Bug 1819 - Add AGP 8x support to radeon
Summary: Add AGP 8x support to radeon
Status: RESOLVED FIXED
Alias: None
Product: DRI
Classification: Unclassified
Component: General (show other bugs)
Version: XOrg git
Hardware: x86 (IA32) All
: high normal
Assignee: Default DRI bug account
QA Contact:
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2004-11-10 09:50 UTC by Alex Deucher
Modified: 2005-10-18 22:16 UTC (History)
1 user (show)

See Also:
i915 platform:
i915 features:


Attachments
/var/log/Xorg.0.log (44.56 KB, text/plain)
2004-12-14 14:33 UTC, Sérgio M. Basto
no flags Details

Description Alex Deucher 2004-11-10 09:50:48 UTC
Patch from Hui Yu and Michel Daenzer:
http://penguinppc.org/~daenzer/DRI/radeon-agp8x.diff

At some point we should just commit this patch or a variation on it.  too many
users get bitten by having 8x agp set up by the bios, but only 4x in the radeon
driver.  I seriously doubt it'll break anything.  Unfortunately my mobo only
does 4x or I'd test 8x support.  if it causees problems it can always be
reverted, or better yet, fixed.  Thoughts?
Comment 1 Alex Deucher 2004-12-11 18:02:46 UTC
this patch doesn't seem to cause any regressions on my hardware, but I don't
have an 8x motherboard to test on.  Committed to HEAD.
Comment 2 Michel Dänzer 2004-12-11 18:40:55 UTC
AFAIR Hui didn't consider the patch ready for inclusion... What do you mean by
'get bitten', anyway? I thought you could achieve AGP even without this patch
via Option "AGPMode" "2"?
Comment 3 Alex Deucher 2004-12-12 00:15:22 UTC
(In reply to comment #2)
> AFAIR Hui didn't consider the patch ready for inclusion... What do you mean by
> 'get bitten', anyway? I thought you could achieve AGP even without this patch
> via Option "AGPMode" "2"?

well, I can back it out if think it shouldn't be included.  I hadn't heard that
Hui didn't think it was ready for inclusion.  All comments I ever saw seemed to
indicate it was ok.  By 'get bitten' I mean that I've gotten many reports of 
users having problems if they have agp mode set to 8x or auto in the bios.  I'm
not sure how much of it is related to agp vs. the driver itself.  forcing 4x in
the bios usually fixed the problems.  While I realize it may not really provide
much performance-wise, it'd be nice to formally support 8x mode.  If you or
anyone else feels this patch would cause a problem or experiences any problems
let me know and I can back it out.  Part of the problem is that no one tests the
patches and most things don't shake out until they hit the trunk.
Comment 4 Dave Airlie 2004-12-12 00:31:13 UTC
You don't seemed to have commited the whole patch at the moment or did you
revert it?

the radeon_dri.c portion doesn't seem to be there....

Comment 5 Alex Deucher 2004-12-12 09:32:45 UTC
(In reply to comment #4)
> You don't seemed to have commited the whole patch at the moment or did you
> revert it?
> 
> the radeon_dri.c portion doesn't seem to be there....
> 
> 

Argh!  Sorry... fixed now.
Comment 6 Sérgio M. Basto 2004-12-13 17:39:29 UTC
I just try my Radeon 9200SE AGP 8X with cvs and with Mesa of some minutes ago. 
I compile Mesa and install radeon dri with :
cd Mesa
make linux-dri-x86
cd Mesa/lib/
cp radeon_dri.so /usr/X11R6/lib/modules/dri/radeon_dri.so

Option      "AGPMode" "4" works 
Option      "AGPMode" "8" screen go tottaly crazy on start up 

btw can't enable 
#   Option "EnablePageFlip" "true"
#   Option "AGPFastWrite" "true"
because hangs the computer.

that all for today :)
Comment 7 Alex Deucher 2004-12-13 18:09:27 UTC
(In reply to comment #6)
> I just try my Radeon 9200SE AGP 8X with cvs and with Mesa of some minutes ago. 
> I compile Mesa and install radeon dri with :
> cd Mesa
> make linux-dri-x86
> cd Mesa/lib/
> cp radeon_dri.so /usr/X11R6/lib/modules/dri/radeon_dri.so
> 
> Option      "AGPMode" "4" works 
> Option      "AGPMode" "8" screen go tottaly crazy on start up 
> 
> btw can't enable 
> #   Option "EnablePageFlip" "true"
> #   Option "AGPFastWrite" "true"
> because hangs the computer.
> 
> that all for today :)

