This has been tested with both OpenGL and OpenGL ES on Android.
We can't use dlsym on Android to get all the function pointers since Android's libGLES* libraries are just shim libraries that pass everything through
to the real vendor specific libraries.
Due to this we must grab function pointers entirely through eglGetProcAddress instead of dlsym.
Fixes failure of the X Server with glamor using desktop GL on EGL, and
the associated testcase in the tree. This was clearly just
copy-and-paste failure on my part in
b251e3ae33.
Now that it's split out from epoxy_get_core_proc_address(),
the particular symbols we were testing for before calling
GetCurrentContext() didn't make much sense, plus we were leaking
dlerror()s still.
This fixes crashes on ARM's Mali drivers, where desktop GL isn't
supported, and epoxy didn't notice the error and tried to dlsym() on
libGL.so.1.
An alternative would have been to look at the response from
eglQueryAPI() first, but that would have meant that we still need to
BindAPI in the case that eglQueryAPI returns something like VG or CL,
and so we needed to handle these errors regardless.
Fixes#22
If you happened to do both GLES2 and GLES1 in a process, it meant that
you'd try to look up v1 procs in the v2 API, and lose. Caught by
piglit oes_fixed_point-attribute-arrays.
For performance, I want to be able to make single-context (well,
single-pixel-format-and-device) apps be able to directly call GL
functions through function pointers. Bake that into the ABI now so I
can get a release out the door and fix this up later.
This also fixes the lack of __stdcall annotation on the
PFNWHATEVERPROC typedefs.
For a non-GLX application, at least, we need to pull our extensions
using dlsym, not glXGetProcAddress(). The symptom was a segfault in
glXGetProcAddress().
In addition to the failing testcase, there were a couple of
regressions in piglit's attribs test: one from glBegin_unwrapped vs
glBegin confusion in the __asm__ directives we were generating, and
one where the function pointers apparently were just getting mixed up
at application runtime.
This is going to change for macos and win32, and this will be easier
than trying to spread that logic through the python code and into the
generated code.
The dlopen bits are left in place, but the functions required the
types, and in the case of EGL, the types require that the platform
header actually exists.
This totally replaces the getprocaddress and dlsym code, which was
basically just stubs up until now. The is_glx/is_egl() is also
dropped -- they weren't doing anything, and the only false answer they
could give is if the dlopen were to fail.
Unfortunately, for GLX 1.4+ entrypoints (just glxGetProcAddress
currently) or extensions, if there isn't a context bound then we don't
have a dpy and screen available to provide useful debug messages. Oh
well.
Chun-wei Fan
Eric Anholt
Ryan Houdek