When glClear is called in direct rendering move (DR), we usually
have to skip the call altogether because clearing out transparency
would erase the pixels in the evas backbuffer. This means Evas
would not be able to blend an RGBA GLView on top of other objects.
But COPY mode should allow Evas GL to poke holes in a window
backbuffer.
Thanks @spacegrapher for the review :)
NOTE: Elm GLView also needs to pass the render op to its Evas.Image.
@fix
evas_gl_native_context_get is an internal function
passed around from an evas engine to evas_gl so that we can
implement evasglCreateImageForContext without exposing
any evas engine internal structure to evas_gl.
It's all a ittle bit ugly but the previous solution with
dlsym(DEFAULT) didn't work.
If the TLS variable was not initialized, Evas GL can't get a pointer
to a valid EGLDisplay which is required by eglImageDestroy.
So, we keep track of the dpy used at creation time and use that
if we're in another thread.
This reverts commit 5beb47aa4d.
Revert "image_savers/jpeg: fix undefined behavior of using sigsetjmp on jmp_buf"
This reverts commit 84c7751e19.
these end up with efl simply not compiling. efl tree does not build at
all now and this warrants a revert.
lib/evas/.libs/libevas.so: undefined reference to sigjmp'
collect2: error: ld returned 1 exit status
Makefile:19321: recipe for target 'bin/evas/evas_cserve2' failed
.. etc.
i was runing perf top and noticed that evas_image_load_file_data_eet(0
was being called. in fact - it was #1 on the list of functions being
called. why? it didn't make sense. i found out. just a blinking cursor
in terminology was causing the background to be unloaded and
re-loaded. the new "actually unload" changes for 1.15 made this happen
and thus we kept sucking in new data all the time even if the
scalecache already had the data - and that was the problem.
so now calcecache prepare tells you if you don't have cached data and
if you likely then have to ensure the data is loaded. this cuts down
quite a bit of work.
while i'm at it... we definitely need to clean house on the internals
of evas. a decade+ of features, mess, optimizations needs to be fixed.
i mean really house-cleaned. rewritten clenl;y re-using existing code
where appropriate.
Those two functions were doing exactly the same thing[1], which
is free an image, so this commit only attempts to simplify the code
a little bit.
[1] Actually image_map_surface_free() might even not have worked
properly with cserve2 sw (calling unload instead of close).
Summary: Gcc suggested parens around the truth value here, however
as some developers don't like overly used parens, just modify the code
slightly to not use an if on the assignment line.
@fix
Signed-off-by: Chris Michael <cp.michael@samsung.com>
@fix
before we mapped these segmentsa read+write for the user or read+write
for EVERYONE if system. this now creates the file as r+w for the user
and +ro for everyone only IF system, and clients voluntarily map
read-only to avoid possible memory corrupting of pixels from the
client side. not more secure for clients, but nicer. defintiely more
secure for system services.
Otherwise there would be conflicts in certain circumstances.
This also requires adding const on many existing functions,
and similar work is necessary in Elementary.
@fix
Summary:
Add NULL type check in output_free of evas fb engine.
If engine setup is failed,
Render_Engine wil be NULL so output_free also need to
handling NULL check.
Test Plan:
It needs specific condition to reproduce,
engine of ecore_evas is set to fb, and setup is failed,
then Render_Engine is NULL, but ecore_evas_free will call
output_free in fb engine's evas_engine.c
Reviewers: raster, cedric, Hermet
Reviewed By: Hermet
Subscribers: cedric, seoz, eagleeye, singh.amitesh
Differential Revision: https://phab.enlightenment.org/D2743
Summary:
Ecore_Buffer is abstraction of graphic buffer.
it supports backend of shm, x11_dri2 and x11_dri3 for now,
and this library also provides method to share buffers between processes.
