Mesa before version 8.0.2 has a slow read pixels path that gets used
with the Intel driver where it converts all of the pixels into a
floating point representation and back even if the data is being read
into exactly the same format. There is however a faster path using the
blitter when reading into a PBO with BGRA format. It works out faster
to read into a PBO and then memcpy back out into the application's
buffer even though it adds an extra memcpy. This patch adds a
workaround in cogl_framebuffer_read_pixels_into_bitmap when it detects
this situation. In that case it will create a temporary CoglBitmap
using cogl_bitmap_new_with_size, read into it and then memcpy the data
back out.
The main impetus for this patch is that Gnome Shell has implemented
this workaround directly using GL calls but it seems like the kind of
thing that would sit better at the Cogl layer.
Reviewed-by: Robert Bragg <robert@linux.intel.com>
This adds a CoglGpuInfo struct to the CoglContext which contains some
enums describing the GL driver in use. This currently includes the
driver package (ie, is it Mesa) the version number of the package and
the vendor of the GPU (ie, is it by Intel). There is also a bitmask
which will contain the workarounds that we should do for that
particular driver configuration. The struct is initialised on context
creation by using a series of string comparisons on the strings
returned from glGetString.
Reviewed-by: Robert Bragg <robert@linux.intel.com>
The Intel driver currently has an optimisation when calling
glReadPixels into a PBO so that it will use a blit instead of the Mesa
fallback path. However this only works if the GL_PACK_ALIGNMENT is
exactly 1, even if this would be equivalent to a higher alignment
value because the bpp*width is already aligned. To make it more likely
to hit this fast path, we now detect this situation and explicitly use
an alignment of 1. To make this work the texture driver needs to be
passed down the bpp*width as well as the rowstride when configuring
the alignment.
Reviewed-by: Robert Bragg <robert@linux.intel.com>
Instead of having a series of if-statements this adds an inline
function to calculate the alignment directly using ffs which is
probably slightly faster. Admittedly this is a pointless
micro-optimisation but I think it makes the code looks a bit neater
anyway.
Reviewed-by: Robert Bragg <robert@linux.intel.com>
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>
Two of the meta texture constructors which take a flags parameter were
ignoring the COGL_TEXTURE_NO_AUTO_MIPMAP flag when creating an
underlying CoglTexture2D. These have now been fixed to call
cogl_primitive_texture_set_auto_mipmap after constructing the texture.
Reviewed-by: Robert Bragg <robert@linux.intel.com>
This adds public constructors which take a CoglBitmap to all primitive
texture types. This constructor should be considered the canonical
constructor for initializing the texture with data because it should
be possible to wrap any type of data in a CoglBitmap. Having at least
this single constructor avoids the need to have an explosion of
constructors such as new_from_data, new_from_pixel_buffer and
new_from_file etc.
The already available internal bitmap constructor for CoglTexture2D
has had its flags parameter removed under the assumption that flags do
not make sense for primitive textures. The meta constructor
cogl_texture_new_from_bitmap now just explicitly calls set_auto_mipmap
after constructing the texture depending on the value of the
COGL_TEXTURE_NO_AUTO_MIPMAP flag.
Reviewed-by: Robert Bragg <robert@linux.intel.com>
This interface represents any textures that are backed by a single
texture in GL and that can be used directly with the
cogl_framebuffer_draw_attributes family of functions. This currently
equates to CoglTexture2D, CoglTexture3D and CoglTextureRectangle.
The interface currently has only one method called
cogl_primitive_set_auto_mipmap. This replaces the
COGL_TEXTURE_NO_AUTO_MIPMAP flag from the CoglTextureFlags parameter
in the constructors. None of the other flags in CoglTextureFlags make
sense for primitive textures so it doesn't seem like a good idea to
need them for primitive constructors.
There is a boolean in the vtable to mark whether a texture type is
primitive which the new cogl_is_primitive function uses. There is also
a new texture virtual called set_auto_mipmap which is only required to
be implemented for primitive textures.
