mutter/clutter/glx/clutter-glx-texture-pixmap.c

935 lines
26 KiB
C
Raw Normal View History

/*
* Clutter.
*
* An OpenGL based 'interactive canvas' library.
*
* Authored By Johan Bilien <johan.bilien@nokia.com>
* Matthew Allum <mallum@o-hand.com>
* Robert Bragg <bob@o-hand.com>
*
* Copyright (C) 2007 OpenedHand
*
* This library 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 of the License, or (at your option) any later version.
*
* This library 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 Lesser General Public
* License along with this library. If not, see <http://www.gnu.org/licenses/>.
*
*
*/
/* TODO:
* - Automagically handle named pixmaps, and window resizes (i.e
* essentially handle window id's being passed in) ?
*/
/**
* SECTION:clutter-glx-texture-pixmap
* @short_description: A texture which displays the content of an X Pixmap.
*
* #ClutterGLXTexturePixmap is a class for displaying the content of an
* X Pixmap as a ClutterActor. Used together with the X Composite extension,
* it allows to display the content of X Windows inside Clutter.
*
* The class requires the GLX_EXT_texture_from_pixmap OpenGL extension
* (http://people.freedesktop.org/~davidr/GLX_EXT_texture_from_pixmap.txt)
*
* The GL_ARB_texture_non_power_of_two extension will be used if it is
* available. Otherwise it will try to use the
* GL_ARB_texture_rectangle extension. If both are available you can
* force it to prefer the rectangle extension by setting the
* CLUTTER_PIXMAP_TEXTURE_RECTANGLE to 'force'. To prevent it ever
* using the rectangle extension you can set it to 'disable'.
*/
#ifdef HAVE_CONFIG_H
#include "config.h"
#endif
#include <string.h>
#include "../x11/clutter-x11-texture-pixmap.h"
#include "clutter-glx-texture-pixmap.h"
#include "clutter-glx.h"
#include "clutter-backend-glx.h"
#include "../clutter-util.h"
#include "../clutter-debug.h"
Adds initial clipped redraw support to Clutter A new (internal only currently) API, _clutter_actor_queue_clipped_redraw can be used to queue a redraw along with a clip rectangle in actor coordinates. This clip rectangle propagates up to the stage and clutter backend which may optionally use the information to optimize stage redraws. The GLX backend in particular may scissor the next redraw to the clip rectangle and use GLX_MESA_copy_sub_buffer to present the stage subregion. The intention is that any actors that can naturally determine the bounds of updates should queue clipped redraws to reduce the cost of updating small regions of the screen. Notes: » If GLX_MESA_copy_sub_buffer isn't available then the GLX backend ignores any clip rectangles. » queuing multiple clipped redraws will result in the bounding box of each clip rectangle being used. » If a clipped redraw has a height > 300 pixels then it's promoted into a full stage redraw, so that the GPU doesn't end up blocking too long waiting for the vsync to reach the optimal position to avoid tearing. » Note: no empirical data was used to come up with this threshold so we may need to tune this. » Currently only ClutterX11TexturePixmap makes use of this new API. This is done via a new "queue-damage-redraw" signal that is emitted when the pixmap is updated. The default handler queues a clipped redraw with the assumption that the pixmap is being painted as a rectangle covering the actors transformed allocation. If you subclass ClutterX11TexturePixmap and change how it's painted you now also need to override the signal handler and queue your own redraw. Technically this is a semantic break, but it's assumed that no one is currently doing this. This still leaves a few unsolved issues with regards to optimizing sub stage redraws that need to be addressed in further work so this can only be considered a stepping stone a this point: » Because we have no reliable way to determine if the painting of any given actor is being modified any optimizations implemented using _clutter_actor_queue_redraw_with_clip must be overridable by a subclass, and technically must be opt-in for existing classes to avoid a change in semantics. E.g. consider that a user connects to the paint signal for ClutterTexture and paints a circle instead of a rectangle. In this case any original logic to queue clipped redraws would be incorrect. » Currently only the implementation of an actor has enough information with which to queue clipped redraws. E.g. It is not possible for generic code in clutter-actor.c to queue a clipped redraw when hiding an actor because actors have no way to report a "paint box". (remember actors can draw outside their allocation and actors with depth may also be projected outside of their allocation) » The current plan is to add a actor_class->get_paint_cuboid() virtual so actors can report a bounding cube for everything they would draw in their current state and use that to queue clipped redraws against the stage by projecting the paint cube into stage coordinates. » Our heuristics for promoting clipped redraws into full redraws to avoid blocking the GPU while we wait for the vsync need improving: » vsync issues aren't relevant for redirected/composited applications so they should use different heuristics. In this case we instead need to trade off the cost of blitting when using glXCopySubBuffer vs promoting to a full redraw and flipping instead.
