linuxconsole-dev

Re: Can someone point me towards a fix for this...?

[James v. Z. <ja...@dv...>]

> Please try before gdm starts commad
> X -probeonly -layout noprefbusid
> 

I am letting rhgb (graphical X boot) use a three card config for this,
so that's already happenning, all three cards, all OK...
The third X server still fails to produce a login when gdm starts after
rhgb closes. Three X servers work fine if you launch the third one
manually with startx. It doesn't seem to matter the order cards are
init. This is the first time I've seen an issue like it.

I shall try some new nv drivers (still 6106, here) and a few other
configs and post back.

Perhaps I should try a PCI BIOS Primary?

I am fairly certain it isn't a usual "card isn't init" problem as I have
three ruby machines here and have found using rhgb to perform the
multi-card video init very effective, not to mention convenient.

Nevertheless, I shall try dropping a script in to close rhgb early and
perform the -probeonly video init.

J

On Thu, 2004-08-26 at 01:21, Aivils wrote:
> On Monday 23 August 2004 17:14, James van Zeeland wrote:
> > 3 consoles, custom ruby kernel 2.6.7 merged with con kolivas and
> > stanford uni's planet ccrma patchsets plus a heap of usb fixes and
> > whatnot. Fedora Core 2
> > 
> > Athlon 2600+ 7NNXP 1Gb (nForce2)
> > 
> > 1 x AGP console FX5200
> > 2 x PCI console GF4MX440
> > 
> > using three servers defined in gdm.conf, console 1 starts OK, console 2
> > starts OK, console three however does the following.
> > X server starts and we have mouse action, but gdm doesn't start. After a
> > bit of a wait, it closes and restarts. The restart is not frequent
> > enough to get the usual "failed to start, will disable until correctly
> > configured." dialogue
> > 
> > What's interesting is that if I disable the third server in gdm, drop to
> > console and "startx -- :2 -layout PCI-Console2 -audit 0 vt19",
> > everything starts fine, and I have three consoles.
> > 
> > If I change the init order of the PCI X servers, it's always the third
> > server that fails. Problem with gdm?
> > 
> > I am glad for the patch page, I was considering doing that myself...
> 
> Please try before gdm starts commad
> X -probeonly -layout noprefbusid
> where noprefbusid is unrestricted xf86, which can use all video
> adapters. This is necessary for proper multiple Nvidia adapters init.
> 
> Aivils
> 

Re: Can someone point me towards a fix for this...?

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

On Monday 23 August 2004 17:14, James van Zeeland wrote:
> 3 consoles, custom ruby kernel 2.6.7 merged with con kolivas and
> stanford uni's planet ccrma patchsets plus a heap of usb fixes and
> whatnot. Fedora Core 2
> 
> Athlon 2600+ 7NNXP 1Gb (nForce2)
> 
> 1 x AGP console FX5200
> 2 x PCI console GF4MX440
> 
> using three servers defined in gdm.conf, console 1 starts OK, console 2
> starts OK, console three however does the following.
> X server starts and we have mouse action, but gdm doesn't start. After a
> bit of a wait, it closes and restarts. The restart is not frequent
> enough to get the usual "failed to start, will disable until correctly
> configured." dialogue
> 
> What's interesting is that if I disable the third server in gdm, drop to
> console and "startx -- :2 -layout PCI-Console2 -audit 0 vt19",
> everything starts fine, and I have three consoles.
> 
> If I change the init order of the PCI X servers, it's always the third
> server that fails. Problem with gdm?
> 
> I am glad for the patch page, I was considering doing that myself...

Please try before gdm starts commad
X -probeonly -layout noprefbusid
where noprefbusid is unrestricted xf86, which can use all video
adapters. This is necessary for proper multiple Nvidia adapters init.

Aivils

Re: Ruby-2.6.8

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

On Wednesday 25 August 2004 09:47, Helge Hafting wrote:
> Aivils wrote:
> 
> >Hi all!
> >
> >	subject reach public cvs :)
> >fbcon will run only on my vesafb-tng. "multiuser" fbcon is untested.
> >fbcon is simplified because do not contain "multiheading".
> >Feel free and applay various subversions like .1 .2 and so on by yourself.
> >  
> >
> Multiuser fbcon failed.  The pc just sat there, the framebuffers loaded
> but no consoles.  The keyboard that usually gets the first console
> responded to the numlock/capslock keys.  The other keyboard was
> completely dead.  Seems it hung after attaching the first keyboard,
> before attaching the second.  sysrq+B worked on the "working" keyboard,
> not the other.
Seems linux-ruby is more half-complete as ordinarily :(
In the office space computer with Intel 82845 / nVidia [GeForce4 MX 440]
hardware linux-ruby behave as You describe above.