Reviewed-by: Robert Bragg <robert@linux.intel.com>
This adds a context member to CoglBitmap which stores the context it
was created with. That way it can be used in texture constructors
which use a bitmap. There is also an internal private function to get
the context out of the bitmap which all of the texture constructors
now use. _cogl_texture_3d_new_from_bitmap has had its context
parameter removed so that it more closely matches the other bitmap
constructors.
Reviewed-by: Robert Bragg <robert@linux.intel.com>
The previous code to initialise the depth state on the default
pipeline wasn't initialising the magic number. If you later tried to
retrieve the depth state using cogl_pipeline_get_depth_state you would
end up with an invalid depth state struct and you would just get
warnings if you tried to use it for anything. This patch just replaces
the initialisation with a call to cogl_depth_state_init because it
uses the same values anyway.
Reviewed-by: Robert Bragg <robert@linux.intel.com>
get_texture_bits_via_offscreen does not check the return value of
cogl_framebuffer_read_pixels_into_bitmap which results into never
using the fallback path texture_get_cb.
cogl_framebuffer_read_pixels_into_bitmap does not check whether the framebuffer
is properly allocated though; so fix that as well.
https://bugzilla.gnome.org/show_bug.cgi?id=673137
Reviewed-by: Neil Roberts <neil@linux.intel.com>
This extension lets you upload texture data from a subregion of a
buffer by passing GL_UNPACK_ROW_LENGTH, GL_UNPACK_SKIP_PIXELS and
GL_UNPACK_SKIP_ROWS to glPixelStore. When this extension is available
the GLES texture driver will now avoid making a copy of the bitmap
when a subregion is used.
Note that Mesa doesn't currently advertise this extension but I've
made a patch to propose it:
http://lists.freedesktop.org/archives/mesa-dev/2012-March/020191.html
This extension allows an application to upload data in BGRA format. We
can use this to avoid a conversion in Cogl whenever it is given BGRA
data. This is quite useful when uploading data generated by Cairo
because at least on little-endian architectures that ends up as BGRA.
The patch just makes the pixel_format_to_gl implementation return
GL_BGRA_EXT for the data format and internal format whenever
COGL_PIXEL_FORMAT_BGRA_8888{,_PRE} is used.
A small caveat with this patch is that once a texture is created as
GL_BGRA, when later using glTexSubImage2D to update the texture it
must always be given data as GL_BGRA. Currently this just works out
because we store the internal format of a texture as a CoglPixelFormat
and we already swizzle the data if it does not match exactly on GLES.
However if we later switch to using a different enum for internal
formats then we might lose the ability to store the component ordering
so we'll have to think of another way to do this.
Cogl already had a vtable for the texture driver. This ended up being
used for some things that are not strictly related to texturing such
as converting between pixel formats and GL enums. Some other functions
that are driver dependent such as updating the features were not
indirected through a vtable but instead switched directly by looking
at the ctx->driver enum value. This patch normalises to the two uses
by adding a separate vtable for driver functions not related to
texturing and moves the pixel format conversion functions to it from
the texture driver vtable. It also adds a context parameter to all of
the functions in the new driver vtable so that they won't have to rely
on the global context.
Because the wayland-client-protocol.h header defines symbols that
collide with the wayland-server-protocol.h header we allow applications
to explicitly ensure that they are only including one at a time by
exposing corresponding <cogl/cogl-wayland-client.h> and
<cogl/cogl-wayland-server.h> headers. This also adds a missing guard to
cogl-texture-2d.h that it isn't included directly.
Reviewed-by: Neil Roberts <neil@linux.intel.com>
The fallback code using stb-image.c was missed out in the upgrade to
cogl_bitmap_new_for_data from commit d18b59d9e6 so it wouldn't
compile.
Reviewed-by: Robert Bragg <robert@linux.intel.com>
• The documentation for the framebuffer and texture interfaces had a
mis-matching open and close <note> tag so DocBook got upset and the
whole documentation disappeared.
