diff mbox

[RFC,4/8] drm/i915: Insert commands for capturing timestamps in the ring

Message ID 1438754144-20435-5-git-send-email-sourab.gupta@intel.com (mailing list archive)
State New, archived
Headers show

Commit Message

sourab.gupta@intel.com Aug. 5, 2015, 5:55 a.m. UTC 
From: Sourab Gupta <sourab.gupta@intel.com>

This patch adds the routines through which one can insert commands in the
ringbuf for capturing timestamps, which are used to insert these commands
around the batchbuffer.

While inserting the commands, we keep a reference of associated request.
This will be released when we are forwarding the samples to userspace
(or when the event is being destroyed).
Also, an active reference of the destination buffer is taken here, so that
we can be assured that the buffer is freed up only after GPU is done with
it, even if the local reference of the buffer is released.

v2: Changes (as suggested by Chris):
    - Passing in 'request' struct for emit report function
    - Removed multiple calls to i915_gem_obj_to_ggtt(). Keeping hold of
      pinned vma from start and using when required.
    - Better nomenclature, and error handling.

Signed-off-by: Sourab Gupta <sourab.gupta@intel.com>
---
 drivers/gpu/drm/i915/i915_drv.h     |  2 +
 drivers/gpu/drm/i915/i915_oa_perf.c | 77 +++++++++++++++++++++++++++++++++++++
 drivers/gpu/drm/i915/i915_reg.h     |  2 +
 3 files changed, 81 insertions(+)

Comments

Chris Wilson Aug. 5, 2015, 9:30 a.m. UTC | #1 
On Wed, Aug 05, 2015 at 11:25:40AM +0530, sourab.gupta@intel.com wrote:
> From: Sourab Gupta <sourab.gupta@intel.com>
> 
> This patch adds the routines through which one can insert commands in the
> ringbuf for capturing timestamps, which are used to insert these commands
> around the batchbuffer.
> 
> While inserting the commands, we keep a reference of associated request.
> This will be released when we are forwarding the samples to userspace
> (or when the event is being destroyed).
> Also, an active reference of the destination buffer is taken here, so that
> we can be assured that the buffer is freed up only after GPU is done with
> it, even if the local reference of the buffer is released.
> 
> v2: Changes (as suggested by Chris):
>     - Passing in 'request' struct for emit report function
>     - Removed multiple calls to i915_gem_obj_to_ggtt(). Keeping hold of
>       pinned vma from start and using when required.
>     - Better nomenclature, and error handling.
> 
> @@ -919,6 +993,7 @@ static int init_gen_pmu_buffer(struct perf_event *event)
>  	dev_priv->gen_pmu.buffer.obj = bo;
>  	dev_priv->gen_pmu.buffer.gtt_offset =
>  				i915_gem_obj_ggtt_offset(bo);
> +	dev_priv->gen_pmu.buffer.vma = i915_gem_obj_to_ggtt(bo);
>  	dev_priv->gen_pmu.buffer.addr = vmap_oa_buffer(bo);
>  	INIT_LIST_HEAD(&dev_priv->gen_pmu.node_list);

Still calling i915_gem_obj_to_ggtt(bo) twice! With pmu_buffer.vma you
can drop pmu_buffer.gtt_offset and never be confused again.
-Chris
sourab.gupta@intel.com Aug. 5, 2015, 9:54 a.m. UTC | #2 
On Wed, 2015-08-05 at 09:30 +0000, Chris Wilson wrote:
> On Wed, Aug 05, 2015 at 11:25:40AM +0530, sourab.gupta@intel.com wrote:

> > From: Sourab Gupta <sourab.gupta@intel.com>

> > 

> > This patch adds the routines through which one can insert commands in the

> > ringbuf for capturing timestamps, which are used to insert these commands

> > around the batchbuffer.

> > 

> > While inserting the commands, we keep a reference of associated request.

> > This will be released when we are forwarding the samples to userspace

> > (or when the event is being destroyed).

> > Also, an active reference of the destination buffer is taken here, so that

> > we can be assured that the buffer is freed up only after GPU is done with

> > it, even if the local reference of the buffer is released.

> > 

> > v2: Changes (as suggested by Chris):

> >     - Passing in 'request' struct for emit report function

> >     - Removed multiple calls to i915_gem_obj_to_ggtt(). Keeping hold of

> >       pinned vma from start and using when required.

> >     - Better nomenclature, and error handling.

