linuxconsole-dev

Re: New fbcon attempt with test11bk9 & cvs ruby

[Svetoslav S. <sv...@gm...>]

> I compiled a new kernel today.  2.6.0-test11-bk9 with
> today's cvs ruby-2.6
> 
> I first tried a monolithic kernel with G550 framebuffer and fbcon.
> Unlike previous fbcon attempts this didn't oops!
> I got all the normal kernel initialization output,
> then init took over and ran the usual scripts.
> Everything was fine until it attempted to start xdm.
> Nothing happened after that.  I also didn't get any gettys
> so no opportunities to log in and no console switching.
> 
> I also made a kernel without fbcon.  The same thing happened,
> it booted (this time with no console output after the framebuffer
> driver loaded) and got stuck later.  X didn't start.
> 
> My guess is some sort of Vt problem, as that prevents both
> X and getty from starting.  I also noticed that the cvs update
> updated vt.c

have you tried booting with only vga console built in -- no fbdrivers ?

i haven't tried latest cvs, but latest cvs with the attached patch reversed
/* a bit older cvs */ does work for me in case it's a modular build 
both fbcon and matroxfb 

best,

svetljo

-- 
+++ GMX - die erste Adresse für Mail, Message, More +++
Neu: Preissenkung für MMS und FreeMMS! http://www.gmx.net

Re: New fbcon attempt with test11bk9 & cvs ruby

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

>> > I first tried a monolithic kernel with G550 framebuffer and fbcon.
>> > Unlike previous fbcon attempts this didn't oops!
>> > I got all the normal kernel initialization output,
>> > then init took over and ran the usual scripts.
>> > Everything was fine until it attempted to start xdm.
>> > Nothing happened after that.  I also didn't get any gettys
>> > so no opportunities to log in and no console switching.
>> >
>> > I also made a kernel without fbcon.  The same thing happened,
>> > it booted (this time with no console output after the framebuffer
>> > driver loaded) and got stuck later.  X didn't start.
>

Today 2.6.0-ruby reach public CVS.
With this version of ruby Your tests results should become
good replayable. Good or bad results - who knows.
Scientific touch needs for replayable results.

Aivils Stoss

fbcon & ruby-2.6.0 works better

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

On Wed, Dec 24, 2003 at 10:16:16AM +0200, Aivils Stoss wrote:
> 
> Today 2.6.0-ruby reach public CVS.
> With this version of ruby Your tests results should become
> good replayable. Good or bad results - who knows.
> Scientific touch needs for replayable results.
> 
2.6.0+ruby works better, but not yet perfect.

It boots with fbcon without crashing now.
I haven't yet succeeded in running two xservers or
switch between X and console yet though.  

I don't yet know wether it is configuration problems
or something else. Starting X on vt17 (the other keyboard)
just goes into D-state. openvt on vt17 also fails.

So far I have succeeded starting up in single-user mode
with framebuffer and fbcon compiled in.  I can then
use openvt and switch between several framebuffer consoles
on /dev/fb0.  

/dev/fb1 works, (I can cat a file into it and get something displayed)
but there don't seem to be a console running there.  The
second keyboard doesn't seem to go anywhere.  It can be used
for sysrq sequences and nothing else.

X runs fine using vt7 and the mga driver.  Framebuffer X fails, probably
because of vt17 trouble.

Helge Hafting

Re: New fbcon attempt with test11bk9 & cvs ruby

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

On Mon, Dec 15, 2003 at 06:50:21PM +0100, Svetoslav Slavtchev wrote:
> > I compiled a new kernel today.  2.6.0-test11-bk9 with
> > today's cvs ruby-2.6
> > 
> > I first tried a monolithic kernel with G550 framebuffer and fbcon.
> > Unlike previous fbcon attempts this didn't oops!
> > I got all the normal kernel initialization output,
> > then init took over and ran the usual scripts.
> > Everything was fine until it attempted to start xdm.
> > Nothing happened after that.  I also didn't get any gettys
> > so no opportunities to log in and no console switching.
> > 
> > I also made a kernel without fbcon.  The same thing happened,
> > it booted (this time with no console output after the framebuffer
> > driver loaded) and got stuck later.  X didn't start.
> > 
> > My guess is some sort of Vt problem, as that prevents both
> > X and getty from starting.  I also noticed that the cvs update
> > updated vt.c
> 
> have you tried booting with only vga console built in -- no fbdrivers ?
> 
I tried 2.6.0 with ruby, fbcon and matroxfb as modules.
It booted, and vgacon came up as usual.  It ended with
"starting xdm" then nothing more happened.  No X, and
no login prompt.  
I could type on the vga console with the wrong keyboard, :-(
but there were no getty.

Helge Hafting