mutter/cogl/cogl-pipeline-layer.c

937 lines
32 KiB
C
Raw Normal View History

/*
* Cogl
*
* An object oriented GL/GLES Abstraction/Utility Layer
*
* Copyright (C) 2008,2009,2010 Intel Corporation.
*
* 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/>.
*
*
*
* Authors:
* Robert Bragg <robert@linux.intel.com>
*/
#ifdef HAVE_CONFIG_H
#include "config.h"
#endif
#include "cogl-util.h"
#include "cogl-context-private.h"
#include "cogl-texture-private.h"
#include "cogl-pipeline.h"
#include "cogl-pipeline-layer-private.h"
#include "cogl-pipeline-layer-state-private.h"
#include "cogl-pipeline-layer-state.h"
#include "cogl-node-private.h"
#include "cogl-pipeline-opengl-private.h"
#include "cogl-context-private.h"
#include "cogl-texture-private.h"
#include <string.h>
static void
_cogl_pipeline_layer_free (CoglPipelineLayer *layer);
/* This type was made deprecated before the cogl_is_pipeline_layer
function was ever exposed in the public headers so there's no need
to make the cogl_is_pipeline_layer function public. We use INTERNAL
so that the cogl_is_* function won't get defined */
COGL_OBJECT_INTERNAL_DEFINE (PipelineLayer, pipeline_layer);
CoglPipelineLayer *
_cogl_pipeline_layer_get_authority (CoglPipelineLayer *layer,
unsigned long difference)
{
CoglPipelineLayer *authority = layer;
while (!(authority->differences & difference))
authority = _cogl_pipeline_layer_get_parent (authority);
return authority;
}
int
_cogl_pipeline_layer_get_unit_index (CoglPipelineLayer *layer)
{
CoglPipelineLayer *authority =
_cogl_pipeline_layer_get_authority (layer, COGL_PIPELINE_LAYER_STATE_UNIT);
return authority->unit_index;
}
CoglBool
_cogl_pipeline_layer_has_alpha (CoglPipelineLayer *layer)
{
CoglPipelineLayer *combine_authority =
_cogl_pipeline_layer_get_authority (layer,
COGL_PIPELINE_LAYER_STATE_COMBINE);
CoglPipelineLayerBigState *big_state = combine_authority->big_state;
CoglPipelineLayer *tex_authority;
CoglPipelineLayer *snippets_authority;
/* has_alpha maintains the alpha status for the GL_PREVIOUS layer */
/* For anything but the default texture combine we currently just
* assume it may result in an alpha value < 1
*
* FIXME: we could do better than this. */
if (big_state->texture_combine_alpha_func !=
COGL_PIPELINE_COMBINE_FUNC_MODULATE ||
big_state->texture_combine_alpha_src[0] !=
COGL_PIPELINE_COMBINE_SOURCE_PREVIOUS ||
big_state->texture_combine_alpha_op[0] !=
COGL_PIPELINE_COMBINE_OP_SRC_ALPHA ||
big_state->texture_combine_alpha_src[1] !=
COGL_PIPELINE_COMBINE_SOURCE_TEXTURE ||
big_state->texture_combine_alpha_op[1] !=
COGL_PIPELINE_COMBINE_OP_SRC_ALPHA)
{
return TRUE;
}
/* NB: A layer may have a combine mode set on it but not yet
* have an associated texture which would mean we'd fallback
* to the default texture which doesn't have an alpha component
*/
tex_authority =
_cogl_pipeline_layer_get_authority (layer,
COGL_PIPELINE_LAYER_STATE_TEXTURE_DATA);
if (tex_authority->texture &&
cogl_texture_get_format (tex_authority->texture) & COGL_A_BIT)
{
return TRUE;
}
/* All bets are off if the layer contains any snippets */
snippets_authority = _cogl_pipeline_layer_get_authority
(layer, COGL_PIPELINE_LAYER_STATE_VERTEX_SNIPPETS);
if (snippets_authority->big_state->vertex_snippets.entries != NULL)
return TRUE;
snippets_authority = _cogl_pipeline_layer_get_authority
(layer, COGL_PIPELINE_LAYER_STATE_FRAGMENT_SNIPPETS);
if (snippets_authority->big_state->fragment_snippets.entries != NULL)
return TRUE;
return FALSE;
}
unsigned int
_cogl_get_n_args_for_combine_func (CoglPipelineCombineFunc func)
{
switch (func)
{
case COGL_PIPELINE_COMBINE_FUNC_REPLACE:
return 1;
case COGL_PIPELINE_COMBINE_FUNC_MODULATE:
case COGL_PIPELINE_COMBINE_FUNC_ADD:
case COGL_PIPELINE_COMBINE_FUNC_ADD_SIGNED:
case COGL_PIPELINE_COMBINE_FUNC_SUBTRACT:
case COGL_PIPELINE_COMBINE_FUNC_DOT3_RGB:
case COGL_PIPELINE_COMBINE_FUNC_DOT3_RGBA:
return 2;
case COGL_PIPELINE_COMBINE_FUNC_INTERPOLATE:
return 3;
}
return 0;
}
void
_cogl_pipeline_layer_copy_differences (CoglPipelineLayer *dest,
CoglPipelineLayer *src,
unsigned long differences)
{
CoglPipelineLayerBigState *big_dest, *big_src;
if ((differences & COGL_PIPELINE_LAYER_STATE_NEEDS_BIG_STATE) &&
!dest->has_big_state)
{
dest->big_state = g_slice_new (CoglPipelineLayerBigState);
dest->has_big_state = TRUE;
}
big_dest = dest->big_state;
big_src = src->big_state;
dest->differences |= differences;
while (differences)
{
int index = _cogl_util_ffs (differences) - 1;
differences &= ~(1 << index);
/* This convoluted switch statement is just here so that we'll
* get a warning if a new state is added without handling it
* here */
switch (index)
{
case COGL_PIPELINE_LAYER_STATE_COUNT:
case COGL_PIPELINE_LAYER_STATE_UNIT_INDEX:
g_warn_if_reached ();
break;
case COGL_PIPELINE_LAYER_STATE_TEXTURE_TYPE_INDEX:
