gtk/gsk/ngl/gsknglrenderer.c

314 lines
9.5 KiB
C
Raw Normal View History

gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
/* gsknglrenderer.c
*
* Copyright 2020 Christian Hergert <chergert@redhat.com>
*
* This file is free software; you can redistribute it and/or modify it under
* the terms of the GNU Lesser General Public License as published by the Free
* Software Foundation; either version 2.1 of the License, or (at your option)
* any later version.
*
* This file is distributed in the hope that it will be useful, but WITHOUT
* ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
* FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public
* License for more details.
*
* You should have received a copy of the GNU General Public License along
* with this program. If not, see <http://www.gnu.org/licenses/>.
*
* SPDX-License-Identifier: LGPL-2.1-or-later
*/
#include "config.h"
#include <gdk/gdkprofilerprivate.h>
#include <gdk/gdkdisplayprivate.h>
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
#include <gsk/gskdebugprivate.h>
#include <gsk/gskrendererprivate.h>
#include "gsknglcommandqueueprivate.h"
#include "gskngldriverprivate.h"
#include "gsknglprogramprivate.h"
#include "gsknglrenderjobprivate.h"
#include "gsknglrendererprivate.h"
struct _GskNglRendererClass
{
GskRendererClass parent_class;
};
struct _GskNglRenderer
{
GskRenderer parent_instance;
/* This context is used to swap buffers when we are rendering directly
* to a GDK surface. It is also used to locate the shared driver for
* the display that we use to drive the command queue.
*/
GdkGLContext *context;
/* Our command queue is private to this renderer and talks to the GL
* context for our target surface. This ensure that framebuffer 0 matches
* the surface we care about. Since the context is shared with other
* contexts from other renderers on the display, texture atlases,
* programs, and other objects are available to them all.
*/
GskNglCommandQueue *command_queue;
/* The driver manages our program state and command queues. It also
* deals with caching textures, shaders, shadows, glyph, and icon
* caches through various helpers.
*/
GskNglDriver *driver;
};
G_DEFINE_TYPE (GskNglRenderer, gsk_ngl_renderer, GSK_TYPE_RENDERER)
/**
* gsk_ngl_renderer_new:
*
* Creates a new `GskRenderer` using the new OpenGL renderer.
*
* Returns: a new NGL renderer
*
* Since: 4.2
*/
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
GskRenderer *
gsk_ngl_renderer_new (void)
{
return g_object_new (GSK_TYPE_NGL_RENDERER, NULL);
}
static gboolean
gsk_ngl_renderer_realize (GskRenderer *renderer,
GdkSurface *surface,
GError **error)
{
G_GNUC_UNUSED gint64 start_time = GDK_PROFILER_CURRENT_TIME;
GskNglRenderer *self = (GskNglRenderer *)renderer;
GdkGLContext *context = NULL;
GskNglDriver *driver = NULL;
gboolean ret = FALSE;
gboolean debug_shaders = FALSE;
g_assert (GSK_IS_NGL_RENDERER (self));
g_assert (GDK_IS_SURFACE (surface));
if (self->context != NULL)
return TRUE;
g_assert (self->driver == NULL);
g_assert (self->context == NULL);
g_assert (self->command_queue == NULL);
if (!(context = gdk_surface_create_gl_context (surface, error)) ||
!gdk_gl_context_realize (context, error))
goto failure;
#ifdef G_ENABLE_DEBUG
if (GSK_RENDERER_DEBUG_CHECK (GSK_RENDERER (self), SHADERS))
debug_shaders = TRUE;
#endif
if (!(driver = gsk_ngl_driver_for_display (gdk_surface_get_display (surface), debug_shaders, error)))
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
goto failure;
self->command_queue = gsk_ngl_driver_create_command_queue (driver, context);
self->context = g_steal_pointer (&context);
self->driver = g_steal_pointer (&driver);
gsk_ngl_command_queue_set_profiler (self->command_queue,
gsk_renderer_get_profiler (renderer));
ret = TRUE;
failure:
g_clear_object (&driver);
g_clear_object (&context);
gdk_profiler_end_mark (start_time, "realize GskNglRenderer", NULL);
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
return ret;
}
static void
gsk_ngl_renderer_unrealize (GskRenderer *renderer)
{
GskNglRenderer *self = (GskNglRenderer *)renderer;
g_assert (GSK_IS_NGL_RENDERER (renderer));
gdk_gl_context_make_current (self->context);
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
g_clear_object (&self->driver);
g_clear_object (&self->command_queue);
g_clear_object (&self->context);
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
}
static cairo_region_t *
get_render_region (GdkSurface *surface,
GdkGLContext *context)
{
const cairo_region_t *damage;
GdkRectangle whole_surface;
GdkRectangle extents;
g_assert (GDK_IS_SURFACE (surface));
g_assert (GDK_IS_GL_CONTEXT (context));
whole_surface.x = 0;
whole_surface.y = 0;
whole_surface.width = gdk_surface_get_width (surface);
whole_surface.height = gdk_surface_get_height (surface);
/* Damage does not have scale factor applied so we can compare it to
* @whole_surface which also doesn't have the scale factor applied.
