gallium/tegra: Add debugging messagesstaging/tegra

Signed-off-by: Thierry Reding <treding@nvidia.com>

4 files changed, 411 insertions, 25 deletions

diff --git a/src/gallium/drivers/tegra/tegra_context.c b/src/gallium/drivers/tegra/tegra_context.c
index a7a7690ec7b..389bc8bc6a5 100644
--- a/src/gallium/drivers/tegra/tegra_context.c
+++ b/src/gallium/drivers/tegra/tegra_context.c
@@ -35,8 +35,12 @@ tegra_destroy(struct pipe_context *pcontext)
 {
 	struct tegra_context *context = to_tegra_context(pcontext);
 
+	debug_printf("> %s(pcontext=%p)\n", __func__, pcontext);
+
 	context->gpu->destroy(context->gpu);
 	free(context);
+
+	debug_printf("< %s()\n", __func__);
 }
 
 static void
@@ -46,13 +50,24 @@ tegra_draw_vbo(struct pipe_context *pcontext,
 	struct tegra_context *context = to_tegra_context(pcontext);
 	struct pipe_draw_info info;
 
+	debug_printf("> %s(pcontext=%p, pinfo=%p)\n", __func__, pcontext,
+		     pinfo);
+	debug_printf("  pinfo:\n");
+	debug_printf("    indexed: %d\n", pinfo->indexed);
+	debug_printf("    mode: %x\n", pinfo->mode);
+	debug_printf("    start: %u\n", pinfo->start);
+	debug_printf("    count: %u\n", pinfo->count);
+
 	if (pinfo && pinfo->indirect) {
+		debug_printf("  unwrapping pipe_draw_info\n");
 		memcpy(&info, pinfo, sizeof(info));
 		info.indirect = tegra_resource_unwrap(info.indirect);
 		pinfo = &info;
 	}
 
 	context->gpu->draw_vbo(context->gpu, pinfo);
+
+	debug_printf("< %s()\n", __func__);
 }
 
 static void *
@@ -60,8 +75,14 @@ tegra_create_blend_state(struct pipe_context *pcontext,
 			 const struct pipe_blend_state *cso)
 {
 	struct tegra_context *context = to_tegra_context(pcontext);
+	void *so;
+
+	debug_printf("> %s(pcontext=%p, cso=%p)\n", __func__, pcontext, cso);
+
+	so = context->gpu->create_blend_state(context->gpu, cso);
 
-	return context->gpu->create_blend_state(context->gpu, cso);
+	debug_printf("< %s() = %p\n", __func__, so);
+	return so;
 }
 
 static void
@@ -70,7 +91,11 @@ tegra_bind_blend_state(struct pipe_context *pcontext,
 {
 	struct tegra_context *context = to_tegra_context(pcontext);
 
+	debug_printf("> %s(pcontext=%p, so=%p)\n", __func__, pcontext, so);
+
 	context->gpu->bind_blend_state(context->gpu, so);
+
+	debug_printf("< %s()\n", __func__);
 }
 
 static void
@@ -79,7 +104,11 @@ tegra_delete_blend_state(struct pipe_context *pcontext,
 {
 	struct tegra_context *context = to_tegra_context(pcontext);
 
+	debug_printf("> %s(pcontext=%p, so=%p)\n", __func__, pcontext, so);
+
 	context->gpu->delete_blend_state(context->gpu, so);
+
+	debug_printf("< %s()\n", __func__);
 }
 
 static void *
@@ -87,8 +116,14 @@ tegra_create_sampler_state(struct pipe_context *pcontext,
 			   const struct pipe_sampler_state *cso)
 {
 	struct tegra_context *context = to_tegra_context(pcontext);
+	void *so;
+
+	debug_printf("> %s(pcontext=%p, cso=%p)\n", __func__, pcontext, cso);
+
+	so = context->gpu->create_sampler_state(context->gpu, cso);
 
-	return context->gpu->create_sampler_state(context->gpu, cso);
+	debug_printf("< %s() = %p\n", __func__, so);
+	return so;
 }
 
 static void
@@ -100,8 +135,14 @@ tegra_bind_sampler_states(struct pipe_context *pcontext,
 {
 	struct tegra_context *context = to_tegra_context(pcontext);
 
+	debug_printf("> %s(pcontext=%p, shader=%u, start_slot=%u, num_samplers=%u, samplers=%p)\n",
+		     __func__, pcontext, shader, start_slot, num_samplers,
+		     samplers);
+
 	context->gpu->bind_sampler_states(context->gpu, shader, start_slot,
 					  num_samplers, samplers);
+
+	debug_printf("< %s()\n", __func__);
 }
 
 static void
@@ -110,7 +151,11 @@ tegra_delete_sampler_state(struct pipe_context *pcontext,
 {
 	struct tegra_context *context = to_tegra_context(pcontext);
 
+	debug_printf("> %s(pcontext=%p, so=%p)\n", __func__, pcontext, so);
+
 	context->gpu->delete_sampler_state(context->gpu, so);
+
+	debug_printf("< %s()\n", __func__);
 }
 
 static void *
@@ -118,8 +163,14 @@ tegra_create_rasterizer_state(struct pipe_context *pcontext,
 			      const struct pipe_rasterizer_state *cso)
 {
 	struct tegra_context *context = to_tegra_context(pcontext);
+	void *so;
+
+	debug_printf("> %s(pcontext=%p, cso=%p)\n", __func__, pcontext, cso);
+
+	so = context->gpu->create_rasterizer_state(context->gpu, cso);
 
