Hardware acceleration for pixmap

Hi All,
      The create pixmap call (gdk_pixmap_new) will allocate the memory
in xserver and will return an ID to client but i want to access the
memory address so that the location can be given to support hardware
acceleration , so is there a way where i can tap into the function
call in xserver and manipulate the memory address to suit the
requirements (Address where sgx can render onto)and pass on memory
address as well to the client along with the id?
If not is there any way this can be achieved without using shared
memory between xclient and xserver?

You can't get the memory for a pixmap. The X Server manages the
memory. It can (and does) move the data between system and video
memory. This is perfectly valid whlist the client interacts with the
pixmap through an ID. Think of the pixmap ID indexing a table to get
the actal address of the bitmap. That mapping tables can change.


You probably ought to be using OpenGL or the RENDER extension (probably
through cairo) if you believe this sort of optimization is necessary.

- Ben