• A lot of symbols from the cogl_framebuffer_* interface were missing
from the cogl-2.0-experimental-sections.txt file.
• cogl_framebuffer_frustum had the wrong version in its Since tag:
Reviewed-by: Robert Bragg <robert@linux.intel.com>
_cogl_framebuffer_draw_multitextured_rectangles had a typo in the
function name in the declaration so it was generating a lot of
compile warnings.
Reviewed-by: Robert Bragg <robert@linux.intel.com>
This adds experimental 2.0 api replacements for the cogl_rectangle[_*]
functions that don't depend on having a current pipeline set on the
context via cogl_{set,push}_source() or having a current framebuffer set
on the context via cogl_push_framebuffer(). The aim for 2.0 is to switch
away from having a statefull context that affects drawing to having
framebuffer drawing apis that are explicitly passed a framebuffer and
pipeline.
To test this change several of the conformance tests were updated to use
this api instead of cogl_rectangle and
cogl_rectangle_with_texture_coords. Since it's quite laborious going
through all of the conformance tests the opportunity was taken to make
other clean ups in the conformance tests to replace other uses of
1.x api with experimental 2.0 api so long as that didn't affect what was
being tested.
This was causing the DocBook for the documentation to be invalid so
all of the framebuffer documentation disappeared.
Reviewed-by: Robert Bragg <robert@linux.intel.com>
This adds a public convenience wrapper around
cogl_framebuffer_read_pixels_into_bitmap which allocates a temporary
CoglBitmap to read into the application's own buffer. This can only be
used for the 99% common case where the rowstride is exactly the
bpp*width and the source is the color buffer.
Reviewed-by: Robert Bragg <robert@linux.intel.com>
In theory none of the journal flushing code should be using anything
that relies on the global framebuffer stack because it should all be
using the new 2.0-style API which explicitly mentions the target
framebuffer. Eventually we want to get rid of the framebuffer stack so
we might as well remove the push and pop now.
Reviewed-by: Robert Bragg <robert@linux.intel.com>
Previously when adding a quad to the journal it would assume the
journal belongs to the framebuffer at the top of the framebuffer stack
and store a reference to that. We eventually want to get rid of the
framebuffer stack so we should avoid using it here. The journal now
takes a pointer back to the framebuffer in its constructor and it
always retains the pointer. As was done previously, the journal still
does not take a reference on the framebuffer unless it is non-empty so
it does not create a permanent circular reference.
Reviewed-by: Robert Bragg <robert@linux.intel.com>
_cogl_shader_compile_real spews a warning when
shader compilation fails if COGL_GL_DEBUG is
defined. This warning is never freed.
https://bugzilla.gnome.org/show_bug.cgi?id=672243
Reviewed-by: Neil Roberts <neil@linux.intel.com>
The 1.0 wrapper for cogl_path_curve_to was using the wrong value for
y_1 so it wouldn't work.
The patch was written by Dénes Almási.
https://bugzilla.gnome.org/show_bug.cgi?id=672174
Reviewed-by: Neil Roberts <neil@linux.intel.com>
We initially assumed that copy_sub_buffer is synchronized on
which is only the case for a subset of GPUs for example it is not
synchronized on INTEL gen6 and gen7, so we remove this assumption
for now.
We should have a specific driver / GPU whitelist if we want to enable
this.
https://bugzilla.gnome.org/show_bug.cgi?id=669122
Reviewed-by: Robert Bragg <robert@linux.intel.com>
Reviewed-by: Neil Roberts <neil@linux.intel.com>
None of the other cogl_is_blah functions have a const pointer so this
is just for consistency. It helps if someone is trying to have an
array of type-check function pointers to determine the Cogl object
type because in that case all of the functions would have to have the
same prototype.
Reviewed-by: Robert Bragg <robert@linux.intel.com>
This creates a CoglBitmap which points into an existing buffer in
system memory. That way it can be used to create a texture or to read
pixel data into. The function replaces the existing internal function
_cogl_bitmap_new_from_data but removes the destroy notify call back.