2009-11-30 17:47:55 +00:00
#include "../clutter-private.h"
#include "cogl/cogl.h"
typedef void (*BindTexImage) (Display *display,
GLXDrawable drawable,
int buffer,
int *attribList);
typedef void (*ReleaseTexImage) (Display *display,
GLXDrawable drawable,
int buffer);
typedef void (*GenerateMipmap) (GLenum target);
typedef enum
{
CLUTTER_GLX_RECTANGLE_DISALLOW,
CLUTTER_GLX_RECTANGLE_ALLOW,
CLUTTER_GLX_RECTANGLE_FORCE
} RectangleState;
Adds initial clipped redraw support to Clutter A new (internal only currently) API, _clutter_actor_queue_clipped_redraw can be used to queue a redraw along with a clip rectangle in actor coordinates. This clip rectangle propagates up to the stage and clutter backend which may optionally use the information to optimize stage redraws. The GLX backend in particular may scissor the next redraw to the clip rectangle and use GLX_MESA_copy_sub_buffer to present the stage subregion. The intention is that any actors that can naturally determine the bounds of updates should queue clipped redraws to reduce the cost of updating small regions of the screen. Notes: » If GLX_MESA_copy_sub_buffer isn't available then the GLX backend ignores any clip rectangles. » queuing multiple clipped redraws will result in the bounding box of each clip rectangle being used. » If a clipped redraw has a height > 300 pixels then it's promoted into a full stage redraw, so that the GPU doesn't end up blocking too long waiting for the vsync to reach the optimal position to avoid tearing. » Note: no empirical data was used to come up with this threshold so we may need to tune this. » Currently only ClutterX11TexturePixmap makes use of this new API. This is done via a new "queue-damage-redraw" signal that is emitted when the pixmap is updated. The default handler queues a clipped redraw with the assumption that the pixmap is being painted as a rectangle covering the actors transformed allocation. If you subclass ClutterX11TexturePixmap and change how it's painted you now also need to override the signal handler and queue your own redraw. Technically this is a semantic break, but it's assumed that no one is currently doing this. This still leaves a few unsolved issues with regards to optimizing sub stage redraws that need to be addressed in further work so this can only be considered a stepping stone a this point: » Because we have no reliable way to determine if the painting of any given actor is being modified any optimizations implemented using _clutter_actor_queue_redraw_with_clip must be overridable by a subclass, and technically must be opt-in for existing classes to avoid a change in semantics. E.g. consider that a user connects to the paint signal for ClutterTexture and paints a circle instead of a rectangle. In this case any original logic to queue clipped redraws would be incorrect. » Currently only the implementation of an actor has enough information with which to queue clipped redraws. E.g. It is not possible for generic code in clutter-actor.c to queue a clipped redraw when hiding an actor because actors have no way to report a "paint box". (remember actors can draw outside their allocation and actors with depth may also be projected outside of their allocation) » The current plan is to add a actor_class->get_paint_cuboid() virtual so actors can report a bounding cube for everything they would draw in their current state and use that to queue clipped redraws against the stage by projecting the paint cube into stage coordinates. » Our heuristics for promoting clipped redraws into full redraws to avoid blocking the GPU while we wait for the vsync need improving: » vsync issues aren't relevant for redirected/composited applications so they should use different heuristics. In this case we instead need to trade off the cost of blitting when using glXCopySubBuffer vs promoting to a full redraw and flipping instead.
2009-11-30 17:47:55 +00:00
enum
{
PROP_0,
PROP_AUTO_REDRAW,
};
static BindTexImage _gl_bind_tex_image = NULL;
static ReleaseTexImage _gl_release_tex_image = NULL;
static GenerateMipmap _gl_generate_mipmap = NULL;
static gboolean _have_tex_from_pixmap_ext = FALSE;
static gboolean _ext_check_done = FALSE;
static gboolean _have_tex_rectangle = FALSE;
static RectangleState _rectangle_state = CLUTTER_GLX_RECTANGLE_ALLOW;
struct _ClutterGLXTexturePixmapPrivate
{
GLXPixmap glx_pixmap;
gboolean use_fallback;
gboolean bound;
gint can_mipmap;
gboolean mipmap_generate_queued;
gboolean bind_tex_image_queued;
gboolean using_rectangle;
};
static void
clutter_glx_texture_pixmap_update_area (ClutterX11TexturePixmap *texture,
gint x,
gint y,
gint width,
gint height);
static void
clutter_glx_texture_pixmap_create_glx_pixmap (ClutterGLXTexturePixmap *tex);
static ClutterX11TexturePixmapClass *parent_class = NULL;
G_DEFINE_TYPE (ClutterGLXTexturePixmap, \
clutter_glx_texture_pixmap, \
CLUTTER_X11_TYPE_TEXTURE_PIXMAP);
static void
bind_texture (ClutterGLXTexturePixmap *tex)
{
GLuint handle = 0;
GLenum target = 0;
CoglHandle cogl_tex =
clutter_texture_get_cogl_texture (CLUTTER_TEXTURE (tex));
if (!cogl_texture_get_gl_texture (cogl_tex, &handle, &target))
g_warning ("Failed to pluck out GL handle from cogl texture to bind");