dest->texture_type = src->texture_type;
break;
case COGL_PIPELINE_LAYER_STATE_TEXTURE_DATA_INDEX:
dest->texture = src->texture;
if (dest->texture)
cogl_object_ref (dest->texture);
break;
case COGL_PIPELINE_LAYER_STATE_SAMPLER_INDEX:
dest->sampler_cache_entry = src->sampler_cache_entry;
break;
case COGL_PIPELINE_LAYER_STATE_COMBINE_INDEX:
{
CoglPipelineCombineFunc func;
int n_args, i;
func = big_src->texture_combine_rgb_func;
big_dest->texture_combine_rgb_func = func;
n_args = _cogl_get_n_args_for_combine_func (func);
for (i = 0; i < n_args; i++)
{
big_dest->texture_combine_rgb_src[i] =
big_src->texture_combine_rgb_src[i];
big_dest->texture_combine_rgb_op[i] =
big_src->texture_combine_rgb_op[i];
}
func = big_src->texture_combine_alpha_func;
big_dest->texture_combine_alpha_func = func;
n_args = _cogl_get_n_args_for_combine_func (func);
for (i = 0; i < n_args; i++)
{
big_dest->texture_combine_alpha_src[i] =
big_src->texture_combine_alpha_src[i];
big_dest->texture_combine_alpha_op[i] =
big_src->texture_combine_alpha_op[i];
}
}
break;
case COGL_PIPELINE_LAYER_STATE_COMBINE_CONSTANT_INDEX:
memcpy (big_dest->texture_combine_constant,
big_src->texture_combine_constant,
sizeof (big_dest->texture_combine_constant));
break;
case COGL_PIPELINE_LAYER_STATE_POINT_SPRITE_COORDS_INDEX:
big_dest->point_sprite_coords = big_src->point_sprite_coords;
break;
case COGL_PIPELINE_LAYER_STATE_VERTEX_SNIPPETS_INDEX:
_cogl_pipeline_snippet_list_copy (&big_dest->vertex_snippets,
&big_src->vertex_snippets);
break;
case COGL_PIPELINE_LAYER_STATE_FRAGMENT_SNIPPETS_INDEX:
_cogl_pipeline_snippet_list_copy (&big_dest->fragment_snippets,
&big_src->fragment_snippets);
break;
}
}
}
static void
_cogl_pipeline_layer_init_multi_property_sparse_state (
CoglPipelineLayer *layer,
CoglPipelineLayerState change)
{
CoglPipelineLayer *authority;
/* Nothing to initialize in these cases since they are all comprised
* of one member which we expect to immediately be overwritten. */
if (!(change & COGL_PIPELINE_LAYER_STATE_MULTI_PROPERTY))
return;
authority = _cogl_pipeline_layer_get_authority (layer, change);
switch (change)
{
/* XXX: avoid using a default: label so we get a warning if we
* don't explicitly handle a newly defined state-group here. */
case COGL_PIPELINE_LAYER_STATE_UNIT:
case COGL_PIPELINE_LAYER_STATE_TEXTURE_TYPE:
case COGL_PIPELINE_LAYER_STATE_TEXTURE_DATA:
case COGL_PIPELINE_LAYER_STATE_POINT_SPRITE_COORDS:
case COGL_PIPELINE_LAYER_STATE_USER_MATRIX:
case COGL_PIPELINE_LAYER_STATE_COMBINE_CONSTANT:
Use GL_ARB_sampler_objects GL_ARB_sampler_objects provides a GL object which overrides the sampler state part of a texture object with different values. The sampler state that Cogl currently exposes is the wrap modes and filters. Cogl exposes the state as part of the pipeline layer state but without this extension GL only exposes it as part of the texture object state. This means that it won't work to use a single texture multiple times in one primitive with different sampler states. It also makes switching between different sampler states with a single texture not terribly efficient because it has to change the texture object state every time. This patch adds a cache for sampler states in a shared hash table attached to the CoglContext. The entire set of parameters for the sampler state is used as the key for the hash table. When a unique state is encountered the sampler cache will create a new entry, otherwise it will return a const pointer to an existing entry. That means we can have a single pointer to represent any combination of sampler state. Pipeline layers now just store this single pointer rather than storing all of the sampler state. The two separate state flags for wrap modes and filters have now been combined into one. It should be faster to compare the sampler state now because instead of comparing each value it can just compare the pointers to the cached sampler entries. The hash table of cached sampler states should only need to perform its more expensive hash on the state when a property is changed on a pipeline, not every time it is flushed. When the sampler objects extension is available each cached sampler state will also get a sampler object to represent it. The common code to flush the GL state will now simply bind this object to a unit instead of flushing the state though the CoglTexture when possible. Reviewed-by: Robert Bragg <robert@linux.intel.com>
2012-04-04 17:20:04 -04:00
case COGL_PIPELINE_LAYER_STATE_SAMPLER:
g_return_if_reached ();
/* XXX: technically we could probably even consider these as
* single property state-groups from the pov that currently the
* corresponding property setters always update all of the values
* at the same time. */
case COGL_PIPELINE_LAYER_STATE_COMBINE:
