linuxconsole-dev

Virtual console documentation

[Boris B. <bo...@ki...>]

Hi all.

I am trying to put together a documentation of how the Virtual Console works under GNU/Linux system as part of LKDP project at http://www.freesoftware.fsf.org/lkdp/ . I decided to go for a 2.5/2.6 kernel code.

So far i have put together preliminary ToC, which will likely change as the code evolves.

For start i will dig through Input code and maybe through Unix98 PTY as these two parts will not change much (am i wrong ?? ).

I would like to hear your opinions/suggestions on this.


Chapters:

	Virtual Console overview [*]
		How the kernel starts the console
		Console data structures
		Device operations(tty)

	Virtual Console control [*]
		Describes how VC are switched between, how another driver takes			over control of console, changing key codes, changing fonts,			console blanking etc.

	Screen control [*]
		Describes screen control functions, (char) attributes,				cursor functions, char printing, scrooling, etc.	

	Output drivers [*]
		Dummy console
		Video mode selection support
		VGA text console
		MDA text console (dual-headed)
		Framebuffer console
		SGI Newport console
		SPARC PROM console

	Input layer
		Input core
		Keyboard interface
		USB HID
		Mouse interface
		Joystick interface
		Event interface
		Touchscreen interface
		Gameport support
		Serial i/o support (PC keyboard, PS/2 mouse...)
		Joysticks
		*Parallel port keyboard [+ other input controllers?]
		Input device drivers [maybe a chapter?]
			Keyboards
			Mice
			Joysticks
			Touchscreens

	Misc
		Speaker
		Console on line printer
		Unix98 PTY support
------------------------------------------------------------------------
Chapters marked with [*] are on hold.

-- 
With best regards,

		Boris B.

ruby CVS compile error

[Boris B. <bo...@ki...>]

Hi all.

During compile i get this error. I choosed AT keyb as module..
Anyone else got this or is it just me ?

BUILD_BASENAME=atkbd   -c -o atkbd.o atkbd.c
atkbd.c:123: field `tq' has incomplete type
atkbd.c: In function `atkbd_interrupt':
atkbd.c:164: warning: implicit declaration of function `queue_task'
atkbd.c:164: `tq_immediate_R0da0dcd1' undeclared (first use in this function)
atkbd.c:164: (Each undeclared identifier is reported only once
atkbd.c:164: for each function it appears in.)
make[3]: *** [atkbd.o] Error 1

-- 
With best regards,

		Boris B.

Re: pretty sad + Re: HELP with aiptek and keybdev

[Svetoslav S. <ga...@st...>]

Quoting Aivils Stoss <Aiv...@un...>:

> 
> >needed hints for debuging
> >i was thinking about uml linux, but i'm not sure that this can help
> me.
> >and my timezone GMT +1
> 
> Very simple is serial console. Turn on this in character devices.

and what does this mean

Unable to handle kernel NULL pointer dereference*pde = 000
00000
Oops: 0000
CPU:    1
EIP:    0010:[<c0216111>]    Not tainted
EFLAGS: 00010297
eax: 0000007e   ebx: 00000000   ecx: 0000007e   edx: 00000001
esi: c2ad1010   edi: 0000007e   ebp: c12d7d58   esp: c12d7d40
ds: 0018   es: 0018   ss: 0018
Process swapper (pid: 0, stackpage=c12d7000)
Stack: c12d7d58 00000001 0000007e 00000001 c2ad1010 0000007e c12d7d6c c021635a
       00000000 0000007e 00000001 c12d7d9c c0265ffd c2269ce0 00000001 0000007e
       00000001 00000000 33323130 c2269ce0 c24e2460 00000001 c24e2500 c12d7dcc
Call Trace: [<c021635a>] [<c0265ffd>] [<c0265b98>] [<c02629ba>] [<c0262ae7>]
   [<c0240a28>] [<c0262d14>] [<c0262d38>] [<c28e31b2>] [<c28e357e>] [<c28e3621>]

   [<c010a004>] [<c010a34b>] [<c010c638>] [<c0106eb4>] [<c0106ef7>] [<c011b2a8>]

   [<c011b112>] [<c0105000>]

