Re: [Dri-users] 32-bit clients on amd64 kernel?

[Alex Deucher <alexdeucher@gm...>]

On Tue, 16 Nov 2004 20:47:04 +0200, Mikko Rauhala
<mjrauhal@...> wrote:
> Hi
> 
> I was wondering what was the status of running 32-bit x86 DRI clients on
> a Linux/AMD64 kernel? I managed to find some devel mails on the issue
> from early autumn, indicating that it was being tackled, but I failed to
> come up with what's the current situation.
> 
> I was mainly wondering, after installing Ubuntu and trying out some
> stuff in a 32-bit chroot, if updating X or dri/drm modules would make
> that work.
> 

Egbert Eich wrote a patch a while back (it's in fd.o bugzilla) to
support that, however it is still being reviewed.

Alex

> Thanks.
> 
> --
> Mikko Rauhala   - mjr@...     - <URL:http://www.iki.fi/mjr/>;
> Transhumanist   - WTA member     - <URL:http://www.transhumanism.org/>;
> Singularitarian - SIAI supporter - <URL:http://www.singinst.org/>;
>

[Dri-users] 32-bit clients on amd64 kernel?

[Mikko Rauhala <mjrauhal@cc...>]

Hi

I was wondering what was the status of running 32-bit x86 DRI clients on
a Linux/AMD64 kernel? I managed to find some devel mails on the issue
from early autumn, indicating that it was being tackled, but I failed to
come up with what's the current situation.

I was mainly wondering, after installing Ubuntu and trying out some
stuff in a 32-bit chroot, if updating X or dri/drm modules would make
that work.

Thanks.

-- 
Mikko Rauhala   - mjr@...     - <URL:http://www.iki.fi/mjr/>;
Transhumanist   - WTA member     - <URL:http://www.transhumanism.org/>;
Singularitarian - SIAI supporter - <URL:http://www.singinst.org/>;

Re: [Dri-users] DRI support/reliability on older kernels

[Alex Deucher <alexdeucher@gm...>]

On Tue, 16 Nov 2004 11:01:54 -0600, Sylla, Craig <csylla@...> wrote:
> Hi -
> 
> Just looking to find out if anyone is running DRI on Redhat 7.3, kernel rev
> 2.4.18-3.
> 
> I'm trying to use it on an 855GME, and it seems to be more-or-less stable,
> but I can't get dual-head output to work.  It *does* work on 2.4.20 (RH 9.0)
> (XFree86 4.4 in both cases, should I use 4.3 instead?).
> 
> I did have to block out some of the chipset/PCI ID's because they weren't in
> the older kernel headers.  This didn't seem to otherwise affect the compile
> (agpgart).
> 
> I'm using the DRI from Intel's website atm (the 915G and down version).
> 
> Also, is there a way to get 3D acceleration on both screens in dual head
> mode?  Xinerama kills *all* 3D accel, and I get only one screen w/it
> otherwise.  (To be fair the windows drivers do the same thing - if dual head
> is on, no 3D accel at all)

In order to use dualhead with 3d someone needs to add mergedfb support
to the intel DDX.  Short term you could enable clone mode then create
virtual desktop big enough for both heads, then hack the adjustframe()
function change the offset of crtc2 and lock the offset of crtc1. 
That should give you dualhead with 3d, however, you will not have any
of the nice features of true mergedfb support.

Alex

> 
> Also, what's the best way to get the right headers for the OpenGL .so in the
> DRI?  Should I just quit messing with the Intel snapshot (which has a .so,
> but no headers) and switch to using the DRI CVS?
> 
> Thanks much for the help.
>

[Dri-users] DRI support/reliability on older kernels

[Sylla, Craig <CSylla@wm...>]

Hi -

Just looking to find out if anyone is running DRI on Redhat 7.3, kernel rev
2.4.18-3.

I'm trying to use it on an 855GME, and it seems to be more-or-less stable,
but I can't get dual-head output to work.  It *does* work on 2.4.20 (RH 9.0)
(XFree86 4.4 in both cases, should I use 4.3 instead?).

I did have to block out some of the chipset/PCI ID's because they weren't in
the older kernel headers.  This didn't seem to otherwise affect the compile
(agpgart).

I'm using the DRI from Intel's website atm (the 915G and down version).

Also, is there a way to get 3D acceleration on both screens in dual head
mode?  Xinerama kills *all* 3D accel, and I get only one screen w/it
otherwise.  (To be fair the windows drivers do the same thing - if dual head
is on, no 3D accel at all)

Also, what's the best way to get the right headers for the OpenGL .so in the
DRI?  Should I just quit messing with the Intel snapshot (which has a .so,
but no headers) and switch to using the DRI CVS?

Thanks much for the help.