Merge branch 'gles2-scissor' into 'master'

render/gles2: scissor region on render

See merge request wlroots/wlroots!5294

render/gles2/pass.c

@@ -90,6 +90,10 @@ static void render(const struct wlr_box *box, const pixman_region32_t *clip, GLi
 		return;
 	}
 
+	glEnable(GL_SCISSOR_TEST);
+	glScissor(region.extents.x1, region.extents.y1, region.extents.x2 - region.extents.x1,
+		region.extents.y2 - region.extents.y1);
+
 	glEnableVertexAttribArray(attrib);
 
 	for (int i = 0; i < rects_len;) {
@@ -120,6 +124,7 @@ static void render(const struct wlr_box *box, const pixman_region32_t *clip, GLi
 	}
 
 	glDisableVertexAttribArray(attrib);
+	glDisable(GL_SCISSOR_TEST);
 
 	pixman_region32_fini(&region);
 }

render/gles2/renderer.c

@@ -326,7 +326,6 @@ static int gles2_get_render_time(struct wlr_render_timer *wlr_timer) {
 	renderer->procs.glGetQueryObjectivEXT(timer->id,
 		GL_QUERY_RESULT_AVAILABLE_EXT, &available);
 	if (!available) {
-		wlr_log(WLR_ERROR, "timer was read too early, gpu isn't done!");
 		wlr_egl_restore_context(&prev_ctx);
 		return -1;
 	}

test/bench_render_pass.c

@@ -0,0 +1,94 @@
+#include <assert.h>
+#include <drm_fourcc.h>
+#include <pixman.h>
+#include <wlr/backend/headless.h>
+#include <wlr/render/allocator.h>
+#include <wlr/render/drm_format_set.h>
+#include <wlr/render/swapchain.h>
+#include <wlr/render/wlr_renderer.h>
+#include <wlr/util/log.h>
+
+int main(int argc, char *argv[]) {
+	struct wl_event_loop *ev = wl_event_loop_create();
+	assert(ev);
+
+	wlr_log_init(WLR_INFO, NULL);
+
+	struct wlr_backend *backend = wlr_headless_backend_create(ev);
+	assert(backend);
+
+	struct wlr_renderer *renderer = wlr_renderer_autocreate(backend);
+	assert(renderer);
+
+	struct wlr_allocator *allocator = wlr_allocator_autocreate(backend, renderer);
+	assert(allocator);
+
+	const struct wlr_drm_format_set *formats = wlr_renderer_get_texture_formats(renderer,
+		allocator->buffer_caps);
+
+	// TODO: swapchain size from argument
+	// TODO: swapchain format from argument?
+	struct wlr_swapchain *swapchain = wlr_swapchain_create(allocator, 1920, 1080,
+		wlr_drm_format_set_get(formats, DRM_FORMAT_ARGB8888));
+	assert(swapchain);
+
+	const int iters = 16;
+	for (int i = 1; i <= iters; ++i) {
+		pixman_region32_t clip;
+		pixman_region32_init(&clip);
+
+		for (int ii = 0; ii < i * 10; ++ii) {
+			pixman_region32_union_rect(&clip, &clip, 0, ii, ii, 1);
+		}
+
+		struct wlr_render_timer *timer = wlr_render_timer_create(renderer);
+		assert(timer);
+
+		const struct wlr_buffer_pass_options options = {
+			.timer = timer,
+		};
+
+		struct wlr_buffer *buffer = wlr_swapchain_acquire(swapchain);
+		assert(buffer);
+
+		struct wlr_render_pass *pass = wlr_renderer_begin_buffer_pass(renderer, buffer, &options);
+
+		struct wlr_render_rect_options data = {
+			.box = {
+				.x = 0, .y = 0,
+				.width = i * 10, .height = i * 10,
+			},
+			.clip = &clip,
+			.color = { .r = 1, .g = 0.5, .b = 0.1, .a = 0.5 },
+		};
+
+		const int rects = 1 << i;
+
+		for (int j = 0; j < rects; ++j) {
+			wlr_render_pass_add_rect(pass, &data);
+		}
+
+		wlr_render_pass_submit(pass);
+
+		wlr_buffer_unlock(buffer);
+
+		int duration = -1;
+		while ((duration = wlr_render_timer_get_duration_ns(timer)) == -1);
+
+		wlr_log(WLR_INFO, "wlr_render_pass iteration %d/%d (%d rects): %d ns", i, iters, rects,
+			duration);
+
+		wlr_render_timer_destroy(timer);
+
+		pixman_region32_fini(&clip);
+	}
+
+	wlr_swapchain_destroy(swapchain);
+	wlr_allocator_destroy(allocator);
+	wlr_renderer_destroy(renderer);
+	wlr_backend_destroy(backend);
+
+	wl_event_loop_destroy(ev);
+
+	return 0;
+}

test/meson.build

@@ -8,3 +8,9 @@ benchmark(
 	executable('bench-scene', 'bench_scene.c', dependencies: wlroots),
 	timeout: 30,
 )
+
+benchmark(
+	'render-pass',
+	executable('bench-render-pass', 'bench_render_pass.c', dependencies: wlroots),
+	timeout: 30,
+)