glBindTexture (target, handle);
}
static void
release_tex_image (ClutterGLXTexturePixmap *texture)
{
ClutterGLXTexturePixmapPrivate *priv = texture->priv;
Display *dpy;
if (!priv->glx_pixmap || !priv->bound)
return;
dpy = clutter_x11_get_default_display ();
bind_texture (texture);
clutter_x11_trap_x_errors ();
(_gl_release_tex_image) (dpy,
priv->glx_pixmap,
GLX_FRONT_LEFT_EXT);
XSync (dpy, FALSE);
if (clutter_x11_untrap_x_errors ())
CLUTTER_NOTE (TEXTURE, "Failed to release?");
CLUTTER_NOTE (TEXTURE, "Destroyed pxm: %li", priv->glx_pixmap);
priv->bound = FALSE;
}
static void
on_glx_texture_pixmap_pre_paint (ClutterGLXTexturePixmap *texture,
gpointer user_data)
{
ClutterGLXTexturePixmapPrivate *priv = texture->priv;
gboolean tex_bound = FALSE;
Display *dpy = clutter_x11_get_default_display();
if (priv->bind_tex_image_queued)
{
CLUTTER_NOTE (TEXTURE, "Really updating via GLX");
bind_texture (CLUTTER_GLX_TEXTURE_PIXMAP (texture));
tex_bound = TRUE;
clutter_x11_trap_x_errors ();
(_gl_bind_tex_image) (dpy,
priv->glx_pixmap,
GLX_FRONT_LEFT_EXT,
NULL);
XSync (dpy, FALSE);
/* Note above fires X error for non name pixmaps - but
* things still seem to work - i.e pixmap updated
*/
if (clutter_x11_untrap_x_errors ())
CLUTTER_NOTE (TEXTURE, "Update bind_tex_image failed");
priv->bound = TRUE;
if (clutter_texture_get_filter_quality (CLUTTER_TEXTURE (texture))
== CLUTTER_TEXTURE_QUALITY_HIGH)
{
priv->mipmap_generate_queued++;
}
priv->bind_tex_image_queued = FALSE;
}
if (_gl_generate_mipmap &&
priv->can_mipmap &&
priv->mipmap_generate_queued)
{
GLuint handle = 0;
GLenum target = 0;
CoglHandle cogl_tex;
cogl_tex = clutter_texture_get_cogl_texture
(CLUTTER_TEXTURE(texture));
bind_texture (texture);
tex_bound = TRUE;
cogl_texture_get_gl_texture (cogl_tex, &handle, &target);
_gl_generate_mipmap (target);
}
priv->mipmap_generate_queued = FALSE;
[cogl] Move the texture filters to be a property of the material layer The texture filters are now a property of the material layer rather than the texture object. Whenever a texture is painted with a material it sets the filters on all of the GL textures in the Cogl texture. The filter is cached so that it won't be changed unnecessarily. The automatic mipmap generation has changed so that the mipmaps are only generated when the texture is painted instead of every time the data changes. Changing the texture sets a flag to mark that the mipmaps are dirty. This works better if the FBO extension is available because we can use glGenerateMipmap. If the extension is not available it will temporarily enable automatic mipmap generation and reupload the first pixel of each slice. This requires tracking the data for the first pixel. The COGL_TEXTURE_AUTO_MIPMAP flag has been replaced with COGL_TEXTURE_NO_AUTO_MIPMAP so that it will default to auto-mipmapping. The mipmap generation is now effectively free if you are not using a mipmap filter mode so you would only want to disable it if you had some special reason to generate your own mipmaps. ClutterTexture no longer has to store its own copy of the filter mode. Instead it stores it in the material and the property is directly set and read from that. This fixes problems with the filters getting out of sync when a cogl handle is set on the texture directly. It also avoids the mess of having to rerealize the texture if the filter quality changes to HIGH because Cogl will take of generating the mipmaps if needed.
2009-06-04 15:04:57 +00:00
/* Disable mipmaps if we can't support them */
if (clutter_texture_get_filter_quality (CLUTTER_TEXTURE (texture))
== CLUTTER_TEXTURE_QUALITY_HIGH
&& !priv->can_mipmap)
[cogl] Move the texture filters to be a property of the material layer The texture filters are now a property of the material layer rather than the texture object. Whenever a texture is painted with a material it sets the filters on all of the GL textures in the Cogl texture. The filter is cached so that it won't be changed unnecessarily. The automatic mipmap generation has changed so that the mipmaps are only generated when the texture is painted instead of every time the data changes. Changing the texture sets a flag to mark that the mipmaps are dirty. This works better if the FBO extension is available because we can use glGenerateMipmap. If the extension is not available it will temporarily enable automatic mipmap generation and reupload the first pixel of each slice. This requires tracking the data for the first pixel. The COGL_TEXTURE_AUTO_MIPMAP flag has been replaced with COGL_TEXTURE_NO_AUTO_MIPMAP so that it will default to auto-mipmapping. The mipmap generation is now effectively free if you are not using a mipmap filter mode so you would only want to disable it if you had some special reason to generate your own mipmaps. ClutterTexture no longer has to store its own copy of the filter mode. Instead it stores it in the material and the property is directly set and read from that. This fixes problems with the filters getting out of sync when a cogl handle is set on the texture directly. It also avoids the mess of having to rerealize the texture if the filter quality changes to HIGH because Cogl will take of generating the mipmaps if needed.