If the application wants notification of destruction it can just use
the cogl_object_set_user_data function as normal. Internally there is
now a convenience function to create a bitmap for system memory and
automatically free the buffer using that mechanism.
The name of the function is inspired by
cairo_image_surface_create_for_data which has similar semantics.
Reviewed-by: Robert Bragg <robert@linux.intel.com>
On GLES, when reading texture data back it may need to allocate a
temporary CoglBitmap if the requested format is not supported by the
driver. Previously it would then copy this temporary buffer back into
the user's buffer by calling _cogl_bitmap_convert which would allocate
a second temporary buffer. It would then copy that data into the
user's buffer. This patch changes it to create a CoglBitmap which
points to the user's data and then convert directly into that buffer
using the new _cogl_bitmap_convert_into_bitmap.
This also fixes a small leak where target_bmp would not get freed if
the target format and the closest supported format do match.
Reviewed-by: Robert Bragg <robert@linux.intel.com>
The 2.0 API for querying features (cogl_has_feature etc) does not
conflict with the old 1.0 API (cogl_features_available) so we might as
well enable it when the experimental API is requested without
requesting the 2.0-only API.
Reviewed-by: Robert Bragg <robert@linux.intel.com>
The if-undefined fallback declaration for GL_PACK_INVERT_MESA was
originally added in cogl.c along with code to use it (as part of commit
6f79eb8a5a). Later on, commit
10a38bb14f moved the code that used it to
cogl-framebuffer.c but didn't move the define along with it. Do that
now.
https://bugzilla.gnome.org/show_bug.cgi?id=672038
Reviewed-by: Neil Roberts <neil@linux.intel.com>
-Removed checks for COGL_ENABLE_EXPERIMENTAL_API since these APIs are
always built into the shared library
-Re-organised the API listing a bit so that they are in alphabetical order
Reviewed-by: Neil Roberts <neil@linux.intel.com>
This API was re-added into COGL for the 1.10.x release as of commit
361bd516f. This will be removed once we branch into the 1.11.x development
cycle.
Reviewed-by: Neil Roberts <neil@linux.intel.com>
If the matrix was reallocated we would use values from the stack
for the matrix parameters. This fixes that and also uses the
function instead of out of lining the same code.
https://bugzilla.gnome.org/show_bug.cgi?id=671985
Reviewed-by: Robert Bragg <robert@linux.intel.com>
Reviewed-by: Neil Roberts <neil@linux.intel.com>
This is only used internally when debugging is enabled to give a
human-readable name to a GL error so we shouldn't be exporting it
outside of the library. This just adds an underscore to the symbol
name. This shouldn't end up removing any public symbols from the 1.9.8
release because by default a non-git build disables debug so it wasn't
exported anyway.
Reviewed-by: Robert Bragg <robert@linux.intel.com>
The cleanup in 185630085 removed some symbols that were previously
exported as public experimental API in Cogl 1.9.8. That release is
already well after the point where we were meant to freeze the ABI so
we probably shouldn't be breaking it again. This patch adds the
removed functions back in so that for 1.9.10 we won't have to bump the
soname. The symbols are bundled together in a new file called
cogl2-compatibility.c so that they will be easy to remove again after
we can break ABI. It is expected that we will revert this patch
immediately after branching for Cogl 1.10.
Reviewed-by: Robert Bragg <robert@linux.intel.com>
I don't think there's really any point in this cache because the
pipeline code completely owns the point size state. Pipelines are
already compared for whether their point size state is different
before setting it so it shouldn't result in any extra calls to
glPointSize apart from maybe when the first pipeline is initially
flushed.
