Just to stave off the inevitable I don't have in mind a library, and i don't
have in mind normal ipc/rpc. What I'm looking at is the paper at
http://reality.sgi.com/opengl/d11/d11.html in particular
http://reality.sgi.com/opengl/d11/node7.html, where the author suggest that the
X server could be implemented using "active contexts". Where a client X app can
context switch to become the X server to execute X calls. Short circuiting the
normal transport mechanism and giving a potential speedup for X.
Now I haven't been able to figure out how to actually implement even a simple
mechanism to achieve this, but then i'm not all that good. But seeing as a
kernel can achieve this, i reckon this is a good a forum as any to get some
pointers on the practicalities of doing something like this, though isn't the
mechanism it uses only possible for the kernel alone ?
The D11 concept seemed to be a goody, until I rapidly hit the edges of my
knowledge, anyone have any ideas. Just in the general case of achieving
something like this. Whether its an appropiate mechanism to use, or
whether or not GGI/DRI/DGA/Berlin/Y is better to implement is outside the
scope of whats puzzling me.
C.
Real Life: Caolan McNamara * Doing: MSc in HCI
Work: Caolan.McNamara@ul.ie * Phone: +353-86-8790257
URL: http://www.csn.ul.ie/~caolan * Sig: an oblique strategy
What are the sections sections of?
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/