*/
damage = gdk_draw_context_get_frame_region (GDK_DRAW_CONTEXT (context));
if (cairo_region_contains_rectangle (damage, &whole_surface) == CAIRO_REGION_OVERLAP_IN)
return NULL;
/* If the extents match the full-scene, do the same as above */
gsk: add OpenGL based GskNglRenderer The primary goal here was to cleanup the current GL renderer to make maintenance easier going forward. Furthermore, it tracks state to allow us to implement more advanced renderer features going forward. Reordering This renderer will reorder batches by render target to reduce the number of times render targets are changed. In the future, we could also reorder by program within the render target if we can determine that vertices do not overlap. Uniform Snapshots To allow for reordering of batches all uniforms need to be tracked for the programs. This allows us to create the full uniform state when the batch has been moved into a new position. Some care was taken as it can be performance sensitive. Attachment Snapshots Similar to uniform snapshots, we need to know all of the texture attachments so that we can rebind them when necessary. Render Jobs To help isolate the process of creating GL commands from the renderer abstraction a render job abstraction was added. This could be extended in the future if we decided to do tiling. Command Queue Render jobs create batches using the command queue. The command queue will snapshot uniform and attachment state so that it can reorder batches right before executing them. Currently, the only reordering done is to ensure that we only visit each render target once. We could extend this by tracking vertices, attachments, and others. This code currently uses an inline array helper to reduce overhead from GArray which was showing up on profiles. It could be changed to use GdkArray without too much work, but had roughly double the instructions. Cycle counts have not yet been determined. GLSL Programs This was simplified to use XMACROS so that we can just extend one file (gskglprograms.defs) instead of multiple places. The programs are added as fields in the driver for easy access. Driver The driver manages textures, render targets, access to atlases, programs, and more. There is one driver per display, by using the shared GL context. Some work could be done here to batch uploads so that we make fewer calls to upload when sending icon theme data to the GPU. We'd need to keep a copy of the atlas data for such purposes.
