSH_COHERENT_LINES));
+
 	/* WaEnablePreemptionGranularityControlByUMD:skl,bxt */
 	ret= wa_ring_whitelist_reg(ring, GEN8_CS_CHICKEN1);
 	if (ret)
Signature · 6 lines
1.9.1

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Reviewed-by on v2 8/8
#11 lore raw

Re: [PATCH v2 8/8] drm/i915/gen9: Add WaOCLCoherentLineFlush

Quoted 4 lines · Arun Siluvery

This is mainly required for preemption.

Quoted 10 lines
---
  drivers/gpu/drm/i915/intel_ringbuffer.c | 4 ++++
  1 file changed, 4 insertions(+)
drivers/gpu/drm/i915/intel_ringbuffer.c +0 -0
index e91fb70..f26f274 100644

drivers/gpu/drm/i915/intel_ringbuffer.c static int gen9_init_workarounds(struct intel_engine_cs *ring) · line 988

Quoted 10 lines
drivers/gpu/drm/i915/intel_ringbuffer.c +4 -0
@@ -979,6 +979,10 @@ static int gen9_init_workarounds(struct intel_engine_cs *ring)
  	/* WaDisableSTUnitPowerOptimization:skl,bxt */
  	WA_SET_BIT_MASKED(HALF_SLICE_CHICKEN2, GEN8_ST_PO_DISABLE);
 
+	/* WaOCLCoherentLineFlush:skl,bxt */
+	I915_WRITE(GEN8_L3SQCREG4, (I915_READ(GEN8_L3SQCREG4) |
+				    GEN8_LQSC_FLUSH_COHERENT_LINES));
+
  	/* WaEnablePreemptionGranularityControlByUMD:skl,bxt */
  	ret= wa_ring_whitelist_reg(ring, GEN8_CS_CHICKEN1);
  	if (ret)

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

#12 lore raw

✗ Fi.CI.BAT: warning for Gen9 HW whitelist and Preemption WA patches (rev2)

In reply to Arun Siluvery
== Summary ==

== Summary ==

Built on 8fe9e785ae04fa7c37f7935cff12d62e38054b60 drm-intel-nightly: 2016y-01m-21d-11h-02m-42s UTC integration manifest

Test kms_flip:

        Subgroup basic-flip-vs-dpms:
                pass       -> DMESG-WARN (ilk-hp8440p)
bdw-nuci7        total:140  pass:131  dwarn:0   dfail:0   fail:0   skip:9  
bdw-ultra        total:143  pass:137  dwarn:0   dfail:0   fail:0   skip:6  
bsw-nuc-2        total:143  pass:119  dwarn:0   dfail:0   fail:0   skip:24 
byt-nuc          total:143  pass:128  dwarn:0   dfail:0   fail:0   skip:15 
hsw-brixbox      total:143  pass:136  dwarn:0   dfail:0   fail:0   skip:7  
hsw-gt2          total:143  pass:139  dwarn:0   dfail:0   fail:0   skip:4  
ilk-hp8440p      total:143  pass:103  dwarn:2   dfail:0   fail:0   skip:38 
ivb-t430s        total:143  pass:137  dwarn:0   dfail:0   fail:0   skip:6  
skl-i5k-2        total:143  pass:134  dwarn:1   dfail:0   fail:0   skip:8  
skl-i7k-2        total:143  pass:134  dwarn:1   dfail:0   fail:0   skip:8  
snb-dellxps      total:143  pass:129  dwarn:0   dfail:0   fail:0   skip:14 
snb-x220t        total:143  pass:129  dwarn:0   dfail:0   fail:1   skip:13 

Results at /archive/results/CI_IGT_test/Patchwork_1238/

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

#13 lore raw

Re: ✗ Fi.CI.BAT: warning for Gen9 HW whitelist and Preemption WA patches (rev2)

Quoted 21 lines · Patchwork

== Summary ==

Built on 8fe9e785ae04fa7c37f7935cff12d62e38054b60 drm-intel-nightly: 2016y-01m-21d-11h-02m-42s UTC integration manifest

Test kms_flip:

         Subgroup basic-flip-vs-dpms:
                 pass       -> DMESG-WARN (ilk-hp8440p)
bdw-nuci7        total:140  pass:131  dwarn:0   dfail:0   fail:0   skip:9
bdw-ultra        total:143  pass:137  dwarn:0   dfail:0   fail:0   skip:6
bsw-nuc-2        total:143  pass:119  dwarn:0   dfail:0   fail:0   skip:24
byt-nuc          total:143  pass:128  dwarn:0   dfail:0   fail:0   skip:15
hsw-brixbox      total:143  pass:136  dwarn:0   dfail:0   fail:0   skip:7
hsw-gt2          total:143  pass:139  dwarn:0   dfail:0   fail:0   skip:4
ilk-hp8440p      total:143  pass:103  dwarn:2   dfail:0   fail:0   skip:38
ivb-t430s        total:143  pass:137  dwarn:0   dfail:0   fail:0   skip:6
skl-i5k-2        total:143  pass:134  dwarn:1   dfail:0   fail:0   skip:8
skl-i7k-2        total:143  pass:134  dwarn:1   dfail:0   fail:0   skip:8
snb-dellxps      total:143  pass:129  dwarn:0   dfail:0   fail:0   skip:14
snb-x220t        total:143  pass:129  dwarn:0   dfail:0   fail:1   skip:13
Quoted 2 lines

Results at /archive/results/CI_IGT_test/Patchwork_1238/

Known issue, this test has been failing on ilk-hp8440p even before,

Bug 93787 - [BAT ILK] sporadic fifo underruns in igt@kms_flip@basic-flip-vs-* on ilk-hp8440p
https://bugs.freedesktop.org/show_bug.cgi?id=93787

regards
Arun

Quoted 1 line

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

#14 lore raw
In reply to Arun Siluvery
Quoted 11 lines · Arun Siluvery

Resending all patches as told by Daniel because I didn't use correct message-id while replying updated version of Patch1 before which means patchwork won't pickup and we won't have CI results. Previous updates regarding Patch1 are available at https://patchwork.freedesktop.org/patch/70527/

Some of the WA patches are now reviewed so I added r-b tags for them.

Reviewed: 1, 2, 3, 5, 6
Updated: 4 (to address review comment)
To be reviewed: 4, 7, 8

Quoted 9 lines

Arun Siluvery (8):
drm/i915/gen9: Add framework to whitelist specific GPU registers drm/i915/gen9: Add GEN8_CS_CHICKEN1 to HW whitelist drm/i915/gen9: Add HDC_CHICKEN1 to HW whitelist drm/i915/bxt: Add GEN9_CS_DEBUG_MODE1 to HW whitelist drm/i915/bxt: Add GEN8_L3SQCREG4 to HW whitelist drm/i915/skl: Add GEN8_L3SQCREG4 to HW whitelist drm/i915/skl: Enable Per context Preemption granularity control drm/i915/gen9: Add WaOCLCoherentLineFlush

With bland changelogs, it is impossible to assess the severity of the issues being addressed. How many of these do we need to apply to kernel release to address potential user impact? Which do backporters need? Are they critical stability fixes or power tuning or enabling for future features?
-Chris

Signature · 5 lines
Chris Wilson, Intel Open Source Technology Centre
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
#15 lore raw
Quoted 22 lines · Chris Wilson
Quoted 11 lines · Arun Siluvery

Resending all patches as told by Daniel because I didn't use correct message-id while replying updated version of Patch1 before which means patchwork won't pickup and we won't have CI results. Previous updates regarding Patch1 are available at https://patchwork.freedesktop.org/patch/70527/

Some of the WA patches are now reviewed so I added r-b tags for them.

Reviewed: 1, 2, 3, 5, 6
Updated: 4 (to address review comment)
To be reviewed: 4, 7, 8

Quoted 9 lines

Arun Siluvery (8):
drm/i915/gen9: Add framework to whitelist specific GPU registers drm/i915/gen9: Add GEN8_CS_CHICKEN1 to HW whitelist drm/i915/gen9: Add HDC_CHICKEN1 to HW whitelist drm/i915/bxt: Add GEN9_CS_DEBUG_MODE1 to HW whitelist drm/i915/bxt: Add GEN8_L3SQCREG4 to HW whitelist drm/i915/skl: Add GEN8_L3SQCREG4 to HW whitelist drm/i915/skl: Enable Per context Preemption granularity control drm/i915/gen9: Add WaOCLCoherentLineFlush