Aivils

Re: Trouble trying to run 2.6.7-ruby and X on Radeon+Matrox

[Helge H. <hel...@hi...>]

Cédric Rivard wrote:

>Le dimanche 22 Août 2004 16:14, Helge Hafting wrote :
>  
>
>>Unfortunately, I haven't even succeeded in running an xserver on the pci
>>card, which I find strange.
>>I haven't been able to run a xserver on the radeon though.
>>The server starts, but the radeon screen turns into vertical stripes
>>and the mouse cursor is also a bunch of stripes.  I can move it around,
>>so the xserver is running but not in a useable state.
>>    
>>
>
>Dual radeon setup usually works with the PCI card configured in the bios as 
>the primary display controller. Maybe could you try this way.
>  
>
Thanks for the tip.  I can't find any such option in the bios though, I hope
that the next board will have it. (If it ever arrives.)

That particular radeon turned out to be broken.  I finally pulled the 
AGP card
so the PCI radeon became the only one.  The pc booted but there were 
vertical
stripes only on the screen - the colors were recogniseable but that was it!

So it looks like the int10 thing in xfree works - but the card itself is 
broken
and I am exchanging it.  Seems I have real bad luck with hardware.

Helge Hafting

Re: Ruby-2.6.8

[Helge H. <hel...@hi...>]

Aivils wrote:

>Hi all!
>
>	subject reach public cvs :)
>fbcon will run only on my vesafb-tng. "multiuser" fbcon is untested.
>fbcon is simplified because do not contain "multiheading".
>Feel free and applay various subversions like .1 .2 and so on by yourself.
>  
>
Multiuser fbcon failed.  The pc just sat there, the framebuffers loaded
but no consoles.  The keyboard that usually gets the first console
responded to the numlock/capslock keys.  The other keyboard was
completely dead.  Seems it hung after attaching the first keyboard,
before attaching the second.  sysrq+B worked on the "working" keyboard,
not the other.

Helge Hafting

Re: Ruby-2.6.8

[Andreas S. <an...@sc...>]

* Aivils (ai...@un...) [040824 09:37]:
> 	subject reach public cvs :)
> fbcon will run only on my vesafb-tng. "multiuser" fbcon is untested.
> fbcon is simplified because do not contain "multiheading".
> Feel free and applay various subversions like .1 .2 and so on by yourself.

i am getting compile errors here:

...
 CC      arch/i386/kernel/setup.o
arch/i386/kernel/setup.c: In function `parse_cmdline_early':
arch/i386/kernel/setup.c:661: error: `saved_command_line' undeclared (first use in this function)
arch/i386/kernel/setup.c:661: error: (Each undeclared identifier is reported only once
arch/i386/kernel/setup.c:661: error: for each function it appears in.)
In file included from arch/i386/kernel/setup.c:1380:
include/asm-i386/mach-default/setup_arch_post.h: At top level:
include/asm-i386/mach-default/setup_arch_post.h:10: warning: `machine_specific_memory_setup' declared inline after being called
make[2]: *** [arch/i386/kernel/setup.o] Error 1

i guessed it was my large memory support causeing this, but
disabling that option did not fix it.

does anyone else see this? how do i avoid it?

Re: Trouble trying to run 2.6.7-ruby and X on Radeon+Matrox

[<ce...@ri...>]

Le dimanche 22 Ao=FBt 2004 16:14, Helge Hafting wrote=A0:
> Unfortunately, I haven't even succeeded in running an xserver on the pci
> card, which I find strange.
> I haven't been able to run a xserver on the radeon though.
> The server starts, but the radeon screen turns into vertical stripes
> and the mouse cursor is also a bunch of stripes.  I can move it around,
> so the xserver is running but not in a useable state.

Dual radeon setup usually works with the PCI card configured in the bios as=
=20
the primary display controller. Maybe could you try this way.

Cheers,
C=E9dric

Re: Ruby-2.6.8

[Helge H. <hel...@hi...>]

Aivils wrote:

>Hi all!
>
>	subject reach public cvs :)
>  
>
Great!

