[PATCH 0/5] Add GuC ADS (Addition Data Structure)

[PATCH 0/5] Add GuC ADS (Addition Data Structure)

[yu.dai]

From: Alex Dai <yu.dai@intel.com>

The GuC firmware uses this for various purposes. The ADS itself is a chunk of
memory created by driver to share with GuC. This series creates the GuC ADS
object and setup some basic settings for it.

Alex Dai (4):
  drm/i915/guc: Add GuC ADS (Addition Data Structure) - allocation
  drm/i915/guc: Add GuC ADS - scheduler policies
  drm/i915/guc: Add GuC ADS - MMIO reg state
  drm/i915/guc: Add GuC ADS - enabling ADS

Dave Gordon (1):
  drm/i915/guc: Expose (intel)_lr_context_size()

 drivers/gpu/drm/i915/i915_guc_reg.h        |   1 +
 drivers/gpu/drm/i915/i915_guc_submission.c |  99 +++++++++++++++++++++++++
 drivers/gpu/drm/i915/intel_guc.h           |   2 +
 drivers/gpu/drm/i915/intel_guc_fwif.h      | 113 ++++++++++++++++++++++++++++-
 drivers/gpu/drm/i915/intel_guc_loader.c    |   7 ++
 drivers/gpu/drm/i915/intel_lrc.c           |   4 +-
 drivers/gpu/drm/i915/intel_lrc.h           |   1 +
 7 files changed, 224 insertions(+), 3 deletions(-)

-- 
2.5.0

_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

[PATCH 1/5] drm/i915/guc: Expose (intel)_lr_context_size()

[yu.dai]

From: Dave Gordon <david.s.gordon@intel.com>

The GuC code needs to know the size of a logical context, so we
expose get_lr_context_size(), renaming it intel_lr_context__size()
to fit the naming conventions for nonstatic functions.

For: VIZ-2021
Signed-off-by: Dave Gordon <david.s.gordon@intel.com>
Signed-off-by: Alex Dai <yu.dai@intel.com>
---
 drivers/gpu/drm/i915/intel_lrc.c | 4 ++--
 drivers/gpu/drm/i915/intel_lrc.h | 1 +
 2 files changed, 3 insertions(+), 2 deletions(-)

diff --git a/drivers/gpu/drm/i915/intel_lrc.c b/drivers/gpu/drm/i915/intel_lrc.c
index cd232d2..bbdcd5d 100644
--- a/drivers/gpu/drm/i915/intel_lrc.c
+++ b/drivers/gpu/drm/i915/intel_lrc.c
@@ -2482,7 +2482,7 @@ void intel_lr_context_free(struct intel_context *ctx)
 	}
 }
 
-static uint32_t get_lr_context_size(struct intel_engine_cs *ring)
+uint32_t intel_lr_context_size(struct intel_engine_cs *ring)
 {
 	int ret = 0;
 
@@ -2550,7 +2550,7 @@ int intel_lr_context_deferred_alloc(struct intel_context *ctx,
 	WARN_ON(ctx->legacy_hw_ctx.rcs_state != NULL);
 	WARN_ON(ctx->engine[ring->id].state);
 
-	context_size = round_up(get_lr_context_size(ring), 4096);
+	context_size = round_up(intel_lr_context_size(ring), 4096);
 
 	/* One extra page as the sharing data between driver and GuC */
 	context_size += PAGE_SIZE * LRC_PPHWSP_PN;
diff --git a/drivers/gpu/drm/i915/intel_lrc.h b/drivers/gpu/drm/i915/intel_lrc.h
index fd1b6b4..9b2b9bd 100644
--- a/drivers/gpu/drm/i915/intel_lrc.h
+++ b/drivers/gpu/drm/i915/intel_lrc.h
@@ -84,6 +84,7 @@ static inline void intel_logical_ring_emit_reg(struct intel_ringbuffer *ringbuf,
 #define LRC_STATE_PN	(LRC_PPHWSP_PN + 1)
 
 void intel_lr_context_free(struct intel_context *ctx);
+uint32_t intel_lr_context_size(struct intel_engine_cs *ring);
 int intel_lr_context_deferred_alloc(struct intel_context *ctx,
 				    struct intel_engine_cs *ring);
 void intel_lr_context_unpin(struct drm_i915_gem_request *req);
-- 
2.5.0

_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

[PATCH 2/5] drm/i915/guc: Add GuC ADS (Addition Data Structure) - allocation

[yu.dai]

From: Alex Dai <yu.dai@intel.com>

The GuC firmware uses this for various purposes. The ADS itself is
a chunk of memory created by driver to share with GuC. Its members
are usually addresses telling where GuC to access them, including
things like scheduler policies, register list that will be saved
and restored during reset etc.

This is the first patch of a series to enable GuC ADS. For now, we
only create the ADS obj whilst keep it disabled.

Signed-off-by: Alex Dai <yu.dai@intel.com>
---
 drivers/gpu/drm/i915/i915_guc_reg.h        |  1 +
 drivers/gpu/drm/i915/i915_guc_submission.c | 50 ++++++++++++++++++++++++++++++
 drivers/gpu/drm/i915/intel_guc.h           |  2 ++
 drivers/gpu/drm/i915/intel_guc_fwif.h      | 31 +++++++++++++++++-
 4 files changed, 83 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/i915/i915_guc_reg.h b/drivers/gpu/drm/i915/i915_guc_reg.h
index 90a84b4..8d27c09 100644
--- a/drivers/gpu/drm/i915/i915_guc_reg.h
+++ b/drivers/gpu/drm/i915/i915_guc_reg.h
@@ -40,6 +40,7 @@
 #define   GS_MIA_CORE_STATE		  (1 << GS_MIA_SHIFT)
 
 #define SOFT_SCRATCH(n)			_MMIO(0xc180 + (n) * 4)
+#define SOFT_SCRATCH_COUNT		16
 