{
int n_args;
int i;
CoglPipelineLayerBigState *src_big_state = authority->big_state;
CoglPipelineLayerBigState *dest_big_state = layer->big_state;
GLint func = src_big_state->texture_combine_rgb_func;
dest_big_state->texture_combine_rgb_func = func;
n_args = _cogl_get_n_args_for_combine_func (func);
for (i = 0; i < n_args; i++)
{
dest_big_state->texture_combine_rgb_src[i] =
src_big_state->texture_combine_rgb_src[i];
dest_big_state->texture_combine_rgb_op[i] =
src_big_state->texture_combine_rgb_op[i];
}
func = src_big_state->texture_combine_alpha_func;
dest_big_state->texture_combine_alpha_func = func;
n_args = _cogl_get_n_args_for_combine_func (func);
for (i = 0; i < n_args; i++)
{
dest_big_state->texture_combine_alpha_src[i] =
src_big_state->texture_combine_alpha_src[i];
dest_big_state->texture_combine_alpha_op[i] =
src_big_state->texture_combine_alpha_op[i];
}
break;
}
case COGL_PIPELINE_LAYER_STATE_VERTEX_SNIPPETS:
_cogl_pipeline_snippet_list_copy (&layer->big_state->vertex_snippets,
&authority->big_state->
vertex_snippets);
break;
case COGL_PIPELINE_LAYER_STATE_FRAGMENT_SNIPPETS:
_cogl_pipeline_snippet_list_copy (&layer->big_state->fragment_snippets,
&authority->big_state->
fragment_snippets);
break;
}
}
/* NB: If a layer has descendants we can't modify the layer
* NB: If the layer is owned and the owner has descendants we can't
* modify the layer.
*
* This function will allocate a new derived layer if you are trying
* to change the state of a layer with dependants (as described above)
* so you must always check the return value.
*
* If a new layer is returned it will be owned by required_owner.
* (NB: a layer is always modified with respect to a pipeline - the
* "required_owner")
*
* required_owner can only by NULL for new, currently unowned layers
* with no dependants.
*/
CoglPipelineLayer *
_cogl_pipeline_layer_pre_change_notify (CoglPipeline *required_owner,
CoglPipelineLayer *layer,
CoglPipelineLayerState change)
{
CoglTextureUnit *unit;
/* Identify the case where the layer is new with no owner or
* dependants and so we don't need to do anything. */
Use the Wayland embedded linked list implementation instead of BSD's This removes cogl-queue.h and adds a copy of Wayland's embedded list implementation. The advantage of the Wayland model is that it is much simpler and so it is easier to follow. It also doesn't require defining a typedef for every list type. The downside is that there is only one list type which is a doubly-linked list where the head has a pointer to both the beginning and the end. The BSD implementation has many more combinations some of which we were taking advantage of to reduce the size of critical structs where we didn't need a pointer to the end of the list. The corresponding changes to uses of cogl-queue.h are: • COGL_STAILQ_* was used for onscreen the list of events and dirty notifications. This makes the size of the CoglContext grow by one pointer. • COGL_TAILQ_* was used for fences. • COGL_LIST_* for CoglClosures. In this case the list head now has an extra pointer which means CoglOnscreen will grow by the size of three pointers, but this doesn't seem like a particularly important struct to optimise for size anyway. • COGL_LIST_* was used for the list of foreign GLES2 offscreens. • COGL_TAILQ_* was used for the list of sub stacks in a CoglMemoryStack. • COGL_LIST_* was used to track the list of layers that haven't had code generated yet while generating a fragment shader for a pipeline. • COGL_LIST_* was used to track the pipeline hierarchy in CoglNode. The last part is a bit more controversial because it increases the size of CoglPipeline and CoglPipelineLayer by one pointer in order to have the redundant tail pointer for the list head. Normally we try to be very careful about the size of the CoglPipeline struct. Because CoglPipeline is slice-allocated, this effectively ends up adding two pointers to the size because GSlice rounds up to the size of two pointers. Reviewed-by: Robert Bragg <robert@linux.intel.com> (cherry picked from commit 13abf613b15f571ba1fcf6d2eb831ffc6fa31324) Conflicts: cogl/cogl-context-private.h cogl/cogl-context.c cogl/driver/gl/cogl-pipeline-fragend-glsl.c doc/reference/cogl-2.0-experimental/Makefile.am
2013-06-08 18:03:25 -04:00
if (_cogl_list_empty (&COGL_NODE (layer)->children) &&
layer->owner == NULL)
goto init_layer_state;
/* We only allow a NULL required_owner for new layers */
_COGL_RETURN_VAL_IF_FAIL (required_owner != NULL, layer);
/* Chain up:
* A modification of a layer is indirectly also a modification of
* its owner so first make sure to flush the journal of any
* references to the current owner state and if necessary perform
* a copy-on-write for the required_owner if it has dependants.