-	return context->gpu->create_rasterizer_state(context->gpu, cso);
+	debug_printf("< %s() = %p\n", __func__, so);
+	return so;
 }
 
 static void
@@ -128,7 +179,11 @@ tegra_bind_rasterizer_state(struct pipe_context *pcontext,
 {
 	struct tegra_context *context = to_tegra_context(pcontext);
 
+	debug_printf("> %s(pcontext=%p, so=%p)\n", __func__, pcontext, so);
+
 	context->gpu->bind_rasterizer_state(context->gpu, so);
+
+	debug_printf("< %s()\n", __func__);
 }
 
 static void
@@ -137,7 +192,11 @@ tegra_delete_rasterizer_state(struct pipe_context *pcontext,
 {
 	struct tegra_context *context = to_tegra_context(pcontext);
 
+	debug_printf("> %s(pcontext=%p, so=%p)\n", __func__, pcontext, so);
+
 	context->gpu->delete_rasterizer_state(context->gpu, so);
+
+	debug_printf("< %s()\n", __func__);
 }
 
 static void *
@@ -145,9 +204,15 @@ tegra_create_depth_stencil_alpha_state(struct pipe_context *pcontext,
 				       const struct pipe_depth_stencil_alpha_state *cso)
 {
 	struct tegra_context *context = to_tegra_context(pcontext);
+	void *so;
+
+	debug_printf("> %s(pcontext=%p, cso=%p)\n", __func__, pcontext, cso);
+
+	so = context->gpu->create_depth_stencil_alpha_state(context->gpu,
+							    cso);
 
-	return context->gpu->create_depth_stencil_alpha_state(context->gpu,
-							      cso);
+	debug_printf("< %s() = %p\n", __func__, so);
+	return so;
 }
 
 static void
@@ -156,7 +221,11 @@ tegra_bind_depth_stencil_alpha_state(struct pipe_context *pcontext,
 {
 	struct tegra_context *context = to_tegra_context(pcontext);
 
+	debug_printf("> %s(pcontext=%p, so=%p)\n", __func__, pcontext, so);
+
 	context->gpu->bind_depth_stencil_alpha_state(context->gpu, so);
+
+	debug_printf("< %s()\n", __func__);
 }
 
 static void
@@ -165,7 +234,11 @@ tegra_delete_depth_stencil_alpha_state(struct pipe_context *pcontext,
 {
 	struct tegra_context *context = to_tegra_context(pcontext);
 
+	debug_printf("> %s(pcontext=%p, so=%p)\n", __func__, pcontext, so);
+
 	context->gpu->delete_depth_stencil_alpha_state(context->gpu, so);
+
+	debug_printf("< %s()\n", __func__);
 }
 
 static void *
@@ -173,8 +246,14 @@ tegra_create_fs_state(struct pipe_context *pcontext,
 		      const struct pipe_shader_state *cso)
 {
 	struct tegra_context *context = to_tegra_context(pcontext);
+	void *so;
+
+	debug_printf("> %s(pcontext=%p, cso=%p)\n", __func__, pcontext, cso);
+
+	so = context->gpu->create_fs_state(context->gpu, cso);
 
-	return context->gpu->create_fs_state(context->gpu, cso);
+	debug_printf("< %s() = %p\n", __func__, so);
+	return so;
 }
 
 static void
@@ -183,7 +262,11 @@ tegra_bind_fs_state(struct pipe_context *pcontext,
 {
 	struct tegra_context *context = to_tegra_context(pcontext);
 
+	debug_printf("> %s(pcontext=%p, so=%p)\n", __func__, pcontext, so);
+
 	context->gpu->bind_fs_state(context->gpu, so);
+
+	debug_printf("< %s()\n", __func__);
 }
 
 static void
@@ -192,7 +275,11 @@ tegra_delete_fs_state(struct pipe_context *pcontext,
 {
 	struct tegra_context *context = to_tegra_context(pcontext);
 
+	debug_printf("> %s(pcontext=%p, so=%p)\n", __func__, pcontext, so);
+
 	context->gpu->delete_fs_state(context->gpu, so);
+
+	debug_printf("< %s()\n", __func__);
 }
 
 static void *
@@ -200,8 +287,14 @@ tegra_create_vs_state(struct pipe_context *pcontext,
 		      const struct pipe_shader_state *cso)
 {
 	struct tegra_context *context = to_tegra_context(pcontext);
+	void *so;
+
+	debug_printf("> %s(pcontext=%p, cso=%p)\n", __func__, pcontext, cso);
+
+	so = context->gpu->create_vs_state(context->gpu, cso);
 
-	return context->gpu->create_vs_state(context->gpu, cso);
+	debug_printf("< %s() = %p\n", __func__, so);
+	return so;
 }
 
 static void
@@ -210,7 +303,11 @@ tegra_bind_vs_state(struct pipe_context *pcontext,
 {
 	struct tegra_context *context = to_tegra_context(pcontext);
 
+	debug_printf("> %s(pcontext=%p, so=%p)\n", __func__, pcontext, so);
+
 	context->gpu->bind_vs_state(context->gpu, so);
+
+	debug_printf("< %s()\n", __func__);
 }
 
 static void
@@ -219,7 +316,11 @@ tegra_delete_vs_state(struct pipe_context *pcontext,
 {
 	struct tegra_context *context = to_tegra_context(pcontext);
 
+	debug_printf("> %s(pcontext=%p, so=%p)\n", __func__, pcontext, so);
+
 	context->gpu->delete_vs_state(context->gpu, so);
+
+	debug_printf("< %s()\n", __func__);
 }
 
 static void *
@@ -227,8 +328,14 @@ tegra_create_gs_state(struct pipe_context *pcontext,
 		      const struct pipe_shader_state *cso)
 {
 	struct tegra_context *context = to_tegra_context(pcontext);
+	void *so;
+
+	debug_printf("> %s(pcontext=%p, cso=%p)\n", __func__, pcontext, cso);
+
+	so = context->gpu->create_gs_state(context->gpu, cso);
 
