[PATCH v6 03/11] drm/xe: Emit ctx timestamp copy in ring ops

[Matthew Brost <matthew.brost@intel.com>]

Copy ctx timestamp at beginning of every GPU job to a saved location.
Used to determine how long a job has been running on the hardware.

v2:
 - - s/ctx_timestamp_job/ctx_job_timestamp

Signed-off-by: Matthew Brost <matthew.brost@intel.com>
Reviewed-by: Jonathan Cavitt <jonathan.cavitt@intel.com>
---
 drivers/gpu/drm/xe/xe_ring_ops.c | 21 +++++++++++++++++++++
 1 file changed, 21 insertions(+)

diff --git a/drivers/gpu/drm/xe/xe_ring_ops.c b/drivers/gpu/drm/xe/xe_ring_ops.c
index db630d27beba..0be4f489d3e1 100644
--- a/drivers/gpu/drm/xe/xe_ring_ops.c
+++ b/drivers/gpu/drm/xe/xe_ring_ops.c
@@ -224,6 +224,19 @@ static u32 get_ppgtt_flag(struct xe_sched_job *job)
 	return job->q->vm ? BIT(8) : 0;
 }
 
+static int emit_copy_timestamp(struct xe_lrc *lrc, u32 *dw, int i)
+{
+	dw[i++] = MI_COPY_MEM_MEM | MI_COPY_MEM_MEM_SRC_GGTT |
+		MI_COPY_MEM_MEM_DST_GGTT;
+	dw[i++] = xe_lrc_ctx_job_timestamp_ggtt_addr(lrc);
+	dw[i++] = 0;
+	dw[i++] = xe_lrc_ctx_timestamp_ggtt_addr(lrc);
+	dw[i++] = 0;
+	dw[i++] = MI_NOOP;
+
+	return i;
+}
+
 /* for engines that don't require any special HW handling (no EUs, no aux inval, etc) */
 static void __emit_job_gen12_simple(struct xe_sched_job *job, struct xe_lrc *lrc,
 				    u64 batch_addr, u32 seqno)
@@ -232,6 +245,8 @@ static void __emit_job_gen12_simple(struct xe_sched_job *job, struct xe_lrc *lrc
 	u32 ppgtt_flag = get_ppgtt_flag(job);
 	struct xe_gt *gt = job->q->gt;
 
+	i = emit_copy_timestamp(lrc, dw, i);
+
 	if (job->ring_ops_flush_tlb) {
 		dw[i++] = preparser_disable(true);
 		i = emit_flush_imm_ggtt(xe_lrc_start_seqno_ggtt_addr(lrc),
@@ -283,6 +298,8 @@ static void __emit_job_gen12_video(struct xe_sched_job *job, struct xe_lrc *lrc,
 	struct xe_device *xe = gt_to_xe(gt);
 	bool decode = job->q->class == XE_ENGINE_CLASS_VIDEO_DECODE;
 
+	i = emit_copy_timestamp(lrc, dw, i);
+
 	dw[i++] = preparser_disable(true);
 
 	/* hsdes: 1809175790 */
@@ -332,6 +349,8 @@ static void __emit_job_gen12_render_compute(struct xe_sched_job *job,
 	bool lacks_render = !(gt->info.engine_mask & XE_HW_ENGINE_RCS_MASK);
 	u32 mask_flags = 0;
 
+	i = emit_copy_timestamp(lrc, dw, i);
+
 	dw[i++] = preparser_disable(true);
 	if (lacks_render)
 		mask_flags = PIPE_CONTROL_3D_ARCH_FLAGS;
@@ -375,6 +394,8 @@ static void emit_migration_job_gen12(struct xe_sched_job *job,
 {
 	u32 dw[MAX_JOB_SIZE_DW], i = 0;
 
+	i = emit_copy_timestamp(lrc, dw, i);
+
 	i = emit_store_imm_ggtt(xe_lrc_start_seqno_ggtt_addr(lrc),
 				seqno, dw, i);
 
-- 
2.34.1