I have been doing some research on running 2 independent displays off of 1
machine (ie 2 keyboards, 2 mice, 2 vid cards, 2 monitors).. there are some
hacks out there now that "sort of" work.... but nothing stable and official..
it's all hacks.... I have read that support for this is planned for
2.5/2.6, and would like to know what progress has been done. I am willing to
help where I can. I am a good C/C++ programmer, but have not done any kernel
work so far.
> I have been doing some research on running 2 independent displays off of 1
> machine (ie 2 keyboards, 2 mice, 2 vid cards, 2 monitors).. there are some
> hacks out there now that "sort of" work.... but nothing stable and official..
> it's all hacks.... I have read that support for this is planned for
> 2.5/2.6, and would like to know what progress has been done. I am willing to
> help where I can. I am a good C/C++ programmer, but have not done any kernel
> work so far.
You are talking about the linuxconsole project. Yes with alot of work
we got a multi-desktop system working. We even got several X servers with
several patches running on different desktops even tho they where working
out of one box. The main problem with this research was the console system
level of code was intertwine in each input and display driver. In 2.5.X
you see the moving of the console keyboards etc to the input api which can
function indepenedent of the console layer. You also had the same effect
with the new framebuffer layer. This was done to make driver writing easy
and to help the embedded space as well as prepare for the future
multi-desktop of linux.
What has not been done is true multi-desktop support. I like to work on
this in the future but due to recent events in my life I have to abandon
such research :-(
> > I have been doing some research on running 2 independent displays
> > off of 1 machine (ie 2 keyboards, 2 mice, 2 vid cards, 2
> > monitors).. there are some hacks out there now that "sort of"
> > work.... but nothing stable and official..
> > it's all hacks.... I have read that support for this is planned
> > for 2.5/2.6, and would like to know what progress has been done.
> > I am willing to help where I can. I am a good C/C++ programmer,
> > but have not done any kernel work so far.
>
> You are talking about the linuxconsole project. Yes with alot of work
> we got a multi-desktop system working. We even got several X servers with
> several patches running on different desktops even tho they where working
> out of one box. The main problem with this research was the console system
> level of code was intertwine in each input and display driver. In 2.5.X
> you see the moving of the console keyboards etc to the input api which can
> function indepenedent of the console layer. You also had the same effect
> with the new framebuffer layer. This was done to make driver writing easy
> and to help the embedded space as well as prepare for the future
> multi-desktop of linux.
There was a thread some months ago, called something like 'Linux as a
minicomputer', where we discussed this kind of thing.
John.
> There was a thread some months ago, called something like 'Linux as a
> minicomputer', where we discussed this kind of thing.
Yeap. This is my dream for linux. Someday :-)
> > There was a thread some months ago, called something like 'Linux as a
> > minicomputer', where we discussed this kind of thing.
>
> Yeap. This is my dream for linux. Someday :-)
Heh, yeah, a low-power RISC cpu based Linux box running of a
battery-backed 12v solar panel connected to 9600 bps serial terminals
with LCD screens all over the house is a future spare-time project for
me :-).
John
On Wed, 12 Feb 2003, James Simmons wrote:
> > I have been doing some research on running 2 independent displays off of 1
> > machine (ie 2 keyboards, 2 mice, 2 vid cards, 2 monitors).. there are some
> > hacks out there now that "sort of" work.... but nothing stable and official..
> > it's all hacks.... I have read that support for this is planned for
> > 2.5/2.6, and would like to know what progress has been done. I am willing to
> > help where I can. I am a good C/C++ programmer, but have not done any kernel
> > work so far.
>
> You are talking about the linuxconsole project. Yes with alot of work
> we got a multi-desktop system working. We even got several X servers with
> several patches running on different desktops even tho they where working
> out of one box. The main problem with this research was the console system
> level of code was intertwine in each input and display driver. In 2.5.X
> you see the moving of the console keyboards etc to the input api which can
> function indepenedent of the console layer. You also had the same effect
> with the new framebuffer layer. This was done to make driver writing easy
> and to help the embedded space as well as prepare for the future
> multi-desktop of linux.
> What has not been done is true multi-desktop support. I like to work on
> this in the future but due to recent events in my life I have to abandon
> such research :-(
>
There is another project dealing with this: KGI. This project not
only deals with Linux, but also with NetBSD and Freew:BSD. Together with
the
Linux input layer KGI should support as many keyboards / mice as you can
buy.... Multiple desktop consoles runs fluently, support for full 3D
accellerated desktops (with help from an user space lib: GGI) is almost
there.
I'm still looking at how to connect multiple keyboards, for now I use a
microcontroller on a serial port which gives me 4 keyboards extra.
Unfortunately we're still missing a programmer willing to take a look at
the multiple keyboard / mice issue. The code is almost there, though not
functional yet. For info about KGI please drop by on
http://kgi-wip.sourceforge.net or irc.freenode.net #kgi
Jos
P.S. For those thinking KGI is still that crazy project doing graphics
accelleration in the kernel trough ioctls, please look again !