From: Caolan McNamara (Caolan.McNamara at domain ul.ie)
Date: Mon 09 Aug 1999 - 10:52:07 IST
On 06-Aug-99 Paul Jakma wrote:
>>
>> Yes, like a kernel as each app calls a X11 call, it becomes the Xserver
>> for the duration of that call. The same was a read call from a userlevel
>> program runs the kernel code directly.
>
>
>You can't call a function in another process, and you can't do with
>libraries - cause you need to keep cross-process state. Only the kernel has
>cross-process context, and i don't think linux would allow that kind of
>stuff in-kernel. (and anyway, in that case ggi already exists).
Well fair nuff, some digging throws up the showstopper problem that I can't
actually do it the 100% natural way, but have to instead fall back to some
standard (ish) ipc route to communicate between the Server and Client (though
doors, see later on), do appear to do most of what D11 would require. First
though i draw your attention to some of the mach work and the Lightweight RPC
stuff. Theres a mach paper at
ftp://cs.cmu.edu/afs/cs/project/mach/public/doc/published/mach-in-x11.ps
which talks about their replacement of the socket transport mechanism with mach
ipc, and it waffles on the joys of it all (to be taken with a grain of salt
i imagine as it only compares mach to mach, rather than mach to anything else)
Another ipc mech I was pointed towards was doors (new to me I have to say), some
info on this one is at http://www.zdjournals.com/sun/s_sun/9810/sun98a2.htm,
with a linux implementation that i'll have to play with for a bit at
http://www.rampant.org/doors Though the zdjournal article is more positive that
my brief scan at the benchmarks in the linux doors paper.
It seems somewhat possible that the mach Xserver tweaks could be implemented for
linux using the doors stuff. A lot less clean that originally theorized, but hey
close enough.
C.
Real Life: Caolan McNamara * Doing: MSc in HCI
Work: Caolan.McNamara at domain ul.ie * Phone: +353-86-8790257
URL: http://www.csn.ul.ie/~caolan * Sig: an oblique strategy
What are the sections sections of?
This archive was generated by hypermail 2.1.6 : Thu 06 Feb 2003 - 13:04:26 GMT