 #define UOS_RSA_SCRATCH(i)		_MMIO(0xc200 + (i) * 4)
 #define   UOS_RSA_SCRATCH_MAX_COUNT	  64
diff --git a/drivers/gpu/drm/i915/i915_guc_submission.c b/drivers/gpu/drm/i915/i915_guc_submission.c
index 7554d16..66d85c3 100644
--- a/drivers/gpu/drm/i915/i915_guc_submission.c
+++ b/drivers/gpu/drm/i915/i915_guc_submission.c
@@ -842,6 +842,51 @@ static void guc_create_log(struct intel_guc *guc)
 	guc->log_flags = (offset << GUC_LOG_BUF_ADDR_SHIFT) | flags;
 }
 
+static void guc_create_ads(struct intel_guc *guc)
+{
+	struct drm_i915_private *dev_priv = guc_to_i915(guc);
+	struct drm_i915_gem_object *obj;
+	struct guc_ads *ads;
+	struct intel_engine_cs *ring;
+	struct page *page;
+	u32 size, i;
+
+	/* The ads obj includes the struct itself and buffers passed to GuC */
+	size = sizeof(struct guc_ads);
+
+	obj = guc->ads_obj;
+	if (!obj) {
+		obj = gem_allocate_guc_obj(dev_priv->dev, PAGE_ALIGN(size));
+		if (!obj)
+			return;
+
+		guc->ads_obj = obj;
+	}
+
+	page = i915_gem_object_get_page(obj, 0);
+	ads = kmap_atomic(page);
+	if (!ads) {
+		guc->ads_obj = NULL;
+		gem_release_guc_obj(obj);
+		return;
+	}
+
+	/*
+	 * The GuC requires a "Golden Context" when it reinitialises
+	 * engines after a reset. Here we use the Render ring default
+	 * context, which must already exist and be pinned in the GGTT,
+	 * so its address won't change after we've told the GuC where
+	 * to find it.
+	 */
+	ring = &dev_priv->ring[RCS];
+	ads->golden_context_lrca = ring->status_page.gfx_addr;
+
+	for_each_ring(ring, dev_priv, i)
+		ads->eng_state_size[i] = intel_lr_context_size(ring);
+
+	kunmap_atomic(ads);
+}
+
 /*
  * Set up the memory resources to be shared with the GuC.  At this point,
  * we require just one object that can be mapped through the GGTT.
@@ -868,6 +913,8 @@ int i915_guc_submission_init(struct drm_device *dev)
 
 	guc_create_log(guc);
 
+	guc_create_ads(guc);
+
 	return 0;
 }
 
@@ -906,6 +953,9 @@ void i915_guc_submission_fini(struct drm_device *dev)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_guc *guc = &dev_priv->guc;
 
+	gem_release_guc_obj(dev_priv->guc.ads_obj);
+	guc->ads_obj = NULL;
+
 	gem_release_guc_obj(dev_priv->guc.log_obj);
 	guc->log_obj = NULL;
 
diff --git a/drivers/gpu/drm/i915/intel_guc.h b/drivers/gpu/drm/i915/intel_guc.h
index 5cf555d..5c9f894 100644
--- a/drivers/gpu/drm/i915/intel_guc.h
+++ b/drivers/gpu/drm/i915/intel_guc.h
@@ -89,6 +89,8 @@ struct intel_guc {
 	uint32_t log_flags;
 	struct drm_i915_gem_object *log_obj;
 
+	struct drm_i915_gem_object *ads_obj;
+
 	struct drm_i915_gem_object *ctx_pool_obj;
 	struct ida ctx_ids;
 
diff --git a/drivers/gpu/drm/i915/intel_guc_fwif.h b/drivers/gpu/drm/i915/intel_guc_fwif.h
index eaa50a4..76ecc85 100644
--- a/drivers/gpu/drm/i915/intel_guc_fwif.h
+++ b/drivers/gpu/drm/i915/intel_guc_fwif.h
@@ -81,11 +81,14 @@
 #define GUC_CTL_CTXINFO			0
 #define   GUC_CTL_CTXNUM_IN16_SHIFT	0
 #define   GUC_CTL_BASE_ADDR_SHIFT	12
+
 #define GUC_CTL_ARAT_HIGH		1
 #define GUC_CTL_ARAT_LOW		2
+
 #define GUC_CTL_DEVICE_INFO		3
 #define   GUC_CTL_GTTYPE_SHIFT		0
 #define   GUC_CTL_COREFAMILY_SHIFT	7
+
 #define GUC_CTL_LOG_PARAMS		4
 #define   GUC_LOG_VALID			(1 << 0)
 #define   GUC_LOG_NOTIFY_ON_HALF_FULL	(1 << 1)
@@ -97,9 +100,12 @@
 #define   GUC_LOG_ISR_PAGES		3
 #define   GUC_LOG_ISR_SHIFT		9
 #define   GUC_LOG_BUF_ADDR_SHIFT	12
+
 #define GUC_CTL_PAGE_FAULT_CONTROL	5
+
 #define GUC_CTL_WA			6
 #define   GUC_CTL_WA_UK_BY_DRIVER	(1 << 3)
+
 #define GUC_CTL_FEATURE			7
 #define   GUC_CTL_VCS2_ENABLED		(1 << 0)
 #define   GUC_CTL_KERNEL_SUBMISSIONS	(1 << 1)
@@ -109,6 +115,7 @@
 #define   GUC_CTL_PREEMPTION_LOG	(1 << 5)
 #define   GUC_CTL_ENABLE_SLPC		(1 << 7)
 #define   GUC_CTL_RESET_ON_PREMPT_FAILURE	(1 << 8)
+
 #define GUC_CTL_DEBUG			8
 #define   GUC_LOG_VERBOSITY_SHIFT	0
 #define   GUC_LOG_VERBOSITY_LOW		(0 << GUC_LOG_VERBOSITY_SHIFT)
@@ -118,9 +125,19 @@
 /* Verbosity range-check limits, without the shift */
 #define	  GUC_LOG_VERBOSITY_MIN		0
 #define	  GUC_LOG_VERBOSITY_MAX		3
+#define	  GUC_LOG_VERBOSITY_MASK	0x0000000f
+#define	  GUC_LOG_DESTINATION_MASK	(3 << 4)
+#define   GUC_LOG_DISABLED		(1 << 6)
+#define   GUC_PROFILE_ENABLED		(1 << 7)
+#define   GUC_WQ_TRACK_ENABLED		(1 << 8)
+#define   GUC_ADS_ENABLED		(1 << 9)
+#define   GUC_DEBUG_RESERVED		(1 << 10)
+#define   GUC_ADS_ADDR_SHIFT		11
+#define   GUC_ADS_ADDR_MASK		0xfffff800
+
 #define GUC_CTL_RSRVD			9
 
