Re: Secondary console does not start properly, int10 problem?

[Aivils S. <ai...@un...>]

On Pirmdiena, 14. Novembris 2005 13:59, Friedrich W. H. Kossebau wrote:
> Am Montag 14 November 2005 10:17, schrieb Aivils Stoss:
> > On Sv=E7tdiena, 13. Novembris 2005 23:23, Friedrich W. H. Kossebau wrot=
e:
> > > PCI ViRGE/DX-graphic cards (s3virge driver) with a LCD
> > > PCI TGUI 9440 graphic cards (trident driver) with a CRT
> >
> > Old is gold?
>
> Rather, old did not cost money ;)
> But they look rather rusty right now, in terms of multi terminal :/
>
> > > 2 PS/2 keyboards :)
> > > 1 serial mouse
> > > 1 USB mouse
> > >
> > > The hardware runs properly in a plain dual head single console setup.
> > > So I read all the docs I found (thanks to those who set them all up :)
> > > and ended with this multi console approach:
> > > * added boot service which probes secondary graphic card (u.
> > > xorg.conf.probe) -> works, second monitor changes from blank to
> > > blinking text mode cursor * added service faketty and start links to
> > > /etc/init.d{,/rc3.d,/rc5.d} -> works, /dev/{fttyN,tty5N} can be opened
> > > and give pressed keys exclusivly* * added some symlinks to X (X0, X1,
> > > why is this needed at all?)
> >
> > Only for killing of right X.
>
> I guess for killall?
>
> > > * edited xorg.conf: one layout for each graphic card/terminal, added
> > > 	Option       "SingleCard" "true"
> > > to every layout,
> > > 	Option       "NoInt10"
> > > to every device.
> > > So the setup looks complete, doesn't it? (kdmrc left out for the
> > > moment)
> > >
> > > Well, it works only so far as the console on the primary is started
> > > (both with vt7 and vt51). Then the X-Server starts up nicely only on
> > > this screen and runs a heavy KDE session with no problems. Shutting
> > > down also goes smoothly. But when starting the secondary console (with
> > > or without the first running), e.g.
> > > 	$ X1 -layout "TGUI console" :1 vt52
> > > it fails. The screen stays in textmode, last line printed
> > > 	(=3D=3D) Using config file: "/etc/X11/xorg.conf"
> > > Resetting per Ctrl-Alt-Backspace does not work, I have to login
> > > remotely to kill -9 the X process. Before doing so listing the
> > > processes returns 5072 ?        Rs     2:02 X1 :1 -layout TGUI console
> > > vt52
> > > Then I get back to the shell prompt locally (sometimes lost in random
> > > character printing, synchronous to key presses), only to have the who=
le
> > > computer hangup completly after some further shell operations (like
> > > trying to reboot).
> >
> > faketty runs correctly if is possible see output of this:
> > # cat /dev/fttyXX
>
> It is, but only for ftty{1,2}. ftty0 stays silent.
>
> > Also full truth about Linux input, ftty included
> > # cat /proc/bus/input/devices
>
> It is, ah, funny, the pc speaker got ftty0, now, for that I have a hard
> time to produce input :P
>
> I: Bus=3D0010 Vendor=3D001f Product=3D0001 Version=3D0100
> N: Name=3D"PC Speaker"
> P: Phys=3Disa0061/input0
> H: Handlers=3Dkbd event0 ftty0
> B: EV=3D40001
> B: SND=3D6
>
> I: Bus=3D0011 Vendor=3D0001 Product=3D0002 Version=3Dab83
> N: Name=3D"AT Raw Set 2 keyboard"
> P: Phys=3Disa0060/serio1/input0
> H: Handlers=3Dkbd event1 ftty1
> B: EV=3D120013
> B: KEY=3D4 2000000 3802078 f840d001 f2ffffdf ffefffff ffffffff fffffffe
> B: MSC=3D10
> B: LED=3D7
>
> I: Bus=3D0011 Vendor=3D0001 Product=3D0001 Version=3Dab41
> N: Name=3D"AT Translated Set 2 keyboard"
> P: Phys=3Disa0060/serio0/input0
> H: Handlers=3Dkbd event2 ftty2
> B: EV=3D120013
> B: KEY=3D4 2000000 3802078 f840d001 f2ffffdf ffefffff ffffffff fffffffe
> B: MSC=3D10
> B: LED=3D7
>
> I: Bus=3D0003 Vendor=3D05fe Product=3D0001 Version=3D047b
> N: Name=3D"Cypress Sem Cypress USB Mouse"
> P: Phys=3Dusb-0000:00:07.2-2/input0
> H: Handlers=3Dmouse0 event3
> B: EV=3D7
> B: KEY=3D70000 0 0 0 0 0 0 0 0
> B: REL=3D3
>
> > > Funny enough: Having started the secondary console when pressing some
> > > keys random characters are added both to the primary (where the shell
> > > is) _and_ the secondary screen. Even better, when some screensaver
> > > starts (no idea which) _both_ screens are blanked and return again
> > > after some keypress. Sometimes even the prompt changes to the second
> > > screen.
> >
> > Not all video adapters or X drivers have capability to share resources
> > (VGA,PCI) between two independ X servers. Looks like one of X driver
> > steer both adapters, because both adapters may be use same resources.
>
> Hm. Is there any output I could investigate to find numbers which could
> point to shared resources? Like Xorg.{0,1}.log? Where to look best? Which
> resources are there at all? What do you mean by VGA-resource? In the BIOS?

stay away from it, as you stay away from a "serpent"!

> And does this mean that the single card option is not fully supported by
> all drivers? Where do/can they driver ignore it?

1st X occupy PCI resource A to B. 2nd X does not have info about this and
think A to B is free and occupy it again. I have very foggy image , how to
works memory mapped video adapter registers, like VGA. X do it all automati=
c=20
without any query and without any preconfigured condition.

Aivils