*/
_cogl_pipeline_pre_change_notify (required_owner,
COGL_PIPELINE_STATE_LAYERS,
NULL,
TRUE);
/* Unlike pipelines; layers are simply considered immutable once
* they have dependants - either direct children, or another
* pipeline as an owner.
*/
Use the Wayland embedded linked list implementation instead of BSD's This removes cogl-queue.h and adds a copy of Wayland's embedded list implementation. The advantage of the Wayland model is that it is much simpler and so it is easier to follow. It also doesn't require defining a typedef for every list type. The downside is that there is only one list type which is a doubly-linked list where the head has a pointer to both the beginning and the end. The BSD implementation has many more combinations some of which we were taking advantage of to reduce the size of critical structs where we didn't need a pointer to the end of the list. The corresponding changes to uses of cogl-queue.h are: • COGL_STAILQ_* was used for onscreen the list of events and dirty notifications. This makes the size of the CoglContext grow by one pointer. • COGL_TAILQ_* was used for fences. • COGL_LIST_* for CoglClosures. In this case the list head now has an extra pointer which means CoglOnscreen will grow by the size of three pointers, but this doesn't seem like a particularly important struct to optimise for size anyway. • COGL_LIST_* was used for the list of foreign GLES2 offscreens. • COGL_TAILQ_* was used for the list of sub stacks in a CoglMemoryStack. • COGL_LIST_* was used to track the list of layers that haven't had code generated yet while generating a fragment shader for a pipeline. • COGL_LIST_* was used to track the pipeline hierarchy in CoglNode. The last part is a bit more controversial because it increases the size of CoglPipeline and CoglPipelineLayer by one pointer in order to have the redundant tail pointer for the list head. Normally we try to be very careful about the size of the CoglPipeline struct. Because CoglPipeline is slice-allocated, this effectively ends up adding two pointers to the size because GSlice rounds up to the size of two pointers. Reviewed-by: Robert Bragg <robert@linux.intel.com> (cherry picked from commit 13abf613b15f571ba1fcf6d2eb831ffc6fa31324) Conflicts: cogl/cogl-context-private.h cogl/cogl-context.c cogl/driver/gl/cogl-pipeline-fragend-glsl.c doc/reference/cogl-2.0-experimental/Makefile.am
2013-06-08 18:03:25 -04:00
if (!_cogl_list_empty (&COGL_NODE (layer)->children) ||
layer->owner != required_owner)
{
CoglPipelineLayer *new = _cogl_pipeline_layer_copy (layer);
if (layer->owner == required_owner)
_cogl_pipeline_remove_layer_difference (required_owner, layer, FALSE);
_cogl_pipeline_add_layer_difference (required_owner, new, FALSE);
cogl_object_unref (new);
layer = new;
goto init_layer_state;
}
/* Note: At this point we know there is only one pipeline dependant on
* this layer (required_owner), and there are no other layers
* dependant on this layer so it's ok to modify it. */
/* NB: Although layers can have private state associated with them
* by multiple backends we know that a layer can't be *changed* if
* it has multiple dependants so if we reach here we know we only
* have a single owner and can only be associated with a single
* backend that needs to be notified of the layer change...
*/
if (required_owner->progend != COGL_PIPELINE_PROGEND_UNDEFINED)
{
const CoglPipelineProgend *progend =
_cogl_pipeline_progends[required_owner->progend];
const CoglPipelineFragend *fragend =
_cogl_pipeline_fragends[progend->fragend];
const CoglPipelineVertend *vertend =
_cogl_pipeline_vertends[progend->vertend];
if (fragend->layer_pre_change_notify)
fragend->layer_pre_change_notify (required_owner, layer, change);
if (vertend->layer_pre_change_notify)
vertend->layer_pre_change_notify (required_owner, layer, change);
if (progend->layer_pre_change_notify)
progend->layer_pre_change_notify (required_owner, layer, change);
}
/* If the layer being changed is the same as the last layer we
* flushed to the corresponding texture unit then we keep a track of
* the changes so we can try to minimize redundant OpenGL calls if
* the same layer is flushed again.
*/
unit = _cogl_get_texture_unit (_cogl_pipeline_layer_get_unit_index (layer));
if (unit->layer == layer)
unit->layer_changes_since_flush |= change;
init_layer_state:
if (required_owner)
required_owner->age++;
if (change & COGL_PIPELINE_LAYER_STATE_NEEDS_BIG_STATE &&
!layer->has_big_state)
{
layer->big_state = g_slice_new (CoglPipelineLayerBigState);
layer->has_big_state = TRUE;
}
/* Note: conceptually we have just been notified that a single
* property value is about to change, but since some state-groups
* contain multiple properties and 'layer' is about to take over
* being the authority for the property's corresponding state-group
* we need to maintain the integrity of the other property values
* too.
*
* To ensure this we handle multi-property state-groups by copying
* all the values from the old-authority to the new...
*
* We don't have to worry about non-sparse property groups since
* we never take over being an authority for such properties so
* they automatically maintain integrity.
