From: Caolan McNamara (Caolan.McNamara at domain ul.ie)
Date: Mon 09 Aug 1999 - 12:54:46 IST
On 09-Aug-99 Paul Jakma wrote:
>in that case, why not forget D11, and use GGI, which is also backwards
>compatible with X11 (i think). kgi is in kernel -> so graphics via system
>calls,. Which was the original idea for D11.
The Xserver for ggi Xggi ? or whatever it is called still uses the X protocol
as does the framebuffer one, so the problem is still there. While the bottom
end graphics calls would reside in the kernel the Xserver logic and structures
that keep track of what has been drawn and so on still reside in userspace and
communicate with the clients through sockets, so you still have the same issues.
Being able to make system calls direct to the drawing commands would create a
faster Xserver, but the Xserver would have to make those calls, I reckon that
allowing the clients to do the drawing calls direct bypasses the X functionality
and X wouldn't know that the calls had been made, so that won't hold up unless
you put the Xserver itself into the kernel and make systems calls to the
Xserver, but thats what NT does i believe so lets not go there, X is a bit big
and complex to ever have a chance of sneaking it into the kernel i reckon. Id
need a lot more knowledge of the XServer internals to see if it would be
possible to do some kind of hybrid messing with a userspace Xserver and short
circuited ggi calls from the client, seems a bit unlikely, dunno though.
>> http://www.rampant.org/doors Though the zdjournal article is more positive
>> than my brief scan at the benchmarks in the linux doors paper.
>IIRC Linus doesn't like doors one bit. (don't remember why). So no chance of
>it being included in linux.
Gagh, its one thing after another isn't it.
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
Do we need holes?
This archive was generated by hypermail 2.1.6 : Thu 06 Feb 2003 - 13:04:26 GMT