In wich Kerneltree will this implented ?
2.4.x or 2.5.x ?
Ati-Drivers will not install or Run on 2.5.70 (its clear ;) )
and 2.4.20 and 2.4.21-pre7
At the End of the XFree Log will come this (2.4.21-pre7) :
(II) fglrx(0): Desc: ATI Fire GL DRM kernel module
(II) fglrx(0): Kernel Module version matches driver.
(II) fglrx(0): Kernel Module Build Time Information:
(II) fglrx(0): Build-Kernel UTS_RELEASE: 2.4.20-9
(II) fglrx(0): Build-Kernel MODVERSIONS: yes
(II) fglrx(0): Build-Kernel __SMP__: no
(II) fglrx(0): Build-Kernel PAGE_SIZE: 0x1000
(II) fglrx(0): [drm] register handle = 0xe9000000
(EE) fglrx(0): [agp] unable to acquire AGP, error "xf86_ENODEV"
(EE) fglrx(0): cannot init AGP
The init will not or can not detect the AGP-Adapture Size.
Regards
Gregor Essers
PS: System is
AMD XP 2000+
Radeon 9700Pro
1 GB Ram
Redhat 9
Epox 8K9A2 Mainboard (KT400)
> In wich Kerneltree will this implented ?
> 2.4.x or 2.5.x ?
>
> Ati-Drivers will not install or Run on 2.5.70 (its clear ;) )
> and 2.4.20 and 2.4.21-pre7
I've had this problem as well.
What I've been able to do is to use a backport for one of the 2.4.21-pre*
series, and move the code forward to the current 2.4.21-rc's .
Here's info on the relevant patch:
http://lists.insecure.org/lists/linux-kernel/2003/Mar/3999.html
The Radeon 9700 Pro now functions with the ATI binary-only drivers, BUT
only with 2D acceleration. In other words, I have to set the "DisableDRI"
option in XF86Config-4 to yes.
So I get no 3D acceleration, but at least I get 2D acceleration and don't
have to run it in framebuffer mode like before.
I wish ATI would either open-source their drivers or come out with a fix
ASAP. :-/
-----
James Sellman -- ISU CoE-CS/ISLUG Linux Lab Admin |"Lum, did you just see
----------------------------------------------------| a hentai rabbit flying
[email protected] | // A4000/604e/60 128M| through the air?"
[email protected] | \X/ A500/20 3M | - Miyake Shinobu
On Wed, Jun 11, 2003 at 03:14:02AM +0200, Gregor Essers wrote:
> In wich Kerneltree will this implented ?
> 2.4.x or 2.5.x ?
2.5
I've not had the time to do a 2.4 backport. Several other folks
have tried, though as the 2.5 code is still constantly moving,
they tend to fall behind.
> Ati-Drivers will not install or Run on 2.5.70 (its clear ;) )
> and 2.4.20 and 2.4.21-pre7
Sadly, there are no fully open drivers for any AGP x8 cards still.
I'm still hoping this will change over time.
Dave
On Wed, Jun 11, 2003 at 02:28:24AM -0600, I Am Falling I Am Fading wrote:
> I've had this problem as well.
>
> What I've been able to do is to use a backport for one of the 2.4.21-pre*
> series, and move the code forward to the current 2.4.21-rc's .
That's not a proper fix. The agp code in 2.4.21pre supports the KT400
only in AGP2.0 mode. When you put an AGP3.0 (x8) card in the slot,
the chipset configures itself into AGP3 mode, and registers change
meaning.
> Here's info on the relevant patch:
> http://lists.insecure.org/lists/linux-kernel/2003/Mar/3999.html
Very, very dated now. Many fixes have gone into the agp code since
2.5.64, on which that backport is based.
Dave
Actually I believe the new ati 9200 cards support agp 8x and they work
with the current DRI r200 driver.
Alex
---------------------
> Ati-Drivers will not install or Run on 2.5.70 (its clear ;) )
> and 2.4.20 and 2.4.21-pre7
Sadly, there are no fully open drivers for any AGP x8 cards still.
I'm still hoping this will change over time.
Dave
__________________________________
Do you Yahoo!?
Yahoo! Calendar - Free online calendar with sync to Outlook(TM).
http://calendar.yahoo.com
On Wed, Jun 11, 2003 at 08:06:55AM -0700, Alex Deucher wrote:
> Actually I believe the new ati 9200 cards support agp 8x and they work
> with the current DRI r200 driver.
Excellent. It may be worthwhile anyone who buys one of these
actually letting ATI know the reason for doing so was the
fact that it's the only AGP x8 card supported by _fully_ open drivers.