2009-06-04 15:04:57 +00:00
{
CoglHandle material
= clutter_texture_get_cogl_material (CLUTTER_TEXTURE (texture));
cogl_material_set_layer_filters (material, 0,
COGL_MATERIAL_FILTER_LINEAR,
COGL_MATERIAL_FILTER_LINEAR);
}
}
static void
clutter_glx_texture_pixmap_init (ClutterGLXTexturePixmap *self)
{
ClutterGLXTexturePixmapPrivate *priv;
priv = self->priv =
G_TYPE_INSTANCE_GET_PRIVATE (self,
CLUTTER_GLX_TYPE_TEXTURE_PIXMAP,
ClutterGLXTexturePixmapPrivate);
g_signal_connect (CLUTTER_ACTOR(self),
"paint", G_CALLBACK (on_glx_texture_pixmap_pre_paint),
NULL);
if (_ext_check_done == FALSE)
{
const char *gl_extensions = NULL;
const gchar *glx_extensions = NULL;
const char *rect_env;
glx_extensions =
glXQueryExtensionsString (clutter_x11_get_default_display (),
clutter_x11_get_default_screen ());
/* Check for the texture from pixmap extension */
if (_cogl_check_extension ("GLX_EXT_texture_from_pixmap",
glx_extensions))
{
_gl_bind_tex_image =
(BindTexImage)cogl_get_proc_address ("glXBindTexImageEXT");
_gl_release_tex_image =
(ReleaseTexImage)cogl_get_proc_address ("glXReleaseTexImageEXT");
if (_gl_bind_tex_image && _gl_release_tex_image)
_have_tex_from_pixmap_ext = TRUE;
}
_gl_generate_mipmap =
(GenerateMipmap)cogl_get_proc_address ("glGenerateMipmapEXT");
gl_extensions = (char *) glGetString (GL_EXTENSIONS);
_have_tex_rectangle = _cogl_check_extension ("GL_ARB_texture_rectangle",
gl_extensions);
if ((rect_env = g_getenv ("CLUTTER_PIXMAP_TEXTURE_RECTANGLE")))
{
if (g_ascii_strcasecmp (rect_env, "force") == 0)
_rectangle_state = CLUTTER_GLX_RECTANGLE_FORCE;
else if (g_ascii_strcasecmp (rect_env, "disable") == 0)
_rectangle_state = CLUTTER_GLX_RECTANGLE_DISALLOW;
else if (rect_env[0])
g_warning ("Unknown value for CLUTTER_PIXMAP_TEXTURE_RECTANGLE, "
"should be 'force' or 'disable'");
}
_ext_check_done = TRUE;
}
}
static void
clutter_glx_texture_pixmap_free_rectangle (ClutterGLXTexturePixmap *texture)
{
ClutterGLXTexturePixmapPrivate *priv = texture->priv;
CoglHandle cogl_tex;
/* Cogl won't free the GL texture resource if it was created with
new_from_foreign so we need to free it manually */
if (priv->using_rectangle)
{
cogl_tex = clutter_texture_get_cogl_texture (CLUTTER_TEXTURE (texture));
if (cogl_tex != COGL_INVALID_HANDLE)
{
GLuint gl_handle;
GLenum gl_target;
cogl_texture_get_gl_texture (cogl_tex, &gl_handle, &gl_target);
if (gl_target == GL_TEXTURE_RECTANGLE_ARB)
glDeleteTextures (1, &gl_handle);
}
priv->using_rectangle = FALSE;
}
}
static void
clutter_glx_texture_pixmap_dispose (GObject *object)
{
ClutterGLXTexturePixmap *texture = CLUTTER_GLX_TEXTURE_PIXMAP (object);
ClutterGLXTexturePixmapPrivate *priv = texture->priv;
clutter_glx_texture_pixmap_free_rectangle (texture);
if (priv->glx_pixmap != None)
{
clutter_x11_trap_x_errors ();
glXDestroyPixmap (clutter_x11_get_default_display(),
priv->glx_pixmap);
XSync (clutter_x11_get_default_display(), FALSE);
clutter_x11_untrap_x_errors ();
priv->glx_pixmap = None;
}
G_OBJECT_CLASS (clutter_glx_texture_pixmap_parent_class)->dispose (object);
}
static void
clutter_glx_texture_pixmap_notify (GObject *object, GParamSpec *pspec)
{
if (g_str_equal (pspec->name, "pixmap"))
{
ClutterGLXTexturePixmap *texture = CLUTTER_GLX_TEXTURE_PIXMAP (object);
if (CLUTTER_ACTOR_IS_REALIZED (texture))
clutter_glx_texture_pixmap_create_glx_pixmap (texture);
}
}
static gboolean
should_use_rectangle (void)
{
/* Use the rectangle only if it is available and either:
the CLUTTER_PIXMAP_TEXTURE_RECTANGLE environment variable is
set to 'force'
*or*
the env var is set to 'allow' (which is the default) and NPOTs
textures are not available */
return _have_tex_rectangle
&& ((_rectangle_state == CLUTTER_GLX_RECTANGLE_ALLOW
&& !clutter_feature_available (CLUTTER_FEATURE_TEXTURE_NPOT))
|| _rectangle_state == CLUTTER_GLX_RECTANGLE_FORCE);
}
static gboolean
create_cogl_texture (ClutterTexture *texture,
guint width,
guint height)
{
ClutterGLXTexturePixmap *texture_glx = CLUTTER_GLX_TEXTURE_PIXMAP (texture);
ClutterGLXTexturePixmapPrivate *priv = texture_glx->priv;
CoglHandle handle;
gboolean using_rectangle;
GLint gl_format;
CoglPixelFormat cogl_format = COGL_PIXEL_FORMAT_RGBA_8888;
guint depth;
g_object_get (G_OBJECT (texture_glx), "pixmap-depth", &depth, NULL);
if (depth == 32)