-	return context->gpu->create_gs_state(context->gpu, cso);
+	debug_printf("< %s() = %p\n", __func__, so);
+	return so;
 }
 
 static void
@@ -237,7 +344,11 @@ tegra_bind_gs_state(struct pipe_context *pcontext,
 {
 	struct tegra_context *context = to_tegra_context(pcontext);
 
+	debug_printf("> %s(pcontext=%p, so=%p)\n", __func__, pcontext, so);
+
 	context->gpu->bind_gs_state(context->gpu, so);
+
+	debug_printf("< %s()\n", __func__);
 }
 
 static void
@@ -246,7 +357,11 @@ tegra_delete_gs_state(struct pipe_context *pcontext,
 {
 	struct tegra_context *context = to_tegra_context(pcontext);
 
+	debug_printf("> %s(pcontext=%p, so=%p)\n", __func__, pcontext, so);
+
 	context->gpu->delete_gs_state(context->gpu, so);
+
+	debug_printf("< %s()\n", __func__);
 }
 
 static void *
@@ -255,10 +370,17 @@ tegra_create_vertex_elements_state(struct pipe_context *pcontext,
 				   const struct pipe_vertex_element *elements)
 {
 	struct tegra_context *context = to_tegra_context(pcontext);
+	void *so;
+
+	debug_printf("> %s(pcontext=%p, num_elements=%u, elements=%p)\n",
+		     __func__, pcontext, num_elements, elements);
 
-	return context->gpu->create_vertex_elements_state(context->gpu,
-							  num_elements,
-							  elements);
+	so = context->gpu->create_vertex_elements_state(context->gpu,
+							num_elements,
+							elements);
+
+	debug_printf("< %s() = %p\n", __func__, so);
+	return so;
 }
 
 static void
@@ -267,7 +389,11 @@ tegra_bind_vertex_elements_state(struct pipe_context *pcontext,
 {
 	struct tegra_context *context = to_tegra_context(pcontext);
 
+	debug_printf("> %s(pcontext=%p, so=%p)\n", __func__, pcontext, so);
+
 	context->gpu->bind_vertex_elements_state(context->gpu, so);
+
+	debug_printf("< %s()\n", __func__);
 }
 
 static void
@@ -276,7 +402,11 @@ tegra_delete_vertex_elements_state(struct pipe_context *pcontext,
 {
 	struct tegra_context *context = to_tegra_context(pcontext);
 
+	debug_printf("> %s(pcontext=%p, so=%p)\n", __func__, pcontext, so);
+
 	context->gpu->delete_vertex_elements_state(context->gpu, so);
+
+	debug_printf("< %s()\n", __func__);
 }
 
 static void
@@ -288,13 +418,20 @@ tegra_set_constant_buffer(struct pipe_context *pcontext,
 	struct tegra_context *context = to_tegra_context(pcontext);
 	struct pipe_constant_buffer buffer;
 
+	debug_printf("> %s(pcontext=%p, shader=%u, index=%u, buf=%p)\n",
+		     __func__, pcontext, shader, index, buf);
+
 	if (buf && buf->buffer) {
 		memcpy(&buffer, buf, sizeof(buffer));
+		debug_printf("  buffer: %p -> %p\n", buffer.buffer,
+			     tegra_resource_unwrap(buffer.buffer));
 		buffer.buffer = tegra_resource_unwrap(buffer.buffer);
 		buf = &buffer;
 	}
 
 	context->gpu->set_constant_buffer(context->gpu, shader, index, buf);
+
+	debug_printf("< %s()\n", __func__);
 }
 
 static void
@@ -305,21 +442,29 @@ tegra_set_framebuffer_state(struct pipe_context *pcontext,
 	struct pipe_framebuffer_state state;
 	unsigned i;
 
+	debug_printf("> %s(pcontext=%p, fb=%p)\n", __func__, pcontext, fb);
+
 	if (fb) {
 		memcpy(&state, fb, sizeof(state));
 
-		for (i = 0; i < fb->nr_cbufs; i++)
+		for (i = 0; i < fb->nr_cbufs; i++) {
 			state.cbufs[i] = tegra_surface_unwrap(fb->cbufs[i]);
+			debug_printf("  %u: %p -> %p\n", i, fb->cbufs[i],
+				     state.cbufs[i]);
+		}
 
 		while (i < PIPE_MAX_COLOR_BUFS)
 			state.cbufs[i++] = NULL;
 
 		state.zsbuf = tegra_surface_unwrap(fb->zsbuf);
+		debug_printf("  zsbuf: %p -> %p\n", fb->zsbuf, state.zsbuf);
 
 		fb = &state;
 	}
 
 	context->gpu->set_framebuffer_state(context->gpu, fb);
+
+	debug_printf("< %s()\n", __func__);
 }
 
 static void
@@ -328,7 +473,12 @@ tegra_set_polygon_stipple(struct pipe_context *pcontext,
 {
 	struct tegra_context *context = to_tegra_context(pcontext);
 
+	debug_printf("> %s(pcontext=%p, stipple=%p)\n", __func__, pcontext,
+		     stipple);
+
 	context->gpu->set_polygon_stipple(context->gpu, stipple);
+
+	debug_printf("< %s()\n", __func__);
 }
 
 static void
@@ -339,8 +489,13 @@ tegra_set_scissor_states(struct pipe_context *pcontext,
 {
 	struct tegra_context *context = to_tegra_context(pcontext);
 