>fbcon will run only on my vesafb-tng. "multiuser" fbcon is untested.
>fbcon is simplified because do not contain "multiheading".
>Feel free and applay various subversions like .1 .2 and so on by yourself.
>
>  
>
I applied ruby-2.6.8 on top of linux-2.6.8.1.  It compiled fine,
including fbcon which didn't compile in Zoltan's version.
I'll boot it later today, when I get home, we'll see how the
fbcon fares with two consoles.

Helge Hafting

Ruby-2.6.8

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

Hi all!

	subject reach public cvs :)
fbcon will run only on my vesafb-tng. "multiuser" fbcon is untested.
fbcon is simplified because do not contain "multiheading".
Feel free and applay various subversions like .1 .2 and so on by yourself.

Aivils

Can someone point me towards a fix for this...?

[James v. Z. <ja...@dv...>]

3 consoles, custom ruby kernel 2.6.7 merged with con kolivas and
stanford uni's planet ccrma patchsets plus a heap of usb fixes and
whatnot. Fedora Core 2

Athlon 2600+ 7NNXP 1Gb (nForce2)

1 x AGP console FX5200
2 x PCI console GF4MX440

using three servers defined in gdm.conf, console 1 starts OK, console 2
starts OK, console three however does the following.
X server starts and we have mouse action, but gdm doesn't start. After a
bit of a wait, it closes and restarts. The restart is not frequent
enough to get the usual "failed to start, will disable until correctly
configured." dialogue

What's interesting is that if I disable the third server in gdm, drop to
console and "startx -- :2 -layout PCI-Console2 -audit 0 vt19",
everything starts fine, and I have three consoles.

If I change the init order of the PCI X servers, it's always the third
server that fails. Problem with gdm?

I am glad for the patch page, I was considering doing that myself...

James

Re: Ruby-2.6.8[.1/.2]?

[Andrew C. <an...@di...>]

On Mon, 2004-08-23 at 07:39, Helge Hafting wrote:
> Aivils wrote:

[snip]

> >
> >i dont trust in 2.6.XX
> >  
> >
> Why not?  My experience is that 2.6 works well, but of course 2.4 have had
> much more time to stabilize.  Getting ruby-2.6 into the 2.6.x-mm kernel
> would be nice, there'd be much more testers and less to sync up against
> with new releases. 
> 

I agree, for me 2.6 is looking very good. This is very cool stuff and
I'd love to see it merged into mainline, though I don't know how it
relates the "console re-architecture" plan? that's been talked about at
OLS and on various mailing lists.


> Helge Hafting
> 


Andrew

Re: Still having a problem running second X server on framebuffer.

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

On Monday 23 August 2004 10:35, dave giffin wrote:
> I am trying to run a regular XFree86 server on an ATI
> radeon 7000, and a second X server on a framebuffer on
> a SiS 630 card, using ruby, to have both on screen at
> the same time.
> When I boot my kernel with dumbcon=2 I get:
> Console: Colour VGA+ 80x25 vc:1-16
> Console: mono dummy device 80x25 vc:17-17
> Console: mono dummy device 80x25 vc:18-18
> Console: switching to colour ATI Radeon QY  80x30
> vc:1-16
> Console: Colour SIS 730 100x37 vc:19-34
> 
> How does the range of vc's relate to /dev/ttyX
> devices?
> Does /dev/tty17 == vc17?

Yep

> 
> :)
You must edit /etc/inittab

<skip>
# Run gettys in standard runlevels
1:2345:respawn:/sbin/mingetty tty1
2:2345:respawn:/sbin/mingetty tty2
3:2345:respawn:/sbin/mingetty tty3
4:2345:respawn:/sbin/mingetty tty4
5:2345:respawn:/sbin/mingetty tty5
6:2345:respawn:/sbin/mingetty tty6
17:2345:respawn:/sbin/mingetty tty17
<skip>

last line start additional mingetty specialy for secondary fb device
and dumbcon is unnecessary. linux-ruby automatic bind
fbdev-keyboard pairs as You can see in the dmesg. After editing
inittab, removing dumbcon commandline, You must be able
login via secondary fb device. If You start multiple mingetty
on tty17 tty18 tty19 ... , then You must be able use secondary
keyboard Alt-F1 Alt-F2 Alt-F3 ... keys for vt switching.


Aivils

Re: Problem compiling ruby kernel. Urgent. Please help!

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

On Friday 20 August 2004 00:36, dave giffin wrote:
> I am trying to get multiple independent framebuffer
> support. Apparently, that is only available in a patch
> to the 2.5 kernel series.

All linuxconsole project administrators are retired.
So web space and CVS is lightly useless :)
Do You think about, James?