Code: 8b 33 74 12 0f b6 45 f0 00 c0 30 d0 0f b6 c0 50 e8 ac 49 ff
 Kernel panic: Aiee, killing interrupt handler!
In interrupt handler - not syncing

ruby questions

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

Hi James

     One weird bug pop up in my ruby mod.
In file drivers/char/tty_io.c
function init_dev()
line 836:      tp_loc = &driver->termios[idx];

Later *tp_loc used as tty->termios. In my end driver->termios[idx] return
varios garbage. I couldn't understand where is initialized tty_driver
member
     struct termios **termios;

Kernel always made oops in n_tty_set_termios() if driver->termios[idx]
return value different form NULL, but tty is uninicialized. Sample returned
value is 0x00000001 and kernel die.
This may pop up in CVS ruby to.

Ruby tty_io.c very easy diff from linus tree and I do not understand why
all work under linus tree.

This is gcc or ANSI C or kernel freak ?

Aivils Stoss

pretty sad + Re: HELP with aiptek and keybdev

[Svetoslav S. <ga...@st...>]

Hi got just the same oopses in sawapper both with usb and second PS2 keyboard,
i'll write them out tomorrow ant put them on 
http://varna.demon.co.uk/~svetlio/

needed hints for debuging 
i was thinking about uml linux, but i'm not sure that this can help me.
and my timezone GMT +1 

kind regards

svetljo

Re: HELP with aiptek and keybdev

[Svetoslav S. <ga...@st...>]

and hm,
i did uploaded my "/proc" if someone is interested, one could browse it
at http://varna.demon.co.uk/~svetlio/proc

thanks again

svetljo
 

Re: HELP with aiptek and keybdev

[Svetoslav S. <ga...@st...>]

Hi 
thanks for the fast reply
> All linus-tree stuff is rewrited here. Keyboard interface put back in
> ruby
> CVS when
> ruby code ported into linux-dj. IMHO very hard get back keyboard
> interface.
> 
> 
> If You lost keyboard , You can play with kbd_connect() keyboard.c
> Any correctly detected keyboard call this to find free VT.
> 
i'll take a look , just upladed diff between your and my drivers/char/keyboard.c
and smth else the patch you made about XFree (and this one as well)
shouldn't it be smth like :
 --- linux.orig/drivers/char/keyboard.c Wed Apr 10 01:32:01 2002
 +++ linux/drivers/char/keyboard.c     Thu Jun 27 11:12:01 2002
and not :
> --- drivers/char/keyboard.c.org Wed Apr 10 01:32:01 2002
> +++ drivers/char/keyboard.c     Thu Jun 27 11:12:01 2002
what if it didn't find keyboard.c.orig
> @@ -1089,16 +1089,19 @@
>                  if(vt->next && !vt->next->keyboard ) {
>                         vt = vt->next;
>                         continue;
>                 }
>                  if (!vt->keyboard ) {
>                          vt->keyboard = handle;
>                         handle->private = vt;
>                         vt_map_input(vt);
> +/* Place here Your code like */
> +/* printk("KBD_CONNECT: same keyboard mapped with %08x\n", vt); */
> +/* vt has no name */
>                         break;
>                 }
>                 vt = vt->next;
>         }
> 
>         /* If we have more keyboards than VTs we still register the
> handler.
>          * It is possible someone might add a graphics card thus
> needing
> the
>           * keyboard later */
> 
> 
> I try start Your kernel and get nasty old bug.
which one :)
> tty_struct tty->termios are destroyed or uninitialized or filled with
> garbage.
> IMHO  my backstreet-ruby has same bug, but that do not pop up.
> First I should find this, then do merging with CVS ruby.
> My knowlege is to few for fast work.
> 
> Unfortunately I must finish one comercial project.
> 
> Best regards
> Aivils
thanks once again for the fast reply

svetljo

Re: HELP with aiptek and keybdev

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

HI Svetljo

>Hi i need help again :(
>i probably went too far updating
>Aivils Stoss's backstreet ruby from ruby's cvs
>and now i think i need the keyboard interface in order to use keyboard,
but i can >not compile keybdev :
>depmod: *** Unresolved symbols in
>/lib/modules/2.4.19-pre10aa2-jam2-backstreet2-2406kbd/kernel/drivers/input/keybdev.o

>depmod:         handle_scancode
>depmod:         keyboard_tasklet
>depmod:         kbd_ledfunc
>make: *** [_modinst_post] Error 1

All linus-tree stuff is rewrited here. Keyboard interface put back in ruby
CVS when
ruby code ported into linux-dj. IMHO very hard get back keyboard interface.