Reviewed-by: Robert Bragg <robert@linux.intel.com>
When using the GLSL vertend on GL, the point size was being flushed in
_cogl_pipeline_vertend_glsl_start. However, this function bails out
early if the pipeline already has a usable program so it would not hit
the code to flush the point size in that case. This patch moves the
code to _cogl_pipeline_vertend_glsl_end so that it will always be
flushed if it is different. That is the same place that is flushed for
the fixed vertend.
Reviewed-by: Robert Bragg <robert@linux.intel.com>
This option to GCC makes it give a warning whenever a global function
is defined without a declaration. This should catch cases were we've
defined a function but forgot to put it in a header. In that case it
is either only used within one file so we should make it static or we
should declare it in a header.
The following changes where made to fix problems:
• Some functions were made static
• cogl-path.h (the one containing the 1.0 API) was split into two
files, one defining the functions and one defining the enums so that
cogl-path.c can include the enum and function declarations from the
2.0 API as well as the function declarations from the 1.0 API.
• cogl2-clip-state has been removed. This only had one experimental
function called cogl_clip_push_from_path but as this is unstable we
might as well remove it favour of the equivalent cogl_framebuffer_*
API.
• The GLX, SDL and WGL winsys's now have a private header to define
their get_vtable function instead of directly declaring in the C
file where it is called.
• All places that were calling COGL_OBJECT_DEFINE need to have the
cogl_is_whatever function declared so these have been added either
as a public function or in a private header.
• Some files that were not including the header containing their
function declarations have been fixed to do so.
• Any unused error quark functions have been removed. If we later want
them we should add them back one by one and add a declaration for
them in a header.
• _cogl_is_framebuffer has been renamed to cogl_is_framebuffer and
made a public function with a declaration in cogl-framebuffer.h
• Similarly for CoglOnscreen.
• cogl_vdraw_indexed_attributes is called
cogl_framebuffer_vdraw_indexed_attributes in the header. The
definition has been changed to match the header.
• cogl_index_buffer_allocate has been removed. This had no declaration
and I'm not sure what it's supposed to do.
• CoglJournal has been changed to use the internal CoglObject macro so
that it won't define an exported cogl_is_journal symbol.
• The _cogl_blah_pointer_from_handle functions have been removed.
CoglHandle isn't used much anymore anyway and in the few places
where it is used I think it's safe to just use the implicit cast
from void* to the right type.
• The test-utils.h header for the conformance tests explicitly
disables the -Wmissing-declaration option using a pragma because all
of the tests declare their main function without a header. Any
mistakes relating to missing declarations aren't really important
for the tests.
• cogl_quaternion_init_from_quaternion and init_from_matrix have been
given declarations in cogl-quaternion.h
Reviewed-by: Robert Bragg <robert@linux.intel.com>
We are in the process of removing all _EXP suffix mangling for
experimental APIs (Ref: c6528c4b6c) and adding missing gtk-doc
comments so that we can instead rely on the "Stability: unstable"
markers in the gtk-doc comments. This patch tackles the
cogl-wayland-renderer api symbols.
Reviewed-by: Neil Roberts <neil@linux.intel.com>
Some of the state when flushing a pipeline depends on the current
framebuffer being used. These are:
• The matrix stack, so that it can flip vertically or not depending on
whether the framebuffer is offscreen.
• The colormask. This is combined with the framebuffer's color mask.
• The cull face mode. If the framebuffer is offscreen then backface
culling is translated to frontface culling and vice-versa.
These states were not working if the new framebuffer draw_primitive
API was used because in that case the framebuffer is not pushed to the
framebuffer stack so it would use the wrong one. This patch changes it
to use ctx->current_draw_buffer which is a pointer to the framebuffer
whose state was last flushed.
https://bugzilla.gnome.org/show_bug.cgi?id=670793
Reviewed-by: Robert Bragg <robert@linux.intel.com>
This adds a small cogl_bitmap_get_buffer public function. Note that
this can return NULL if the bitmap was not created with a pixel
buffer. It might be nice to change this eventually so that all bitmaps
have a pixel buffer.
Reviewed-by: Robert Bragg <robert@linux.intel.com>
