They were missed, and Copy is very high in profile here. It doesn't block the GPU,
but it stalls the driver thread. So with our bad GL instructions, this might block quite a while.
Those implementations are quite costly, so there is no need to inline them to the caller.
Ressource deletion is often a performance bug, so in this way, we support to add breakpoints to them.
The state object isn't used anywhere else, so there
is no need to revert the state.
And the comment is just wrong: It doesn't matter
which textures are bound on framebuffer binding, it
only matters at draw time. And we reset all bindings
before the draw call. So let's use gl_state as it is
designed to avoid flipping states.
Framebuffer blits depends on pixel tests:
Ownership (is fine)
Scissor (is broken on the last commit)
Masking (is broken on master for a while)
So let's be honest and start with a clean state in
those helper functions.
There is no need to disable this state after the draw call,
gl_state will handle this for us if needed. This kind of
redundant state changes are bad for the driver overhead,
as flipping bits will invalidate the driver state.
This replaces the glBufferData logic with the shared stream buffer code.
The new code doesn't need a temporary staging buffer any more, so the
performance should imrpove quite a bit.
This OpenGL call synchronize the worker thread of the nvidia blob.
It can be verified on linux with the __GL_THREADED_OPTIMIZATIONS=1 environment variable.
Those errors should not happen on tested drivers.
It was used as a workaround for https://bugs.freedesktop.org/show_bug.cgi?id=94148