>so can smbd please help me to get keyboard working without keybdev or  to
get >keybdev working
>i did booted the kernel, and i wasn't that surprised when i found out that
i got no >repsonse from the keyboard
>( the same was with the -dj tree )
>while cat /proc/bus/input/devices tells me
>I: Bus=0003 Vendor=05fe Product=0011 Version=0000
>N: Name="Cypress Sem. PS2/USB Browser Combo Mouse"
>P: Phys=usb1:1/input0
>D: Drivers=event2 mouse1
>B: EV=6
>B: KEY=1f0000 0 0 0 0 0 0 0 0
>B: REL=103
>
>I: Bus=0011 Vendor=0002 Product=0002 Version=0100
>N: Name="PS2++ Logitech Mouse"
>P: Phys=isa0060/serio1/input0
>D: Drivers=mouse0 event0
>B: EV=6
>B: KEY=70000 0 0 0 0 0 0 0 0
>B: REL=3
>
>I: Bus=0011 Vendor=0001 Product=0002 Version=ab02
>N: Name="AT Set 2 keyboard"
>P: Phys=isa0060/serio0/input0
>D: Drivers=event1 kbd
>B: EV=120002
>B: KEY=41f9 fff9df20 efdfffdf ffefffff ffffffff fffffffe
>B: LED=7

If You lost keyboard , You can play with kbd_connect() keyboard.c
Any correctly detected keyboard call this to find free VT.

--- drivers/char/keyboard.c.org Wed Apr 10 01:32:01 2002
+++ drivers/char/keyboard.c     Thu Jun 27 11:12:01 2002
@@ -1089,16 +1089,19 @@
                 if(vt->next && !vt->next->keyboard ) {
                        vt = vt->next;
                        continue;
                }
                 if (!vt->keyboard ) {
                         vt->keyboard = handle;
                        handle->private = vt;
                        vt_map_input(vt);
+/* Place here Your code like */
+/* printk("KBD_CONNECT: same keyboard mapped with %08x\n", vt); */
+/* vt has no name */
                        break;
                }
                vt = vt->next;
        }

        /* If we have more keyboards than VTs we still register the
handler.
         * It is possible someone might add a graphics card thus needing
the
          * keyboard later */


I try start Your kernel and get nasty old bug.
tty_struct tty->termios are destroyed or uninitialized or filled with
garbage.
IMHO  my backstreet-ruby has same bug, but that do not pop up.
First I should find this, then do merging with CVS ruby.
My knowlege is to few for fast work.

Unfortunately I must finish one comercial project.

Best regards
Aivils

HELP with aiptek and keybdev

[svetljo <sv...@ly...>]

Hi i need help again :(
i probably went too far updating 
Aivils Stoss's backstreet ruby from ruby's cvs
and now i think i need the keyboard interface in order to use keyboard, but i can not compile keybdev :
depmod: *** Unresolved symbols in /lib/modules/2.4.19-pre10aa2-jam2-backstreet2-2406kbd/kernel/drivers/input/keybdev.o
depmod:         handle_scancode
depmod:         keyboard_tasklet
depmod:         kbd_ledfunc
make: *** [_modinst_post] Error 1

so can smbd please help me to get keyboard working without keybdev or  to get keybdev working
i did booted the kernel, and i wasn't that surprised when i found out that i got no repsonse from the keyboard
( the same was with the -dj tree )
while cat /proc/bus/input/devices tells me
I: Bus=0003 Vendor=05fe Product=0011 Version=0000
N: Name="Cypress Sem. PS2/USB Browser Combo Mouse"
P: Phys=usb1:1/input0
D: Drivers=event2 mouse1
B: EV=6
B: KEY=1f0000 0 0 0 0 0 0 0 0
B: REL=103