-#define GUC_CTL_MAX_DWORDS		(GUC_CTL_RSRVD + 1)
+#define GUC_CTL_MAX_DWORDS		(SOFT_SCRATCH_COUNT - 2) /* [1..14] */
 
 /**
  * DOC: GuC Firmware Layout
@@ -299,6 +316,18 @@ struct guc_context_desc {
 #define GUC_POWER_D2		3
 #define GUC_POWER_D3		4
 
+/* GuC Additional Data Struct */
+
+struct guc_ads {
+	u32 reg_state_addr;
+	u32 reg_state_buffer;
+	u32 golden_context_lrca;
+	u32 scheduler_policies;
+	u32 reserved0[3];
+	u32 eng_state_size[I915_NUM_RINGS];
+	u32 reserved2[4];
+} __packed;
+
 /* This Action will be programmed in C180 - SOFT_SCRATCH_O_REG */
 enum host2guc_action {
 	HOST2GUC_ACTION_DEFAULT = 0x0,
-- 
2.5.0

_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

[PATCH 3/5] drm/i915/guc: Add GuC ADS - scheduler policies

[yu.dai]

From: Alex Dai <yu.dai@intel.com>

GuC supports different scheduling policies for its four internal
queues. Currently these have been set to the same default values
as KMD_NORMAL queue.

Particularly POLICY_MAX_NUM_WI is set to 15 to match GuC internal
maximum submit queue numbers to avoid an out-of-space problem.
This value indicates max number of work items allowed to be queued
for one DPC process. A smaller value will let GuC schedule more
frequently while a larger number may increase chances to optimize
cmds (such as collapse cmds from same lrc) with risks that keeps
CS idle.

Signed-off-by: Alex Dai <yu.dai@intel.com>
---
 drivers/gpu/drm/i915/i915_guc_submission.c | 31 +++++++++++++++++++-
 drivers/gpu/drm/i915/intel_guc_fwif.h      | 45 ++++++++++++++++++++++++++++++
 2 files changed, 75 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/i915/i915_guc_submission.c b/drivers/gpu/drm/i915/i915_guc_submission.c
index 66d85c3..a5c555c 100644
--- a/drivers/gpu/drm/i915/i915_guc_submission.c
+++ b/drivers/gpu/drm/i915/i915_guc_submission.c
@@ -842,17 +842,39 @@ static void guc_create_log(struct intel_guc *guc)
 	guc->log_flags = (offset << GUC_LOG_BUF_ADDR_SHIFT) | flags;
 }
 
+static void init_guc_policies(struct guc_policies *policies)
+{
+	struct guc_policy *policy;
+	u32 p, i;
+
+	policies->dpc_promote_time = 500000;
+	policies->max_num_work_items = POLICY_MAX_NUM_WI;
+
+	for (p = 0; p < GUC_CTX_PRIORITY_NUM; p++)
+	for (i = 0; i < I915_NUM_RINGS; i++) {
+		policy = &policies->policy[p][i];
+
+		policy->execution_quantum = 1000000;
+		policy->preemption_time = 500000;
+		policy->fault_time = 250000;
+		policy->policy_flags = 0;
+	}
+
+	policies->is_valid = 1;
+}
+
 static void guc_create_ads(struct intel_guc *guc)
 {
 	struct drm_i915_private *dev_priv = guc_to_i915(guc);
 	struct drm_i915_gem_object *obj;
 	struct guc_ads *ads;
+	struct guc_policies *policies;
 	struct intel_engine_cs *ring;
 	struct page *page;
 	u32 size, i;
 
 	/* The ads obj includes the struct itself and buffers passed to GuC */
-	size = sizeof(struct guc_ads);
+	size = sizeof(struct guc_ads) + sizeof(struct guc_policies);
 
 	obj = guc->ads_obj;
 	if (!obj) {
@@ -884,6 +906,13 @@ static void guc_create_ads(struct intel_guc *guc)
 	for_each_ring(ring, dev_priv, i)
 		ads->eng_state_size[i] = intel_lr_context_size(ring);
 
+	/* GuC scheduling policies */
+	policies = (void *)ads + sizeof(struct guc_ads);
+	init_guc_policies(policies);
+
+	ads->scheduler_policies = i915_gem_obj_ggtt_offset(obj) +
+			sizeof(struct guc_ads);
+
 	kunmap_atomic(ads);
 }
 
diff --git a/drivers/gpu/drm/i915/intel_guc_fwif.h b/drivers/gpu/drm/i915/intel_guc_fwif.h
index 76ecc85..0cc17c7 100644
--- a/drivers/gpu/drm/i915/intel_guc_fwif.h
+++ b/drivers/gpu/drm/i915/intel_guc_fwif.h
@@ -39,6 +39,7 @@
 #define GUC_CTX_PRIORITY_HIGH		1
 #define GUC_CTX_PRIORITY_KMD_NORMAL	2
 #define GUC_CTX_PRIORITY_NORMAL		3
+#define GUC_CTX_PRIORITY_NUM		4
 