{
gl_format = GL_RGBA;
cogl_format = COGL_PIXEL_FORMAT_RGBA_8888;
}
else if (depth == 24)
{
gl_format = GL_RGB;
cogl_format = COGL_PIXEL_FORMAT_RGB_888;
}
else
{
g_critical ("Can't create a TFP cogl texture for pixmap with depth < 24");
return FALSE;
}
/* We want to use the GL_ARB_texture_rectangle extension on some
chipsets because GL_ARB_texture_non_power_of_two is not always
supported or might be buggy */
if (should_use_rectangle ())
{
GLuint tex = 0;
using_rectangle = TRUE;
glGenTextures (1, &tex);
glBindTexture (GL_TEXTURE_RECTANGLE_ARB, tex);
glTexImage2D (GL_TEXTURE_RECTANGLE_ARB, 0,
gl_format, width, height,
0, GL_RGB, GL_UNSIGNED_BYTE, NULL);
handle = cogl_texture_new_from_foreign (tex, GL_TEXTURE_RECTANGLE_ARB,
width, height,
0, 0,
cogl_format | COGL_BGR_BIT |
COGL_PREMULT_BIT);
}
else
{
handle
= cogl_texture_new_with_size (width, height,
[cogl] Remove max_waste argument from Texture ctors The CoglTexture constructors expose the "max-waste" argument for controlling the maximum amount of wasted areas for slicing or, if set to -1, disables slicing. Slicing is really relevant only for large images that are never repeated, so it's a useful feature only in controlled use cases. Specifying the amount of wasted area is, on the other hand, just a way to mess up this feature; 99% the times, you either pull this number out of thin air, hoping it's right, or you try to do the right thing and you choose the wrong number anyway. Instead, we can use the CoglTextureFlags to control whether the texture should not be sliced (useful for Clutter-GST and for the texture-from-pixmap actors) and provide a reasonable value for enabling the slicing ourself. At some point, we might even provide a way to change the default at compile time or at run time, for particular platforms. Since max_waste is gone, the :tile-waste property of ClutterTexture becomes read-only, and it proxies the cogl_texture_get_max_waste() function. Inside Clutter, the only cases where the max_waste argument was not set to -1 are in the Pango glyph cache (which is a POT texture anyway) and inside the test cases where we want to force slicing; for the latter we can create larger textures that will be bigger than the threshold we set. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com> Signed-off-by: Robert Bragg <robert@linux.intel.com> Signed-off-by: Neil Roberts <neil@linux.intel.com>
2009-05-23 18:18:18 +00:00
COGL_TEXTURE_NO_SLICING,
cogl_format | COGL_BGR_BIT |
COGL_PREMULT_BIT);
using_rectangle = FALSE;
}
if (handle)
{
clutter_glx_texture_pixmap_free_rectangle (texture_glx);
priv->using_rectangle = using_rectangle;
clutter_texture_set_cogl_texture (texture, handle);
cogl_handle_unref(handle);
return TRUE;
}
return FALSE;
}
static void
clutter_glx_texture_pixmap_realize (ClutterActor *actor)
{
ClutterGLXTexturePixmap *texture;
ClutterGLXTexturePixmapPrivate *priv;
texture = CLUTTER_GLX_TEXTURE_PIXMAP (actor);
priv = texture->priv;
clutter_glx_texture_pixmap_create_glx_pixmap (texture);
if (priv->use_fallback)
{
CLUTTER_ACTOR_CLASS (clutter_glx_texture_pixmap_parent_class)->
realize (actor);
return;
}
CLUTTER_NOTE (TEXTURE, "texture pixmap realised");
}
static void
clutter_glx_texture_pixmap_unrealize (ClutterActor *actor)
{
ClutterGLXTexturePixmap *texture = CLUTTER_GLX_TEXTURE_PIXMAP (actor);
Display *dpy;
dpy = clutter_x11_get_default_display();
clutter_glx_texture_pixmap_free_rectangle (texture);
if (!_have_tex_from_pixmap_ext)
{
CLUTTER_ACTOR_CLASS (clutter_glx_texture_pixmap_parent_class)->
unrealize (actor);
return;
}
if (!CLUTTER_ACTOR_IS_REALIZED (actor))
return;
release_tex_image (texture);
CLUTTER_ACTOR_UNSET_FLAGS (actor, CLUTTER_ACTOR_REALIZED);
}
static GLXFBConfig *
get_fbconfig_for_depth (ClutterGLXTexturePixmap *texture, guint depth)
{
GLXFBConfig *fbconfigs, *ret = NULL;
int n_elements, i, found;
Display *dpy;
int db, stencil, alpha, mipmap, rgba, value;
static GLXFBConfig *cached_config = NULL;
static gboolean have_cached_config = FALSE;
static int cached_mipmap = 0;
if (have_cached_config)
{
texture->priv->can_mipmap = cached_mipmap;
return cached_config;
}
dpy = clutter_x11_get_default_display ();
fbconfigs = glXGetFBConfigs (dpy,
clutter_x11_get_default_screen (),
&n_elements);