I: Bus=0011 Vendor=0002 Product=0002 Version=0100
N: Name="PS2++ Logitech Mouse"
P: Phys=isa0060/serio1/input0
D: Drivers=mouse0 event0
B: EV=6
B: KEY=70000 0 0 0 0 0 0 0 0
B: REL=3

I: Bus=0011 Vendor=0001 Product=0002 Version=ab02
N: Name="AT Set 2 keyboard"
P: Phys=isa0060/serio0/input0
D: Drivers=event1 kbd
B: EV=120002
B: KEY=41f9 fff9df20 efdfffdf ffefffff ffffffff fffffffe
B: LED=7
dummy me, i didn't run dmesg or dmesg | ksymoops
but i did copied the major part of /proc
so i hope i have some needed information
i did that trught ssh ( my keyboard wasn't working )

i'll put patches and the kernel as usual on
http://varna.demon.co.uk/~svetlio

and about Aiptek i got
line 300 structure aiptek doesn't have member dev.number

looking forward for your help

regards

svetljo
---
Svetoslav Dimitrov Slavtschev

sv...@ly...
sve...@we...



____________________________________________________________
Win a first-class trip to New Orleans and vacation Elvis Style!.
Enter NOW!
http://r.lycos.com/r/sagel_mail/http://www.elvis.lycos.com/sweepstakes/

Re: (no subject)

[Brad H. <bh...@bi...>]

On Sat, 15 Jun 2002 00:19, Aivils Stoss wrote:
<snip>
> Hi,
> i'm uploading new test kernels the patch from Aivils Stoss +
> 2.4.19-pre10aa2jam2
> + alsa + freeswan + ...   on http://varna.demon.co.uk/~svetlio .
> Help is needed to test the kernels and fix some glitches, and may be to
> backport
> some changes from the current ruby.

I have already started backporting some of the input subsystem changes. 
Marcelo will probably take them in 2.4.20-pre. Do you want them now?

Brad

-- 
http://conf.linux.org.au. 22-25Jan2003. Perth, Australia. Birds in Black.

(no subject)

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

----- Forwarded by Aivils Stoss/PARVALDE/UNIBANKA/LV on 2002.06.14 17:16
-----
                                                                                                        
                    Svetoslav Slavtchev                                                                 
                    <ga...@st...w.uni-er        To:     Aiv...@un...                  
                    langen.de>                        cc:                                               
                                                      Subject:                                          
                    2002.06.14 16:20                                                                    
                                                                                                        
                                                                                                        




Hi
could you forward this message to the list,
i don't get it , but the list doesn't accept  my messages for a week now.
thanks

svetljo
------------------------------------------

Hi,
i'm uploading new test kernels the patch from Aivils Stoss +
2.4.19-pre10aa2jam2
+ alsa + freeswan + ...   on http://varna.demon.co.uk/~svetlio .
Help is needed to test the kernels and fix some glitches, and may be to
backport
some changes from the current ruby.

svetljo
------------------------------------------

Re: problems configuring/compile the kernel

[Svetoslav S. <ga...@st...>]

>
>
>I couldn't start ALSA sound drivers , because these
>confict with USB keyboard.
>Gameport support is modified, untested , may be
>uncompatible with ALSA.
>Ask Vojtech Pavlik how to use ALSA with ruby.
>
>Svjatoslav to made big work http://varna.demon.co.uk/~svetlio
>Here ruby with included ALSA drivers.
>  
>
:( but it's completly unstable

alsa oopses in /proc/asound which was fixed from the alsa developers
and i'm getting oopses in swaper as soon as i press a key on the usb 
keyboard
i'll try to put a new one as soon as -jp14 is out

svetljo

Re: problems configuring/compile the kernel

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

>Any suggestions on getting things to compile?   I don't know what the
>"warning: implicit declaration of function" message means.  Possibly this
is
>okay I just need to set a compiler flag to skip this warning?  I am using
>RedHat 7.3 with kernel 2.4.18-3. Thanks.

You can use older ruby clone http://startx.times.lv.
This one should be compatible with 2.4.18.
I never try RedHat altered kernel.
If will not work grab standard linux-2.4.18.tar.gz.
I couldn't start ALSA sound drivers , because these
confict with USB keyboard.
Gameport support is modified, untested , may be
uncompatible with ALSA.
Ask Vojtech Pavlik how to use ALSA with ruby.

