For the componentAlpha with PictOpOver, we use two pass
rendering to implement it. Previous implementation call
two times the glamor_composite_... independently which is
very inefficient. Now we change the control flow, and do
the two pass internally and avoid duplicate works.
For the x11perf -rgb10text, this optimization can get about
30% improvement.
Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
To split a glyph's extent region to three sub-boxes
as below.
left box 2 x h
center box (w-4) x h
right box 2 x h
Take a simple glyph A as an example:
*
__* *__
*****
* *
~~ ~~
The left box and right boxes are both 2 x 2. The center box
is 2 x 4.
The left box has two bitmaps 0001'b and 0010'b to indicate
the real inked area.
The right box also has two bitmaps 0010'b and 0001'b.
And then we can check the inked area in left and right boxes with
previous glyph. If the direction is from left to right, then we
need to check the previous right bitmap with current left bitmap.
And if we found the center box has overlapped or we overlap with
not only the previous glyph, we will treat it as real overlapped
and will render the glyphs via mask.
If we only intersect with previous glyph on the left/right edge.
Then we further compute the real overlapped bits. We set a loose
check criteria here, if it has less than two pixel overlapping, we
treat it as non-overlapping.
With this patch, The aa10text boost fom 1660000 to 320000.
Almost double the performance! And the cairo test result is the
same as without this patch.
Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
Because when mask depth is 1, there is no Anti-Alias at all,
in this case, the directly render can work well and it is faseter.
Signed-off-by: Junyan He <junyan.he@linux.intel.com>
We firstly get the render area by clipping the trapezoid
with the clip rect, then split the clipped area into small
triangles and use the composite logic to generate the result
directly. This manner is fast but have the problem that
some implementation of GL do not implement the Anti-Alias
of triangles fill, so the edge sometimes has sawtooth. It is
not acceptable when use trapezoid to approximate circles and
wide lines.
Signed-off-by: Junyan He <junyan.he@linux.intel.com>
We seperate the composite to two phases, firstly to
select the shader according to source type and logic
op, setting the right parameters. Then we emit the
vertex array to generate the dest result.
The reason why we do this is that the shader may be
used to composite no only rect, trapezoid and triangle
render function can also use it to render triangles and
polygens. The old function glamor_composite_with_shader
do the whole two phases work and can not match the
new request.
Signed-off-by: Junyan He <junyan.he@linux.intel.com>
The old manner of trapezoid render uses pixman to
generate a mask pixmap and upload it to the GPU.
This effect the performance. We now use shader to
generate the temp trapezoid mask to avoid the
uploading of this pixmap.
We implement a anti-alias manner in the shader
according to pixman, which will caculate the area
inside the trapezoid dividing total area for every
pixel and assign it to the alpha value of that pixel.
The pixman use a int-to-fix manner to approximate but
the shader use float, so the result may have some
difference.
Because the array in the shader has optimization problem,
we need to emit the vertex of every trapezoid every
time, which will effect the performance a lot. Need to
improve it.
Signed-off-by: Junyan He <junyan.he@linux.intel.com>
Create the file glamor_trapezoid.c, extract the logic
relating to trapezoid from glamor_render.c to this file.
Signed-off-by: Junyan He <junyan.he@linux.intel.com>
If the source and destination are the same pixmap/fbo, and we
need to split the copy to small pieces. Then we do need to
consider the sequence of the small pieces when the copy area
has overlaps. This commit take the reverse/upsidedown into
the clipping function, thus it can generate correct sequence
and avoid corruption self copying.
Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
The simplest way to support large pixmap's self compositing
is to just clone a pixmap private data structure, and change
the fbo and box to point to the correct postions. Don't need
to copy a new box.
Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
Added infrastructure for largepixmap, this commit implemented:
1. Create/Destroy large pixmap.
2. Upload/Download large pixmap.
3. Implement basic repeat normal support.
3. tile/fill/copyarea large pixmap get supported.
The most complicated part glamor_composite still not implemented.
Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
This is the first commit to add support for large pixmap.
The large here means a pixmap is larger than the texutre's
size limitation thus can't fit into one single texutre.
The previous implementation will simply fallback to use a
in memory pixmap to contain the large pixmap which is
very slow in practice.
The basic idea here is to use an array of texture to hold
the large pixmap. And when we need to get a specific area
of the pixmap, we just need to compute/clip the correct
region and find the corresponding fbo.
We need to implement some auxiliary routines to clip every
rendering operations into small pieces which can fit into
one texture.