db = G_MAXSHORT;
stencil = G_MAXSHORT;
mipmap = 0;
rgba = 0;
found = n_elements;
for (i = 0; i < n_elements; i++)
{
XVisualInfo *vi;
int visual_depth;
vi = glXGetVisualFromFBConfig (dpy,
fbconfigs[i]);
if (vi == NULL)
continue;
visual_depth = vi->depth;
XFree (vi);
if (visual_depth != depth)
continue;
glXGetFBConfigAttrib (dpy,
fbconfigs[i],
GLX_ALPHA_SIZE,
&alpha);
glXGetFBConfigAttrib (dpy,
fbconfigs[i],
GLX_BUFFER_SIZE,
&value);
if (value != depth && (value - alpha) != depth)
continue;
value = 0;
if (depth == 32)
{
glXGetFBConfigAttrib (dpy,
fbconfigs[i],
GLX_BIND_TO_TEXTURE_RGBA_EXT,
&value);
if (value)
rgba = 1;
}
if (!value)
{
if (rgba)
continue;
glXGetFBConfigAttrib (dpy,
fbconfigs[i],
GLX_BIND_TO_TEXTURE_RGB_EXT,
&value);
if (!value)
continue;
}
glXGetFBConfigAttrib (dpy,
fbconfigs[i],
GLX_DOUBLEBUFFER,
&value);
if (value > db)
continue;
db = value;
glXGetFBConfigAttrib (dpy,
fbconfigs[i],
GLX_STENCIL_SIZE,
&value);
if (value > stencil)
continue;
stencil = value;
if (_gl_generate_mipmap)
{
glXGetFBConfigAttrib (dpy,
fbconfigs[i],
GLX_BIND_TO_MIPMAP_TEXTURE_EXT,
&value);
if (value < mipmap)
continue;
mipmap = value;
}
found = i;
}
if (found != n_elements)
{
ret = g_malloc (sizeof (GLXFBConfig));
*ret = fbconfigs[found];
}
if (n_elements)
XFree (fbconfigs);
have_cached_config = TRUE;
cached_config = ret;
texture->priv->can_mipmap = cached_mipmap = mipmap;
return ret;
}
static void
clutter_glx_texture_pixmap_free_glx_pixmap (ClutterGLXTexturePixmap *texture)
{
ClutterGLXTexturePixmapPrivate *priv = texture->priv;
Display *dpy;
dpy = clutter_x11_get_default_display ();
release_tex_image (texture);
clutter_x11_trap_x_errors ();
if (priv->glx_pixmap)
glXDestroyPixmap (dpy, priv->glx_pixmap);
XSync (dpy, FALSE);
clutter_x11_untrap_x_errors ();
priv->glx_pixmap = None;
}
static void
clutter_glx_texture_pixmap_create_glx_pixmap (ClutterGLXTexturePixmap *texture)
{
ClutterGLXTexturePixmapPrivate *priv = texture->priv;
GLXPixmap glx_pixmap = None;
int attribs[7], i = 0, mipmap = 0;
GLXFBConfig *fbconfig;
Display *dpy;
guint depth;
Pixmap pixmap = None;
guint pixmap_width = 0, pixmap_height = 0;
ClutterBackendGLX *backend_glx;
ClutterTextureQuality quality;
backend_glx = CLUTTER_BACKEND_GLX(clutter_get_default_backend ());
dpy = clutter_x11_get_default_display ();
if (!_have_tex_from_pixmap_ext)
goto cleanup;
g_object_get (texture,
"pixmap-width", &pixmap_width,
"pixmap-height", &pixmap_height,
"pixmap-depth", &depth,
"pixmap", &pixmap,
NULL);
if (!pixmap)
{
goto cleanup;
}
CLUTTER_NOTE (TEXTURE, "Creating GLXPixmap");
fbconfig = get_fbconfig_for_depth (texture, depth);
if (!fbconfig)
{
g_warning ("Could not find an FBConfig for selected pixmap");
goto cleanup;
}
attribs[i++] = GLX_TEXTURE_FORMAT_EXT;
if (depth == 24)
{
attribs[i++] = GLX_TEXTURE_FORMAT_RGB_EXT;
}
else if (depth == 32)
{
attribs[i++] = GLX_TEXTURE_FORMAT_RGBA_EXT;
}
else
{
g_warning ("Pixmap with depth below 24 are not supported");
goto cleanup;
}
quality = clutter_texture_get_filter_quality (CLUTTER_TEXTURE (texture));
if (quality == CLUTTER_TEXTURE_QUALITY_HIGH && priv->can_mipmap)
mipmap = priv->can_mipmap;
attribs[i++] = GLX_MIPMAP_TEXTURE_EXT;
attribs[i++] = mipmap;
attribs[i++] = GLX_TEXTURE_TARGET_EXT;
attribs[i++] = (should_use_rectangle ()
? GLX_TEXTURE_RECTANGLE_EXT : GLX_TEXTURE_2D_EXT);
attribs[i++] = None;
/* Note: some drivers (e.g. Nvidia) get upset if you effectivly create a glx
* pixmap for the same server side object, even though you might have unique
* client side names, we currently avoid hitting this problem by destroying
* the current glx pixmap first */
if (priv->glx_pixmap)
clutter_glx_texture_pixmap_free_glx_pixmap (texture);
clutter_x11_trap_x_errors ();
glx_pixmap = glXCreatePixmap (dpy,
*fbconfig,
pixmap,
attribs);
XSync (dpy, FALSE);
if (clutter_x11_untrap_x_errors ())
{
CLUTTER_NOTE (TEXTURE, "Failed to create GLXPixmap");
/* Make sure we don't think the call actually succeeded */
glx_pixmap = None;
}
if (!create_cogl_texture (CLUTTER_TEXTURE (texture),
pixmap_width, pixmap_height))
{
CLUTTER_NOTE (TEXTURE, "Unable to create texture for pixmap");
glXDestroyPixmap (dpy, glx_pixmap);
glx_pixmap = None;
goto cleanup;
}
cleanup:
if (priv->glx_pixmap)
clutter_glx_texture_pixmap_free_glx_pixmap (texture);