Svjatoslav to made big work http://varna.demon.co.uk/~svetlio
Here ruby with included ALSA drivers.

Aivils

Re: problems configuring/compile the kernel

[Johann D. <joh...@la...>]

Christopher Burke wrote:
> [...]
> make[2]: *** [printk.o] Error 1
> make[2]: Leaving directory `/usr/src/linux-2.4.18-3/kernel'
> 
> Any suggestions on getting things to compile?   I don't know what the 
> "warning: implicit declaration of function" message means.  Possibly this is 
> okay I just need to set a compiler flag to skip this warning?  I am using 
> RedHat 7.3 with kernel 2.4.18-3. Thanks.
> 
> 

You are using the wrong kernel. In ruby/linux, you should find a file 
named AGAINST-xxx, where xxx indicates the version of the kernel (from 
_linus_, not a redhat-modified one). You can only use this kernel.


-- 
Johann Deneux

gameport integrated into motherboard supported?

[Christopher B. <cj...@as...>]

On my computer the gameport and sound are integrated on the motherboard.  I am 
wondering whether this type of gameport is supported?  Should I try the ISA 
or PnP gameport selection during kernel configuration?  My motherboard uses 
the VIA 8233A Southbridge.  The diagram of my motherboard in the manual shows 
a chip connected to the southbridge labeled ITE 8705 that controls the COM 
ports, PS/2 kb/mouse, LPT Port, Floppy, and Gameport.
Thanks for your help.
 

problems configuring/compile the kernel

[Christopher B. <cj...@as...>]

I am a new user of the linuxconsole project
I am trying to get my Interact HammerHead joystick working under linux.
So, far I have encountered several problems trying to configure/compile the 
kernel.
I downloaded the ruby kernel source through the CVS.
During configuration of the kernel I ran into two problems in 
/ruby/linux/arch/i386/config.in (Rev. 1.42)
Configure failed when it could not find init/Config.in
This doesn't exist, so I commented out
source init/Config.in
You may want to add a check for the file before sourcing it?

An additional failure happened on the following lines
tristate 'Sound card support' CONFIG_SOUND
if [ "$CONFIG_SOUND" != "n" ]; then
   source sound/Config.in
fi
I added the drivers directory
if [ "$CONFIG_SOUND" != "n" ]; then
   source drivers/sound/Config.in
fi

After these two changes I was able to configure the kernel.
Compiling the kernel failed on
/ruby/linux/kernel/printk.c (Rev. 1.29)
with the following error message
printk.c:57: invalid initializer
printk.c:58: conflicting types for `console_printk'
printk.c:57: previous declaration of `console_printk'
printk.c:58: invalid initializer
printk.c:59: conflicting types for `console_printk'
printk.c:58: previous declaration of `console_printk'
printk.c:59: invalid initializer
printk.c:60: conflicting types for `console_printk'
printk.c:59: previous declaration of `console_printk'
printk.c:60: invalid initializer
printk.c: In function `console_conditional_schedule':
printk.c:582: warning: implicit declaration of function `need_resched'
make[2]: *** [printk.o] Error 1
make[2]: Leaving directory `/usr/src/linux-2.4.18-3/kernel'

Any suggestions on getting things to compile?   I don't know what the 
"warning: implicit declaration of function" message means.  Possibly this is 
okay I just need to set a compiler flag to skip this warning?  I am using 
RedHat 7.3 with kernel 2.4.18-3. Thanks.

Re: X compiled for evdev?

[<Jim...@hp...>]

> Sender: lin...@li...
> From: James Simmons <jsi...@tr...>
> Date: Fri, 31 May 2002 14:25:06 -0700 (PDT)
> To: Brad Midgley <bmi...@tu...>
> Cc: lin...@li...
> Subject: Re: X compiled for evdev?
> -----
> > I see a note in linux/Documentation/input/input.txt about using X with
> > the current input stuff for getting multihead. Has anyone built X with
> > evdev support or is it standard via XFConfig manipulation?
> >
> > either way I'd like to document this.
> 
> Jim Getty was working on supporting evdev in X windows but I haven't heard
> how far he has gotten.