+	debug_printf("> %s(pcontext=%p, start_slot=%u, num_scissors=%u, scissors=%p)\n",
+		     __func__, pcontext, start_slot, num_scissors, scissors);
+
 	context->gpu->set_scissor_states(context->gpu, start_slot,
 					 num_scissors, scissors);
+
+	debug_printf("< %s()\n", __func__);
 }
 
 static void
@@ -351,8 +506,14 @@ tegra_set_viewport_states(struct pipe_context *pcontext,
 {
 	struct tegra_context *context = to_tegra_context(pcontext);
 
+	debug_printf("> %s(pcontext=%p, start_slot=%u, num_viewports=%u, viewports=%p)\n",
+		     __func__, pcontext, start_slot, num_viewports,
+		     viewports);
+
 	context->gpu->set_viewport_states(context->gpu, start_slot,
 					  num_viewports, viewports);
+
+	debug_printf("< %s()\n", __func__);
 }
 
 static void
@@ -366,11 +527,17 @@ tegra_set_sampler_views(struct pipe_context *pcontext,
 	struct tegra_context *context = to_tegra_context(pcontext);
 	unsigned i;
 
+	debug_printf("> %s(pcontext=%p, shader=%u, start_slot=%u, num_views=%u, pviews=%p)\n",
+		     __func__, pcontext, shader, start_slot, num_views,
+		     pviews);
+
 	for (i = 0; i < num_views; i++)
 		views[i] = tegra_sampler_view_unwrap(pviews[i]);
 
 	context->gpu->set_sampler_views(context->gpu, shader, start_slot,
 					num_views, views);
+
+	debug_printf("< %s()\n", __func__);
 }
 
 static void
@@ -381,8 +548,13 @@ tegra_set_shader_resources(struct pipe_context *pcontext,
 {
 	struct tegra_context *context = to_tegra_context(pcontext);
 
+	debug_printf("> %s(pcontext=%p, startt=%u, count=%u, resources=%p)\n",
+		     __func__, pcontext, start, count, resources);
+
 	context->gpu->set_shader_resources(context->gpu, start, count,
 					   resources);
+
+	debug_printf("< %s()\n", __func__);
 }
 
 static void
@@ -395,17 +567,33 @@ tegra_set_vertex_buffers(struct pipe_context *pcontext,
 	struct pipe_vertex_buffer buf[PIPE_MAX_SHADER_INPUTS];
 	unsigned i;
 
+	debug_printf("> %s(pcontext=%p, start_slot=%u, num_buffers=%u, buffers=%p)\n",
+		     __func__, pcontext, start_slot, num_buffers, buffers);
+
 	if (num_buffers && buffers) {
+		for (i = 0; i < num_buffers; i++) {
+			debug_printf("  %u:\n", i);
+			debug_printf("    stride: %u\n", buffers[i].stride);
+			debug_printf("    offset: %u\n", buffers[i].buffer_offset);
+			debug_printf("    buffer: %p\n", buffers[i].buffer);
+			debug_printf("    user: %p\n", buffers[i].user_buffer);
+		}
+
 		memcpy(buf, buffers, num_buffers * sizeof(struct pipe_vertex_buffer));
 
-		for (i = 0; i < num_buffers; i++)
+		for (i = 0; i < num_buffers; i++) {
+			debug_printf("  %u: %p -> %p\n", i, buf[i].buffer,
+				     tegra_resource_unwrap(buf[i].buffer));
 			buf[i].buffer = tegra_resource_unwrap(buf[i].buffer);
+		}
 
 		buffers = buf;
 	}
 
 	context->gpu->set_vertex_buffers(context->gpu, start_slot,
 					 num_buffers, buffers);
+
+	debug_printf("< %s()\n", __func__);
 }
 
 static void
@@ -415,13 +603,20 @@ tegra_set_index_buffer(struct pipe_context *pcontext,
 	struct tegra_context *context = to_tegra_context(pcontext);
 	struct pipe_index_buffer buf;
 
+	debug_printf("> %s(pcontext=%p, buffer=%p)\n", __func__, pcontext,
+		     buffer);
+
 	if (buffer) {
 		memcpy(&buf, buffer, sizeof(buf));
+		debug_printf("  buffer: %p -> %p\n", buf.buffer,
+			     tegra_resource_unwrap(buf.buffer));
 		buf.buffer = tegra_resource_unwrap(buf.buffer);
 		buffer = &buf;
 	}
 
 	context->gpu->set_index_buffer(context->gpu, buffer);
+
+	debug_printf("< %s()\n", __func__);
 }
 
 static struct pipe_stream_output_target *
@@ -432,11 +627,19 @@ tegra_create_stream_output_target(struct pipe_context *pcontext,
 {
 	struct tegra_resource *resource = to_tegra_resource(presource);
 	struct tegra_context *context = to_tegra_context(pcontext);
+	struct pipe_stream_output_target *target;
 
-	return context->gpu->create_stream_output_target(context->gpu,
-							 resource->gpu,
-							 buffer_offset,
-							 buffer_size);
+	debug_printf("> %s(pcontext=%p, presource=%p, buffer_offset=%u, buffer_size=%u)\n",
+		     __func__, pcontext, presource, buffer_offset,
+		     buffer_size);
+
+	target = context->gpu->create_stream_output_target(context->gpu,
+							   resource->gpu,
+							   buffer_offset,
+							   buffer_size);
+
+	debug_printf("< %s() = %p\n", __func__, target);
+	return target;
 }
 
 static void
@@ -445,7 +648,12 @@ tegra_stream_output_target_destroy(struct pipe_context *pcontext,
 {
 	struct tegra_context *context = to_tegra_context(pcontext);
 