Dave
Although I don't know that anyone has actually tried it in 8x mode.
does anyone know if the radeon driver even has an AGPMode "8x" option?
I haven't looked myself...
Alex
--- Dave Jones <[email protected]> wrote:
> On Wed, Jun 11, 2003 at 08:06:55AM -0700, Alex Deucher wrote:
> > Actually I believe the new ati 9200 cards support agp 8x and they
> work
> > with the current DRI r200 driver.
>
> Excellent. It may be worthwhile anyone who buys one of these
> actually letting ATI know the reason for doing so was the
> fact that it's the only AGP x8 card supported by _fully_ open
> drivers.
>
> Dave
>
__________________________________
Do you Yahoo!?
Yahoo! Calendar - Free online calendar with sync to Outlook(TM).
http://calendar.yahoo.com
Dana sreda 11. jun 2003. 23:26, Alex Deucher je napisao/la:
> Although I don't know that anyone has actually tried it in 8x mode.
> does anyone know if the radeon driver even has an AGPMode "8x" option?
> I haven't looked myself...
Nope, 4x max.
--
Pozdrav,
Tanasković Toplica
On Thu, Jun 12, 2003 at 12:36:21AM +0200, Toplica Tanaskovi?? wrote:
> Dana sreda 11. jun 2003. 23:26, Alex Deucher je napisao/la:
> > Although I don't know that anyone has actually tried it in 8x mode.
> > does anyone know if the radeon driver even has an AGPMode "8x" option?
> > I haven't looked myself...
>
> Nope, 4x max.
That's likely an X limitation. Someone with X-fu needs to hack that up
so it passes the right things through to agpgart. Would be nice to have
that in place for the next X release, so that when distros come to start
shipping 2.6, userspace is up to speed.
Dave
Dana četvrtak 12. jun 2003. 00:53, Dave Jones je napisao/la:
> That's likely an X limitation. Someone with X-fu needs to hack that up
> so it passes the right things through to agpgart. Would be nice to have
> that in place for the next X release, so that when distros come to start
> shipping 2.6, userspace is up to speed.
It's going to be difficult one, because, there are no hw. acc. drivers for
nVidia built in X, and there is no R300 accel. Don't know about nv, but for
Radeon chips only R300 can engage 8x transfer, so until ATI releases docs, or
some genius figures how R300 work in 3D, IMHO there is no point of hacking X.
--
Pozdrav,
Tanasković Toplica
On Wed, 11 Jun 2003, Dave Jones wrote:
> On Wed, Jun 11, 2003 at 02:28:24AM -0600, I Am Falling I Am Fading wrote:
>
> > I've had this problem as well.
> >
> > What I've been able to do is to use a backport for one of the 2.4.21-pre*
> > series, and move the code forward to the current 2.4.21-rc's .
>
> That's not a proper fix. The agp code in 2.4.21pre supports the KT400
> only in AGP2.0 mode. When you put an AGP3.0 (x8) card in the slot,
> the chipset configures itself into AGP3 mode, and registers change
> meaning.
>
> > Here's info on the relevant patch:
> > http://lists.insecure.org/lists/linux-kernel/2003/Mar/3999.html
>
> Very, very dated now. Many fixes have gone into the agp code since
> 2.5.64, on which that backport is based.
I'm aware that it doesn't work right, unfortunately getting AGP3.0 working
at all under the 2.4 kernel doesn't have any real solution, with the
exception of nVidia's binary-only drivers for the nForce2.
This solution at least initializes some stuff and allows me to use the ATI
drivers for 2D acceleration on my KT400-based board, whcih didn't work
before I applied that patch. Consequently, it's the best solution I have
now for people who want to use their R3xx-series ATI cards on AGP3.0
systems.
The only other solution is to kick your card down into AGP 2.0 mode, which
most BIOSes do not allow you to do in software. Instead what you have to
do is cut/unsolder traces on your video card for the pins used for AGP 3.0
detection. This is a near-permanent and horrible solution but it does get
everything working. :-/
Dave, if you have a better solution for the 2.4 kernel I'd be very, very
happy to see it. :-( Unfortunately I'm not a good enough coder to write
proper AGP3.0 support for the 2.4 kernel myself, and can only shuffle
other people's code around.
Someday 2.6 is going to fix all this but it's just not ready yet, and
while it's fun to run a 2.5 kernel it's not something I would recommend to
someone else (aside from this, it wouldn't fix his situation since the ATI
binary-only drivers puke on 2.5).