 #define GUC_MAX_GPU_CONTEXTS		1024
 #define	GUC_INVALID_CTX_ID		GUC_MAX_GPU_CONTEXTS
@@ -316,6 +317,50 @@ struct guc_context_desc {
 #define GUC_POWER_D2		3
 #define GUC_POWER_D3		4
 
+/* Scheduling policy settings */
+
+/* Reset engine upon preempt failure */
+#define POLICY_RESET_ENGINE		(1<<0)
+/* Preempt to idle on quantum expiry */
+#define POLICY_PREEMPT_TO_IDLE		(1<<1)
+
+#define POLICY_MAX_NUM_WI		15
+
+struct guc_policy {
+	/* Time for one workload to execute. (in micro seconds) */
+	u32 execution_quantum;
+	u32 reserved1;
+
+	/* Time to wait for a preemption request to completed before issuing a
+	 * reset. (in micro seconds). */
+	u32 preemption_time;
+
+	/* How much time to allow to run after the first fault is observed.
+	 * Then preempt afterwards. (in micro seconds) */
+	u32 fault_time;
+
+	u32 policy_flags;
+	u32 reserved[2];
+} __packed;
+
+struct guc_policies {
+	struct guc_policy policy[GUC_CTX_PRIORITY_NUM][I915_NUM_RINGS];
+
+	/* In micro seconds. How much time to allow before DPC processing is
+	 * called back via interrupt (to prevent DPC queue drain starving).
+	 * Typically 1000s of micro seconds (example only, not granularity). */
+	u32 dpc_promote_time;
+
+	/* Must be set to take these new values. */
+	u32 is_valid;
+
+	/* Max number of WIs to process per call. A large value may keep CS
+	 * idle. */
+	u32 max_num_work_items;
+
+	u32 reserved[19];
+} __packed;
+
 /* GuC Additional Data Struct */
 
 struct guc_ads {
-- 
2.5.0

_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

[PATCH 4/5] drm/i915/guc: Add GuC ADS - MMIO reg state

[yu.dai]

From: Alex Dai <yu.dai@intel.com>

GuC needs to know which registers and how they will be saved and
restored during event such as engine reset or power state changes.
For now only the base address of reg state is initialized. The
detail register table probably will be setup in future GuC TDR or
Preemption patch series.

Signed-off-by: Alex Dai <yu.dai@intel.com>
---
 drivers/gpu/drm/i915/i915_guc_submission.c | 22 +++++++++++++++++-
 drivers/gpu/drm/i915/intel_guc_fwif.h      | 37 ++++++++++++++++++++++++++++++
 2 files changed, 58 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/i915/i915_guc_submission.c b/drivers/gpu/drm/i915/i915_guc_submission.c
index a5c555c..92b8a34 100644
--- a/drivers/gpu/drm/i915/i915_guc_submission.c
+++ b/drivers/gpu/drm/i915/i915_guc_submission.c
@@ -869,12 +869,15 @@ static void guc_create_ads(struct intel_guc *guc)
 	struct drm_i915_gem_object *obj;
 	struct guc_ads *ads;
 	struct guc_policies *policies;
+	struct guc_mmio_reg_state *reg_state;
 	struct intel_engine_cs *ring;
 	struct page *page;
 	u32 size, i;
 
 	/* The ads obj includes the struct itself and buffers passed to GuC */
-	size = sizeof(struct guc_ads) + sizeof(struct guc_policies);
+	size = sizeof(struct guc_ads) + sizeof(struct guc_policies) +
+			sizeof(struct guc_mmio_reg_state) +
+			GUC_S3_SAVE_SPACE_PAGES * PAGE_SIZE;
 
 	obj = guc->ads_obj;
 	if (!obj) {
@@ -913,6 +916,23 @@ static void guc_create_ads(struct intel_guc *guc)
 	ads->scheduler_policies = i915_gem_obj_ggtt_offset(obj) +
 			sizeof(struct guc_ads);
 
+	/* MMIO reg state */
+	reg_state = (void *)policies + sizeof(struct guc_policies);
+
+	for (i = 0; i < I915_NUM_RINGS; i++) {
+		reg_state->mmio_white_list[i].mmio_start =
+			dev_priv->ring[i].mmio_base + GUC_MMIO_WHITE_LIST_START;
+
+		/* Nothing to be saved or restored for now. */
+		reg_state->mmio_white_list[i].count = 0;
+	}
+
+	ads->reg_state_addr = ads->scheduler_policies +
+			sizeof(struct guc_policies);
+
+	ads->reg_state_buffer = ads->reg_state_addr +
+			sizeof(struct guc_mmio_reg_state);
+
 	kunmap_atomic(ads);
 }
 
diff --git a/drivers/gpu/drm/i915/intel_guc_fwif.h b/drivers/gpu/drm/i915/intel_guc_fwif.h
index 0cc17c7..1bb6410 100644
--- a/drivers/gpu/drm/i915/intel_guc_fwif.h
+++ b/drivers/gpu/drm/i915/intel_guc_fwif.h
@@ -361,6 +361,43 @@ struct guc_policies {
 	u32 reserved[19];
 } __packed;
 