+	debug_printf("> %s(pcontext=%p, target=%p)\n", __func__, pcontext,
+		     target);
+
 	context->gpu->stream_output_target_destroy(context->gpu, target);
+
+	debug_printf("< %s()\n", __func__);
 }
 
 static void
@@ -456,8 +664,13 @@ tegra_set_stream_output_targets(struct pipe_context *pcontext,
 {
 	struct tegra_context *context = to_tegra_context(pcontext);
 
+	debug_printf("> %s(pcontext=%p, num_targets=%u, targets=%p, offsets=%p)\n",
+		     __func__, pcontext, num_targets, targets, offsets);
+
 	context->gpu->set_stream_output_targets(context->gpu, num_targets,
 						targets, offsets);
+
+	debug_printf("< %s()\n", __func__);
 }
 
 static void
@@ -467,6 +680,9 @@ tegra_blit(struct pipe_context *pcontext,
 	struct tegra_context *context = to_tegra_context(pcontext);
 	struct pipe_blit_info info;
 
+	debug_printf("> %s(pcontext=%p, pinfo=%p)\n", __func__, pcontext,
+		     pinfo);
+
 	if (pinfo) {
 		memcpy(&info, pinfo, sizeof(info));
 		info.dst.resource = tegra_resource_unwrap(info.dst.resource);
@@ -475,6 +691,8 @@ tegra_blit(struct pipe_context *pcontext,
 	}
 
 	context->gpu->blit(context->gpu, pinfo);
+
+	debug_printf("< %s()\n", __func__);
 }
 
 static void
@@ -486,7 +704,12 @@ tegra_clear(struct pipe_context *pcontext,
 {
 	struct tegra_context *context = to_tegra_context(pcontext);
 
+	debug_printf("> %s(pcontext=%p, buffers=%x, color=%p, depth=%f, stencil=%u)\n",
+		     __func__, pcontext, buffers, color, depth, stencil);
+
 	context->gpu->clear(context->gpu, buffers, color, depth, stencil);
+
+	debug_printf("< %s()\n", __func__);
 }
 
 static void
@@ -496,7 +719,12 @@ tegra_flush(struct pipe_context *pcontext,
 {
 	struct tegra_context *context = to_tegra_context(pcontext);
 
+	debug_printf("> %s(pcontext=%p, fence=%p, flags=%x)\n", __func__,
+		     pcontext, fence, flags);
+
 	context->gpu->flush(context->gpu, fence, flags);
+
+	debug_printf("< %s()\n", __func__);
 }
 
 static struct pipe_sampler_view *
@@ -508,6 +736,9 @@ tegra_create_sampler_view(struct pipe_context *pcontext,
 	struct tegra_context *context = to_tegra_context(pcontext);
 	struct tegra_sampler_view *view;
 
+	debug_printf("> %s(pcontext=%p, ptexture=%p, template=%p)\n",
+		     __func__, pcontext, ptexture, template);
+
 	view = calloc(1, sizeof(*view));
 	if (!view)
 		return NULL;
@@ -523,6 +754,7 @@ tegra_create_sampler_view(struct pipe_context *pcontext,
 	pipe_resource_reference(&view->base.texture, ptexture);
 	view->base.context = pcontext;
 
+	debug_printf("< %s() = %p\n", __func__, &view->base);
 	return &view->base;
 }
 
@@ -532,9 +764,14 @@ tegra_sampler_view_destroy(struct pipe_context *pcontext,
 {
 	struct tegra_sampler_view *view = to_tegra_sampler_view(pview);
 
+	debug_printf("> %s(pcontext=%p, view=%p)\n", __func__, pcontext,
+		     view);
+
 	pipe_resource_reference(&view->base.texture, NULL);
 	pipe_sampler_view_reference(&view->gpu, NULL);
 	free(view);
+
+	debug_printf("< %s()\n", __func__);
 }
 
 static void
@@ -544,7 +781,12 @@ tegra_flush_resource(struct pipe_context *pcontext,
 	struct tegra_resource *resource = to_tegra_resource(presource);
 	struct tegra_context *context = to_tegra_context(pcontext);
 
+	debug_printf("> %s(pcontext=%p, presource=%p)\n", __func__, pcontext,
+		     presource);
+
 	context->gpu->flush_resource(context->gpu, resource->gpu);
+
+	debug_printf("< %s()\n", __func__);
 }
 
 static void *
@@ -559,6 +801,10 @@ tegra_transfer_map(struct pipe_context *pcontext,
 	struct tegra_context *context = to_tegra_context(pcontext);
 	struct tegra_transfer *transfer;
 
+	debug_printf("> %s(pcontext=%p, presource=%p, level=%u, usage=%x, box=%p, ptransfer=%p)\n",
+		     __func__, pcontext, presource, level, usage, box,
+		     ptransfer);
+
 	transfer = calloc(1, sizeof(*transfer));
 	if (!transfer)
 		return NULL;
@@ -570,11 +816,14 @@ tegra_transfer_map(struct pipe_context *pcontext,
 						   box,
 						   &transfer->gpu);
 	memcpy(&transfer->base, transfer->gpu, sizeof(*transfer->gpu));
+	if (transfer->base.resource != NULL)
+		debug_printf("  resource: %p\n", transfer->base.resource);
 	transfer->base.resource = NULL;
 	pipe_resource_reference(&transfer->base.resource, presource);
 
 	*ptransfer = &transfer->base;
 