*/
if (change & COGL_PIPELINE_LAYER_STATE_ALL_SPARSE &&
!(layer->differences & change))
{
_cogl_pipeline_layer_init_multi_property_sparse_state (layer, change);
layer->differences |= change;
}
return layer;
}
static void
_cogl_pipeline_layer_unparent (CoglNode *layer)
{
/* Chain up */
_cogl_pipeline_node_unparent_real (layer);
}
static void
_cogl_pipeline_layer_set_parent (CoglPipelineLayer *layer,
CoglPipelineLayer *parent)
{
/* Chain up */
_cogl_pipeline_node_set_parent_real (COGL_NODE (layer),
COGL_NODE (parent),
_cogl_pipeline_layer_unparent,
TRUE);
}
CoglPipelineLayer *
_cogl_pipeline_layer_copy (CoglPipelineLayer *src)
{
CoglPipelineLayer *layer = g_slice_new (CoglPipelineLayer);
_cogl_pipeline_node_init (COGL_NODE (layer));
layer->owner = NULL;
layer->index = src->index;
layer->differences = 0;
layer->has_big_state = FALSE;
_cogl_pipeline_layer_set_parent (layer, src);
return _cogl_pipeline_layer_object_new (layer);
}
/* XXX: This is duplicated logic; the same as for
* _cogl_pipeline_prune_redundant_ancestry it would be nice to find a
* way to consolidate these functions! */
void
_cogl_pipeline_layer_prune_redundant_ancestry (CoglPipelineLayer *layer)
{
CoglPipelineLayer *new_parent = _cogl_pipeline_layer_get_parent (layer);
/* walk up past ancestors that are now redundant and potentially
* reparent the layer. */
while (_cogl_pipeline_layer_get_parent (new_parent) &&
(new_parent->differences | layer->differences) ==
layer->differences)
new_parent = _cogl_pipeline_layer_get_parent (new_parent);
_cogl_pipeline_layer_set_parent (layer, new_parent);
}
/* Determine the mask of differences between two layers.
*
* XXX: If layers and pipelines could both be cast to a common Tree
* type of some kind then we could have a unified
* compare_differences() function.
*/
unsigned long
_cogl_pipeline_layer_compare_differences (CoglPipelineLayer *layer0,
CoglPipelineLayer *layer1)
{
GSList *head0 = NULL;
GSList *head1 = NULL;
CoglPipelineLayer *node0;
CoglPipelineLayer *node1;
int len0 = 0;
int len1 = 0;
int count;
GSList *common_ancestor0;
GSList *common_ancestor1;
unsigned long layers_difference = 0;
/* Algorithm:
*
* 1) Walk the ancestors of each layer to the root node, adding a
* pointer to each ancester node to two linked lists
*
* 2) Compare the lists to find the nodes where they start to
* differ marking the common_ancestor node for each list.
*
* 3) For each list now iterate starting after the common_ancestor
* nodes ORing each nodes ->difference mask into the final
* differences mask.
*/
for (node0 = layer0; node0; node0 = _cogl_pipeline_layer_get_parent (node0))
{
GSList *link = alloca (sizeof (GSList));
link->next = head0;
link->data = node0;
head0 = link;
len0++;
}
for (node1 = layer1; node1; node1 = _cogl_pipeline_layer_get_parent (node1))
{
GSList *link = alloca (sizeof (GSList));
link->next = head1;
link->data = node1;
head1 = link;
len1++;
}
/* NB: There's no point looking at the head entries since we know both layers
* must have the same default layer as their root node. */
common_ancestor0 = head0;
common_ancestor1 = head1;
head0 = head0->next;
head1 = head1->next;
count = MIN (len0, len1) - 1;
while (count--)
{
if (head0->data != head1->data)
break;
common_ancestor0 = head0;
common_ancestor1 = head1;
head0 = head0->next;
head1 = head1->next;
}
for (head0 = common_ancestor0->next; head0; head0 = head0->next)
{
node0 = head0->data;
layers_difference |= node0->differences;
}
for (head1 = common_ancestor1->next; head1; head1 = head1->next)
{
node1 = head1->data;
layers_difference |= node1->differences;
}
return layers_difference;
}
static CoglBool
layer_state_equal (CoglPipelineLayerStateIndex state_index,
CoglPipelineLayer **authorities0,
CoglPipelineLayer **authorities1,
CoglPipelineLayerStateComparitor comparitor)
{
return comparitor (authorities0[state_index], authorities1[state_index]);
}
void
_cogl_pipeline_layer_resolve_authorities (CoglPipelineLayer *layer,
unsigned long differences,
CoglPipelineLayer **authorities)
{
unsigned long remaining = differences;
CoglPipelineLayer *authority = layer;
do
{
unsigned long found = authority->differences & remaining;
int i;
if (found == 0)
continue;
for (i = 0; TRUE; i++)
{
unsigned long state = (1L<<i);
if (state & found)
authorities[i] = authority;
else if (state > found)
break;
}
remaining &= ~found;
if (remaining == 0)
return;
}
while ((authority = _cogl_pipeline_layer_get_parent (authority)));
g_assert (remaining == 0);
}
CoglBool
_cogl_pipeline_layer_equal (CoglPipelineLayer *layer0,
CoglPipelineLayer *layer1,
unsigned long differences_mask,
CoglPipelineEvalFlags flags)
{
unsigned long layers_difference;
CoglPipelineLayer *authorities0[COGL_PIPELINE_LAYER_STATE_SPARSE_COUNT];
CoglPipelineLayer *authorities1[COGL_PIPELINE_LAYER_STATE_SPARSE_COUNT];
if (layer0 == layer1)
return TRUE;
layers_difference =
_cogl_pipeline_layer_compare_differences (layer0, layer1);
/* Only compare the sparse state groups requested by the caller... */
layers_difference &= differences_mask;
_cogl_pipeline_layer_resolve_authorities (layer0,
layers_difference,
authorities0);
_cogl_pipeline_layer_resolve_authorities (layer1,
layers_difference,
authorities1);
if (layers_difference & COGL_PIPELINE_LAYER_STATE_TEXTURE_TYPE)