> > 

> > @@ -919,6 +993,7 @@ static int init_gen_pmu_buffer(struct perf_event *event)

> >  	dev_priv->gen_pmu.buffer.obj = bo;

> >  	dev_priv->gen_pmu.buffer.gtt_offset =

> >  				i915_gem_obj_ggtt_offset(bo);

> > +	dev_priv->gen_pmu.buffer.vma = i915_gem_obj_to_ggtt(bo);

> >  	dev_priv->gen_pmu.buffer.addr = vmap_oa_buffer(bo);

> >  	INIT_LIST_HEAD(&dev_priv->gen_pmu.node_list);

> 

> Still calling i915_gem_obj_to_ggtt(bo) twice! With pmu_buffer.vma you

> can drop pmu_buffer.gtt_offset and never be confused again.

> -Chris

> 

Sorry, I'll have the one pertaining to gtt offset removed, and derive
the gtt offset field from the vma only.
diff mbox

Patch

diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index 5717cb0..46ece85 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -1663,6 +1663,7 @@  enum i915_oa_event_state {
 
 enum i915_profile_mode {
 	I915_PROFILE_OA = 0,
+	I915_PROFILE_TS,
 	I915_PROFILE_MAX,
 };
 
@@ -2007,6 +2008,7 @@  struct drm_i915_private {
 		struct {
 			struct drm_i915_gem_object *obj;
 			u32 gtt_offset;
+			struct i915_vma *vma;
 			u8 *addr;
 			u32 node_size;
 			u32 node_count;
diff --git a/drivers/gpu/drm/i915/i915_oa_perf.c b/drivers/gpu/drm/i915/i915_oa_perf.c
index 06645f0..2cf7f1b 100644
--- a/drivers/gpu/drm/i915/i915_oa_perf.c
+++ b/drivers/gpu/drm/i915/i915_oa_perf.c
@@ -113,6 +113,79 @@  static void i915_oa_emit_perf_report(struct drm_i915_gem_request *req,
 	i915_vma_move_to_active(dev_priv->oa_pmu.oa_rcs_buffer.vma, ring);
 }
 
+/*
+ * Emits the commands to capture timestamps, into the CS
+ */
+static void i915_gen_emit_ts_data(struct drm_i915_gem_request *req,
+			u32 global_ctx_id, u32 tag)
+{
+	struct intel_engine_cs *ring = req->ring;
+	struct drm_i915_private *dev_priv = ring->dev->dev_private;
+	struct drm_i915_gem_object *obj = dev_priv->gen_pmu.buffer.obj;
+	struct i915_gen_pmu_node *entry;
+	u32 addr = 0;
+	int ret;
+
+	entry = kzalloc(sizeof(*entry), GFP_KERNEL);
+	if (entry == NULL) {
+		DRM_ERROR("alloc failed\n");
+		return;
+	}
+
+	ret = intel_ring_begin(ring, 6);
+	if (ret) {
+		kfree(entry);
+		return;
+	}
+
+	entry->ctx_id = global_ctx_id;
+	i915_gem_request_assign(&entry->req, ring->outstanding_lazy_request);
+
+	spin_lock(&dev_priv->gen_pmu.lock);
+	if (list_empty(&dev_priv->gen_pmu.node_list))
+		entry->offset = 0;
+	else {
+		struct i915_gen_pmu_node *last_entry;
+		int max_offset = dev_priv->gen_pmu.buffer.node_count *
+				dev_priv->gen_pmu.buffer.node_size;
+
+		last_entry = list_last_entry(&dev_priv->gen_pmu.node_list,
+					struct i915_gen_pmu_node, head);
+		entry->offset = last_entry->offset +
+				dev_priv->gen_pmu.buffer.node_size;
+
+		if (entry->offset > max_offset)
+			entry->offset = 0;
+	}
+	list_add_tail(&entry->head, &dev_priv->gen_pmu.node_list);
+	spin_unlock(&dev_priv->gen_pmu.lock);
+
+	addr = dev_priv->gen_pmu.buffer.gtt_offset + entry->offset;
+
+	if (ring->id == RCS) {
+		intel_ring_emit(ring, GFX_OP_PIPE_CONTROL(5));
+		intel_ring_emit(ring,
+				PIPE_CONTROL_GLOBAL_GTT_IVB |
+				PIPE_CONTROL_TIMESTAMP_WRITE);
+		intel_ring_emit(ring, addr | PIPE_CONTROL_GLOBAL_GTT);
+		intel_ring_emit(ring, 0);
+		intel_ring_emit(ring, 0);
+		intel_ring_emit(ring, MI_NOOP);
+	} else {
+		intel_ring_emit(ring,
+				(MI_FLUSH_DW + 1) | MI_FLUSH_DW_OP_STAMP);
+		intel_ring_emit(ring, addr | MI_FLUSH_DW_USE_GTT);
+		intel_ring_emit(ring, 0);
+		intel_ring_emit(ring, 0);
+		intel_ring_emit(ring, MI_NOOP);
+		intel_ring_emit(ring, MI_NOOP);
+	}
+	intel_ring_advance(ring);
+
+	obj->base.write_domain = I915_GEM_DOMAIN_RENDER;
+	i915_vma_move_to_active(dev_priv->gen_pmu.buffer.vma, ring);
+}
+
 static void forward_one_oa_snapshot_to_event(struct drm_i915_private *dev_priv,
 					     u8 *snapshot,
 					     struct perf_event *event)
@@ -738,6 +811,7 @@  out:
 	spin_lock(&dev_priv->gen_pmu.lock);
 	dev_priv->gen_pmu.buffer.obj = NULL;
 	dev_priv->gen_pmu.buffer.gtt_offset = 0;
+	dev_priv->gen_pmu.buffer.vma = NULL;
 	dev_priv->gen_pmu.buffer.addr = NULL;
 	spin_unlock(&dev_priv->gen_pmu.lock);
 }
@@ -919,6 +993,7 @@  static int init_gen_pmu_buffer(struct perf_event *event)
 	dev_priv->gen_pmu.buffer.obj = bo;
 	dev_priv->gen_pmu.buffer.gtt_offset =
 				i915_gem_obj_ggtt_offset(bo);
+	dev_priv->gen_pmu.buffer.vma = i915_gem_obj_to_ggtt(bo);
 	dev_priv->gen_pmu.buffer.addr = vmap_oa_buffer(bo);
 	INIT_LIST_HEAD(&dev_priv->gen_pmu.node_list);
 