Ecore_Buffer_Provider and Ecore_Buffer_Consumer is for this, sharing buffer.
provider draws something in to Ecore_Buffer, and consumer receives and displays it.
the binary, bq_mgr is a connection maker for buffer provider and consumer.
it can be included Enlightenment as a deamon later.
@feature
Test Plan:
1. Configure with --enable-ecore-buffer and --enable-always-build-examples to build examples.
2. Run bq_mgr, it connects consumer and provider.
3. Run ecore_buffer_provider_example and ecore_buffer_consumer_example
Reviewers: lsj119, gwanglim, cedric, zmike, jpeg, raster, devilhorns
Subscribers: cedric
Differential Revision: https://phab.enlightenment.org/D2197
Summary:
After this commit 684c51360f it didn't work
Miss bind color pick texture to drawable data
@fix
Reviewers: Hermet, cedric
Reviewed By: cedric
Subscribers: cedric, Ievgen, se.osadchy
Differential Revision: https://phab.enlightenment.org/D2724
Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
Summary: This cleans up nasty compiler warnings when building the
gl_drm engine. If we do not define EGL_EGLEXT_PROTOTYPES then we end
up with implicit declaration warnings
@fix
Signed-off-by: Chris Michael <cp.michael@samsung.com>
Summary: Seems some platforms (my setup here at least) don't have EGL
configs which suppose GBM_FORMAT_XRGB8888, so default to ARGB
@fix
Signed-off-by: Chris Michael <cp.michael@samsung.com>
Summary: This change adds support for using EGL_MESA_platform_gbm to
get the display directly from gbm and to create a window surface
(usable by egl) directly from the gbm surface.
Signed-off-by: Chris Michael <cp.michael@samsung.com>
Summary: Add null check for glGetString version check and evgl_engine
Test Plan: Local tests
Reviewers: jpeg
Reviewed By: jpeg
Subscribers: wonsik, cedric
Differential Revision: https://phab.enlightenment.org/D2699
Summary:
When evas is recreated, this static variable value remains,
so it references already freed memory.
Test Plan: Local tests
Reviewers: cedric, jpeg
Reviewed By: jpeg
Subscribers: wonsik, mer.kim, cedric
Differential Revision: https://phab.enlightenment.org/D2678
Summary: This cleans up the gl-drm engine code a bit and avoids the
need to constantly destroy and recreate the gbm device itself. This
also adds checks during Outbuf create to handle the swap_mode as per
previously added enum.
Signed-off-by: Chris Michael <cp.michael@samsung.com>
Summary: This adds an enum we can use for setting/determining the
swap_mode to use for the gl_drm engine
Signed-off-by: Chris Michael <cp.michael@samsung.com>
Summary:
<CID 1304559: Logically dead code>
The dead code is only valid for GLES backend, so move if statement
to be used for GLES backend only.
<CID 1304560: Bad bit shift operation>
When calculating depth bit, bit shifting could be done with negative values.
@fix
Test Plan: Local tests
Reviewers: jpeg
Reviewed By: jpeg
Subscribers: wonsik, cedric
Differential Revision: https://phab.enlightenment.org/D2654
In on-demand show mode, input panel will be shown in case of only an user's explicit Mouse Up event.
It doesn't request to show the input panel even though the Input Method Context has focus.
Evas_Image_Orient is an enum starting at 0 thus we never go below 0. Remove
unneeded checks.
modules/evas/engines/software_generic/evas_engine.c:1558:4: warning: comparison of unsigned expression >= 0 is always true [-Wtype-limits]
modules/evas/engines/software_generic/evas_engine.c:1560:8: warning: comparison of unsigned expression >= 0 is always true [-Wtype-limits]
Summary: If we fail to lock the front buffer, we will not get a valid
bo returned so we should not be calling other functions which require
a bo.
@fix
Signed-off-by: Chris Michael <cp.michael@samsung.com>
Summary: This fixes a crashing issue in evas image cache due to the
engine returning an improper swap mode.