{
CoglPipelineLayerStateIndex state_index =
COGL_PIPELINE_LAYER_STATE_TEXTURE_TYPE_INDEX;
if (!_cogl_pipeline_layer_texture_type_equal (authorities0[state_index],
authorities1[state_index],
flags))
return FALSE;
}
if (layers_difference & COGL_PIPELINE_LAYER_STATE_TEXTURE_DATA)
{
CoglPipelineLayerStateIndex state_index =
COGL_PIPELINE_LAYER_STATE_TEXTURE_DATA_INDEX;
if (!_cogl_pipeline_layer_texture_data_equal (authorities0[state_index],
authorities1[state_index],
flags))
return FALSE;
}
if (layers_difference & COGL_PIPELINE_LAYER_STATE_COMBINE &&
!layer_state_equal (COGL_PIPELINE_LAYER_STATE_COMBINE_INDEX,
authorities0, authorities1,
_cogl_pipeline_layer_combine_state_equal))
return FALSE;
if (layers_difference & COGL_PIPELINE_LAYER_STATE_COMBINE_CONSTANT &&
!layer_state_equal (COGL_PIPELINE_LAYER_STATE_COMBINE_CONSTANT_INDEX,
authorities0, authorities1,
_cogl_pipeline_layer_combine_constant_equal))
return FALSE;
Use GL_ARB_sampler_objects GL_ARB_sampler_objects provides a GL object which overrides the sampler state part of a texture object with different values. The sampler state that Cogl currently exposes is the wrap modes and filters. Cogl exposes the state as part of the pipeline layer state but without this extension GL only exposes it as part of the texture object state. This means that it won't work to use a single texture multiple times in one primitive with different sampler states. It also makes switching between different sampler states with a single texture not terribly efficient because it has to change the texture object state every time. This patch adds a cache for sampler states in a shared hash table attached to the CoglContext. The entire set of parameters for the sampler state is used as the key for the hash table. When a unique state is encountered the sampler cache will create a new entry, otherwise it will return a const pointer to an existing entry. That means we can have a single pointer to represent any combination of sampler state. Pipeline layers now just store this single pointer rather than storing all of the sampler state. The two separate state flags for wrap modes and filters have now been combined into one. It should be faster to compare the sampler state now because instead of comparing each value it can just compare the pointers to the cached sampler entries. The hash table of cached sampler states should only need to perform its more expensive hash on the state when a property is changed on a pipeline, not every time it is flushed. When the sampler objects extension is available each cached sampler state will also get a sampler object to represent it. The common code to flush the GL state will now simply bind this object to a unit instead of flushing the state though the CoglTexture when possible. Reviewed-by: Robert Bragg <robert@linux.intel.com>
2012-04-04 17:20:04 -04:00
if (layers_difference & COGL_PIPELINE_LAYER_STATE_SAMPLER &&
!layer_state_equal (COGL_PIPELINE_LAYER_STATE_SAMPLER_INDEX,
authorities0, authorities1,
Use GL_ARB_sampler_objects GL_ARB_sampler_objects provides a GL object which overrides the sampler state part of a texture object with different values. The sampler state that Cogl currently exposes is the wrap modes and filters. Cogl exposes the state as part of the pipeline layer state but without this extension GL only exposes it as part of the texture object state. This means that it won't work to use a single texture multiple times in one primitive with different sampler states. It also makes switching between different sampler states with a single texture not terribly efficient because it has to change the texture object state every time. This patch adds a cache for sampler states in a shared hash table attached to the CoglContext. The entire set of parameters for the sampler state is used as the key for the hash table. When a unique state is encountered the sampler cache will create a new entry, otherwise it will return a const pointer to an existing entry. That means we can have a single pointer to represent any combination of sampler state. Pipeline layers now just store this single pointer rather than storing all of the sampler state. The two separate state flags for wrap modes and filters have now been combined into one. It should be faster to compare the sampler state now because instead of comparing each value it can just compare the pointers to the cached sampler entries. The hash table of cached sampler states should only need to perform its more expensive hash on the state when a property is changed on a pipeline, not every time it is flushed. When the sampler objects extension is available each cached sampler state will also get a sampler object to represent it. The common code to flush the GL state will now simply bind this object to a unit instead of flushing the state though the CoglTexture when possible. Reviewed-by: Robert Bragg <robert@linux.intel.com>
2012-04-04 17:20:04 -04:00
_cogl_pipeline_layer_sampler_equal))
return FALSE;
if (layers_difference & COGL_PIPELINE_LAYER_STATE_USER_MATRIX &&
!layer_state_equal (COGL_PIPELINE_LAYER_STATE_USER_MATRIX_INDEX,
authorities0, authorities1,
_cogl_pipeline_layer_user_matrix_equal))
return FALSE;