The complex part is the transformation/repeat/repeatReflect
and repeat pad and their comination. We will support all of
them step by step.
This commit just add some necessary data structure to represent
the large pixmap, and doesn't change any rendering process.
This commit doesn't add real large pixmap support.
Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
1. Extract the logic of gradient from the glamor_render.c
to the file glamor_gradient.c.
2. Modify the logic of gradient pixmap gl draw. Use the
logic like composite before, but the gradient always just
have one rect to render, so no need to set the VB and EB,
replace it with just call glDrawArrays. 3.Kill all the
warning in glamor_render.c
Reviewed-by: Zhigang Gong<zhigang.gong@linux.intel.com>
Signed-off-by: Junyan He <junyan.he@linux.intel.com>
Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
Previous implementation set the whole fbo's width and height as the
viewpoint. This may increase the numerical error as we may only has
a partial region as the valid pixmap. So add a new marco
pixmap_priv_get_dest_scale to get proper scale factor for the
destination pixmap. For the source/mask pixmap, we still need to
consider the whole fbo's size.
Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
Caching texture objects is not necessary based on previous testing.
To keep the code simple, we remove it.
Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
Currently set it to 256MB. If cache pool watermark increases
to this value, then don't push any fbo to this pool, will purge
the fbo directly.
Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
We should use difference calculation for these two repeat mode
when we are a sub region within one texture.
Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
As PVR glsl compiler seems doesn't support external fragment
function, and fails at compile gradient shader. Disable it
for now. We may need to modify gradient shader to don't use
external function.
Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
Fix the problem of memory leak in gradient pixmap
generating. The problem caused by we do not call
glDeleteShader when destroy a shader program. This patch
will split the gradient pixmap generating to three
category. If nstops < 6, we will use the no array version
of the shader, which has the best performance. Else if
nstops < 16, we use array version of the shader, which is
compiled and linked at screen init stage. Else if nstops >
16, we dynamically create a new shader program, and this
program will be cached until bigger nstops.
Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
We have disabled this feature for a long time, and previous
testing shows that this(pending fill) will not bring observed
performance gain. Now remove it.
Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
As a textured_drm pixmap has a drm bo attached to it, and
it's the DDX layer to set it stride value. In some case,
the stride value is not equal to PixmapBytePad(w, depth)
which is used within glamor.
Then if it is the case, we have two choice, one is to set
the GL_PACK_ROW_LENGTH/GL_UNPACK_ROW_LENGTH when we need
to download or upload the pixmap. The other option is to
change the pixmap's devKind to match the one glamor is using
when downloading the pixmap, and restore it to the drm stride
after uploading the pixmap.
We choose the 2nd option, as GLES doesn't support the first
method.
Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
If a pixmap doesn't have a private, then set its type to
GLAMOR_MEMORY, and thus it will get a valid private.
Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
Just as the downloading side, we can upload an sub region data to
a pixmap's specified region. The data could be in memory or in a
pbo buffer.
Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
Introduced two function glamor_get_sub_pixmap/glamor_put_sub_pixmap,
can easily used to get and put sub region of a big textured pixmap.
And it can use pbo if possible.
To support download a big textured pixmap's sub region to another
pixmap's pbo, we introduce a new type of pixmap GLAMOR_MEMORY_MAP.
This type of pixmap has a valid devPrivate.ptr pointer, and that
pointer points to a pbo mapped address.
Now, we are ready to refine those
glamor_prepare_access/glamor_finish_access pairs.
Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
As GLES2 doesn't support LogiOps, we have to fallback
here. GLES2 programing guide's statement is as below:
"In addition, LogicOp is removed as it is very
infrequently used by applications and the OpenGL ES
working group did not get requests from independent
software vendors (ISVs) to support this feature in
OpenGL ES 2.0."
So, I think, fallback here may not a big deal ;).
Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
Added color conversion code to support 1555/2101010
formats,now gles2 can pass the render check with all
formats.
We use 5551 to represent 1555, and do the revertion
if downloading/uploading is needed.
For 2101010, as gles2 doesn't support reading the
identical formats. We have to use 8888 to represent,
thus we may introduce some accurate problem. But anyway,
we can pass the error checking in render check, so that
may not be a big problem.
Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
This patch fixed two major problems when we do the color convesion with
GLES2.
1. lack of necessary formats in FBO pool.
GLES2 has three different possible texture formats, GL_RGBA,
GL_BGRA and GL_ALPHA. Previous implementation only has one bucket
for all the three formats which may reuse a incorrect texture format
when do the cache lookup. After this fix, we can enable fbo safely
when running with GLES2.