+	debug_printf("< %s() = %p\n", __func__, transfer->map);
 	return transfer->map;
 }
 
@@ -585,9 +834,14 @@ tegra_transfer_unmap(struct pipe_context *pcontext,
 	struct tegra_transfer *transfer = to_tegra_transfer(ptransfer);
 	struct tegra_context *context = to_tegra_context(pcontext);
 
+	debug_printf("> %s(pcontext=%p, ptransfer=%p)\n", __func__, pcontext,
+		     ptransfer);
+
 	context->gpu->transfer_unmap(context->gpu, transfer->gpu);
 	pipe_resource_reference(&transfer->base.resource, NULL);
 	free(transfer);
+
+	debug_printf("< %s()\n", __func__);
 }
 
 static void
@@ -603,9 +857,15 @@ tegra_transfer_inline_write(struct pipe_context *pcontext,
 	struct tegra_resource *resource = to_tegra_resource(presource);
 	struct tegra_context *context = to_tegra_context(pcontext);
 
+	debug_printf("> %s(pcontext=%p, presource=%p, level=%u, usage=%x, box=%p, data=%p, stride=%u, layer_stride=%u)\n",
+		     __func__, pcontext, presource, level, usage, box, data,
+		     stride, layer_stride);
+
 	context->gpu->transfer_inline_write(context->gpu, resource->gpu,
 					    level, usage, box, data, stride,
 					    layer_stride);
+
+	debug_printf("< %s()\n", __func__);
 }
 
 struct pipe_context *
@@ -614,6 +874,8 @@ tegra_context_create(struct pipe_screen *pscreen, void *priv)
 	struct tegra_screen *screen = to_tegra_screen(pscreen);
 	struct tegra_context *context;
 
+	debug_printf("> %s(pscreen=%p, priv=%p)\n", __func__, pscreen, priv);
+
 	context = calloc(1, sizeof(*context));
 	if (!context)
 		return NULL;
@@ -695,5 +957,6 @@ tegra_context_create(struct pipe_screen *pscreen, void *priv)
 	context->base.transfer_unmap = tegra_transfer_unmap;
 	context->base.transfer_inline_write = tegra_transfer_inline_write;
 
+	debug_printf("< %s() = %p\n", __func__, &context->base);
 	return &context->base;
 }
diff --git a/src/gallium/drivers/tegra/tegra_resource.c b/src/gallium/drivers/tegra/tegra_resource.c
index 8c5b7d4e41f..3b8accad9f3 100644
--- a/src/gallium/drivers/tegra/tegra_resource.c
+++ b/src/gallium/drivers/tegra/tegra_resource.c
@@ -47,6 +47,23 @@ tegra_resource_create(struct pipe_screen *pscreen,
 	struct tegra_screen *screen = to_tegra_screen(pscreen);
 	struct tegra_resource *resource;
 
+	debug_printf("> %s(pscreen=%p, template=%p)\n", __func__, pscreen,
+		     template);
+	/*
+	debug_printf("  template: %p\n", template);
+	debug_printf("    target: %u\n", template->target);
+	debug_printf("    format: %u\n", template->format);
+	debug_printf("    width: %u\n", template->width0);
+	debug_printf("    height: %u\n", template->height0);
+	debug_printf("    depth: %u\n", template->depth0);
+	debug_printf("    array_size: %u\n", template->array_size);
+	debug_printf("    last_level: %u\n", template->last_level);
+	debug_printf("    nr_samples: %u\n", template->nr_samples);
+	debug_printf("    usage: %u\n", template->usage);
+	debug_printf("    bind: %x\n", template->bind);
+	debug_printf("    flags: %x\n", template->flags);
+	*/
+
 	resource = calloc(1, sizeof(*resource));
 	if (!resource)
 		return NULL;
@@ -104,10 +121,13 @@ tegra_resource_create(struct pipe_screen *pscreen,
 			goto free;
 	}
 
+	debug_printf("  gpu: %p\n", resource->gpu);
+
 	memcpy(&resource->base, resource->gpu, sizeof(*resource->gpu));
 	pipe_reference_init(&resource->base.reference, 1);
 	resource->base.screen = &screen->base;
 
+	debug_printf("< %s() = %p\n", __func__, &resource->base);
 	return &resource->base;
 
 destroy:
@@ -125,6 +145,9 @@ tegra_resource_from_handle(struct pipe_screen *pscreen,
 	struct tegra_screen *screen = to_tegra_screen(pscreen);
 	struct tegra_resource *resource;
 
+	_debug_printf("> %s(pscreen=%p, template=%p, handle=%p)\n", __func__,
+		      pscreen, template);
+
 	resource = calloc(1, sizeof(*resource));
 	if (!resource)
 		return NULL;
@@ -141,6 +164,7 @@ tegra_resource_from_handle(struct pipe_screen *pscreen,
 	pipe_reference_init(&resource->base.reference, 1);
 	resource->base.screen = &screen->base;
 
+	_debug_printf("< %s() = %p\n", __func__, &resource->base);
 	return &resource->base;
 }
 
@@ -153,6 +177,9 @@ tegra_resource_get_handle(struct pipe_screen *pscreen,
 	struct tegra_screen *screen = to_tegra_screen(pscreen);
 	boolean ret = TRUE;
 
+	debug_printf("> %s(pscreen=%p, presource=%p, handle=%p)\n", __func__,
+		     pscreen, presource, handle);
+
 	if (presource->bind & PIPE_BIND_SCANOUT) {
 		handle->handle = resource->handle;
 		handle->stride = resource->stride;
@@ -162,6 +189,7 @@ tegra_resource_get_handle(struct pipe_screen *pscreen,
 						       handle);
 	}
 