if (glx_pixmap != None)
{
priv->use_fallback = FALSE;
priv->glx_pixmap = glx_pixmap;
CLUTTER_NOTE (TEXTURE, "Created GLXPixmap");
clutter_glx_texture_pixmap_update_area (CLUTTER_X11_TEXTURE_PIXMAP (texture),
0, 0,
pixmap_width, pixmap_height);
/* Get ready to queue initial mipmap generation */
if (clutter_texture_get_filter_quality (CLUTTER_TEXTURE (texture))
== CLUTTER_TEXTURE_QUALITY_HIGH)
{
priv->mipmap_generate_queued = TRUE;
}
return;
}
else
{
if (pixmap)
CLUTTER_NOTE (TEXTURE, "Falling back to X11 manual mechansim");
priv->use_fallback = TRUE;
priv->glx_pixmap = None;
}
}
static void
clutter_glx_texture_pixmap_update_area (ClutterX11TexturePixmap *texture,
gint x,
gint y,
gint width,
gint height)
{
ClutterGLXTexturePixmap *texture_glx = CLUTTER_GLX_TEXTURE_PIXMAP (texture);
Adds initial clipped redraw support to Clutter A new (internal only currently) API, _clutter_actor_queue_clipped_redraw can be used to queue a redraw along with a clip rectangle in actor coordinates. This clip rectangle propagates up to the stage and clutter backend which may optionally use the information to optimize stage redraws. The GLX backend in particular may scissor the next redraw to the clip rectangle and use GLX_MESA_copy_sub_buffer to present the stage subregion. The intention is that any actors that can naturally determine the bounds of updates should queue clipped redraws to reduce the cost of updating small regions of the screen. Notes: » If GLX_MESA_copy_sub_buffer isn't available then the GLX backend ignores any clip rectangles. » queuing multiple clipped redraws will result in the bounding box of each clip rectangle being used. » If a clipped redraw has a height > 300 pixels then it's promoted into a full stage redraw, so that the GPU doesn't end up blocking too long waiting for the vsync to reach the optimal position to avoid tearing. » Note: no empirical data was used to come up with this threshold so we may need to tune this. » Currently only ClutterX11TexturePixmap makes use of this new API. This is done via a new "queue-damage-redraw" signal that is emitted when the pixmap is updated. The default handler queues a clipped redraw with the assumption that the pixmap is being painted as a rectangle covering the actors transformed allocation. If you subclass ClutterX11TexturePixmap and change how it's painted you now also need to override the signal handler and queue your own redraw. Technically this is a semantic break, but it's assumed that no one is currently doing this. This still leaves a few unsolved issues with regards to optimizing sub stage redraws that need to be addressed in further work so this can only be considered a stepping stone a this point: » Because we have no reliable way to determine if the painting of any given actor is being modified any optimizations implemented using _clutter_actor_queue_redraw_with_clip must be overridable by a subclass, and technically must be opt-in for existing classes to avoid a change in semantics. E.g. consider that a user connects to the paint signal for ClutterTexture and paints a circle instead of a rectangle. In this case any original logic to queue clipped redraws would be incorrect. » Currently only the implementation of an actor has enough information with which to queue clipped redraws. E.g. It is not possible for generic code in clutter-actor.c to queue a clipped redraw when hiding an actor because actors have no way to report a "paint box". (remember actors can draw outside their allocation and actors with depth may also be projected outside of their allocation) » The current plan is to add a actor_class->get_paint_cuboid() virtual so actors can report a bounding cube for everything they would draw in their current state and use that to queue clipped redraws against the stage by projecting the paint cube into stage coordinates. » Our heuristics for promoting clipped redraws into full redraws to avoid blocking the GPU while we wait for the vsync need improving: » vsync issues aren't relevant for redirected/composited applications so they should use different heuristics. In this case we instead need to trade off the cost of blitting when using glXCopySubBuffer vs promoting to a full redraw and flipping instead.