+/* GuC MMIO reg state struct */
+
+#define GUC_REGSET_FLAGS_NONE		0x0
+#define GUC_REGSET_POWERCYCLE		0x1
+#define GUC_REGSET_MASKED		0x2
+#define GUC_REGSET_ENGINERESET		0x4
+#define GUC_REGSET_SAVE_DEFAULT_VALUE	0x8
+#define GUC_REGSET_SAVE_CURRENT_VALUE	0x10
+
+#define GUC_REGSET_MAX_REGISTERS	20
+#define GUC_MMIO_WHITE_LIST_START	0x24d0
+#define GUC_MMIO_WHITE_LIST_MAX		12
+#define GUC_S3_SAVE_SPACE_PAGES		10
+
+struct guc_mmio_regset {
+	struct __packed {
+		u32 offset;
+		u32 value;
+		u32 flags;
+	} registers[GUC_REGSET_MAX_REGISTERS];
+
+	u32 values_valid;
+	u32 number_of_registers;
+} __packed;
+
+struct guc_mmio_reg_state {
+	struct guc_mmio_regset global_reg;
+	struct guc_mmio_regset engine_reg[I915_NUM_RINGS];
+
+	/* MMIO registers that are set as non privileged */
+	struct __packed {
+		u32 mmio_start;
+		u32 offsets[GUC_MMIO_WHITE_LIST_MAX];
+		u32 count;
+	} mmio_white_list[I915_NUM_RINGS];
+} __packed;
+
 /* GuC Additional Data Struct */
 
 struct guc_ads {
-- 
2.5.0

_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

[PATCH 5/5] drm/i915/guc: Add GuC ADS - enabling ADS

[yu.dai]

From: Alex Dai <yu.dai@intel.com>

Set ADS enabling flag during GuC init.

Signed-off-by: Alex Dai <yu.dai@intel.com>
---
 drivers/gpu/drm/i915/intel_guc_loader.c | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/drivers/gpu/drm/i915/intel_guc_loader.c b/drivers/gpu/drm/i915/intel_guc_loader.c
index 4740949..625272f4 100644
--- a/drivers/gpu/drm/i915/intel_guc_loader.c
+++ b/drivers/gpu/drm/i915/intel_guc_loader.c
@@ -165,6 +165,13 @@ static void set_guc_init_params(struct drm_i915_private *dev_priv)
 			i915.guc_log_level << GUC_LOG_VERBOSITY_SHIFT;
 	}
 
+	if (guc->ads_obj) {
+		u32 ads = (u32)i915_gem_obj_ggtt_offset(guc->ads_obj)
+				>> PAGE_SHIFT;
+		params[GUC_CTL_DEBUG] |= ads << GUC_ADS_ADDR_SHIFT;
+		params[GUC_CTL_DEBUG] |= GUC_ADS_ENABLED;
+	}
+
 	/* If GuC submission is enabled, set up additional parameters here */
 	if (i915.enable_guc_submission) {
 		u32 pgs = i915_gem_obj_ggtt_offset(dev_priv->guc.ctx_pool_obj);
-- 
2.5.0

_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: [PATCH 2/5] drm/i915/guc: Add GuC ADS (Addition Data Structure) - allocation

[Chris Wilson]

On Wed, Dec 16, 2015 at 01:40:52PM -0800, yu.dai@intel.com wrote:
> +	page = i915_gem_object_get_page(obj, 0);
> +	ads = kmap_atomic(page);
> +	if (!ads) {

kmap_atomic() cannot fail.
-Chris

-- 
Chris Wilson, Intel Open Source Technology Centre
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: [PATCH 3/5] drm/i915/guc: Add GuC ADS - scheduler policies

[Chris Wilson]

On Wed, Dec 16, 2015 at 01:40:53PM -0800, yu.dai@intel.com wrote:
> From: Alex Dai <yu.dai@intel.com>
> 
> GuC supports different scheduling policies for its four internal
> queues. Currently these have been set to the same default values
> as KMD_NORMAL queue.
> 
> Particularly POLICY_MAX_NUM_WI is set to 15 to match GuC internal
> maximum submit queue numbers to avoid an out-of-space problem.
> This value indicates max number of work items allowed to be queued
> for one DPC process. A smaller value will let GuC schedule more
> frequently while a larger number may increase chances to optimize
> cmds (such as collapse cmds from same lrc) with risks that keeps
> CS idle.
> 
> Signed-off-by: Alex Dai <yu.dai@intel.com>
> ---
>  drivers/gpu/drm/i915/i915_guc_submission.c | 31 +++++++++++++++++++-
>  drivers/gpu/drm/i915/intel_guc_fwif.h      | 45 ++++++++++++++++++++++++++++++
>  2 files changed, 75 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/gpu/drm/i915/i915_guc_submission.c b/drivers/gpu/drm/i915/i915_guc_submission.c
> index 66d85c3..a5c555c 100644
> --- a/drivers/gpu/drm/i915/i915_guc_submission.c
> +++ b/drivers/gpu/drm/i915/i915_guc_submission.c
> @@ -842,17 +842,39 @@ static void guc_create_log(struct intel_guc *guc)
>  	guc->log_flags = (offset << GUC_LOG_BUF_ADDR_SHIFT) | flags;
>  }
>  
> +static void init_guc_policies(struct guc_policies *policies)
> +{
> +	struct guc_policy *policy;
> +	u32 p, i;
> +
> +	policies->dpc_promote_time = 500000;
> +	policies->max_num_work_items = POLICY_MAX_NUM_WI;
> +
> +	for (p = 0; p < GUC_CTX_PRIORITY_NUM; p++)
> +	for (i = 0; i < I915_NUM_RINGS; i++) {

Please indent this properly.