+	debug_printf("< %s() = %d\n", __func__, ret);
 	return ret;
 }
 
@@ -171,8 +199,13 @@ tegra_resource_destroy(struct pipe_screen *pscreen,
 {
 	struct tegra_resource *resource = to_tegra_resource(presource);
 
+	debug_printf("> %s(pscreen=%p, presource=%p)\n", __func__, pscreen,
+		     presource);
+
 	pipe_resource_reference(&resource->gpu, NULL);
 	free(resource);
+
+	debug_printf("< %s()\n", __func__);
 }
 
 struct pipe_surface *
@@ -184,6 +217,9 @@ tegra_create_surface(struct pipe_context *pcontext,
 	struct tegra_context *context = to_tegra_context(pcontext);
 	struct tegra_surface *surface;
 
+	debug_printf("> %s(pcontext=%p, presource=%p, template=%p)\n",
+		     __func__, pcontext, presource, template);
+
 	surface = calloc(1, sizeof(*surface));
 	if (!surface)
 		return NULL;
@@ -204,6 +240,8 @@ tegra_create_surface(struct pipe_context *pcontext,
 	pipe_resource_reference(&surface->base.texture, presource);
 	surface->base.context = &context->base;
 
+	debug_printf("  gpu: %p\n", surface->gpu);
+	debug_printf("< %s() = %p\n", __func__, &surface->base);
 	return &surface->base;
 }
 
@@ -213,7 +251,12 @@ tegra_surface_destroy(struct pipe_context *pcontext,
 {
 	struct tegra_surface *surface = to_tegra_surface(psurface);
 
+	debug_printf("> %s(pcontext=%p, psurface=%p)\n", __func__, pcontext,
+		     psurface);
+
 	pipe_resource_reference(&surface->base.texture, NULL);
 	pipe_surface_reference(&surface->gpu, NULL);
 	free(surface);
+
+	debug_printf("< %s()\n", __func__);
 }
diff --git a/src/gallium/drivers/tegra/tegra_screen.c b/src/gallium/drivers/tegra/tegra_screen.c
index aa7bf65cb7e..b6394faa0c6 100644
--- a/src/gallium/drivers/tegra/tegra_screen.c
+++ b/src/gallium/drivers/tegra/tegra_screen.c
@@ -54,24 +54,42 @@ static void tegra_screen_destroy(struct pipe_screen *pscreen)
 {
 	struct tegra_screen *screen = to_tegra_screen(pscreen);
 
+	debug_printf("> %s(pscreen=%p)\n", __func__, pscreen);
+
 	screen->gpu->destroy(screen->gpu);
 	free(pscreen);
+
+	debug_printf("< %s()\n", __func__);
 }
 
 static int
 tegra_screen_get_param(struct pipe_screen *pscreen, enum pipe_cap param)
 {
 	struct tegra_screen *screen = to_tegra_screen(pscreen);
+	int ret;
+
+	debug_printf("> %s(pscreen=%p, param=%d)\n", __func__, pscreen,
+		     param);
 
-	return screen->gpu->get_param(screen->gpu, param);
+	ret = screen->gpu->get_param(screen->gpu, param);
+
+	debug_printf("< %s() = %d\n", __func__, ret);
+	return ret;
 }
 
 static float
 tegra_screen_get_paramf(struct pipe_screen *pscreen, enum pipe_capf param)
 {
 	struct tegra_screen *screen = to_tegra_screen(pscreen);
+	float ret;
+
+	debug_printf("> %s(pscreen=%p, param=%d)\n", __func__, pscreen,
+		     param);
+
+	ret = screen->gpu->get_paramf(screen->gpu, param);
 
-	return screen->gpu->get_paramf(screen->gpu, param);
+	debug_printf("< %s() = %f\n", __func__, ret);
+	return ret;
 }
 
 static int
@@ -80,8 +98,15 @@ tegra_screen_get_shader_param(struct pipe_screen *pscreen,
 			      enum pipe_shader_cap param)
 {
 	struct tegra_screen *screen = to_tegra_screen(pscreen);
+	int ret;
 
-	return screen->gpu->get_shader_param(screen->gpu, shader, param);
+	debug_printf("> %s(pscreen=%p, param=%d)\n", __func__, pscreen,
+		     param);
+
+	ret = screen->gpu->get_shader_param(screen->gpu, shader, param);
+
+	debug_printf("< %s() = %d\n", __func__, ret);
+	return ret;
 }
 
 static boolean
@@ -92,9 +117,16 @@ tegra_screen_is_format_supported(struct pipe_screen *pscreen,
 				 unsigned usage)
 {
 	struct tegra_screen *screen = to_tegra_screen(pscreen);
+	boolean ret;
+
+	debug_printf("> %s(pscreen=%p, format=%d, target=%d, sample_count=%u, usage=%x)\n",
+	       __func__, pscreen, format, target, sample_count, usage);
 
-	return screen->gpu->is_format_supported(screen->gpu, format, target,
-						sample_count, usage);
+	ret = screen->gpu->is_format_supported(screen->gpu, format, target,
+					       sample_count, usage);
+
+	debug_printf("< %s() = %d\n", __func__, ret);
+	return ret;
 }
 
 static void
@@ -104,7 +136,12 @@ tegra_fence_reference(struct pipe_screen *pscreen,
 {
 	struct tegra_screen *screen = to_tegra_screen(pscreen);
 