I haven't gotten anywhere yet: been distracted on other things. I hope
to get back to it this summer,  unless someone beats me to the punch.

> 
> > what about order detection? if two usb keyboards are attached, aren't
> > they essentially randomly ordered? a long time ago i wrote an app for
> > sorting this out via "press the 1 key" and "press the 2 key" etc on each
> > head. do we have to do something like that?
> 
> That would require a huge reconstruction of the X server. The core code
> assumes one keyboard.
> 
You can have multiple keyboards:
	o either they get merged and look like one keyboard
	o or, if you use the X Input extension, it appears as another input
device.

There is a built in presumption about there being one primary keyboard
and pointing device; some of these will eventually have to get revisited
in an extension to deal with situations like display walls being used
by multiple people simultaneously.
			- Jim


--
Jim Gettys
Cambridge Research Laboratory
HP Labs, Hewlett-Packard Company
Jim...@hp...

Re: X compiled for evdev?

[James S. <jsi...@tr...>]

> I see a note in linux/Documentation/input/input.txt about using X with
> the current input stuff for getting multihead. Has anyone built X with
> evdev support or is it standard via XFConfig manipulation?
>
> either way I'd like to document this.

Jim Getty was working on supporting evdev in X windows but I haven't heard
how far he has gotten.

> what about order detection? if two usb keyboards are attached, aren't
> they essentially randomly ordered? a long time ago i wrote an app for
> sorting this out via "press the 1 key" and "press the 2 key" etc on each
> head. do we have to do something like that?

That would require a huge reconstruction of the X server. The core code
assumes one keyboard.

Re: input pages

[James S. <jsi...@tr...>]

> were the input pages copied locally and put into cvs so they could be
> modified as part of this project?

Yes.

> if so, that makes sense but i think
> i'd distill the relevant stuff into one or two pages so it's cleaner and
> easier to read/modify.

Could you clean them up please.

writing userland docs

[Brad M. <bmi...@tu...>]

hi

i've looked over the docs and i think there is one big missing piece: 
what userland apps need to do differently to deal with the new console 
system properly and what new userland apps we will need to do stuff like 
console/vt/head assignments. this document would basically explain how 
to do the xfree mods we need.

can someone point me to whatever we have so far and i will work on that?

i will link it under the "Documentaion" section in htdocs/index.html (i 
swear i fixed that typo but it persists :)

brad

[PATCH] fbdev updates.

[James S. <jsi...@tr...>]

Hi!

   I just ported over a few driver to the new api. I also included a few
bug fixes as well as a few new drivers. Please try it out and send me any
patches. Once the testing is done I will ask Linus to included into his
tree. If you look at skeletonfb.c you will see actual info on the new api
to help you port it over to the new api.

BK:

   http://fbdev.bkbits.net:8080/fbdev-2.5

Diff:

   http://www.transvirtual.com/~jsimmons/fbdev.diff

P.S
   Yes with with new api I haven't finished the drawing penguin code so
you don't see a penguin yet. I will but I figured the new api is more
important.

   . ---
   |o_o |
   |:_/ |   Give Micro$oft the Bird!!!!
  //   \ \  Use Linux!!!!
 (|     | )
 /'_   _/`\
 ___)=(___/

Re: multihead patches

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

Hi
I was to lazy test Migel's 2-nd server patch.
USB will not work with my motherboard.
Migel's 1-st server patch allow me run only Voodoo1 + any other video card
pair.

I use this in both (triplet is tested) X  4.2.0 servers with ruby VT

--- xc/programs/Xserver/hw/xfree86/common/xf86pciBus.c.original   Fri Nov
30 14:11:55 2001
+++ xc/programs/Xserver/hw/xfree86/common/xf86pciBus.c  Mon Apr 15 10:10:41
2002
@@ -584,9 +584,11 @@
 #ifdef DEBUG
     ErrorF("pciIoAccessDisable: 0x%05lx\n", *(PCITAG *)arg);
 #endif
+#if 0
     ((pciArg*)arg)->ctrl &= ~SETBITS;
     ((pciArg*)arg)->func(((pciArg*)arg)->tag, PCI_CMD_STAT_REG,
                ((pciArg*)arg)->ctrl);
+#endif
 }