2009-11-30 17:47:55 +00:00
ClutterGLXTexturePixmapPrivate *priv = texture_glx->priv;
CLUTTER_NOTE (TEXTURE, "Updating texture pixmap");
if (!CLUTTER_ACTOR_IS_REALIZED (texture))
return;
if (priv->use_fallback)
{
CLUTTER_NOTE (TEXTURE, "Falling back to X11");
clutter_glx_texture_pixmap_free_rectangle (texture_glx);
parent_class->update_area (texture,
x, y,
width, height);
return;
}
if (priv->glx_pixmap == None)
return;
Adds initial clipped redraw support to Clutter A new (internal only currently) API, _clutter_actor_queue_clipped_redraw can be used to queue a redraw along with a clip rectangle in actor coordinates. This clip rectangle propagates up to the stage and clutter backend which may optionally use the information to optimize stage redraws. The GLX backend in particular may scissor the next redraw to the clip rectangle and use GLX_MESA_copy_sub_buffer to present the stage subregion. The intention is that any actors that can naturally determine the bounds of updates should queue clipped redraws to reduce the cost of updating small regions of the screen. Notes: » If GLX_MESA_copy_sub_buffer isn't available then the GLX backend ignores any clip rectangles. » queuing multiple clipped redraws will result in the bounding box of each clip rectangle being used. » If a clipped redraw has a height > 300 pixels then it's promoted into a full stage redraw, so that the GPU doesn't end up blocking too long waiting for the vsync to reach the optimal position to avoid tearing. » Note: no empirical data was used to come up with this threshold so we may need to tune this. » Currently only ClutterX11TexturePixmap makes use of this new API. This is done via a new "queue-damage-redraw" signal that is emitted when the pixmap is updated. The default handler queues a clipped redraw with the assumption that the pixmap is being painted as a rectangle covering the actors transformed allocation. If you subclass ClutterX11TexturePixmap and change how it's painted you now also need to override the signal handler and queue your own redraw. Technically this is a semantic break, but it's assumed that no one is currently doing this. This still leaves a few unsolved issues with regards to optimizing sub stage redraws that need to be addressed in further work so this can only be considered a stepping stone a this point: » Because we have no reliable way to determine if the painting of any given actor is being modified any optimizations implemented using _clutter_actor_queue_redraw_with_clip must be overridable by a subclass, and technically must be opt-in for existing classes to avoid a change in semantics. E.g. consider that a user connects to the paint signal for ClutterTexture and paints a circle instead of a rectangle. In this case any original logic to queue clipped redraws would be incorrect. » Currently only the implementation of an actor has enough information with which to queue clipped redraws. E.g. It is not possible for generic code in clutter-actor.c to queue a clipped redraw when hiding an actor because actors have no way to report a "paint box". (remember actors can draw outside their allocation and actors with depth may also be projected outside of their allocation) » The current plan is to add a actor_class->get_paint_cuboid() virtual so actors can report a bounding cube for everything they would draw in their current state and use that to queue clipped redraws against the stage by projecting the paint cube into stage coordinates. » Our heuristics for promoting clipped redraws into full redraws to avoid blocking the GPU while we wait for the vsync need improving: » vsync issues aren't relevant for redirected/composited applications so they should use different heuristics. In this case we instead need to trade off the cost of blitting when using glXCopySubBuffer vs promoting to a full redraw and flipping instead.
2009-11-30 17:47:55 +00:00
priv->bind_tex_image_queued = TRUE;
}
static void
clutter_glx_texture_pixmap_class_init (ClutterGLXTexturePixmapClass *klass)
{
GObjectClass *object_class = G_OBJECT_CLASS (klass);
ClutterActorClass *actor_class = CLUTTER_ACTOR_CLASS (klass);
ClutterX11TexturePixmapClass *x11_texture_class =
CLUTTER_X11_TEXTURE_PIXMAP_CLASS (klass);
g_type_class_add_private (klass, sizeof (ClutterGLXTexturePixmapPrivate));
parent_class = g_type_class_peek_parent(klass);
object_class->dispose = clutter_glx_texture_pixmap_dispose;
object_class->notify = clutter_glx_texture_pixmap_notify;
actor_class->realize = clutter_glx_texture_pixmap_realize;
actor_class->unrealize = clutter_glx_texture_pixmap_unrealize;
x11_texture_class->update_area = clutter_glx_texture_pixmap_update_area;
}
/**
* clutter_glx_texture_pixmap_using_extension:
* @texture: A #ClutterGLXTexturePixmap
*
* Checks whether @texture is using the GLX_EXT_texture_from_pixmap
* extension; this extension can be optionally (though it is strongly
* encouraged) implemented as a zero-copy between a GLX pixmap and
* a GL texture.
*
* Return value: %TRUE if the texture is using the
* GLX_EXT_texture_from_pixmap OpenGL extension or falling back to the
* slower software mechanism.
*
* Since: 0.8
*/
gboolean
clutter_glx_texture_pixmap_using_extension (ClutterGLXTexturePixmap *texture)
{
ClutterGLXTexturePixmapPrivate *priv;
priv = CLUTTER_GLX_TEXTURE_PIXMAP (texture)->priv;
return (_have_tex_from_pixmap_ext && !priv->use_fallback);
/* Assume NPOT TFP's are supported even if regular NPOT isn't advertised
* but tfp is. Seemingly some Intel drivers do this ?
*/
/* && clutter_feature_available (COGL_FEATURE_TEXTURE_NPOT)); */
}
/**
* clutter_glx_texture_pixmap_new_with_pixmap:
* @pixmap: the X Pixmap to which this texture should be bound
*
* Creates a new #ClutterGLXTexturePixmap for @pixmap
*
* Return value: A new #ClutterGLXTexturePixmap bound to the given X Pixmap
*
* Since: 0.8
*/
ClutterActor *
clutter_glx_texture_pixmap_new_with_pixmap (Pixmap pixmap)
{
return g_object_new (CLUTTER_GLX_TYPE_TEXTURE_PIXMAP,
"pixmap", pixmap,
NULL);
}
/**
* clutter_glx_texture_pixmap_new_with_window:
* @window: the X window to which this texture should be bound
*
* Creates a new #ClutterGLXTexturePixmap for @window
*
* Return value: A new #ClutterGLXTexturePixmap bound to the given X window
*
* Since: 0.8
**/
ClutterActor *
clutter_glx_texture_pixmap_new_with_window (Window window)
{
return g_object_new (CLUTTER_GLX_TYPE_TEXTURE_PIXMAP,
"window", window,
NULL);
}
/**
* clutter_glx_texture_pixmap_new:
*
* Creates a new, empty #ClutterGLXTexturePixmap
*
* Return value: A new #ClutterGLXTexturePixmap
*
* Since: 0.8
*/
ClutterActor *
clutter_glx_texture_pixmap_new (void)
{
return g_object_new (CLUTTER_GLX_TYPE_TEXTURE_PIXMAP, NULL);
}