Do you have a 8x capable motherboard?  could you post any agp or drm messages
from the kernel log when you start X with 8 set?
Comment 8 Sérgio M. Basto 2004-12-14 14:33:18 UTC
Created attachment 1546 [details]
/var/log/Xorg.0.log

(In reply to comment #7)
yes is AGP 4x/8x
Comment 9 Sérgio M. Basto 2004-12-14 15:06:47 UTC
I went to bios and have AGP setting on auto or 4x or 8x , I try to force 8x but
some results.
after some more tests Option "AGPFastWrite" "true" hangs my computer and 
Option "EnablePageFlip" "true" can be enabled but don't see much different
unless some little better performance sometimes, I will keep it on :).


Comment 10 Dave Airlie 2004-12-14 17:01:05 UTC
Tested on i865 motherboard with Radeon 9200 8x and kernel reports:

agpgart: Found an AGP 3.0 compliant device at 0000:00:00.0.
agpgart: Putting AGP V3 device at 0000:00:00.0 into 8x mode
agpgart: Putting AGP V3 device at 0000:01:00.0 into 8x mode

played ut2004 no problems.
Comment 11 Sérgio M. Basto 2004-12-19 08:35:02 UTC
dmesg | grep -i agp

Linux agpgart interface v0.100 (c) Dave Jones
agpgart: Detected VIA PT800 chipset
agpgart: Maximum main memory to use for agp memory: 439M
agpgart: AGP aperture is 128M @ 0xe0000000
agpgart: Found an AGP 3.5 compliant device at 0000:00:00.0.
agpgart: Putting AGP V3 device at 0000:00:00.0 into 4x mode
agpgart: Putting AGP V3 device at 0000:01:00.0 into 4x mode
agpgart: Found an AGP 3.5 compliant device at 0000:00:00.0.
agpgart: Putting AGP V3 device at 0000:00:00.0 into 8x mode
agpgart: Putting AGP V3 device at 0000:01:00.0 into 8x mode

4x works , 8x no 
Comment 12 Sérgio M. Basto 2004-12-19 08:46:38 UTC
Trying describe the problem with 8x, looks like bad refresh rate with monitor.
I try lower modes and lower depths with no luck. 

Modes    "1024x768" "800x600" "640x480"
Depth     16
Comment 13 Laurent GUERBY 2005-01-30 13:09:27 UTC
AGPFastWrite hard froze my system, as does AGP 8x, here
is a summary of what I got working with 
X.Org 6.8.2 RC3 RPMs for SuSE 9.2 


http://lists.suse.com/archive/suse-xfree86/2005-Jan/0106.html

My motherboard is ASUS A7N8X, here is what dmesg|grep agp said:

agpgart: X passes broken AGP3 flags (1f00420f). Fixed
agpgart: Putting AGP V3 device at 0000:00:00.0 into 8x mode
agpgart: Putting AGP V3 device at 0000:03:00.0 into 8x mode
Comment 14 Alex Deucher 2005-01-30 14:03:59 UTC
(In reply to comment #13)
> AGPFastWrite hard froze my system, as does AGP 8x, here
> is a summary of what I got working with 
> X.Org 6.8.2 RC3 RPMs for SuSE 9.2 
> 

did you apply this patch to the rpms?  6.8.x doesn't have this patch.

> 
> http://lists.suse.com/archive/suse-xfree86/2005-Jan/0106.html
> 
> My motherboard is ASUS A7N8X, here is what dmesg|grep agp said:
> 
> agpgart: X passes broken AGP3 flags (1f00420f). Fixed
> agpgart: Putting AGP V3 device at 0000:00:00.0 into 8x mode
> agpgart: Putting AGP V3 device at 0000:03:00.0 into 8x mode

fastwrites locks up many systems that's why it's disabled by default.  as I said
before did you patch the rpms?  if X and agpgart disagree about the agp mode,
that may cause problems.
Comment 15 Eric Anholt 2005-06-27 16:14:24 UTC
The patch has been committed, and remaining issues are likely agp driver issues
and not xorg, so closing this.


Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.