Dual screen framebuffer development advice

Dual screen framebuffer development advice

[Andrew Walrond]

I need to setup a machine to develop a framebuffer based application. I'd like 
to use two graphics devices; The first an ati agp chipset on the motherboard 
which will run X11 with X11's ati driver, the second a PCI card running a fb 
driver and displaying my framebuffer application.

Is this possible/sensible?

Andrew Walrond


-------------------------------------------------------
This SF.Net email is sponsored by The 2004 JavaOne(SM) Conference
Learn from the experts at JavaOne(SM), Sun's Worldwide Java Developer
Conference, June 28 - July 1 at the Moscone Center in San Francisco, CA
REGISTER AND SAVE! http://java.sun.com/javaone/sf Priority Code NWMGYKND

Re: Dual screen framebuffer development advice

[Antonino A. Daplas]

On Saturday 19 June 2004 21:55, Andrew Walrond wrote:
> I need to setup a machine to develop a framebuffer based application. I'd
> like to use two graphics devices; The first an ati agp chipset on the
> motherboard which will run X11 with X11's ati driver, the second a PCI card
> running a fb driver and displaying my framebuffer application.
>
> Is this possible/sensible?

Sure, it's possible and actually not hard to do.  Just make sure that the 
driver of your primary card is fbdev, and the secondary X11.  X happens to do 
a fine job of initializing non-primary cards whereas fbdev is mostly 
restricted to cards initialized by the firmware.

Note: You wont' be able to see both displays simultaneously, switching from 
one blanks the other.   If you do want multi-seat, see the Ruby project of 
linuxconsole.

As for sensible, that's up to you :-)

Tony




-------------------------------------------------------
This SF.Net email is sponsored by The 2004 JavaOne(SM) Conference
Learn from the experts at JavaOne(SM), Sun's Worldwide Java Developer
Conference, June 28 - July 1 at the Moscone Center in San Francisco, CA
REGISTER AND SAVE! http://java.sun.com/javaone/sf Priority Code NWMGYKND

Re: Dual screen framebuffer development advice

[Andrew Walrond]

I asked this a few months ago:

On Saturday 19 Jun 2004 21:25, Antonino A. Daplas wrote:
> On Saturday 19 June 2004 21:55, Andrew Walrond wrote:
> > I need to setup a machine to develop a framebuffer based application. I'd
> > like to use two graphics devices; The first an ati agp chipset on the
> > motherboard which will run X11 with X11's ati driver, the second a PCI
> > card running a fb driver and displaying my framebuffer application.
> >
> > Is this possible/sensible?
>
> Sure, it's possible and actually not hard to do.  Just make sure that the
> driver of your primary card is fbdev, and the secondary X11.  X happens to
> do a fine job of initializing non-primary cards whereas fbdev is mostly
> restricted to cards initialized by the firmware.
>

I've got this working nicely now. I have a recent agp nvidia card and an old 
pci matrox millenium in this dual opteron machine running 64bit linux. I boot 
into a console on the matrox using the built-in matroxfb driver, then startx 
which runs on the nvidia card using the nv driver. I can then work on my fb 
application in X, viewing the output on the second monitor via the matrox 
card. Nice!

> Note: You wont' be able to see both displays simultaneously, switching from
> one blanks the other.   If you do want multi-seat, see the Ruby project of
> linuxconsole.
>

This is wrong; I can indeed see both displays simultaneously :)

The matrox fb is pretty slow (being on a pci bus); it can only manage 5fps at 
1600x1200-16bpp. The rivafb is much faster (36fps) so I might try switching 
roles (X on matrox, fb on nvidia) at some point. I think X might accelerate 
the matrox card quite well enough for running xemacs etc.

Andrew Walrond


-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now. 
http://productguide.itmanagersjournal.com/

Re: Dual screen framebuffer development advice

[Antonino A. Daplas]

On Tuesday 23 November 2004 05:35, Andrew Walrond wrote:
> I asked this a few months ago:
>
> On Saturday 19 Jun 2004 21:25, Antonino A. Daplas wrote:
> > On Saturday 19 June 2004 21:55, Andrew Walrond wrote:
> > > I need to setup a machine to develop a framebuffer based application.
> > > I'd like to use two graphics devices; The first an ati agp chipset on
> > > the motherboard which will run X11 with X11's ati driver, the second a
> > > PCI card running a fb driver and displaying my framebuffer application.
> > >
> > > Is this possible/sensible?
> >
> > Sure, it's possible and actually not hard to do.  Just make sure that the
> > driver of your primary card is fbdev, and the secondary X11.  X happens
> > to do a fine job of initializing non-primary cards whereas fbdev is
> > mostly restricted to cards initialized by the firmware.
>
> I've got this working nicely now. I have a recent agp nvidia card and an
> old pci matrox millenium in this dual opteron machine running 64bit linux.
> I boot into a console on the matrox using the built-in matroxfb driver,
> then startx which runs on the nvidia card using the nv driver. I can then
> work on my fb application in X, viewing the output on the second monitor
> via the matrox card. Nice!
>
> > Note: You wont' be able to see both displays simultaneously, switching
> > from one blanks the other.   If you do want multi-seat, see the Ruby
> > project of linuxconsole.
>
> This is wrong; I can indeed see both displays simultaneously :)

That sounds very nice :-).  So I was wrong.  I probably should have said
that you cannot display multiple consoles simultaneously.