Aivils

Still having a problem running second X server on framebuffer.

[dave g. <giv...@ya...>]

I am trying to run a regular XFree86 server on an ATI
radeon 7000, and a second X server on a framebuffer on
a SiS 630 card, using ruby, to have both on screen at
the same time.

I have recompiled XFree86 4.3.0 with the PrefBusID
patch, and in the config for the 2nd X server, I have
set it to use the "fbdev" driver and have set the
"fbdev" option to "/dev/fbX".

I am working with vanilla kernel 2.6.7 with the ruby
patch. I have catted files to both framebuffer devices
to verify that they work. I have enabled 'hackvideo'.

When I boot with support for both framebuffers
(radeonfb & sisfb), without any X server running, I
can start the frame buffer X or regular X. 

With both framebuffers, I have set the framebuffer-X
to run on /dev/fb1.

When I start the frame buffer X (onto 2nd display) on
the first display (initted @ boot) I try switching
consoles to login to another shell and run a program
on the X server, the frame buffer X stops updating the
screen. When I run the regular X server first, and
then try starting the fb-X from an xterm, I get an
error message: 
(EE) FBDEV(0): FBIOPAN_DISPLAY: Invalid argument

and the server doesn't display anything.

When I loaded support for only the SiS framebuffer and
disabled support for framebuffer console, I was able
to start the framebuffer X without getting the FBIOPan
error message. (running X on /dev/fb0)

But, when I switched consoles on the display operated
by my ATI card (the card initialized @ boot), the
framebuffer X screen turns blue (it does update).

If I Ctrl+C the fb-X, I can restart it and the screen
returns to normal.

What is going on here? Can I run these 2 servers the
way I have outlined?

Should I try a different kernel version (using 2.6.7)?


(I got the ruby-2.6 patch from this mailing list a few
days ago. the patch is dated 2004-06-16, should I try
to get a new one, if yes, from where?)

My ultimate goal is to use XDirectFB(an X server using
the DirectFB library to access the framebuffer), not
XFree86, for my framebuffer X.

XDirectFB accepts a command-line option as to which
/dev/ttyX device to use (not vt or vc). 

When I boot my kernel with dumbcon=2 I get:
Console: Colour VGA+ 80x25 vc:1-16
Console: mono dummy device 80x25 vc:17-17
Console: mono dummy device 80x25 vc:18-18
Console: switching to colour ATI Radeon QY  80x30
vc:1-16
Console: Colour SIS 730 100x37 vc:19-34

How does the range of vc's relate to /dev/ttyX
devices?
Does /dev/tty17 == vc17?

:)



		
_______________________________
Do you Yahoo!?
Win 1 of 4,000 free domain names from Yahoo! Enter now.
http://promotions.yahoo.com/goldrush

Re: Ruby-2.6.8[.1/.2]?

[Helge H. <hel...@hi...>]

Aivils wrote:

>On Saturday 21 August 2004 05:13, Zoltan Boszormenyi wrote:
>  
>
>>Zoltan Boszormenyi írta:
>>    
>>
>>>Dear Aivil,
>>>
>>>now that Linux-2.6.8.1 is out and the fix to the most important bug,
>>>namely a huge memleak during audio CD/SVCD writing is found and
>>>Linux-2.6.8.2 is just around the corner, can we expect a ruby tree
>>>against them? Or as there is a long weekend in Hungary, will you accept
>>>my work if I volunteer?
>>>
>>>Best regards,
>>>Zoltán Böszörményi
>>>      
>>>
>>I had time to make it, here is the result.
>>Two patches: one is against linux-2.6.8.1,
>>the other is against the ruby/ruby-2.6 CVS tree.
>>
>>WARNING!
>>
>>The FB changes were highly excessive between 2.6.7 and 2.6.8
>>and I don't really know the code there, so fbcon, et.al does not
>>even compile.
>>
>>But the vgacon and dummycon consoles are working so multiple X servers
>>are possible. I am running it now.
>>    
>>
>
>Thank You. I will compare my and your code. I have my own version,
>but without fbcon. fbcon will not work on my Voodoo3. If anyone can
>start vanilla 2.6.8 fbcon, please drop a note to me.
>  
>
I didn't try 2.6.8, but linux 2.6.8.1 works for me with fbcon
on a radeon card.  Zoltans ruby 2.6.8.1 compiles fine, except for the
fbcon module.  I can log in blind and start X though, so
it is useable.  X starting from a boot script works too.
Multiple framebuffers (mga+radeon) does not
work, I get a fb for the mga only.  Ruby 2.6.7 have the same problem.

