The way I see it, pipe_surface is a render target view for
pipe_framebuffer_state. The fact that other functions also accept
pipe_surface is a historic artifact. resource_copy_region used to accept 2
pipe_surfaces before it was changed to 2 pipe_resources + parameters.
Somebody might decide to chan
I don't mind renaming it to invalidate_subsurface() if you prefer
that, but I really prefer that it takes a surface rather than a
resource... pipe_resource is the wrong level to do this.
Even though surfaces are transient, tracking the invalidate state in
the surface works out much easier in the d
Can you please call it invalidate_subresource and inline relevant
pipe_surface variables inside the parameters?
Thanks,
Marek
On Wed, Dec 12, 2018 at 10:48 AM Rob Clark wrote:
> A new API to implement glInvalidateFramebuffer() and friends. It is
> similar to invalidate_resource() but can be us
A new API to implement glInvalidateFramebuffer() and friends. It is
similar to invalidate_resource() but can be used to invalidate a
specific layer/level, so it is suitable to use for user FBOs in addition
to window system framebuffer.
Signed-off-by: Rob Clark
---
src/gallium/docs/source/contex