Quoted 7 lines

With bland changelogs, it is impossible to assess the severity of the issues being addressed. How many of these do we need to apply to kernel release to address potential user impact? Which do backporters need? Are they critical stability fixes or power tuning or enabling for future features?
-Chris

All of these changes are required for Preemption so they affect Gen9 only. The whitelist changes are really to support for preemption WA that are going to be part of UMD. The WA data is sketchy and for some of them hsd links are not available so I don't really know the details.

regards
Arun

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

#16 lore raw
Quoted 30 lines · Arun Siluvery
Quoted 22 lines · Chris Wilson
Quoted 11 lines · Arun Siluvery

Resending all patches as told by Daniel because I didn't use correct message-id while replying updated version of Patch1 before which means patchwork won't pickup and we won't have CI results. Previous updates regarding Patch1 are available at https://patchwork.freedesktop.org/patch/70527/

Some of the WA patches are now reviewed so I added r-b tags for them.

Reviewed: 1, 2, 3, 5, 6
Updated: 4 (to address review comment)
To be reviewed: 4, 7, 8

Quoted 9 lines

Arun Siluvery (8):
drm/i915/gen9: Add framework to whitelist specific GPU registers drm/i915/gen9: Add GEN8_CS_CHICKEN1 to HW whitelist drm/i915/gen9: Add HDC_CHICKEN1 to HW whitelist drm/i915/bxt: Add GEN9_CS_DEBUG_MODE1 to HW whitelist drm/i915/bxt: Add GEN8_L3SQCREG4 to HW whitelist drm/i915/skl: Add GEN8_L3SQCREG4 to HW whitelist drm/i915/skl: Enable Per context Preemption granularity control drm/i915/gen9: Add WaOCLCoherentLineFlush

Quoted 7 lines

With bland changelogs, it is impossible to assess the severity of the issues being addressed. How many of these do we need to apply to kernel release to address potential user impact? Which do backporters need? Are they critical stability fixes or power tuning or enabling for future features?
-Chris

Quoted 5 lines

All of these changes are required for Preemption so they affect Gen9 only. The whitelist changes are really to support for preemption WA that are going to be part of UMD. The WA data is sketchy and for some of them hsd links are not available so I don't really know the details.

Are you happy with adding "Required for future enabling of pre-emptive command execution" to each? It just helps clarify the purpose of the patch, and as I said helps answer the question of "do I need to backport this to stable kernels".
-Chris

Signature · 5 lines
Chris Wilson, Intel Open Source Technology Centre
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
#17 lore raw
Quoted 37 lines · Chris Wilson
Quoted 30 lines · Arun Siluvery
Quoted 22 lines · Chris Wilson
Quoted 11 lines · Arun Siluvery

Resending all patches as told by Daniel because I didn't use correct message-id while replying updated version of Patch1 before which means patchwork won't pickup and we won't have CI results. Previous updates regarding Patch1 are available at https://patchwork.freedesktop.org/patch/70527/

Some of the WA patches are now reviewed so I added r-b tags for them.

Reviewed: 1, 2, 3, 5, 6
Updated: 4 (to address review comment)
To be reviewed: 4, 7, 8

Quoted 9 lines

Arun Siluvery (8):
drm/i915/gen9: Add framework to whitelist specific GPU registers drm/i915/gen9: Add GEN8_CS_CHICKEN1 to HW whitelist drm/i915/gen9: Add HDC_CHICKEN1 to HW whitelist drm/i915/bxt: Add GEN9_CS_DEBUG_MODE1 to HW whitelist drm/i915/bxt: Add GEN8_L3SQCREG4 to HW whitelist drm/i915/skl: Add GEN8_L3SQCREG4 to HW whitelist drm/i915/skl: Enable Per context Preemption granularity control drm/i915/gen9: Add WaOCLCoherentLineFlush

Quoted 7 lines

With bland changelogs, it is impossible to assess the severity of the issues being addressed. How many of these do we need to apply to kernel release to address potential user impact? Which do backporters need? Are they critical stability fixes or power tuning or enabling for future features?
-Chris

Quoted 5 lines

All of these changes are required for Preemption so they affect Gen9 only. The whitelist changes are really to support for preemption WA that are going to be part of UMD. The WA data is sketchy and for some of them hsd links are not available so I don't really know the details.