I despair at the DRI project ever getting back in gear -- it's in horrible
disarray and the development list is now 75% spam. Some good improvements
came thanks to a donation from the Weather Channel, but it's not enough to
provide comprehensive support. :-( Maybe there needs to be a "save DRI"
fund. :-/
-----
James Sellman -- ISU CoE-CS/ISLUG Linux Lab Admin |"Lum, did you just see
----------------------------------------------------| a hentai rabbit flying
[email protected] | // A4000/604e/60 128M| through the air?"
[email protected] | \X/ A500/20 3M | - Miyake Shinobu
On Thu, Jun 12, 2003 at 02:44:45AM +0200, Toplica Tanaskovi?? wrote:
> Dana ??etvrtak 12. jun 2003. 00:53, Dave Jones je napisao/la:
> > That's likely an X limitation. Someone with X-fu needs to hack that up
> > so it passes the right things through to agpgart. Would be nice to have
> > that in place for the next X release, so that when distros come to start
> > shipping 2.6, userspace is up to speed.
>
> It's going to be difficult one, because, there are no hw. acc. drivers for
> nVidia built in X, and there is no R300 accel. Don't know about nv, but for
> Radeon chips only R300 can engage 8x transfer, so until ATI releases docs, or
> some genius figures how R300 work in 3D, IMHO there is no point of hacking X.
The 9200 is an R200 core (actually RV280) with AGP x8 support.
Dave
> The only other solution is to kick your card down into AGP 2.0 mode, which
> most BIOSes do not allow you to do in software. Instead what you have to
> do is cut/unsolder traces on your video card for the pins used for AGP 3.0
> detection. This is a near-permanent and horrible solution but it does get
> everything working. :-/
Insulating tape on certain pins works on ISA cards, but whether it would be
practical on the smaller pins of an AGP card, I'm not sure.
John.
On Thu, 12 Jun 2003, John Bradford wrote:
> > The only other solution is to kick your card down into AGP 2.0 mode, which
> > most BIOSes do not allow you to do in software. Instead what you have to
> > do is cut/unsolder traces on your video card for the pins used for AGP 3.0
> > detection. This is a near-permanent and horrible solution but it does get
> > everything working. :-/
>
> Insulating tape on certain pins works on ISA cards, but whether it would be
> practical on the smaller pins of an AGP card, I'm not sure.
Tried it already... The pins are too small to get adequate purchase for
the tape -- the friction just causes it to slide around in the slot and
gets goo around.
Superglue might be a better solution....
...but I think the solder method is better.
On the Radeon 9700 Pro at least there are a couple jumpers on the
appropriate pins, bridged by 0-ohm surface mount resistors (i.e. simple
conductors). What you can do is just unsolder the bridges and it becomes
an AGP 2.0 card... If you have a very steady hand you can also resolder
them to get your AGP 3.0 back.
Still this is not a fun solution as you can potentially cook your card
(make sure to use a 15 watt iron, nothing higher).
-----
James Sellman -- ISU CoE-CS/ISLUG Linux Lab Admin |"Lum, did you just see
----------------------------------------------------| a hentai rabbit flying
[email protected] | // A4000/604e/60 128M| through the air?"
[email protected] | \X/ A500/20 3M | - Miyake Shinobu
Hi i will look into that with the bridges, i hope that Hercules is so that
they give me the spec?s (Plan of the Card) with the Jupers/Bridges for AGP
2.0.
The minus on Performace is not great, in my eyes.
It?s very SAD that Ati and Nvidia will not give the Specs or an Sourcecode
of the Drivers :/.
Regards
Gregor Essers
----- Original Message -----
From: "I Am Falling I Am Fading" <[email protected]>
To: "John Bradford" <[email protected]>
Cc: <[email protected]>; <[email protected]>;
<[email protected]>
Sent: Thursday, June 12, 2003 1:15 PM
Subject: Re: Via KT400 and AGP 8x Support
> On Thu, 12 Jun 2003, John Bradford wrote:
>
> > > The only other solution is to kick your card down into AGP 2.0 mode,
which
> > > most BIOSes do not allow you to do in software. Instead what you have
to
> > > do is cut/unsolder traces on your video card for the pins used for AGP
3.0
> > > detection. This is a near-permanent and horrible solution but it does
get
> > > everything working. :-/
> >
> > Insulating tape on certain pins works on ISA cards, but whether it would
be
> > practical on the smaller pins of an AGP card, I'm not sure.
>
> Tried it already... The pins are too small to get adequate purchase for
> the tape -- the friction just causes it to slide around in the slot and
> gets goo around.
>
> Superglue might be a better solution....
>
> ...but I think the solder method is better.