2020-12-19 01:36:59 +00:00
cairo_region_get_extents (damage, &extents);
if (gdk_rectangle_equal (&extents, &whole_surface))
return NULL;
/* Draw clipped to the bounding-box of the region. */
return cairo_region_create_rectangle (&extents);
}
static void
gsk_ngl_renderer_render (GskRenderer *renderer,
GskRenderNode *root,
const cairo_region_t *update_area)
{
GskNglRenderer *self = (GskNglRenderer *)renderer;
cairo_region_t *render_region;
graphene_rect_t viewport;
GskNglRenderJob *job;
GdkSurface *surface;
float scale_factor;
g_assert (GSK_IS_NGL_RENDERER (renderer));
g_assert (root != NULL);
surface = gdk_draw_context_get_surface (GDK_DRAW_CONTEXT (self->context));
scale_factor = gdk_surface_get_scale_factor (surface);
viewport.origin.x = 0;
viewport.origin.y = 0;
viewport.size.width = gdk_surface_get_width (surface) * scale_factor;
viewport.size.height = gdk_surface_get_height (surface) * scale_factor;
gdk_gl_context_make_current (self->context);
gdk_draw_context_begin_frame (GDK_DRAW_CONTEXT (self->context), update_area);
/* Must be called *AFTER* gdk_draw_context_begin_frame() */
render_region = get_render_region (surface, self->context);
gsk_ngl_driver_begin_frame (self->driver, self->command_queue);
job = gsk_ngl_render_job_new (self->driver, &viewport, scale_factor, render_region, 0);
#ifdef G_ENABLE_DEBUG
if (GSK_RENDERER_DEBUG_CHECK (GSK_RENDERER (self), FALLBACK))
gsk_ngl_render_job_set_debug_fallback (job, TRUE);
#endif
gsk_ngl_render_job_render (job, root);
gsk_ngl_driver_end_frame (self->driver);
gsk_ngl_render_job_free (job);
gdk_gl_context_make_current (self->context);
gdk_draw_context_end_frame (GDK_DRAW_CONTEXT (self->context));
gsk_ngl_driver_after_frame (self->driver);
cairo_region_destroy (render_region);
}
static GdkTexture *
gsk_ngl_renderer_render_texture (GskRenderer *renderer,
GskRenderNode *root,
const graphene_rect_t *viewport)
{
GskNglRenderer *self = (GskNglRenderer *)renderer;
GskNglRenderTarget *render_target;
GskNglRenderJob *job;
GdkTexture *texture = NULL;
guint texture_id;
int width;
int height;
g_assert (GSK_IS_NGL_RENDERER (renderer));
g_assert (root != NULL);
width = ceilf (viewport->size.width);
height = ceilf (viewport->size.height);
if (gsk_ngl_driver_create_render_target (self->driver,
width, height,
GL_NEAREST, GL_NEAREST,
&render_target))
{
gsk_ngl_driver_begin_frame (self->driver, self->command_queue);
job = gsk_ngl_render_job_new (self->driver, viewport, 1, NULL, render_target->framebuffer_id);
#ifdef G_ENABLE_DEBUG
if (GSK_RENDERER_DEBUG_CHECK (GSK_RENDERER (self), FALLBACK))
gsk_ngl_render_job_set_debug_fallback (job, TRUE);
#endif
gsk_ngl_render_job_render_flipped (job, root);
texture_id = gsk_ngl_driver_release_render_target (self->driver, render_target, FALSE);
texture = gsk_ngl_driver_create_gdk_texture (self->driver, texture_id);
gsk_ngl_driver_end_frame (self->driver);
gsk_ngl_render_job_free (job);
gsk_ngl_driver_after_frame (self->driver);
}
return g_steal_pointer (&texture);
}
static void
gsk_ngl_renderer_dispose (GObject *object)
{
#ifdef G_ENABLE_DEBUG
GskNglRenderer *self = (GskNglRenderer *)object;
g_assert (self->driver == NULL);
#endif
G_OBJECT_CLASS (gsk_ngl_renderer_parent_class)->dispose (object);
}
static void
gsk_ngl_renderer_class_init (GskNglRendererClass *klass)
{
GObjectClass *object_class = G_OBJECT_CLASS (klass);
GskRendererClass *renderer_class = GSK_RENDERER_CLASS (klass);
object_class->dispose = gsk_ngl_renderer_dispose;
renderer_class->realize = gsk_ngl_renderer_realize;
renderer_class->unrealize = gsk_ngl_renderer_unrealize;
renderer_class->render = gsk_ngl_renderer_render;
renderer_class->render_texture = gsk_ngl_renderer_render_texture;
}
static void
gsk_ngl_renderer_init (GskNglRenderer *self)
{
}
gboolean
gsk_ngl_renderer_try_compile_gl_shader (GskNglRenderer *renderer,
GskGLShader *shader,
GError **error)
{
GskNglProgram *program;
g_return_val_if_fail (GSK_IS_NGL_RENDERER (renderer), FALSE);
g_return_val_if_fail (shader != NULL, FALSE);
program = gsk_ngl_driver_lookup_shader (renderer->driver, shader, error);
return program != NULL;
}