>Unfortunately a have not patience enough walk through changes,
>which freezy all at start.
>  
>
Probably a broken fb driver then, linux 2.6.8.1 seems fine with the
radeon fb+fbcon.

>I have expirence, if fb low level driver is broken, then it never comes
>back alive. After i will try vesafb-tng by spock.
>  
>

>  
>
>>I have a question: do you intend to push this thing into Linus' or
>>akpm's kernel? I would be happy to see this in -mm because there are
>>many changes again in 2.6.8.1-mm3 against the areas the Ruby tree
>>touches.
>>    
>>
>
>i dont trust in 2.6.XX
>  
>
Why not?  My experience is that 2.6 works well, but of course 2.4 have had
much more time to stabilize.  Getting ruby-2.6 into the 2.6.x-mm kernel
would be nice, there'd be much more testers and less to sync up against
with new releases. 

Helge Hafting

Re: Ruby-2.6.8[.1/.2]?

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

On Saturday 21 August 2004 05:13, Zoltan Boszormenyi wrote:
> Zoltan Boszormenyi =EDrta:
> > Dear Aivil,
> >=20
> > now that Linux-2.6.8.1 is out and the fix to the most important bug,
> > namely a huge memleak during audio CD/SVCD writing is found and
> > Linux-2.6.8.2 is just around the corner, can we expect a ruby tree
> > against them? Or as there is a long weekend in Hungary, will you accept
> > my work if I volunteer?
> >=20
> > Best regards,
> > Zolt=E1n B=F6sz=F6rm=E9nyi
>=20
> I had time to make it, here is the result.
> Two patches: one is against linux-2.6.8.1,
> the other is against the ruby/ruby-2.6 CVS tree.
>=20
> WARNING!
>=20
> The FB changes were highly excessive between 2.6.7 and 2.6.8
> and I don't really know the code there, so fbcon, et.al does not
> even compile.
>=20
> But the vgacon and dummycon consoles are working so multiple X servers
> are possible. I am running it now.

Thank You. I will compare my and your code. I have my own version,
but without fbcon. fbcon will not work on my Voodoo3. If anyone can
start vanilla 2.6.8 fbcon, please drop a note to me.
Unfortunately a have not patience enough walk through changes,
which freezy all at start.
I have expirence, if fb low level driver is broken, then it never comes
back alive. After i will try vesafb-tng by spock.

> I have a question: do you intend to push this thing into Linus' or
> akpm's kernel? I would be happy to see this in -mm because there are
> many changes again in 2.6.8.1-mm3 against the areas the Ruby tree
> touches.

i dont trust in 2.6.XX

Aivils

Trouble trying to run 2.6.7-ruby and X on Radeon+Matrox

[Helge H. <hel...@ai...>]

I have been running ruby-2.6.7 successfully with one xserver
using Matrox G550 accelerated and another using the framebuffer driver
one the mga's secondary framebuffer.


Hearing about success with 2-user DRI, I got a radeon sapphire pci
card in order to run two accelerated servers (and eventually dual DRI too.)


Unfortunately, I haven't even succeeded in running an xserver on the pci card,
which I find strange.  The first error comes in the boot messages:

fb0 initializing hardware  (The G550 becomes fb0)
PCI enabling  device 0000:00:0d.0  (The radeon)
radeonfb: cannot map fb

Seems strange.  /dev/fb0 is probably taken, but there are plenty
of others?  Anyway, X can work without a framebuffer so it
shouldn't matter.  I compile in framebuffers though, because I
hope to have several gettys on each fbcon someday, when ruby
gets support for VT switching on several consoles.

AGP initializes ok
[drm] for radeon 7000/VE
[drm] for mga ok

Later, the fbcon module loads and tell me I get two fbcons
on the G550 and none on the radeon.  It shouldn't really
matter for X though.


I haven't been able to run a xserver on the radeon though.
The server starts, but the radeon screen turns into vertical stripes
and the mouse cursor is also a bunch of stripes.  I can move it around,
so the xserver is running but not in a useable state.

Starting _only_ the radeon xserver and no others does not help,
so it isn't even a problem of simultaneous servers yet.

Do anybody have any ideas?  I am compiling ruby-2.6.8.1 now,
and I deconfigured the second head support for matrox as that
won't be needed once the radeon works.

Helge Hafting

Re: Error running a second X server via framebuffer. Please help!