2. Refine the format matching method in
glamor_get_tex_format_type_from_pictformat.
If both revertion and swap_rb are needed, for example use GL_RGBA
to represent PICT_b8g8r8a8. Then the downloading and uploading should
be handled differently.
The picture's format is PICT_b8g8r8a8,
Then the expecting color layout is as below (little endian):
0 1 2 3 : address
a r g b
Now the in GLES2 the supported color format is GL_RGBA, type is
GL_UNSIGNED_TYPE, then we need to shuffle the fragment
color as :
frag_color = sample(texture).argb;
before we use glReadPixel to get it back.
For the uploading process, the shuffle is a revert shuffle.
We still use GL_RGBA, GL_UNSIGNED_BYTE to upload the color
to a texture, then let's see
0 1 2 3 : address
a r g b : correct colors
R G B A : GL_RGBA with GL_UNSIGNED_BYTE
Now we need to shuffle again, the mapping rule is
r = G, g = B, b = A, a = R. Then the uploading shuffle is as
below:
frag_color = sample(texture).gbar;
After this commit, gles2 version can pass render check with all
the formats except those 1555/2101010.
Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
I found when enable the gradient shader, the firefox's tab's
background has incorrect rendering result.
Need furthr investigation, for now, just disable it.
Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
Prepare for modification of gradient using shader. The
gradient pixmaps now is generated by pixman and we will
replace them with shader. Add structure fields and
dispatch functions which will be needed. Some auxiliary
macro for vertex convert.
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Zhigang Gong <zhigang.gong@linux.intel.com>
Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
Because the file list.h in xorg/include has changed the
functions and struct names, adding xorg_ prefix before
the original name. So Modify glamor_screen_private
struct and the code which use list's functions in
glamor_fbo.c. We hack at glamor_priv.h avoid the
compile error when using old version xserver header
file.
Signed-off-by: Junyan He <junyan.he@linux.intel.com>
This commit added two APIs to support the DRI swap buffer.
one is glamor_egl_exchange_buffers() which can swap two
pixmaps' underlying KHRimages/fbos/texs. The DDX layer should
exchange the DRM bos to make them consistent to each other.
Another API is glamor_egl_create_textured_screen_ext(), which
extent one more parameters to track the DDX layer's back pixmap
pointer. This is for the triple buffer support. When using triple
buffer, the DDX layer will keep a back pixmap rather then the
front pixmap and the pixmap used by the DRI2 client. And during
the closing screen stage, we have to dereference all the back
pixmap's glamor resources. Thus we have to extent this API to
register it when create new screen.
Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
Then we don't need to fixup the larger pixmap to the exact
size, just need to let the shader to re-calculate the correct
texture coords.
Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
Renaming glamor_priv->dispatch and wrapping the access to
the dispatch table with a function that also ensured the
context was bound.
dispatch = glamor_get_dispatch(glamor_priv);
...
glamor_put_dispatch(glamor_priv);
So that we catch all places where we attempt to call into GL withouta
context. As an optimisation we can then do glamor_get_context();
glamor_put_context() around the rendering entry points to reduce the
frequency of having to restore the old context. (Along with allowing
the context to be recursively acquired and making the old context part of
the glamor_egl state.)
Reviewed-by: Zhigang Gong <zhigang.gong@linux.intel.com>
Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
If we are using MESA as our GL library, then both xserver's
GLX and glamor are link to the same library. As xserver's
GLX has its own _glapi_get/set_context/dispatch etc, and it
is a simplified version derived from mesa thus is not
sufficient for mesa/egl's dri loader which is used by glamor.
Then if glx module is loaded before glamoregl module, the
initialization of mesa/egl/opengl will not be correct, and
will fail at a very early stage, most likely fail to map
the element buffer.
Two methodis to fix this problem, first is to modify the xserver's
glx's glapi.c to fit mesa's requirement. The second is to put
a glamor.conf as below, to the system's xorg.conf path.
Section "Module"
Load "glamoregl"
EndSection
Then glamor will be loaded firstly, and the mesa's libglapi.so
will be used. As current xserver's dispatch table is the same
as mesa's, then the glx's dri loader can work without problem.
We took the second method as it don't need any change to xorg.:)
Although this is not a graceful implementation as it depends
on the xserver's dispatch table and the mesa's dispatch table
is the same and the context set and get is using the same method.
Anyway it works.
As by default, xserver will enable GLX_USE_TLS. But mesa will not
enable it, you may need to enable that when build mesa.