if (layers_difference & COGL_PIPELINE_LAYER_STATE_POINT_SPRITE_COORDS &&
!layer_state_equal (COGL_PIPELINE_LAYER_STATE_POINT_SPRITE_COORDS_INDEX,
authorities0, authorities1,
_cogl_pipeline_layer_point_sprite_coords_equal))
return FALSE;
if (layers_difference & COGL_PIPELINE_LAYER_STATE_VERTEX_SNIPPETS &&
!layer_state_equal (COGL_PIPELINE_LAYER_STATE_VERTEX_SNIPPETS_INDEX,
authorities0, authorities1,
_cogl_pipeline_layer_vertex_snippets_equal))
return FALSE;
if (layers_difference & COGL_PIPELINE_LAYER_STATE_FRAGMENT_SNIPPETS &&
!layer_state_equal (COGL_PIPELINE_LAYER_STATE_FRAGMENT_SNIPPETS_INDEX,
authorities0, authorities1,
_cogl_pipeline_layer_fragment_snippets_equal))
return FALSE;
return TRUE;
}
static void
_cogl_pipeline_layer_free (CoglPipelineLayer *layer)
{
_cogl_pipeline_layer_unparent (COGL_NODE (layer));
if (layer->differences & COGL_PIPELINE_LAYER_STATE_TEXTURE_DATA &&
layer->texture != NULL)
cogl_object_unref (layer->texture);
if (layer->differences & COGL_PIPELINE_LAYER_STATE_VERTEX_SNIPPETS)
_cogl_pipeline_snippet_list_free (&layer->big_state->vertex_snippets);
if (layer->differences & COGL_PIPELINE_LAYER_STATE_FRAGMENT_SNIPPETS)
_cogl_pipeline_snippet_list_free (&layer->big_state->fragment_snippets);
if (layer->differences & COGL_PIPELINE_LAYER_STATE_NEEDS_BIG_STATE)
g_slice_free (CoglPipelineLayerBigState, layer->big_state);
g_slice_free (CoglPipelineLayer, layer);
}
void
_cogl_pipeline_init_default_layers (void)
{
CoglPipelineLayer *layer = g_slice_new0 (CoglPipelineLayer);
CoglPipelineLayerBigState *big_state =
g_slice_new0 (CoglPipelineLayerBigState);
CoglPipelineLayer *new;
_COGL_GET_CONTEXT (ctx, NO_RETVAL);
_cogl_pipeline_node_init (COGL_NODE (layer));
layer->index = 0;
layer->differences = COGL_PIPELINE_LAYER_STATE_ALL_SPARSE;
layer->unit_index = 0;
layer->texture = NULL;
layer->texture_type = COGL_TEXTURE_TYPE_2D;
Use GL_ARB_sampler_objects GL_ARB_sampler_objects provides a GL object which overrides the sampler state part of a texture object with different values. The sampler state that Cogl currently exposes is the wrap modes and filters. Cogl exposes the state as part of the pipeline layer state but without this extension GL only exposes it as part of the texture object state. This means that it won't work to use a single texture multiple times in one primitive with different sampler states. It also makes switching between different sampler states with a single texture not terribly efficient because it has to change the texture object state every time. This patch adds a cache for sampler states in a shared hash table attached to the CoglContext. The entire set of parameters for the sampler state is used as the key for the hash table. When a unique state is encountered the sampler cache will create a new entry, otherwise it will return a const pointer to an existing entry. That means we can have a single pointer to represent any combination of sampler state. Pipeline layers now just store this single pointer rather than storing all of the sampler state. The two separate state flags for wrap modes and filters have now been combined into one. It should be faster to compare the sampler state now because instead of comparing each value it can just compare the pointers to the cached sampler entries. The hash table of cached sampler states should only need to perform its more expensive hash on the state when a property is changed on a pipeline, not every time it is flushed. When the sampler objects extension is available each cached sampler state will also get a sampler object to represent it. The common code to flush the GL state will now simply bind this object to a unit instead of flushing the state though the CoglTexture when possible. Reviewed-by: Robert Bragg <robert@linux.intel.com>
2012-04-04 17:20:04 -04:00
layer->sampler_cache_entry =
_cogl_sampler_cache_get_default_entry (ctx->sampler_cache);
layer->big_state = big_state;
layer->has_big_state = TRUE;
/* Choose the same default combine mode as OpenGL:
* RGBA = MODULATE(PREVIOUS[RGBA],TEXTURE[RGBA]) */
big_state->texture_combine_rgb_func =
COGL_PIPELINE_COMBINE_FUNC_MODULATE;
big_state->texture_combine_rgb_src[0] =
COGL_PIPELINE_COMBINE_SOURCE_PREVIOUS;
big_state->texture_combine_rgb_src[1] =
COGL_PIPELINE_COMBINE_SOURCE_TEXTURE;
big_state->texture_combine_rgb_op[0] =
COGL_PIPELINE_COMBINE_OP_SRC_COLOR;
big_state->texture_combine_rgb_op[1] =
COGL_PIPELINE_COMBINE_OP_SRC_COLOR;
big_state->texture_combine_alpha_func =
COGL_PIPELINE_COMBINE_FUNC_MODULATE;
big_state->texture_combine_alpha_src[0] =
COGL_PIPELINE_COMBINE_SOURCE_PREVIOUS;
big_state->texture_combine_alpha_src[1] =
COGL_PIPELINE_COMBINE_SOURCE_TEXTURE;
big_state->texture_combine_alpha_op[0] =
COGL_PIPELINE_COMBINE_OP_SRC_ALPHA;
big_state->texture_combine_alpha_op[1] =
COGL_PIPELINE_COMBINE_OP_SRC_ALPHA;
big_state->point_sprite_coords = FALSE;
cogl_matrix_init_identity (&big_state->matrix);
ctx->default_layer_0 = _cogl_pipeline_layer_object_new (layer);
/* TODO: we should make default_layer_n comprise of two
* descendants of default_layer_0:
* - the first descendant should change the texture combine
* to what we expect is most commonly used for multitexturing
* - the second should revert the above change.