[Helge H. <hel...@ai...>]

On Fri, Aug 20, 2004 at 11:26:32PM -0700, dave giffin wrote:
> I'm trying to run a second X server using the
> framebuffer.
> 
> I applied the ruby-2.6 patch to vanilla kernel version
> 2.6.7. I recompiled XFree86 4.3.0 with the prefbusid
> patch. 
> 
> I start up my 1st X server (on an ATI card, not using
> framebuffer) without a problem. Then from an xterm on
> that server, I try starting the 2nd framebuffer X
> server on a SiS card, and get the error below. 
> 

You are sure you _have_ a working framebuffer?
I.e.
$ cat somefile > /dev/fbX
results in some garbage on that screen? (Replace
X with the correct number. 0 if it is the only framebuffer,
1 if the ATI card also claims a framebuffer.  Or test to see
what framebuffer it really is.  Surprises happens here . . . )

> The 2nd server continues to run (though the screen it
> is connected to is blank). When I ctrl+c it, the
> screen my 1st server is on goes blank and is replaced
> by a garbled image of what it used to look like (the
> image doesn't change, maybe the 2nd X server accessed
> the framebuffer on the 1st card).

Well, perhaps the second xserver accessed the wrong framebuffer?
If the Device section of the XF86Config.fb file specifies
Driver "fbdev"
then you also need to specify what framebuffer to use, i.e.
Option "fbdev"  "/dev/fbX"
I believe the framebuffer driver don't use pci at all, so
prefbusid and such doesn't affect it.  Make sure you're trying
to use the correct fb, and that that fb actually works.

> 
> I can still press CTRL+Alt+F1 and then CTRL+ALT+F7 to
> get back to my 1st X server.
> 
Such switching rep}aints the screen, and also ensures that the right
console is working.

> I am  using the "prefbusid" option in the config files
> for both servers. I also activated the 'hackvideo'
> extension. (I ran echo 1 > /proc/bus/pci/hackvideo ).
> 
Not necessary for the fb server, but it may be necessary for the other one
so it don't naively turn off your SiS card.  

> My System:
> ATI Radeon 7000 (1st X server, no framebuffer)
> SiS 630 (2nd X server, using framebuffer)
> Redhat 9 (i686)
> gcc 3.2.2
> XFree86 4.3.0
> kernel 2.6.7 + ruby & SiS & ATI framebuffer drivers
> 
Ah, you have a ATI framebuffer.  Then you probably have
both /dev/fb0 and /dev/fb1.  Make sure your fb-X is
using the right one!
> 
> 
> What can I do about this problem? 
> 
> The error message:
> 
> XFree86 Version 4.3.0
> Release Date: 27 February 2003
> X Protocol Version 11, Revision 0, Release 6.6
> Build Operating System: Linux 2.6.7-ruby i686 [ELF]
> Build Date: 20 August 2004
>         Before reporting problems, check
> http://www.XFree86.Org/
>         to make sure that you have the latest version.
> Module Loader present
> Markers: (--) probed, (**) from config file, (==)
> default setting,
>          (++) from command line, (!!) notice, (II)
> informational,
>          (WW) warning, (EE) error, (NI) not
> implemented, (??) unknown.
> (==) Log file: "/var/log/XFree86.1.0.log", Time: Fri
> Aug 20 23:01:06 2004
> (++) Using config file: "/etc/X11/XF86Config.fb"
> (EE) FBDEV(0): FBIOPAN_DISPLAY: Invalid argument
> 
Is this FBIOPAN-thing something from the XF86Config.fb file?
If so, try commenting it out.  Or perhaps it is an error
indicating you tried to use a "busy" fb, if you happened to use
the wrong one?

Helge Hafting

patches from 2.6.1 to 2.6.7

[Daniel S. <dan...@ep...>]

Hi!

I have created a simple page where the "history" of ruby from kernel
2.6.1 to 2.6.7 is documented as bzip2-compressed patches against the
vanilla linux kernel. 

http://www.saring.de/derdani/ruby/index.html

I don't know whether it is possible to fetch the state of the repository
from sourceforge for every kernel version. Anyway, as I am not a
developer I missed a place where I could just download a kernel patch.
So I created one :-)

Daniel

Error running a second X server via framebuffer. Please help!

[dave g. <giv...@ya...>]

I'm trying to run a second X server using the
framebuffer.

I applied the ruby-2.6 patch to vanilla kernel version
2.6.7. I recompiled XFree86 4.3.0 with the prefbusid
patch. 

