I use a mostly stable amd64 box (main unstable packages are KDE 4.3.1 and some scientific packages), including the amd64 stable gentoo-sources kernel.
I own a laptop with 4GB RAM and an ATI HD4570/512MB.
I can't use the recently stable ati-drivers-9.9-r2, because I get a black screen, but I can still login through ssh. However, ati-drivers-9.8 work.
I have attached the xorg.0.log for both ati-drivers versions. You can see the device identifier is not the same (I don't know if that's relevant). Also, the log for the 9.9-r2 version is much shorter: after the following line
(II) fglrx(0): Using adapter: 2:0.0.
the log for the 9.9 version is empty!!
What can I do? I'm worried because I know I won't be able to use the 9.8 version with kernel >=2.6.31
Steps to Reproduce:
1. emerge the stable amd64 ati-drivers
The drivers work
My emerge --info:
Portage 220.127.116.11 (default/linux/amd64/2008.0/desktop, gcc-4.3.2, glibc-2.9_p20081201-r2, 2.6.30-gentoo-r5 x86_64)
System uname: Linux-2.6.30-gentoo-r5-x86_64-AMD_Athlon-tm-_X2_Dual-Core_QL-65-with-gentoo-18.104.22.168
Timestamp of tree: Sun, 04 Oct 2009 20:30:01 +0000
sys-devel/autoconf: 2.13, 2.63-r1
sys-devel/automake: 1.7.9-r1, 1.8.5-r3, 1.9.6-r2, 1.10.2
CFLAGS="-march=athlon64-sse3 -mtune=athlon64-sse3 -O2 -pipe"
CONFIG_PROTECT="/etc /usr/share/config /var/lib/hsqldb"
CONFIG_PROTECT_MASK="/etc/ca-certificates.conf /etc/env.d /etc/env.d/java/ /etc/fonts/fonts.conf /etc/gconf /etc/revdep-rebuild /etc/sandbox.d /etc/terminfo /etc/texmf/language.dat.d /etc/texmf/language.def.d /etc/texmf/updmap.d /etc/texmf/web2c /etc/udev/rules.d"
CXXFLAGS="-march=athlon64-sse3 -mtune=athlon64-sse3 -O2 -pipe"
FEATURES="distlocks fixpackages parallel-fetch protect-owned sandbox sfperms strict unmerge-orphans userfetch"
LINGUAS="es_ES es en_US en"
PORTAGE_RSYNC_OPTS="--recursive --links --safe-links --perms --times --compress --force --whole-file --delete --stats --timeout=180 --exclude=/distfiles --exclude=/local --exclude=/packages"
PORTDIR_OVERLAY="/usr/portage/local /usr/local/portage/layman/science /usr/local/portage/layman/roslin /usr/local/portage/layman/sunrise /usr/local/portage/layman/pro-audio /usr/portage"
Unset: CPPFLAGS, CTARGET, EMERGE_DEFAULT_OPTS, FFLAGS, INSTALL_MASK, LC_ALL, PORTAGE_COMPRESS, PORTAGE_COMPRESS_FLAGS, PORTAGE_RSYNC_EXTRA_OPTS
Created attachment 206070 [details]
xorg.0.log for ati-drivers-9.8
Created attachment 206071 [details]
xorg.0.log for ati-drivers-9.9-r2
I forgot. This is the relevant part in xorg.conf:
Option "XAANoOffscreenPixmaps" "true"
The BusID line was added by "aticonfig --initial".
PLease attach the complete xorg.conf
Check wether it is a duplicate off bug 282256 and the solution from there fixes it.
Created attachment 206691 [details]
(In reply to comment #5)
> Check wether it is a duplicate off bug 282256 and the solution from there fixes
No, it isn't. The first thing I did was to check every open or fixed bug related to ati-drivers.
Thanks for your help.
Nothing yet? Am I the only one suffering from this?
I have 4GB RAM, and I read somew time ago there were some problems with 4GB RAM and ATI drivers (versions lower than 9.9). Some people decreased RAM to 2 GB and everything worked again. Could this be he cause?
Or maybe a udev problem? I use the stable 141.
New version 9.10 solves the problem. May I suggest it to go stable? 9.9 version was a headache for too many people.
(In reply to comment #9)
> New version 9.10 solves the problem. May I suggest it to go stable? 9.9 version
> was a headache for too many people.
9.10 still doesn't work for me. But i'm not sure if my problem is the same.
I submitted the bug:
(In reply to comment #10)
> 9.10 still doesn't work for me. But i'm not sure if my problem is the same.
I can imagine how desperate you are, because I was. Please tell me if I can be of any help. Although I'm just a regular user, I can provide you with the .config for my kernel or any other config file you may ask for, the list of my installed packages, my hardware specs... Just anything that could be useful for comparison purposes between your box and mine.
I think I posted all the configs to the bug report there.
HW spec is:
- Radeon HD 4850
- Athlon X2 3800+
- 6GiB RAM
- Asus mother board wih nvidia chipset
And I have xinerama (which might be the problem).
I also used my brothers computer as a comparison, and couldn't pint point the problem. I would also theorize that memory corruption is not connected to the ati-drivers.
I think that memory corruption occurs on hard-shutdown of the machine. But I don't know what causes it.