@fix
Signed-off-by: Chris Michael <cp.michael@samsung.com>
Summary: The pre/post swap callbacks were forgotten when doing the
engine rewrite, so readd them
@fix
Signed-off-by: Chris Michael <cp.michael@samsung.com>
Summary: Previous Ecore_Evas drm code would create the gbm device and
surface inside the gl_drm function. As these are now handled inside
the evas-gl-drm engine itself, this code is not needed here.
@fix
Signed-off-by: Chris Michael <cp.michael@samsung.com>
Summary: Previous gl_drm evas engine code did not work properly (or at
all really). This reworks/refactors the gl_drm engine code to work
again with the changes made to ecore_drm.
@fix
Signed-off-by: Chris Michael <cp.michael@samsung.com>
Summary:
Evas has quite a number of shaders, and resetting the projection matrix
for all shaders takes some time, which is an overhead when target surface
is continuously being changed (ie. when using proxy).
So here we reset the projection matrix for shaders that are actually used.
Summary:
Evas GL now supports surfaceless make current, where
evas_gl_make_current can be called with sfc parameter NULL.
This closely resembles EGL_KHR_surfaceless_context extension,
where applications that only want to render to client API targets
can make current to NULL surface instead of creating a dummy egl surface.
@feature
Summary:
When direct rendering is enabled, FBO configuration should match
window surface configuration as FBO will be used in fallback cases.
So create FBO with configuration from window surface.
@fix
Summary:
When default framebuffer(0) is bound, attachment should contain
COLOR, DEPTH or STENCIL for glDiscardFramebufferEXT.
When a framebuffer object is bound, attachment should contain
COLOR_ATTACHMENT0, DEPTH_ATTACHMENT or STENCIL_ATTACHMENT.
This should be correctly taken into account for indirect rendering,
where internal FBO is used.
@fix
Summary:
Separate EGL extensions from GL/GLES extension list, since
we have extension list for each GL version, and we do not want to
check EGL extensions differently when different GL versions are used.
This also simplifies extension string get function as we just need to
concatenate EGL and GL extensions rathan than keeping track of
GL extensions only.
Summary:
We should make current to an Evas GL context before calling _surface_buffers_destroy,
otherwise resources from GL backend will be removed.
@fix
Summary:
When either FBO or EGL image from texture extension is not supported,
we can use pixmap surface as indirect surface fallback.
Since native pixmaps have (0,0) in the upper left while
FBOs have (0,0) in the lower left, we should invert the y coordinates
when native pixmaps are used as the render target.
To accomodate run-time y-invert check we add a new callback for
EVAS_NATIVE_SURFACE_EVASGL type.
Reviewers: cedric, jpeg
Subscribers: wonsik, mer.kim, cedric
Summary:
Remove surface buffer create/allocate in pbuffer_surface_create.
In evgl_make_current, we should use direct_mem_opt, direct_override values
set in EVGL_Surface struct, not the global value set by env variables.
@fix
Summary:
We should destroy egl image when surface is destroyed
to avoid memory leak.
@fix
Test Plan: Evas GL test cases
Reviewers: jpeg
Subscribers: mer.kim, mythri, wonsik, cedric
Differential Revision: https://phab.enlightenment.org/D2357
Summary:
Now we can support EVAS_GL_GLES_1_X version for GLX backend
with both direct and indirect rendering.
Refactored api get functions to have similar code path for each version.
@feature
Test Plan: Evas GL test case
Reviewers: cedric, jpeg
Subscribers: cedric, mer.kim, mythri, wonsik
Differential Revision: https://phab.enlightenment.org/D2342
Thanks Dongyeon for finding out this solution. Now that was
one nasty bug :)
Somehow the currently bound texture id would not match what
Evas expected, so Evas would not call glBindTexture when
required. As a result it was drawing black (sampling from tex 0).
@fix
Summary: These fields are not needed in the outbuf structure as they
are never used, so just remove them.