I start up my 1st X server (on an ATI card, not using
framebuffer) without a problem. Then from an xterm on
that server, I try starting the 2nd framebuffer X
server on a SiS card, and get the error below. 

The 2nd server continues to run (though the screen it
is connected to is blank). When I ctrl+c it, the
screen my 1st server is on goes blank and is replaced
by a garbled image of what it used to look like (the
image doesn't change, maybe the 2nd X server accessed
the framebuffer on the 1st card).

I can still press CTRL+Alt+F1 and then CTRL+ALT+F7 to
get back to my 1st X server.

I am  using the "prefbusid" option in the config files
for both servers. I also activated the 'hackvideo'
extension. (I ran echo 1 > /proc/bus/pci/hackvideo ).

My System:
ATI Radeon 7000 (1st X server, no framebuffer)
SiS 630 (2nd X server, using framebuffer)
Redhat 9 (i686)
gcc 3.2.2
XFree86 4.3.0
kernel 2.6.7 + ruby & SiS & ATI framebuffer drivers



What can I do about this problem? 

The error message:

XFree86 Version 4.3.0
Release Date: 27 February 2003
X Protocol Version 11, Revision 0, Release 6.6
Build Operating System: Linux 2.6.7-ruby i686 [ELF]
Build Date: 20 August 2004
        Before reporting problems, check
http://www.XFree86.Org/
        to make sure that you have the latest version.
Module Loader present
Markers: (--) probed, (**) from config file, (==)
default setting,
         (++) from command line, (!!) notice, (II)
informational,
         (WW) warning, (EE) error, (NI) not
implemented, (??) unknown.
(==) Log file: "/var/log/XFree86.1.0.log", Time: Fri
Aug 20 23:01:06 2004
(++) Using config file: "/etc/X11/XF86Config.fb"
(EE) FBDEV(0): FBIOPAN_DISPLAY: Invalid argument



		
__________________________________
Do you Yahoo!?
Yahoo! Mail - 50x more storage than other providers!
http://promotions.yahoo.com/new_mail

[PATCH] Ruby-2.6.8.1, part2, was:Re: Ruby-2.6.8[.1/.2]?

[Zoltan B. <zb...@fr...>]

The second patch is against the latest Ruby-2.6 CVS tree,
a.k.a Ruby-2.6.7.

[PATCH] Ruby-2.6.8.1, part1, was:Re: Ruby-2.6.8[.1/.2]?

[Zoltan B. <zb...@fr...>]

The list settings for attachments is set to 150K, so my work did not
get through. :-( I am sending it again, one patch at a time, compressed
with bzip2 insted of gzip. The first patch is against Linux-2.6.8.1.

Zoltan Boszormenyi =EDrta:
> Dear Aivil,
>=20
> now that Linux-2.6.8.1 is out and the fix to the most important bug,
> namely a huge memleak during audio CD/SVCD writing is found and
> Linux-2.6.8.2 is just around the corner, can we expect a ruby tree
> against them? Or as there is a long weekend in Hungary, will you accept
> my work if I volunteer?
>=20
> Best regards,
> Zolt=E1n B=F6sz=F6rm=E9nyi

I had time to make it, here is the result.
Two patches: one is against linux-2.6.8.1,
the other is against the ruby/ruby-2.6 CVS tree.

WARNING!

The FB changes were highly excessive between 2.6.7 and 2.6.8
and I don't really know the code there, so fbcon, et.al does not
even compile.

But the vgacon and dummycon consoles are working so multiple X servers
are possible. I am running it now.

I have a question: do you intend to push this thing into Linus' or
akpm's kernel? I would be happy to see this in -mm because there are
many changes again in 2.6.8.1-mm3 against the areas the Ruby tree
touches.

Best regards,
Zolt=E1n B=F6sz=F6rm=E9nyi

Ruby-2.6.8[.1/.2]?

[Zoltan B. <zb...@fr...>]

Dear Aivil,

now that Linux-2.6.8.1 is out and the fix to the most important bug,
namely a huge memleak during audio CD/SVCD writing is found and
Linux-2.6.8.2 is just around the corner, can we expect a ruby tree
against them? Or as there is a long weekend in Hungary, will you accept
my work if I volunteer?

Best regards,
Zolt=E1n B=F6sz=F6rm=E9nyi

Re: Problem compiling ruby kernel. Urgent. Please help!

[Zoltan B. <zb...@fr...>]

Hi,

dave giffin =EDrta:
> I am trying to get multiple independent framebuffer
> support. Apparently, that is only available in a patch
> to the 2.5 kernel series.
>=20
> I downloaded the source for the 2.5.59 kernel and
> checked out the contents of ruby/linux from the
> linuxconsole CVS tree.

