[Intel-gfx] [PATCH 4/5] drm/i915: Add basic execlist info to error state

[Ben Widawsky]

Sample output:
...
  waiting: yes
  ring->head: 0x00000000
  ring->tail: 0x00000c50
  ring->next_context_status_buffer: 0x5
  CSB Pointer: 0x00000405
    Context 0 Status: 0x0000000000000001
    Context 1 Status: 0x0000009d00000018
    Context 2 Status: 0x0000000000000001
    Context 3 Status: 0x0000000100000018
    Context 4 Status: 0x0000000000000001
    Context 5 Status: 0x0000009d00000018
  hangcheck: hung [40]
bsd command stream:
  START: 0x00039000
  HEAD:  0x00000018
...

Signed-off-by: Ben Widawsky <benjamin.widawsky at intel.com>
---
 drivers/gpu/drm/i915/i915_drv.h       |  7 ++++++-
 drivers/gpu/drm/i915/i915_gpu_error.c | 23 +++++++++++++++++++++++
 drivers/gpu/drm/i915/intel_lrc.c      | 10 +++++-----
 drivers/gpu/drm/i915/intel_lrc.h      |  4 ++++
 4 files changed, 38 insertions(+), 6 deletions(-)

diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index c6dd4db..c79e869 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -526,6 +526,7 @@ struct drm_i915_error_state {
 		u32 cpu_ring_tail;
 
 		u32 semaphore_seqno[I915_NUM_RINGS - 1];
+		u32 semaphore_mboxes[I915_NUM_RINGS - 1];
 
 		/* Register state */
 		u32 start;
@@ -545,7 +546,11 @@ struct drm_i915_error_state {
 		u32 fault_reg;
 		u64 faddr;
 		u32 rc_psmi; /* sleep state */
-		u32 semaphore_mboxes[I915_NUM_RINGS - 1];
+
+		/* execlist state */
+		u32 csb_ptr;
+		u8 next_context_status_buffer;
+		u64 context_status[GEN8_CSB_ENTRIES];
 
 		struct drm_i915_error_object {
 			int page_count;
diff --git a/drivers/gpu/drm/i915/i915_gpu_error.c b/drivers/gpu/drm/i915/i915_gpu_error.c
index 06ca408..20a5daa 100644
--- a/drivers/gpu/drm/i915/i915_gpu_error.c
+++ b/drivers/gpu/drm/i915/i915_gpu_error.c
@@ -301,6 +301,18 @@ static void i915_ring_error_state(struct drm_i915_error_state_buf *m,
 	err_printf(m, "  waiting: %s\n", yesno(ring->waiting));
 	err_printf(m, "  ring->head: 0x%08x\n", ring->cpu_ring_head);
 	err_printf(m, "  ring->tail: 0x%08x\n", ring->cpu_ring_tail);
+
+	if (i915.enable_execlists) {
+		int j;
+		err_printf(m, "  ring->next_context_status_buffer: 0x%d\n",
+			      ring->next_context_status_buffer);
+		err_printf(m, "  CSB Pointer: 0x%08x\n", ring->csb_ptr);
+		for (j = 0; j < GEN8_CSB_ENTRIES; j++) {
+			err_printf(m, "    Context %d Status: 0x%016llx\n",
+				           j, ring->context_status[j]);
+		}
+	}
+
 	err_printf(m, "  hangcheck: %s [%d]\n",
 		   hangcheck_action_to_str(ring->hangcheck_action),
 		   ring->hangcheck_score);
@@ -1042,6 +1054,8 @@ static void i915_gem_record_rings(struct drm_device *dev,
 		}
 
 		if (i915.enable_execlists) {
+			int j;
+
 			/* TODO: This is only a small fix to keep basic error
 			 * capture working, but we need to add more information
 			 * for it to be useful (e.g. dump the context being
@@ -1051,6 +1065,15 @@ static void i915_gem_record_rings(struct drm_device *dev,
 				rbuf = request->ctx->engine[ring->id].ringbuf;
 			else
 				rbuf = ring->default_context->engine[ring->id].ringbuf;
+
+			error->ring[i].csb_ptr = I915_READ(RING_CONTEXT_STATUS_PTR(ring));
+			error->ring[i].next_context_status_buffer = ring->next_context_status_buffer;
+			for (j = 0; j < GEN8_CSB_ENTRIES; j++) {
+				u32 status, id;
+				intel_lrc_get_context_status(ring, j, &status, &id);
+				error->ring[i].context_status[j] = ((__u64)id<<32)|(__u64)status;
+			}
+
 		} else
 			rbuf = ring->buffer;
 
diff --git a/drivers/gpu/drm/i915/intel_lrc.c b/drivers/gpu/drm/i915/intel_lrc.c
index 23839ff..a118146 100644
--- a/drivers/gpu/drm/i915/intel_lrc.c
+++ b/drivers/gpu/drm/i915/intel_lrc.c
@@ -496,9 +496,9 @@ static bool execlists_check_remove_request(struct intel_engine_cs *ring,
 	return false;
 }
 
-static void get_context_status(struct intel_engine_cs *ring,
-			       u8 read_pointer,
-			       u32 *status, u32 *context_id)
+void intel_lrc_get_context_status(struct intel_engine_cs *ring,
+				  u8 read_pointer,
+				  u32 *status, u32 *context_id)
 {
 	struct drm_i915_private *dev_priv = ring->dev->dev_private;
 
@@ -537,8 +537,8 @@ void intel_lrc_irq_handler(struct intel_engine_cs *ring)
 
 	while (read_pointer < write_pointer) {
 
-		get_context_status(ring, ++read_pointer % GEN8_CSB_ENTRIES,
-				   &status, &status_id);
+		intel_lrc_get_context_status(ring, ++read_pointer % GEN8_CSB_ENTRIES,
+					     &status, &status_id);
 
 		if (status & GEN8_CTX_STATUS_IDLE_ACTIVE)
 			continue;
diff --git a/drivers/gpu/drm/i915/intel_lrc.h b/drivers/gpu/drm/i915/intel_lrc.h
index de41ad6..82c87f9 100644
--- a/drivers/gpu/drm/i915/intel_lrc.h
+++ b/drivers/gpu/drm/i915/intel_lrc.h
@@ -118,4 +118,8 @@ u32 intel_execlists_ctx_id(struct drm_i915_gem_object *ctx_obj);
 void intel_lrc_irq_handler(struct intel_engine_cs *ring);
 void intel_execlists_retire_requests(struct intel_engine_cs *ring);
 
+void intel_lrc_get_context_status(struct intel_engine_cs *ring,
+				  u8 read_pointer,
+				  u32 *status, u32 *context_id);
+
 #endif /* _INTEL_LRC_H_ */
-- 
2.6.4