>
> On the Radeon 9700 Pro at least there are a couple jumpers on the
> appropriate pins, bridged by 0-ohm surface mount resistors (i.e. simple
> conductors). What you can do is just unsolder the bridges and it becomes
> an AGP 2.0 card... If you have a very steady hand you can also resolder
> them to get your AGP 3.0 back.
>
> Still this is not a fun solution as you can potentially cook your card
> (make sure to use a 15 watt iron, nothing higher).
>
> -----
> James Sellman -- ISU CoE-CS/ISLUG Linux Lab Admin |"Lum, did you just
see
> ----------------------------------------------------| a hentai rabbit
flying
> [email protected] | // A4000/604e/60 128M| through the air?"
> [email protected] | \X/ A500/20 3M | - Miyake Shinobu
>
>
On Wed, 11 Jun 2003, Gregor Essers wrote:
> Hi i will look into that with the bridges, i hope that Hercules is so that
> they give me the spec?s (Plan of the Card) with the Jupers/Bridges for AGP
> 2.0.
Follow the traces from pins A3 and A11 on the edge connector, they'll lead
you to the 0 ohm resistor jumpers (Just look for the little rectangular
things that have a 0 printed on them where others have numbers).
Make sure they are the ones for A3 and A11, and remove 'em. Voila, an AGP
2.0 card.
(These are on the ATI 9700 reference design, which all manufacturers have
copied so far, your Hercules will probably have them too, although it's
a 9800 I'm assuming)
> It?s very SAD that Ati and Nvidia will not give the Specs or an Sourcecode
> of the Drivers :/.
Agreed. :-(
-----
James Sellman -- ISU CoE-CS/ISLUG Linux Lab Admin |"Lum, did you just see
----------------------------------------------------| a hentai rabbit flying
[email protected] | // A4000/604e/60 128M| through the air?"
[email protected] | \X/ A500/20 3M | - Miyake Shinobu
On Thu, Jun 12, 2003 at 05:15:21AM -0600, I Am Falling I Am Fading wrote:
> > Insulating tape on certain pins works on ISA cards, but whether it would be
> > practical on the smaller pins of an AGP card, I'm not sure.
>
> Tried it already... The pins are too small to get adequate purchase for
> the tape -- the friction just causes it to slide around in the slot and
> gets goo around.
>
> Superglue might be a better solution....
> ...but I think the solder method is better.
So rather than experiment with backporting the 2.5 code to 2.4,
you'd rather risk damaging your hardware ?
I think this way is madness.
Dave
On Thu, 12 Jun 2003, Dave Jones wrote:
> > Tried it already... The pins are too small to get adequate purchase for
> > the tape -- the friction just causes it to slide around in the slot and
> > gets goo around.
> >
> > Superglue might be a better solution....
> > ...but I think the solder method is better.
>
> So rather than experiment with backporting the 2.5 code to 2.4,
> you'd rather risk damaging your hardware ?
>
> I think this way is madness.
Unfortunately even a perfect backport seems to be only a partial solution
-- the ATI binary only drivers don't seem to know how to talk to the 2.5
AGP 3.0 stuff anyway (well, at least they didn't work at all when I tried
them under the 2.5 kernel :-/), and as they are lame binary-only drivers
there is no way to fix that.
There are also no other drivers for the R300-series Radeon GPUs. :-(
This absolutely sucks, but turning the card into an AGP 2.0 card seems to
be the only surefire way to get it to work properly under Linux. :-(
-----
James Sellman -- ISU CoE-CS/ISLUG Linux Lab Admin |"Lum, did you just see
----------------------------------------------------| a hentai rabbit flying
[email protected] | // A4000/604e/60 128M| through the air?"
[email protected] | \X/ A500/20 3M | - Miyake Shinobu
here i have a tip for Slackware Users with ATI Cards (9700pro).
http://www.n3t.net/Infos/Slackware-ATI-Radeon9700Pro.shtml
Regards
Gregor Essers
----- Original Message -----
From: "Gregor Essers" <[email protected]>
To: "I Am Falling I Am Fading" <[email protected]>
Cc: <[email protected]>; <[email protected]>
Sent: Wednesday, June 11, 2003 2:19 PM
Subject: Re: Via KT400 and AGP 8x Support
Hi i will look into that with the bridges, i hope that Hercules is so that
they give me the spec?s (Plan of the Card) with the Jupers/Bridges for AGP
2.0.
The minus on Performace is not great, in my eyes.
It?s very SAD that Ati and Nvidia will not give the Specs or an Sourcecode
of the Drivers :/.