> +		policy = &policies->policy[p][i];
> +
> +		policy->execution_quantum = 1000000;
> +		policy->preemption_time = 500000;
> +		policy->fault_time = 250000;
> +		policy->policy_flags = 0;
> +	}
> +
> +	policies->is_valid = 1;
> +}
> +
>  static void guc_create_ads(struct intel_guc *guc)
>  {
>  	struct drm_i915_private *dev_priv = guc_to_i915(guc);
>  	struct drm_i915_gem_object *obj;
>  	struct guc_ads *ads;
> +	struct guc_policies *policies;
>  	struct intel_engine_cs *ring;
>  	struct page *page;
>  	u32 size, i;
>  
>  	/* The ads obj includes the struct itself and buffers passed to GuC */
> -	size = sizeof(struct guc_ads);
> +	size = sizeof(struct guc_ads) + sizeof(struct guc_policies);
>  
>  	obj = guc->ads_obj;
>  	if (!obj) {
> @@ -884,6 +906,13 @@ static void guc_create_ads(struct intel_guc *guc)
>  	for_each_ring(ring, dev_priv, i)
>  		ads->eng_state_size[i] = intel_lr_context_size(ring);
>  
> +	/* GuC scheduling policies */
> +	policies = (void *)ads + sizeof(struct guc_ads);
> +	init_guc_policies(policies);

Please limit atomic context to only the critical section, i.e. don't
make me have to read every single function to check for violations.
-Chris

> +
> +	ads->scheduler_policies = i915_gem_obj_ggtt_offset(obj) +
> +			sizeof(struct guc_ads);
> +
>  	kunmap_atomic(ads);

-- 
Chris Wilson, Intel Open Source Technology Centre
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: [PATCH 3/5] drm/i915/guc: Add GuC ADS - scheduler policies

[Yu Dai]

On 12/16/2015 11:39 PM, Chris Wilson wrote:
> On Wed, Dec 16, 2015 at 01:40:53PM -0800, yu.dai@intel.com wrote:
> > From: Alex Dai <yu.dai@intel.com>
> >
> > GuC supports different scheduling policies for its four internal
> > queues. Currently these have been set to the same default values
> > as KMD_NORMAL queue.
> >
> > Particularly POLICY_MAX_NUM_WI is set to 15 to match GuC internal
> > maximum submit queue numbers to avoid an out-of-space problem.
> > This value indicates max number of work items allowed to be queued
> > for one DPC process. A smaller value will let GuC schedule more
> > frequently while a larger number may increase chances to optimize
> > cmds (such as collapse cmds from same lrc) with risks that keeps
> > CS idle.
> >
> > Signed-off-by: Alex Dai <yu.dai@intel.com>
> > ---
> >  drivers/gpu/drm/i915/i915_guc_submission.c | 31 +++++++++++++++++++-
> >  drivers/gpu/drm/i915/intel_guc_fwif.h      | 45 ++++++++++++++++++++++++++++++
> >  2 files changed, 75 insertions(+), 1 deletion(-)
> >
> > diff --git a/drivers/gpu/drm/i915/i915_guc_submission.c b/drivers/gpu/drm/i915/i915_guc_submission.c
> > index 66d85c3..a5c555c 100644
> > --- a/drivers/gpu/drm/i915/i915_guc_submission.c
> > +++ b/drivers/gpu/drm/i915/i915_guc_submission.c
> > @@ -842,17 +842,39 @@ static void guc_create_log(struct intel_guc *guc)
> >  	guc->log_flags = (offset << GUC_LOG_BUF_ADDR_SHIFT) | flags;
> >  }
> >
> > +static void init_guc_policies(struct guc_policies *policies)
> > +{
> > +	struct guc_policy *policy;
> > +	u32 p, i;
> > +
> > +	policies->dpc_promote_time = 500000;
> > +	policies->max_num_work_items = POLICY_MAX_NUM_WI;
> > +
> > +	for (p = 0; p < GUC_CTX_PRIORITY_NUM; p++)
> > +	for (i = 0; i < I915_NUM_RINGS; i++) {
>
> Please indent this properly.
>
> > +		policy = &policies->policy[p][i];
> > +
> > +		policy->execution_quantum = 1000000;
> > +		policy->preemption_time = 500000;
> > +		policy->fault_time = 250000;
> > +		policy->policy_flags = 0;
> > +	}
> > +
> > +	policies->is_valid = 1;
> > +}
> > +
> >  static void guc_create_ads(struct intel_guc *guc)
> >  {
> >  	struct drm_i915_private *dev_priv = guc_to_i915(guc);
> >  	struct drm_i915_gem_object *obj;
> >  	struct guc_ads *ads;
> > +	struct guc_policies *policies;
> >  	struct intel_engine_cs *ring;
> >  	struct page *page;
> >  	u32 size, i;
> >
> >  	/* The ads obj includes the struct itself and buffers passed to GuC */
> > -	size = sizeof(struct guc_ads);
> > +	size = sizeof(struct guc_ads) + sizeof(struct guc_policies);
> >
> >  	obj = guc->ads_obj;
> >  	if (!obj) {
> > @@ -884,6 +906,13 @@ static void guc_create_ads(struct intel_guc *guc)
> >  	for_each_ring(ring, dev_priv, i)
> >  		ads->eng_state_size[i] = intel_lr_context_size(ring);
> >
> > +	/* GuC scheduling policies */
> > +	policies = (void *)ads + sizeof(struct guc_ads);
> > +	init_guc_policies(policies);
>
> Please limit atomic context to only the critical section, i.e. don't
> make me have to read every single function to check for violations.

Could you clarify this? I am not sure what's the atomic context and 
critical section you mentioned here.

Alex
> > +
> > +	ads->scheduler_policies = i915_gem_obj_ggtt_offset(obj) +
> > +			sizeof(struct guc_ads);
> > +
> >  	kunmap_atomic(ads);
>

_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

[PATCH v1] drm/i915/guc: Add GuC ADS (Addition Data Structure) - allocation

[yu.dai]

From: Alex Dai <yu.dai@intel.com>

The GuC firmware uses this for various purposes. The ADS itself is
a chunk of memory created by driver to share with GuC. Its members
are usually addresses telling where GuC to access them, including
things like scheduler policies, register list that will be saved
and restored during reset etc.