When I got to the log (uncorrupted :S), I got the same output like you. Stopped at
(II) fglrx(0): Using adapter: 3:0.0.
(In reply to comment #11)
> (In reply to comment #10)
> > 9.10 still doesn't work for me. But i'm not sure if my problem is the same.
> I can imagine how desperate you are, because I was. Please tell me if I can be
> of any help. Although I'm just a regular user, I can provide you with the
> .config for my kernel or any other config file you may ask for, the list of my
> installed packages, my hardware specs... Just anything that could be useful for
> comparison purposes between your box and mine.
I can only ask one question:
Do you really need ati-drivers?
Because basicaly xf86-video-ati should work for anyone and i can give some help with these setups, but i am seriously just commiter&dependency checker for the ati-drivers, nothing more...
Also when there is 2.6.32 kernel 3d will be availible for all ati cards.
I think I'll wait a little while for open source 3D support to mature a little bit (not going to change my graphics card any time soon).
But I would really like to solve this problem :) (especially the memory corruption thingy), since I beleive it's unconnected to the ati drivers.
Some time later I'll try to hard-shutdown my machine to look for memory corruptions so I could confirm my suspicions.
(In reply to comment #13)
> I can only ask one question:
> Do you really need ati-drivers?
> Because basicaly xf86-video-ati should work for anyone and i can give some help
> with these setups, but i am seriously just commiter&dependency checker for the
> ati-drivers, nothing more...
> Also when there is 2.6.32 kernel 3d will be availible for all ati cards.
(In reply to comment #14)
> I think I'll wait a little while for open source 3D support to mature a little
> bit (not going to change my graphics card any time soon).
The sooner you try the OSS drivers, the sooner you can report bugs and the sooner upstream can fix them.
I've been maintaining Intel drivers for more than a year now, and this is the _exact_ pattern that I've seen over and over and over.
Please try those drivers and help upstream make them better faster. It might not be perfect at first, you might get bugs and maybe crashes, but it's _worth_ the trouble.
AMD's binary drivers are a plague. Even AMD seems to have stopped caring about them. They regularly update their public docs and have hired a few X devs to work on the OSS drivers. That and there's absolutely no support from AMD on the binary drivers, we don't have any way of reporting bugs to them.
> But I would really like to solve this problem :) (especially the memory
> corruption thingy), since I beleive it's unconnected to the ati drivers.
There's a track record for binary drivers here. Most of them are just awful at memory management. Even the free ones have issues, but they get fixed really quickly.
Again, I can't stress how important it is to try the free drivers and help improve them.
Tomas, please close this bug CANTFIX since there's nothing we can do.
Right, thats true resolution. We sadly cant help on this matter :/ try forums to get some support from other users, but really try the OSS drivers.
Ok.. I'll try to use the OSS drivers. How is the current state of the r700 3D support? Have to use the code from git? And also, is this wiki up to date?
(In reply to comment #17)
> Ok.. I'll try to use the OSS drivers. How is the current state of the r700 3D
> support? Have to use the code from git? And also, is this wiki up to date?
Eew, most of this is either useless or wrong.
Just add "radeon" to VIDEO_CARDS and build the radeon DRM module in your kernel and you should be fine.
(In reply to comment #18)
> Eew, most of this is either useless or wrong.
> Just add "radeon" to VIDEO_CARDS and build the radeon DRM module in your kernel
> and you should be fine.
I'm kinda confused with radeon and radeonhd names. What are the differences?
2 different drivers. "radeon" installs xf86-video-ati which is the historical ATI OSS driver. "radeonhd" installs xf86-video-radeonhd which is a newer driver meant only for HD-series chips.
For a long time, radeonhd was the only one to support newer cards. But that's no longer the case and xf86-video-ati is the preferred driver for all radeon chips.
Long story short : VIDEO_CARDS="radeon" is what you should be using.
(In reply to comment #20)
> 2 different drivers. "radeon" installs xf86-video-ati which is the historical
> ATI OSS driver. "radeonhd" installs xf86-video-radeonhd which is a newer driver
> meant only for HD-series chips.
> For a long time, radeonhd was the only one to support newer cards. But that's
> no longer the case and xf86-video-ati is the preferred driver for all radeon
> Long story short : VIDEO_CARDS="radeon" is what you should be using.
Currently using radeonhd, radeon switches my two monitors (kde thinks that my smaller screen is the primary one). When I figure out how to renumber the screens (if it is possible) will switch to radeon.
Also, I have to upgrade to 2.6.31 kernel, right?
@Tomàš, your turn to pipe in, you know ATI stuff better than me :)
This is what i wrote for people interested in migration.
Well 3D will be around kernel-2.6.32. Mesa is prepared already :]
(In reply to comment #23)
> This is what i wrote for people interested in migration.
> Well 3D will be around kernel-2.6.32. Mesa is prepared already :]
Merged mesa-7.6 and x11-drm from git (using that wiki page). Got 3D hardware supported but kwin compositing still doesn't want to work :) . Which is odd, GLX_EXT_texture_from_pixmap is supported.
Also I get "IRQ's not enabled, falling back to busy waits: 2 0" warnings. I presume that I should also update some other thing to latest unstable :).
Ok... I just kinda got the compositing working :). Kwin effects are working now :).