Quoted 4 lines

Are you happy with adding "Required for future enabling of pre-emptive command execution" to each? It just helps clarify the purpose of the patch, and as I said helps answer the question of "do I need to backport this to stable kernels".

yes it helps, will add it and resend.

regards
Arun

Quoted 2 lines

-Chris

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

#18 lore raw
In reply to Arun Siluvery
Quoted 30 lines · Arun Siluvery
Quoted 22 lines · Chris Wilson
Quoted 11 lines · Arun Siluvery

Resending all patches as told by Daniel because I didn't use correct message-id while replying updated version of Patch1 before which means patchwork won't pickup and we won't have CI results. Previous updates regarding Patch1 are available at https://patchwork.freedesktop.org/patch/70527/

Some of the WA patches are now reviewed so I added r-b tags for them.

Reviewed: 1, 2, 3, 5, 6
Updated: 4 (to address review comment)
To be reviewed: 4, 7, 8

Quoted 9 lines

Arun Siluvery (8):
drm/i915/gen9: Add framework to whitelist specific GPU registers drm/i915/gen9: Add GEN8_CS_CHICKEN1 to HW whitelist drm/i915/gen9: Add HDC_CHICKEN1 to HW whitelist drm/i915/bxt: Add GEN9_CS_DEBUG_MODE1 to HW whitelist drm/i915/bxt: Add GEN8_L3SQCREG4 to HW whitelist drm/i915/skl: Add GEN8_L3SQCREG4 to HW whitelist drm/i915/skl: Enable Per context Preemption granularity control drm/i915/gen9: Add WaOCLCoherentLineFlush

Quoted 7 lines

With bland changelogs, it is impossible to assess the severity of the issues being addressed. How many of these do we need to apply to kernel release to address potential user impact? Which do backporters need? Are they critical stability fixes or power tuning or enabling for future features?
-Chris

Quoted 4 lines

All of these changes are required for Preemption so they affect Gen9 only. The whitelist changes are really to support for preemption WA that are going to be part of UMD. The WA data is sketchy and for some of them hsd links are not available so I don't really know the details.

Oh, we need pre-empt wa in UMDs? This is going to be a lot of fun to make it work in a backwards-compat fashion ... Have you started digging into what would be needed for mesa&co?
-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
#19 lore raw
Quoted 31 lines · Daniel Vetter
Quoted 30 lines · Arun Siluvery
Quoted 29 lines · Chris Wilson
Quoted 11 lines · Arun Siluvery

Resending all patches as told by Daniel because I didn't use correct message-id while replying updated version of Patch1 before which means patchwork won't pickup and we won't have CI results. Previous updates regarding Patch1 are available at https://patchwork.freedesktop.org/patch/70527/

Some of the WA patches are now reviewed so I added r-b tags for them.

Reviewed: 1, 2, 3, 5, 6
Updated: 4 (to address review comment)
To be reviewed: 4, 7, 8

Quoted 9 lines

Arun Siluvery (8):
drm/i915/gen9: Add framework to whitelist specific GPU registers drm/i915/gen9: Add GEN8_CS_CHICKEN1 to HW whitelist drm/i915/gen9: Add HDC_CHICKEN1 to HW whitelist drm/i915/bxt: Add GEN9_CS_DEBUG_MODE1 to HW whitelist drm/i915/bxt: Add GEN8_L3SQCREG4 to HW whitelist drm/i915/skl: Add GEN8_L3SQCREG4 to HW whitelist drm/i915/skl: Enable Per context Preemption granularity control drm/i915/gen9: Add WaOCLCoherentLineFlush

With bland changelogs, it is impossible to assess the severity of the issues being addressed. How many of these do we need to apply to kernel release to address potential user impact? Which do backporters need? Are they critical stability fixes or power tuning or enabling for future features?
-Chris

Quoted 8 lines
Quoted 4 lines

All of these changes are required for Preemption so they affect Gen9 only. The whitelist changes are really to support for preemption WA that are going to be part of UMD. The WA data is sketchy and for some of them hsd links are not available so I don't really know the details.

Oh, we need pre-empt wa in UMDs? This is going to be a lot of fun to make it work in a backwards-compat fashion ... Have you started digging into what would be needed for mesa&co?

Yes it is, userspace mainly selects the type of preemption and anything HW allows them in the whitelisted registers but I don't know what all would be needed.

There is one WA where something need to be cleared in SIP but neither mesa nor beignet uses that at the moment.

regards
Arun

Quoted 2 lines

-Daniel

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