This is the first patch of a series to enable GuC ADS. For now, we
only create the ADS obj whilst keep it disabled.

v1: remove dead code checking return of kmap_atomic (Chris Wilson)

Signed-off-by: Alex Dai <yu.dai@intel.com>

diff --git a/drivers/gpu/drm/i915/i915_guc_reg.h b/drivers/gpu/drm/i915/i915_guc_reg.h
index 90a84b4..8d27c09 100644
--- a/drivers/gpu/drm/i915/i915_guc_reg.h
+++ b/drivers/gpu/drm/i915/i915_guc_reg.h
@@ -40,6 +40,7 @@
 #define   GS_MIA_CORE_STATE		  (1 << GS_MIA_SHIFT)
 
 #define SOFT_SCRATCH(n)			_MMIO(0xc180 + (n) * 4)
+#define SOFT_SCRATCH_COUNT		16
 
 #define UOS_RSA_SCRATCH(i)		_MMIO(0xc200 + (i) * 4)
 #define   UOS_RSA_SCRATCH_MAX_COUNT	  64
diff --git a/drivers/gpu/drm/i915/i915_guc_submission.c b/drivers/gpu/drm/i915/i915_guc_submission.c
index 7554d16..28531e6 100644
--- a/drivers/gpu/drm/i915/i915_guc_submission.c
+++ b/drivers/gpu/drm/i915/i915_guc_submission.c
@@ -842,6 +842,46 @@ static void guc_create_log(struct intel_guc *guc)
 	guc->log_flags = (offset << GUC_LOG_BUF_ADDR_SHIFT) | flags;
 }
 
+static void guc_create_ads(struct intel_guc *guc)
+{
+	struct drm_i915_private *dev_priv = guc_to_i915(guc);
+	struct drm_i915_gem_object *obj;
+	struct guc_ads *ads;
+	struct intel_engine_cs *ring;
+	struct page *page;
+	u32 size, i;
+
+	/* The ads obj includes the struct itself and buffers passed to GuC */
+	size = sizeof(struct guc_ads);
+
+	obj = guc->ads_obj;
+	if (!obj) {
+		obj = gem_allocate_guc_obj(dev_priv->dev, PAGE_ALIGN(size));
+		if (!obj)
+			return;
+
+		guc->ads_obj = obj;
+	}
+
+	page = i915_gem_object_get_page(obj, 0);
+	ads = kmap_atomic(page);
+
+	/*
+	 * The GuC requires a "Golden Context" when it reinitialises
+	 * engines after a reset. Here we use the Render ring default
+	 * context, which must already exist and be pinned in the GGTT,
+	 * so its address won't change after we've told the GuC where
+	 * to find it.
+	 */
+	ring = &dev_priv->ring[RCS];
+	ads->golden_context_lrca = ring->status_page.gfx_addr;
+
+	for_each_ring(ring, dev_priv, i)
+		ads->eng_state_size[i] = intel_lr_context_size(ring);
+
+	kunmap_atomic(ads);
+}
+
 /*
  * Set up the memory resources to be shared with the GuC.  At this point,
  * we require just one object that can be mapped through the GGTT.
@@ -868,6 +908,8 @@ int i915_guc_submission_init(struct drm_device *dev)
 
 	guc_create_log(guc);
 
+	guc_create_ads(guc);
+
 	return 0;
 }
 
@@ -906,6 +948,9 @@ void i915_guc_submission_fini(struct drm_device *dev)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_guc *guc = &dev_priv->guc;
 
+	gem_release_guc_obj(dev_priv->guc.ads_obj);
+	guc->ads_obj = NULL;
+
 	gem_release_guc_obj(dev_priv->guc.log_obj);
 	guc->log_obj = NULL;
 
diff --git a/drivers/gpu/drm/i915/intel_guc.h b/drivers/gpu/drm/i915/intel_guc.h
index 5cf555d..5c9f894 100644
--- a/drivers/gpu/drm/i915/intel_guc.h
+++ b/drivers/gpu/drm/i915/intel_guc.h
@@ -89,6 +89,8 @@ struct intel_guc {
 	uint32_t log_flags;
 	struct drm_i915_gem_object *log_obj;
 
+	struct drm_i915_gem_object *ads_obj;
+
 	struct drm_i915_gem_object *ctx_pool_obj;
 	struct ida ctx_ids;
 
diff --git a/drivers/gpu/drm/i915/intel_guc_fwif.h b/drivers/gpu/drm/i915/intel_guc_fwif.h
index eaa50a4..76ecc85 100644
--- a/drivers/gpu/drm/i915/intel_guc_fwif.h
+++ b/drivers/gpu/drm/i915/intel_guc_fwif.h
@@ -81,11 +81,14 @@
 #define GUC_CTL_CTXINFO			0
 #define   GUC_CTL_CTXNUM_IN16_SHIFT	0
 #define   GUC_CTL_BASE_ADDR_SHIFT	12
+
 #define GUC_CTL_ARAT_HIGH		1
 #define GUC_CTL_ARAT_LOW		2
+
 #define GUC_CTL_DEVICE_INFO		3
 #define   GUC_CTL_GTTYPE_SHIFT		0
 #define   GUC_CTL_COREFAMILY_SHIFT	7
+
 #define GUC_CTL_LOG_PARAMS		4
 #define   GUC_LOG_VALID			(1 << 0)
 #define   GUC_LOG_NOTIFY_ON_HALF_FULL	(1 << 1)
@@ -97,9 +100,12 @@
 #define   GUC_LOG_ISR_PAGES		3
 #define   GUC_LOG_ISR_SHIFT		9
 #define   GUC_LOG_BUF_ADDR_SHIFT	12
+
 #define GUC_CTL_PAGE_FAULT_CONTROL	5
+
 #define GUC_CTL_WA			6
 #define   GUC_CTL_WA_UK_BY_DRIVER	(1 << 3)
+
 #define GUC_CTL_FEATURE			7
 #define   GUC_CTL_VCS2_ENABLED		(1 << 0)
 #define   GUC_CTL_KERNEL_SUBMISSIONS	(1 << 1)
@@ -109,6 +115,7 @@
 #define   GUC_CTL_PREEMPTION_LOG	(1 << 5)
 #define   GUC_CTL_ENABLE_SLPC		(1 << 7)
 #define   GUC_CTL_RESET_ON_PREMPT_FAILURE	(1 << 8)
+
 #define GUC_CTL_DEBUG			8
 #define   GUC_LOG_VERBOSITY_SHIFT	0
 #define   GUC_LOG_VERBOSITY_LOW		(0 << GUC_LOG_VERBOSITY_SHIFT)
@@ -118,9 +125,19 @@
 /* Verbosity range-check limits, without the shift */
 #define	  GUC_LOG_VERBOSITY_MIN		0
 #define	  GUC_LOG_VERBOSITY_MAX		3
+#define	  GUC_LOG_VERBOSITY_MASK	0x0000000f
+#define	  GUC_LOG_DESTINATION_MASK	(3 << 4)
+#define   GUC_LOG_DISABLED		(1 << 6)
+#define   GUC_PROFILE_ENABLED		(1 << 7)
+#define   GUC_WQ_TRACK_ENABLED		(1 << 8)
+#define   GUC_ADS_ENABLED		(1 << 9)
+#define   GUC_DEBUG_RESERVED		(1 << 10)
+#define   GUC_ADS_ADDR_SHIFT		11
+#define   GUC_ADS_ADDR_MASK		0xfffff800
+
 #define GUC_CTL_RSRVD			9
 