Regards
Gregor Essers
----- Original Message -----
From: "I Am Falling I Am Fading" <[email protected]>
To: "John Bradford" <[email protected]>
Cc: <[email protected]>; <[email protected]>;
<[email protected]>
Sent: Thursday, June 12, 2003 1:15 PM
Subject: Re: Via KT400 and AGP 8x Support
> On Thu, 12 Jun 2003, John Bradford wrote:
>
> > > The only other solution is to kick your card down into AGP 2.0 mode,
which
> > > most BIOSes do not allow you to do in software. Instead what you have
to
> > > do is cut/unsolder traces on your video card for the pins used for AGP
3.0
> > > detection. This is a near-permanent and horrible solution but it does
get
> > > everything working. :-/
> >
> > Insulating tape on certain pins works on ISA cards, but whether it would
be
> > practical on the smaller pins of an AGP card, I'm not sure.
>
> Tried it already... The pins are too small to get adequate purchase for
> the tape -- the friction just causes it to slide around in the slot and
> gets goo around.
>
> Superglue might be a better solution....
>
> ...but I think the solder method is better.
>
> On the Radeon 9700 Pro at least there are a couple jumpers on the
> appropriate pins, bridged by 0-ohm surface mount resistors (i.e. simple
> conductors). What you can do is just unsolder the bridges and it becomes
> an AGP 2.0 card... If you have a very steady hand you can also resolder
> them to get your AGP 3.0 back.
>
> Still this is not a fun solution as you can potentially cook your card
> (make sure to use a 15 watt iron, nothing higher).
>
> -----
> James Sellman -- ISU CoE-CS/ISLUG Linux Lab Admin |"Lum, did you just
see
> ----------------------------------------------------| a hentai rabbit
flying
> [email protected] | // A4000/604e/60 128M| through the air?"
> [email protected] | \X/ A500/20 3M | - Miyake Shinobu
>
>
I have to disagree here. The DRI project is still very much alive.
Over the last few months lots of OpenGL extensions have been added.
Just recently most drivers got YUV texture and NV texture rectangle
support. A new configuration system is in the works and progress
continues on mach64, savage, and trident support. The reason we
haven't seen support for lots of new chips is that no one is funding
development! It is a volunteer group! It takes a lot of work to write
a new driver. There are docs available for quite a few chips, but
someone has to write the driver.
Alex
-------------------------
I despair at the DRI project ever getting back in gear -- it's in
horrible
disarray and the development list is now 75% spam. Some good
improvements
came thanks to a donation from the Weather Channel, but it's not enough
to
provide comprehensive support. :-( Maybe there needs to be a "save DRI"
fund. :-/
__________________________________
Do you Yahoo!?
Yahoo! Calendar - Free online calendar with sync to Outlook(TM).
http://calendar.yahoo.com
On Wed, Jun 11, 2003 at 03:18:51PM +0200, Gregor Essers wrote:
> here i have a tip for Slackware Users with ATI Cards (9700pro).
>
> http://www.n3t.net/Infos/Slackware-ATI-Radeon9700Pro.shtml
That's nice, the latest ATI driver contains 98% of the KT400 AGP
driver I wrote for 2.5. The remainder being 1% copyright/credit
notices that they didn't need, and the other 1% being some breakage
to make sure it can't possibly work in AGP 3 mode.
What a train wreck.
I would look deeper into the source, but I'd rather not take a second
look at my lunch right now.
Go ATI!
Dave
On Thursday 12 June 2003 07:36, I Am Falling I Am Fading wrote:
> On Thu, 12 Jun 2003, Dave Jones wrote:
> > > Tried it already... The pins are too small to get adequate purchase
> > > for the tape -- the friction just causes it to slide around in the
> > > slot and gets goo around.
> > >
> > > Superglue might be a better solution....
> > > ...but I think the solder method is better.
> >
> > So rather than experiment with backporting the 2.5 code to 2.4,
> > you'd rather risk damaging your hardware ?
> >
> > I think this way is madness.
>
> Unfortunately even a perfect backport seems to be only a partial solution
> -- the ATI binary only drivers don't seem to know how to talk to the 2.5
> AGP 3.0 stuff anyway (well, at least they didn't work at all when I tried
> them under the 2.5 kernel :-/), and as they are lame binary-only drivers
> there is no way to fix that.
>
> There are also no other drivers for the R300-series Radeon GPUs. :-(
>
> This absolutely sucks, but turning the card into an AGP 2.0 card seems to
> be the only surefire way to get it to work properly under Linux. :-(
I'm not sure this will help the hardware situation, but you could try a
bus extender (it will make the board stick out of the slot a couple of
inches).
It should allow you the option of either cutting the extender wire traces
or pull jumpers to see if things do work.