@@ -1504,6 +1579,7 @@  static void i915_gen_event_start(struct perf_event *event, int flags)
 
 	spin_lock(&dev_priv->gen_pmu.lock);
 	dev_priv->gen_pmu.event_active = true;
+	dev_priv->emit_profiling_data[I915_PROFILE_TS] = i915_gen_emit_ts_data;
 	spin_unlock(&dev_priv->gen_pmu.lock);
 
 	__hrtimer_start_range_ns(&dev_priv->gen_pmu.timer, ns_to_ktime(PERIOD),
@@ -1523,6 +1599,7 @@  static void i915_gen_event_stop(struct perf_event *event, int flags)
 
 	spin_lock(&dev_priv->gen_pmu.lock);
 	dev_priv->gen_pmu.event_active = false;
+	dev_priv->emit_profiling_data[I915_PROFILE_TS] = NULL;
 	list_for_each_entry(entry, &dev_priv->gen_pmu.node_list, head)
 		entry->discard = true;
 	spin_unlock(&dev_priv->gen_pmu.lock);
diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
index c9955968..f816b08 100644
--- a/drivers/gpu/drm/i915/i915_reg.h
+++ b/drivers/gpu/drm/i915/i915_reg.h
@@ -347,6 +347,7 @@ 
 #define   MI_FLUSH_DW_STORE_INDEX	(1<<21)
 #define   MI_INVALIDATE_TLB		(1<<18)
 #define   MI_FLUSH_DW_OP_STOREDW	(1<<14)
+#define   MI_FLUSH_DW_OP_STAMP		(3<<14)
 #define   MI_FLUSH_DW_OP_MASK		(3<<14)
 #define   MI_FLUSH_DW_NOTIFY		(1<<8)
 #define   MI_INVALIDATE_BSD		(1<<7)
@@ -423,6 +424,7 @@ 
 #define   PIPE_CONTROL_MEDIA_STATE_CLEAR		(1<<16)
 #define   PIPE_CONTROL_QW_WRITE				(1<<14)
 #define   PIPE_CONTROL_POST_SYNC_OP_MASK                (3<<14)
+#define   PIPE_CONTROL_TIMESTAMP_WRITE			(3<<14)
 #define   PIPE_CONTROL_DEPTH_STALL			(1<<13)
 #define   PIPE_CONTROL_WRITE_FLUSH			(1<<12)
 #define   PIPE_CONTROL_RENDER_TARGET_CACHE_FLUSH	(1<<12) /* gen6+ */