-#define GUC_CTL_MAX_DWORDS		(GUC_CTL_RSRVD + 1)
+#define GUC_CTL_MAX_DWORDS		(SOFT_SCRATCH_COUNT - 2) /* [1..14] */
 
 /**
  * DOC: GuC Firmware Layout
@@ -299,6 +316,18 @@ struct guc_context_desc {
 #define GUC_POWER_D2		3
 #define GUC_POWER_D3		4
 
+/* GuC Additional Data Struct */
+
+struct guc_ads {
+	u32 reg_state_addr;
+	u32 reg_state_buffer;
+	u32 golden_context_lrca;
+	u32 scheduler_policies;
+	u32 reserved0[3];
+	u32 eng_state_size[I915_NUM_RINGS];
+	u32 reserved2[4];
+} __packed;
+
 /* This Action will be programmed in C180 - SOFT_SCRATCH_O_REG */
 enum host2guc_action {
 	HOST2GUC_ACTION_DEFAULT = 0x0,
-- 
2.5.0

_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: [PATCH v1] drm/i915/guc: Add GuC ADS (Addition Data Structure) - allocation

[kbuild test robot]

[-- Attachment #1: Type: text/plain, Size: 1419 bytes --]

Hi Alex,

[auto build test ERROR on drm-intel/for-linux-next]
[also build test ERROR on next-20151217]
[cannot apply to v4.4-rc5]

url:    https://github.com/0day-ci/linux/commits/yu-dai-intel-com/drm-i915-guc-Add-GuC-ADS-Addition-Data-Structure-allocation/20151218-030155
base:   git://anongit.freedesktop.org/drm-intel for-linux-next
config: x86_64-nfsroot (attached as .config)
reproduce:
        # save the attached .config to linux build tree
        make ARCH=x86_64 

All errors (new ones prefixed by >>):

   drivers/gpu/drm/i915/i915_guc_submission.c: In function 'guc_create_ads':
>> drivers/gpu/drm/i915/i915_guc_submission.c:870:28: error: implicit declaration of function 'intel_lr_context_size' [-Werror=implicit-function-declaration]
      ads->eng_state_size[i] = intel_lr_context_size(ring);
                               ^
   cc1: some warnings being treated as errors

vim +/intel_lr_context_size +870 drivers/gpu/drm/i915/i915_guc_submission.c

   864		 * to find it.
   865		 */
   866		ring = &dev_priv->ring[RCS];
   867		ads->golden_context_lrca = ring->status_page.gfx_addr;
   868	
   869		for_each_ring(ring, dev_priv, i)
 > 870			ads->eng_state_size[i] = intel_lr_context_size(ring);
   871	
   872		kunmap_atomic(ads);
   873	}

---
0-DAY kernel test infrastructure                Open Source Technology Center
https://lists.01.org/pipermail/kbuild-all                   Intel Corporation

[-- Attachment #2: .config.gz --]
[-- Type: application/octet-stream, Size: 25190 bytes --]

[-- Attachment #3: Type: text/plain, Size: 159 bytes --]

_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: [PATCH 3/5] drm/i915/guc: Add GuC ADS - scheduler policies

[Chris Wilson]

On Thu, Dec 17, 2015 at 10:36:02AM -0800, Yu Dai wrote:
> On 12/16/2015 11:39 PM, Chris Wilson wrote:
> >On Wed, Dec 16, 2015 at 01:40:53PM -0800, yu.dai@intel.com wrote:
> >> From: Alex Dai <yu.dai@intel.com>
> >> @@ -884,6 +906,13 @@ static void guc_create_ads(struct intel_guc *guc)
> >>  	for_each_ring(ring, dev_priv, i)
> >>  		ads->eng_state_size[i] = intel_lr_context_size(ring);
> >>
> >> +	/* GuC scheduling policies */
> >> +	policies = (void *)ads + sizeof(struct guc_ads);
> >> +	init_guc_policies(policies);
> >
> >Please limit atomic context to only the critical section, i.e. don't
> >make me have to read every single function to check for violations.
> 
> Could you clarify this? I am not sure what's the atomic context and
> critical section you mentioned here.

This is inside an kmap_atomic()/kunmap_atomic() section. That means this
code is in atomic context and cannot sleep or be preempted i.e. there
are restrictions placed on what the code can do and what it can call.
-Chris

-- 
Chris Wilson, Intel Open Source Technology Centre
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx