AMD Catalyst 8.9 Gets WINE Fix, RandR 1.2 Support

[storma]

http://ati.amd.com/support/drivers/linux/linux-radeon.html

Release notes: https://a248.e.akamai.net/f/674/9206/0/www2.ati.com/drivers/linux/catalyst_89_linux.html

[Tsume]

Come on ATi, support my 3870x2 already... the thing's price has already halved from when I bought it in April so put that profit to good use!

[grantek]

Woohoo! I broke my linux server at home last night with a pvmove that went awry, but it's almost fully back up and running - now to break the workstation with this :)

Good to see there's a lot of "under the hood" fixes for stuff involving screen corruption.

[LOLatMS]

Anyone notice any changes when it comes to full screen corruption (the dreaded checkerboard)? For me, full screen native apps did NOT have any issues in 8.8, but wine apps and flash full screen videos (e.g. YouTube) did have problems. In 8.9, wine has been fixed, but flash remains a problem. I did not have any full screen corruption problems with any of these several months ago (forgot what Catalyst that was, though), so this is probably a problem with the drivers and not the flash plugin.

[Chris Clearwater]

Anyone notice any changes when it comes to full screen corruption (the dreaded checkerboard)? For me, full screen native apps did NOT have any issues in 8.8, but wine apps and flash full screen videos (e.g. YouTube) did have problems. In 8.9, wine has been fixed, but flash remains a problem. I did not have any full screen corruption problems with any of these several months ago (forgot what Catalyst that was, though), so this is probably a problem with the drivers and not the flash plugin.

My screen corruption problems appear to be solved. It affected mostly wine apps, but there were some native apps that exhibited the problem as well. For instance, the loading screen of nexuiz. I don't have any problem with flash at all. If you use Ubuntu, try enabling backports, which will install a newer (possibly beta) version of flash player.

[yoshi314]

i must say that's the most impressive fglrx changelog i've seen in months. assuming that all the mentioned problems are really resolved.

· A fatal server error no longer occurs when changing the display resolution using Catalyst Control Center Linux Edition and restarting X. Further details can be found in topic number: 737-37419

· Catalyst Control Center Linux Edition: Clicking on the 3D leaf in Page Tree no longer results in 3D Page Tree failing to respond. Further details can be found in topic number: 737-37425

· Catalyst Control Center Linux Edition: A green box at the bottom of the page the TV and Position page is no longer displayed. Further details can be found in topic number: 737-37426

· Default buttons are no longer available in the Catalyst Control Center for driver options that do not have custom settings available. Further details can be found in topic number: 737-37427

· HDMI Audio support for the Linux Operating system is now available for the ATI Radeon HD 3100/3200 series of products. Further details can be found in topic number: 737-37420

· Hot plugging a digital panel to an ATI Radeon HD 46x0 series product no longer results in the digital panel failing to become active. Further details can be found in topic number: 737-37421

· X-Windows no longer fails to launch after the driver installation for the ATI Radeon HD 4870 product. Further details can be found in topic number: 737-37422

· Rebooting the system after completing the driver installation no longer results in X failing to re-start on systems containing an ATI FireGL V7700 series of product. Further details can be found in topic number: 737-37431

· X no longer fails to respond when shutting down in dual head mode. Further details can be found in topic number: 737-37430

· The entire X desktop including the mouse cursor is no longer choppy during playback of video files or suzi test app. Further details can be found in topic number: 737-37423

· The TV size and position adjusting sliders are no longer missing on systems running Linux. Further details can be found in topic number: 737-37424

· Running aticonfig –initial –f on a system with two display devices connected no longer results in dual head mode failing to function. Further details can be found in topic number: 737-37438

· Automatic post install configuration is now functioning. Further details can be found in topic number: 737-37441

· Corruption is no longer noticed on the Linux Desktop when moving an application or browser window. Further details can be found in topic number: 737-37439

· Logging out one of multiple users no longer results in X failing to display an image. Further details can be found in topic number: 737-37440

· Connecting an HDMI display device may result in the audio failing to be heard. Further details can be found in topic number: 737-37428

· Connecting both a CRT and DFP display device no longer results in the CRT failing to display an image if the DFP is disconnected and then reconnected to the system. Further details can be found in topic number: 737-37429

· The resolutions of 540p and 1080p @ 60Hz no longer fail to be added to the Display Manager list of resolutions when they are selected from HDTV -> Formats. Further details can be found in topic number: 737-37435

· Connecting an HDMI display device and connecting an HDMI cable from the graphics card to the display device no longer results in audio failing. Further details can be found in topic number: 737-37436

· Closing fgl_glxgears or glxgears via the toolkit “X” in a multi-head dual-head environment no longer results in a segmentation fault. Further details can be found in topic number: 737-37432

· Running textured video on systems running Linux no longer results in a segmentation fault. Further details can be found in topic number: 737-37433

· GLESX: An unresolved symbol error is no longer noticed in the XOrg log. Further details can be found in topic number: 737-37442

· Playing rotyov no longer results in the system response time becoming slow on SLES10.2. Further details can be found in topic number: 737-37443

· Doom 3: The performance gap between Windows XP and Linux no longer increases when CrossFire is enabled. Further details can be found in topic number: 737-37434

· Doom3: Enabling clone mode and setting the display resolution to 1280x1024 no longer results in corruption being noticed. Further details can be found in topic number: 737-37437

if each fglrx release will get that much attention i can see bright future for linux proprietary ati driver.

than again, the regression list looks quite scary.

[Heiko]

I still have an error on closing glx_gears. In combination with millions of errors reported when using valgrind (memory leak checker) on glxgears. Besides that, glxinfo hanged (no error, it just hangs). This happened only on one of my computers (which uses dual screen on a AMD 3450 with two X screens), my other computer (HD 4870, single screen) does not hang with glxinfo.

Other things I still have to try, I didn't have time for that. For example, with 8.8 I got screen corruption when using Blender in windowed mode (full screen: no problems). I didn't have time yet to test this with 8.9.


Tested blender: screen corruption is less bad as with 8.8, but it still blackens my Gnome panels (panels on top and bottom of screen turn black, when you use the menu, the part of the panel which is used turn normal again.

[GoBBLeS]

Come on ATi, support my 3870x2 already... the thing's price has already halved from when I bought it in April so put that profit to good use!

What he said. They just got xfire, get on with xfireX already(or whatever's needed) :) It's the only reason i'm in Vista atm...

[StefanHamminga]

Undefined symbols may be noticed when trying to load fglrx on Ubuntu 8.10. Further details can be found in topic number: 737-37447

So still no xserver 1.5 support?

[debunk]

Great to see they fixed the checkerboard of death, but I now get a black screen when launching all hl2 games I own (hl2, ep1 and 2 and portal). If I disable pixel shaders in winecfg I do get visible output the startup screen, but the game exits and a popup appears telling me the game needs at least directx 8.0. I tried setting dxlevel to 80, 81 and 90, but I keep getting the same error.
I'll try tweaking the wine registry when I have time later on today to see if I can fix it.

Nevermind, I forgot to reboot after I installed the new drivers.

I'll delete the original post but leave this reply here in case anyone else is as dumb as I am ;)

[Chris Clearwater]

It seems that half-life 2 via wine works the first time I run it. But subsequent runs result in a black screen after the Valve logo. This persists until reboot.

[LOLatMS]

Chris Clearwater:

Thanks. Rather than enable backports, I just installed Flash Player 10 RC manually. It's much better than the beta I tried a month or two months back. The full screen problem is now gone. I guess the problem was on Adobe's side. Not surprised there, but between blaming ATI or blaming Adobe with regards to their Linux support... now that's a hard choice! :)

[NeoBrain]

So still no xserver 1.5 support?

I also had hoped that I could leave 1.5 on my upcomming Fedora 9 installation, guess I'll need to downgrade that one...
Downloading the new driver now, hope it will be better than 8.8.

[LOLatMS]

I also had hoped that I could leave 1.5 on my upcomming Fedora 9 installation, guess I'll need to downgrade that one...
Downloading the new driver now, hope it will be better than 8.8.

I'd say that this release is definitely an improvement, at least for what I do (occasional 3D game, Starcraft and Warcraft 3 in wine, videos in mplayer, flash videos on YouTube... I haven't checked for tearing in videos (that has never gone away for me, even when the release notes claim the issue was fixed.) I use mplayer (generally latest or near-latest svn) xv output.

