[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

RE: Fast thread-local storage for OpenGL drivers



Richard Henderson wrote:
> 
> Indeed, further reduced to 
> 
>         jmpl    *%gs:__foo_dispatch ntpoff

The first entry in the dispatch table could be implemented like that, but in
general you'd have something like:

	glFoo:
		movl %gs:__gl_dispatch ntpoff, %eax
		jmpl *__offset_Foo(%eax)

Not sure how fast that far jump would be, though...

-- 
Gareth Hughes (gareth nvidia com)
OpenGL Developer, NVIDIA Corporation





[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]