*
* why? the documentation for how a new layer is initialized
* doesn't say that layers > 0 have different defaults so unless
* we change the documentation we can't use different defaults,
* but if the user does what we expect and changes the
* texture combine then we can revert the authority to the
* first descendant which means we can maximize the number
* of layers with a common ancestor.
*
* The main problem will be that we'll need to disable the
* optimizations for flattening the ancestry when we make
* the second descendant which reverts the state.
*/
ctx->default_layer_n = _cogl_pipeline_layer_copy (layer);
new = _cogl_pipeline_set_layer_unit (NULL, ctx->default_layer_n, 1);
g_assert (new == ctx->default_layer_n);
/* Since we passed a newly allocated layer we don't expect that
* _set_layer_unit() will have to allocate *another* layer. */
/* Finally we create a dummy dependant for ->default_layer_n which
* effectively ensures that ->default_layer_n and ->default_layer_0
* remain immutable.
*/
ctx->dummy_layer_dependant =
_cogl_pipeline_layer_copy (ctx->default_layer_n);
}
void
_cogl_pipeline_layer_pre_paint (CoglPipelineLayer *layer)
{
CoglPipelineLayer *texture_authority;
texture_authority =
_cogl_pipeline_layer_get_authority (layer,
COGL_PIPELINE_LAYER_STATE_TEXTURE_DATA);
if (texture_authority->texture != NULL)
{
CoglTexturePrePaintFlags flags = 0;
CoglPipelineFilter min_filter;
CoglPipelineFilter mag_filter;
_cogl_pipeline_layer_get_filters (layer, &min_filter, &mag_filter);
if (min_filter == COGL_PIPELINE_FILTER_NEAREST_MIPMAP_NEAREST
|| min_filter == COGL_PIPELINE_FILTER_LINEAR_MIPMAP_NEAREST
|| min_filter == COGL_PIPELINE_FILTER_NEAREST_MIPMAP_LINEAR
|| min_filter == COGL_PIPELINE_FILTER_LINEAR_MIPMAP_LINEAR)
flags |= COGL_TEXTURE_NEEDS_MIPMAP;
_cogl_texture_pre_paint (texture_authority->texture, flags);
}
}
/* Determines if we need to handle the RGB and A texture combining
* separately or is the same function used for both channel masks and
* with the same arguments...
*/
CoglBool
_cogl_pipeline_layer_needs_combine_separate
(CoglPipelineLayer *combine_authority)
{
CoglPipelineLayerBigState *big_state = combine_authority->big_state;
int n_args;
int i;
if (big_state->texture_combine_rgb_func !=
big_state->texture_combine_alpha_func)
return TRUE;
n_args = _cogl_get_n_args_for_combine_func (big_state->texture_combine_rgb_func);
for (i = 0; i < n_args; i++)
{
if (big_state->texture_combine_rgb_src[i] !=
big_state->texture_combine_alpha_src[i])
return TRUE;
/*
* We can allow some variation of the source operands without
* needing a separation...
*
* "A = REPLACE (CONSTANT[A])" + either of the following...
* "RGB = REPLACE (CONSTANT[RGB])"
* "RGB = REPLACE (CONSTANT[A])"
*
* can be combined as:
* "RGBA = REPLACE (CONSTANT)" or
* "RGBA = REPLACE (CONSTANT[A])" or
*
* And "A = REPLACE (1-CONSTANT[A])" + either of the following...
* "RGB = REPLACE (1-CONSTANT)" or
* "RGB = REPLACE (1-CONSTANT[A])"
*
* can be combined as:
* "RGBA = REPLACE (1-CONSTANT)" or
* "RGBA = REPLACE (1-CONSTANT[A])"
*/
switch (big_state->texture_combine_alpha_op[i])
{
case GL_SRC_ALPHA:
switch (big_state->texture_combine_rgb_op[i])
{
case GL_SRC_COLOR:
case GL_SRC_ALPHA:
break;
default:
return FALSE;
}
break;
case GL_ONE_MINUS_SRC_ALPHA:
switch (big_state->texture_combine_rgb_op[i])
{
case GL_ONE_MINUS_SRC_COLOR:
case GL_ONE_MINUS_SRC_ALPHA:
break;
default:
return FALSE;
}
break;
default:
return FALSE; /* impossible */
}
}
return FALSE;
}