The presently the Ruby(-2.6) tree is against 2.6.7.
Please download

http://www.kernel.org/pub/linux/kernel/v2.6/linux-2.6.7.tar.bz2

untar it somewhere and do a

(cd /where you cheched out the ruby-tree/ruby/ruby-2.6 ; tar cvf - .) \
| (cd /where you unpacked the kernel src/linux-2.6.7 | tar xf -)

Or use the attached patch that was made with the same procedure.
Now try again to configure and compile your kernel.

Best regards,
Zolt=E1n B=F6sz=F6rm=E9nyi

Problem compiling ruby kernel. Urgent. Please help!

[dave g. <giv...@ya...>]

I am trying to get multiple independent framebuffer
support. Apparently, that is only available in a patch
to the 2.5 kernel series.

I downloaded the source for the 2.5.59 kernel and
checked out the contents of ruby/linux from the
linuxconsole CVS tree.

I am compiling using gcc 3.2.2 on Redhat9 on an i686.

When I tried to compile it, I got this error (below).
I tried removing -Wall from the Makefile, but the
error remained. 


Perhaps I am compiling the wrong kernel or didn't
patch it with the correct files? 

How can I fix this compile error or compile something
else to get multiple framebuffer support?

Is it possible to get multiple framebuffer support for
a 2.6 series kernel?

:)

The Error message:
  gcc -Wp,-MD,drivers/char/.selection.o.d -D__KERNEL__
-Iinclude -Wstrict-prototypes -Wno-trigraphs -O2
-fno-strict-aliasing -fno-common -pipe
-mpreferred-stack-boundary=2 -march=athlon
-Iinclude/asm-i386/mach-default -nostdinc -iwithprefix
include    -DKBUILD_BASENAME=selection
-DKBUILD_MODNAME=selection -DEXPORT_SYMTAB  -c -o
drivers/char/selection.o drivers/char/selection.c
In file included from include/linux/vt_kern.h:11,
                 from drivers/char/selection.c:23:
include/linux/kbd_kern.h:78: warning: `struct vc_data'
declared inside parameter list
include/linux/kbd_kern.h:78: warning: its scope is
only this definition or declaration, which is probably
not what you want
include/linux/kbd_kern.h:79: warning: `struct vc_data'
declared inside parameter list
  gcc -Wp,-MD,drivers/char/.keyboard.o.d -D__KERNEL__
-Iinclude -Wstrict-prototypes -Wno-trigraphs -O2
-fno-strict-aliasing -fno-common -pipe
-mpreferred-stack-boundary=2 -march=athlon
-Iinclude/asm-i386/mach-default -nostdinc -iwithprefix
include    -DKBUILD_BASENAME=keyboard
-DKBUILD_MODNAME=keyboard -DEXPORT_SYMTAB  -c -o
drivers/char/keyboard.o drivers/char/keyboard.c
In file included from include/linux/vt_kern.h:11,
                 from drivers/char/keyboard.c:39:
include/linux/kbd_kern.h:78: warning: `struct vc_data'
declared inside parameter list
include/linux/kbd_kern.h:78: warning: its scope is
only this definition or declaration, which is probably
not what you want
include/linux/kbd_kern.h:79: warning: `struct vc_data'
declared inside parameter list
drivers/char/keyboard.c:784: conflicting types for
`getledstate'
include/linux/kbd_kern.h:78: previous declaration of
`getledstate'
drivers/char/keyboard.c: In function `getledstate':
drivers/char/keyboard.c:785: structure has no member
named `vc_ledstate'
drivers/char/keyboard.c: At top level:
drivers/char/keyboard.c:789: conflicting types for
`setledstate'
include/linux/kbd_kern.h:79: previous declaration of
`setledstate'
drivers/char/keyboard.c: In function `setledstate':
drivers/char/keyboard.c:791: structure has no member
named `vc_ledioctl'
drivers/char/keyboard.c: In function `getleds':
drivers/char/keyboard.c:816: structure has no member
named `vc_ledioctl'
drivers/char/keyboard.c: In function `kbd_bh':
drivers/char/keyboard.c:858: structure has no member
named `vc_ledstate'
drivers/char/keyboard.c:863: structure has no member
named `vc_ledstate'
make[2]: *** [drivers/char/keyboard.o] Error 1
make[1]: *** [drivers/char] Error 2
make: *** [drivers] Error 2


__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com