Tony




-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now. 
http://productguide.itmanagersjournal.com/

Re: Dual screen framebuffer development advice

[Andrew Walrond]

On Monday 22 Nov 2004 22:14, Antonino A. Daplas wrote:
> On Tuesday 23 November 2004 05:35, Andrew Walrond wrote:
> >
> > I've got this working nicely now. I have a recent agp nvidia card and an
> > old pci matrox millenium in this dual opteron machine running 64bit
> > linux. I boot into a console on the matrox using the built-in matroxfb
> > driver, then startx which runs on the nvidia card using the nv driver. I
> > can then work on my fb application in X, viewing the output on the second
> > monitor via the matrox card. Nice!
>

I've just done a bk pull from linus and build a new kernel (was using post rc2 
before) and now this nice scheme (see above) fails to work:

When I 'startx', the X log complains about no core keyboard, my usb keyboard 
stops responding and both screens go blank. I can only get access by ssh from 
another machine.

Nothing changed other than kernel; using same .config.

If I boot on rivafb console instead of matroxfb, startx and keyboard work 
fine.

Before I start hunting down the problem; Do any recent changes spring to mind?

Thanks

Andrew


-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now. 
http://productguide.itmanagersjournal.com/

Re: Dual screen framebuffer development advice

[Andrew Walrond]

On Monday 22 Nov 2004 22:14, Antonino A. Daplas wrote:
> On Tuesday 23 November 2004 05:35, Andrew Walrond wrote:
> >
> > I've got this working nicely now. I have a recent agp nvidia card and an
> > old pci matrox millenium in this dual opteron machine running 64bit
> > linux. I boot into a console on the matrox using the built-in matroxfb
> > driver, then startx which runs on the nvidia card using the nv driver. I
> > can then work on my fb application in X, viewing the output on the second
> > monitor via the matrox card. Nice!
>

I've just done a bk pull from linus and build a new kernel (was using post rc2 
before) and now this nice scheme (see above) fails to work:

When I 'startx', the X log complains about no core keyboard, my usb keyboard 
stops responding and both screens go blank. I can only get access by ssh from 
another machine.

Nothing changed other than kernel; using same .config.

If I boot on rivafb console instead of matroxfb, startx and keyboard work 
fine.

Before I start hunting down the problem; Do any recent changes spring to mind?

Thanks

Andrew


-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now. 
http://productguide.itmanagersjournal.com/

Re: Dual screen framebuffer development advice

[Antonino A. Daplas]

--- Andrew Walrond <andrew@walrond.org> wrote:

> On Monday 22 Nov 2004 22:14, Antonino A. Daplas
> wrote:
> > On Tuesday 23 November 2004 05:35, Andrew Walrond
> wrote:
> > >
> > > I've got this working nicely now. I have a
> recent agp nvidia card and an
> > > old pci matrox millenium in this dual opteron
> machine running 64bit
> > > linux. I boot into a console on the matrox using
> the built-in matroxfb
> > > driver, then startx which runs on the nvidia
> card using the nv driver. I
> > > can then work on my fb application in X, viewing
> the output on the second
> > > monitor via the matrox card. Nice!
> >
> 
> I've just done a bk pull from linus and build a new
> kernel (was using post rc2 
> before) and now this nice scheme (see above) fails
> to work:
> 
> When I 'startx', the X log complains about no core
> keyboard, my usb keyboard 
> stops responding and both screens go blank. I can
> only get access by ssh from 
> another machine.
> 
> Nothing changed other than kernel; using same
> .config.
> 
> If I boot on rivafb console instead of matroxfb,
> startx and keyboard work 
> fine.
> 
> Before I start hunting down the problem; Do any
> recent changes spring to mind?

-rc3 is bug-fix only, so no, I can't think of any changes in fbdev/fbcon
that will cause your problem, unless someone submitted changes without
passing through this list.

Anyway, here's linus' changelog between rc2 and rc3.

  o fbdev: Fix for using >16 pixel wide font in fb console*** 
  o fbdev: Support for bigger than 16x32 fonts in softcursor
  o fbdev: Support for bigger than 16x32 fonts in rivafb cursor
  o fbcon: Disable fbcon cursor if vt softcursor is enabled***
  o fbdev: Allow mode change even if EDID block is not found
  o fbdev: Fix cursor in doublescan mode in atyfb
  o fbdev: Fix typo in atyfb
  o fbdev: Change the find_mode behavior***
  o rivafb: fix broken burst length calculation
  o fbdev: Fix screen corruption in neofb
  o fbdev: Fix lockup when switching to/from X/console
  o fbdev: Fix module_param in rivafb
  o fbdev: Fix crash if fb_set_var() called before register_framebuffer()***
  o fbdev: fix wrong colors at 16 bpp in tridentfb

*** affects core fbcon/fbdev layer

Tony




-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now. 
http://productguide.itmanagersjournal.com/