 #undef SETBITS
@@ -608,9 +610,11 @@
 #ifdef DEBUG
     ErrorF("pciIo_MemAccessDisable: 0x%05lx\n", *(PCITAG *)arg);
 #endif
+#if 0
     ((pciArg*)arg)->ctrl &= ~SETBITS;
     ((pciArg*)arg)->func(((pciArg*)arg)->tag, PCI_CMD_STAT_REG,
                ((pciArg*)arg)->ctrl);
+#endif
 }

 #undef SETBITS
@@ -632,9 +636,11 @@
 #ifdef DEBUG
     ErrorF("pciMemAccessDisable: 0x%05lx\n", *(PCITAG *)arg);
 #endif
+#if 0
     ((pciArg*)arg)->ctrl &= ~SETBITS;
     ((pciArg*)arg)->func(((pciArg*)arg)->tag, PCI_CMD_STAT_REG,
                ((pciArg*)arg)->ctrl);
+#endif
 }
 #undef SETBITS

VGA output freeze if You start 2-nd X.
Standard RedHat 7.2 do not feel difference is patched or normal X.

Aivils Stoss

Re: multihead patches

[Brad M. <bmi...@tu...>]

hi

>>Have you read the text? How do you suggest receiving scan codes from
>>the
>>second keyboard only?
> 
> no way until ruby works, but when it will work ?

so it sounds like there's a mapping that we don't get until the console 
redesign is finished... we're not going to want to push for an xfree 
patch upstream until it's extremely clean.

there are some changes in pci stuff, leds, ioctls, etc. but hopefully 
that can be selected at runtime or otherwise cleaned up without much 
trouble.

brad

Re: multihead patches

[Svetoslav S. <ga...@st...>]

Hi to all of you
> Hi Brad,
> 
> On Tue, 2002-05-21 at 13:13, Brad Midgley wrote:
> > i would like to get a set of patches such that you can use the same 
> > xfree binary for either head. ideally, this same binary could be used
> by 
> > people not doing multihead so xfree could maybe merge the patch. do
> you 
> > think it could be based on your patch or should it be started from
> scratch?
> 
> I think you can base it on this patch, just adding some parameters to
> xfree86 so it would disable vt handling code and avoid messing with
> some
> pci stuff.

you can probably find helpful some ideas from here
 http://mail-index.netbsd.org/port-i386/2001/09/
thread "multiuser X 4.1.0"
> Note that even joining both patches it still a hack! It is much likely
> to be other problems with some specific video cards/hardware.

I`m haveing the opinion that not all cards can be used at same time from 
different X servers
> > i have a couple of questions from looking at the second-head patch 
> >
> http://cambuca.ldhs.cetuc.puc-rio.br/multiuser/XFree86-4.1.0-2nd-0.92.patch
> > 
> > is there a way to avoid the pc-at keyboard hack code? why is it
> necessary?
> 
> Have you read the text? How do you suggest receiving scan codes from
> the
> second keyboard only?
no way until ruby works, but when it will work ?

and i wanted to report that the part about G450 works with G550 useing fbdev as 
driver, i got some troubles with the mices, and i couldn`t get the first head 
accelerated so i reverted to to gfx cards


Regards

svetljo

Re: multihead patches

[Miguel F. <mi...@ce...>]

Hi Brad,

On Tue, 2002-05-21 at 13:13, Brad Midgley wrote:
> i would like to get a set of patches such that you can use the same 
> xfree binary for either head. ideally, this same binary could be used by 
> people not doing multihead so xfree could maybe merge the patch. do you 
> think it could be based on your patch or should it be started from scratch?

I think you can base it on this patch, just adding some parameters to
xfree86 so it would disable vt handling code and avoid messing with some
pci stuff.

Note that even joining both patches it still a hack! It is much likely
to be other problems with some specific video cards/hardware.

> i have a couple of questions from looking at the second-head patch 
> http://cambuca.ldhs.cetuc.puc-rio.br/multiuser/XFree86-4.1.0-2nd-0.92.patch
> 
> is there a way to avoid the pc-at keyboard hack code? why is it necessary?

Have you read the text? How do you suggest receiving scan codes from the
second keyboard only?


regards,

Miguel