Signed-off-by: Chris Michael <cp.michael@samsung.com>
if yuou use 709 instead of 601 yuv (ycbcr) evas will just be wrong and
use 601. this fixes that and implements 709. it also fixes a scaling
bug for yuv in the gl engine. no one noticed but me, so i won't call
this a bug fix, and it can go into the next efl release - no need to
backport unless it actually bothers peolpe (which it seemingly doesn't)
Summary:
Use fourth component texture. Update mechanism generation pixels, scene renderer
to texture and geting color pixels from texture. Update shader for color pick.
Reviewers: Hermet, raster, cedric
Reviewed By: cedric
Subscribers: Oleksander, cedric
Differential Revision: https://phab.enlightenment.org/D2549
Summary:
Use RGBA texture unit for generation shadow. Use separate framebuffer and
renderbuffer for shadow map.
Refactor shadow map shader to use position of shadow and pack depth value in
RGBA texture.
Refactor shader function for generation shadow factor to unpack depth value from
RGBA shadow map.
Reviewers: cedric, Hermet
Subscribers: cedric
Differential Revision: https://phab.enlightenment.org/D2578
Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
so you know the files should change when you "make" and shf files are
updated, add notes to the .x files generated to indicate they should
be committed back to the repo so it is clear
Summary:
Add alpha test in shader code instead of function glAlphaFunc that doesn't supported
in GLES2.0
Add gag and TODO in shader function that generate shadow. Without it all scene is
shadowed, because shadow effect doesn't work yet.
Reviewers: cedric, Hermet
Subscribers: cedric
Differential Revision: https://phab.enlightenment.org/D2529
Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
Instead of "@in type name;" we now use "@in name: type;". This change
is done because of consistency with the rest of Eolian; pretty much
every other part of Eolian syntax uses the latter form.
This is a big breaking change in the .eo format, so please update your
.eo files accordingly and compile Elementary together with the EFL.
@feature
Summary: add ecore_evas_extn_socket_events_block_set/get
Test Plan: add mouse event callback, and check whether it could get event or not
Reviewers: raster, woohyun, jaehwan, Sergeant_Whitespace
Reviewed By: Sergeant_Whitespace
Subscribers: Sergeant_Whitespace, seoz, cedric
Differential Revision: https://phab.enlightenment.org/D2268
Summary: We should be marking the framebuffer dirty regions Before we
send the pageflip to the kernel so that the kernel can properly
optimize drawing dirty regions before scheduling a pageflip.
NB: Slightly optimizes evas drm rendering a bit more. Thanks to Daniel
Stone for the help :)
@fix
Signed-off-by: Chris Michael <cp.michael@samsung.com>
Summary:
When the ecore_evas of wayland_egl/shm is freed with ecore_evas_free(),
ecore_wl_shutdown() is called by _ecore_evas_wl_common_free().
so ecore_wl_shutdown() after ecore_evas_free() has to be skipped.
@fix
Reviewers: devilhorns, raster, zmike, gwanglim
Subscribers: cedric
Differential Revision: https://phab.enlightenment.org/D2503
EGL might very well not support RGBA read mode, so we
need to check for it first.
Also remove some error logs (see previous commit), and useless
initialization of the Evas GL engine.
@fix
Based on a quick git diff we check that the glsl code has not changed.
This should fix out-of-tree builds and avoid all source modification
unless required.
When compiling from a tarball there should be no git tree (err 129),
or if there's one the files should not be checked in (ie. no diff).
If you changed the glsl files in a tarball... too bad for you.
If this is still not enough to fix the build, then go ahead and disable
the script from Makefile_Evas.am
I would like to note that the auto-generation during make is extremely
useful when working on the shaders, which is why I'd rather keep it enabled.
@fix
Summary:
Bind flag atlas_enable.
Build adjusting matrix only in case getting texture unit with atlas.
Reviewers: Hermet, cedric
Reviewed By: cedric
Subscribers: cedric
Differential Revision: https://phab.enlightenment.org/D2473
Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>