+	debug_printf("> %s(pscreen=%p, ptr=%p, fence=%p)\n", __func__,
+		     pscreen, ptr, fence);
+
 	screen->gpu->fence_reference(screen->gpu, ptr, fence);
+
+	debug_printf("< %s()\n", __func__);
 }
 
 static boolean
@@ -112,8 +149,15 @@ tegra_fence_signalled(struct pipe_screen *pscreen,
 		      struct pipe_fence_handle *fence)
 {
 	struct tegra_screen *screen = to_tegra_screen(pscreen);
+	boolean ret;
 
-	return screen->gpu->fence_signalled(screen->gpu, fence);
+	debug_printf("> %s(pscreen=%p, fence=%p)\n", __func__, pscreen,
+		     fence);
+
+	ret = screen->gpu->fence_signalled(screen->gpu, fence);
+
+	debug_printf("< %s() = %d\n", __func__, ret);
+	return ret;
 }
 
 static boolean
@@ -122,8 +166,15 @@ tegra_fence_finish(struct pipe_screen *pscreen,
 		   uint64_t timeout)
 {
 	struct tegra_screen *screen = to_tegra_screen(pscreen);
+	boolean ret;
+
+	debug_printf("> %s(pscreen=%p, fence=%p, timeout=%llu)\n", __func__,
+		     pscreen, fence, timeout);
 
-	return screen->gpu->fence_finish(screen->gpu, fence, timeout);
+	ret = screen->gpu->fence_finish(screen->gpu, fence, timeout);
+
+	debug_printf("< %s() = %d\n", __func__, ret);
+	return ret;
 }
 
 static struct udev_device *udev_device_new_from_fd(struct udev *udev, int fd)
@@ -151,14 +202,20 @@ static struct udev_device *udev_device_get_root(struct udev_device *device)
 {
 	struct udev_device *parent;
 
+	debug_printf("> %s(device=%p)\n", __func__, device);
+	debug_printf("  syspath: %s\n", udev_device_get_syspath(device));
+
 	while (true) {
 		parent = udev_device_get_parent(device);
 		if (!parent)
 			break;
 
+		debug_printf("  parent: %p\n", parent);
+		debug_printf("    syspath: %s\n", udev_device_get_syspath(parent));
 		device = parent;
 	}
 
+	debug_printf("< %s() = %p\n", __func__, device);
 	return device;
 }
 
@@ -187,6 +244,8 @@ static int tegra_open_render_node(int fd)
 		return -ENODEV;
 	}
 
+	debug_printf("path: %s\n", udev_device_get_devpath(display));
+
 	parent = udev_device_get_parent(display);
 	if (!parent) {
 		udev_device_unref(display);
@@ -194,6 +253,8 @@ static int tegra_open_render_node(int fd)
 		return -ENODEV;
 	}
 
+	debug_printf("parent: %s\n", udev_device_get_syspath(parent));
+
 	display = parent;
 
 	root = udev_device_get_root(display);
@@ -203,6 +264,8 @@ static int tegra_open_render_node(int fd)
 		return -ENODEV;
 	}
 
+	debug_printf("root: %s\n", udev_device_get_syspath(root));
+
 	enumerate = udev_enumerate_new(udev);
 	if (!enumerate) {
 		udev_device_unref(display);
@@ -216,6 +279,8 @@ static int tegra_open_render_node(int fd)
 
 	list = udev_enumerate_get_list_entry(enumerate);
 
+	debug_printf("devices:\n");
+
 	udev_list_entry_foreach(entry, list) {
 		const char *path = udev_list_entry_get_name(entry);
 		struct udev_device *device, *bus;
@@ -248,6 +313,8 @@ static int tegra_open_render_node(int fd)
 
 		/* both devices need to be on the same bus, though */
 		if (udev_device_match(bus, root)) {
+			debug_printf("match found: %s\n", path);
+
 			fd = open(path, O_RDWR);
 			if (fd < 0)
 				fd = -errno;
@@ -267,6 +334,8 @@ tegra_screen_create(int fd)
 {
 	struct tegra_screen *screen;
 
+	debug_printf("> %s()\n", __func__);
+
 	screen = calloc(1, sizeof(*screen));
 	if (!screen)
 		return NULL;
@@ -289,6 +358,9 @@ tegra_screen_create(int fd)
 		return NULL;
 	}
 
+	debug_printf("GPU: %p\n", screen->gpu);
+	debug_printf("  fd: %d\n", screen->gpu_fd);
+
 	screen->base.get_name = tegra_get_name;
 	screen->base.get_vendor = tegra_get_vendor;
 	screen->base.destroy = tegra_screen_destroy;
@@ -307,5 +379,6 @@ tegra_screen_create(int fd)
 	screen->base.fence_signalled = tegra_fence_signalled;
 	screen->base.fence_finish = tegra_fence_finish;
 
+	debug_printf("< %s() = %p\n", __func__, &screen->base);
 	return &screen->base;
 }
diff --git a/src/gallium/winsys/tegra/drm/tegra_drm_winsys.c b/src/gallium/winsys/tegra/drm/tegra_drm_winsys.c
index 99b0e164902..037a043360a 100644
--- a/src/gallium/winsys/tegra/drm/tegra_drm_winsys.c
+++ b/src/gallium/winsys/tegra/drm/tegra_drm_winsys.c
@@ -29,5 +29,12 @@ struct pipe_screen *tegra_drm_screen_create(int fd);
 
 struct pipe_screen *tegra_drm_screen_create(int fd)
 {
-	return tegra_screen_create(fd);
+	struct pipe_screen *screen;
+
+	debug_printf("> %s(fd=%d)\n", __func__, fd);
+
+	screen = tegra_screen_create(fd);
+
+	debug_printf("< %s() = %p\n", __func__, screen);
+	return screen;
 }