Three pre-requirements to make this glamor version work:
0. Make sure xserver has commit 66e603, if not please pull the latest
master branch.
1. Rebuild mesa by enable GLX_USE_TLS.
2. Put the glamor.conf to your system's xorg.conf path and make sure
it loaded prior to glx module.
Preliminary testing shows indirect glxgears works fine.
If user want to use GLES2 for glamor by using MESA, GLX will not
work correctly.
If you are not using normal MESA, for example PVR's private GLES
implementation, then it should be ok to use GLES2 glamor and the
GLX should work as expected. In this commit, I use gbm to check
whether we are using MESA or non-mesa. Maybe not the best way.
Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
We add a new gl_fbo status GLAMOR_FBO_DOWNLOADED to indicate
the fbo was already downloaded to CPU. Then latter the access
to this pixmap will be treated as pure CPU access. In glamor,
if we fallback to DDX/fbXXX, then we fallback everything
currently. We don't support to jump into glamor acceleration
layer between a prepare_access/finish_access. Actually, fbCopyPlane
is such a function which may call to acceleration function within
it. Then we must mark the downloaded pixmap to another state
rather than a normal fbo textured pixmap, and then stick to use
it as a in-memory pixmap.
Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
Tested-by: Peng Li <peng.li@intel.com>
We may change the way to set/get those private data latter.
consolidate to glamor_set_pixmap/screen_private is better
than call those dixSetPrivate directly.
Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
This commit move the calling to glamor_close_screen from
glamor_egl_free_screen to glamor_egl_close_screen, as this
is the right place to do this.
We should detach screen fbo and destroy the corresponding
KHR image at glamor_egl_close_screen stage. As latter
DDX driver will call DestroyPixmap to destroy screen pixmap,
if the fbo and image are still there but glamor screen private
data pointer has been freed, then it causes segfault.
This commit also introduces a new flag GLAMOR_USE_EGL_SCREEN.
if DDX driver is using EGL layer then should set this bit
when call to glamor_init and then both glamor_close_screen
and glamor_egl_close_screen will be registered correctly,
DDX layer will not need to call these two functions manually.
This way is also the preferred method within Xorg domain.
As interfaces changed, bump the version to 0.3.0.
Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
Tested-by: Peng Li <peng.li@intel.com>
Just an initial implementation and disabled by default.
When uploading a pixmap to a texture, we don't really want
to attach the texture to any fbo. So add one fbo type
which doesn't has a gl FBO attached to it.
This commit can increase the cairo-trace's performance by
10-20%. Now the firefox-planet-gnome is 8.3s. SNA is still
the best, only take 3.5s.
Thanks for Chris to point out the A1 pixmap uploading bug.
Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
Fixup three special cases, one is in tile and the other is in
composite. Both cases are due to repeat texture issue. Maybe
we can refine the shader to recalculate texture coords to
support partial texture's repeating.
The third is when upload a memory pixmap to texture, as now
the texture may not have the exact size as the pixmap, we
should not use the full rect coords.
Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
We classify the cache according to the texture's format/width/height.
As openGL doesn't allow us to change a texture's format/width/height
after the internal texture object is already allocated, we can't
just calculate the size and then according ths size to put the
fbo to an bucket which is just like SNA does. We can only put
the fbo to the corresponding format/width/height bucket.
This commit only support the exact size match. The following patch
will remove this restriction, just need to handle the repeat/tile
case when the size is not exactly match.
Should use fls instead of ffs when decide the width/height bucket,
thanks for Chris to point this out.
Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
This is the first patch to implement a fbo/tex pool mechanism which
is like the sna's BO cache list. We firstly need to decopule the
fbo/tex from each pixmap. The new glamor_pixmap_fbo data
structure is for that purpose. It's somehow independent to each
pixmap and can be reused latter by other pixmaps once it's detached
from the current pixmap.
And this commit also slightly change the way to create a
memory pixmap. We will not create a pixmap private data structure
by default, instead we will crete that structure when a memory
pixmap is attaching a fbo to it.
Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
To split a rectangle (0,1,2,3) to two separated triangles need to feed
6 vertices, (0,1,2) and (0,2,3). use glDrawElements can reuse the shared
vertices.
Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
Should check the enable-glamor-gles2 before use the variable.
And should include the config.h as the GLAMOR_GLES2 macro is
defined there.
Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
As we want to take over all the possible GC ops from the DDX
layer, we need to add all the missed functions.
This commit also fixed one bug at polylines.
We simply drop the bugy optimized code now, as it did not
consider of clip info.
Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