Hopefully, you don't have to downgrade. That's just a bummer. Please let us know how it goes.

Overall, thumbs up from me. (I'll save two thumbs up when (if?) we reach parity with the Windows Catalyst drivers... but hey, even those are buggy!)

Random Edit: What's the difference between "thumbs up" and "two thumbs up"? "Thumbs up" is already plural. Perhaps I should give this release a thumb up, but that just sounds awkward.

[NeoBrain]

It seems that half-life 2 via wine works the first time I run it. But subsequent runs result in a black screen after the Valve logo. This persists until reboot.

I once had to turn off shader support in winecfg in order to get into the menu. Not sure if the problem should be fixed now though.

EDIT: oh wait, you are talking about HL 2...
I really should practise reading a bit :P

[Pedric]

Worst release this year. Breaks virtual desktop switching with WoW & Wine. The wm context menue flickers heavily and cannot be used. Alt-Tabbing does not work. Subsequent corruption when attempting to switch desktops. Still no pbuffers fix. Reverted to 8.08 within minutes...

[izual]

I am confused.
On one hand there is a bug fix for:
Hot plugging a digital panel to an ATI Radeon HD 46x0 series product no longer results in the digital panel failing to become active. Further details can be found in topic number: 737-37421

On the other hand there is no support for the 46x0 Series?

[Fenix-TX]

Bad, bad, bad....


jesus@tux-CT:~$ fglrxinfo
Locking assertion failure. Backtrace:
#0 /usr/lib/libxcb-xlib.so.0 [0xb7bc9767]
#1 /usr/lib/libxcb-xlib.so.0(xcb_xlib_unlock+0x31) [0xb7bc98b1]
#2 /usr/lib/libX11.so.6 [0xb7db8421]
#3 /usr/X11R6/lib/libGL.so.1(__glXInitialize+0x5d5) [0xb7e78e75]
#4 /usr/X11R6/lib/libGL.so.1 [0xb7e7507d]
#5 /usr/X11R6/lib/libGL.so.1(glXChooseVisual+0x3a) [0xb7e76c3a]
#6 fglrxinfo [0x8048d14]
#7 fglrxinfo [0x8048a58]
#8 /lib/tls/i686/cmov/libc.so.6(__libc_start_main+0xe0) [0xb7c32450]
#9 fglrxinfo [0x80488f1]
display: :0.0 screen: 0
OpenGL vendor string: ATI Technologies Inc.
OpenGL renderer string: ATI RADEON 9600 Series
OpenGL version string: 2.1.7979 Release


SOLVED: It was something wrong with libGL* links on /usr/X11R6/lib/

[NeoBrain]

Okay, at least nothing broke for me in 8.9.
My experiences:
- screen corruption because of the screen corruption workaround (mode 1361x768), easily fixed though
- switching from X to VT1 and back works now :)
- creating a second X server on VT8 doesn't fail now, but when switching back to the first one the system freezes (Sysrq-able)
- WINE APPLICATIONS still give this damn libGL error, this one exists like about 10 releases now...
- Well, I guess this is a limitation of the X server, but firefox scrolling is usuable, but way slower than in windows.
- And c'mon: Why must the download be 63 MB? It also seems to get bigger with each release, and I often spent 2 hours on downloading & installing just to see that the new driver doesn't offer something useful for me without breaking anything at the same time.

[phoronix]

Phoronix: AMD Catalyst 8.9 Gets WINE Fix, RandR 1.2 Support

Last month with the Catalyst 8.8 Linux driver we finally experienced CrossFire support on Linux as well as OverDrive support. This was one of the most significant driver releases of the year, but it's now time to move on as Catalyst 8.9 has just been released. This new driver update does bring a few exciting changes: the long-awaited WINE fixes, what appears to be RandR 1.2 with accelerated rotation support, and a horde of bug-fixes.

http://www.phoronix.com/vr.php?view=12856

[ArtX]

Mhhh last ati drivers include new features but they come with more problems.
I use only linux and the only app 3d that I use is quake3arena.
So, with 8.7 I've no problems with q3 (also if there is problems with wine, fixed with this realase, and blender); but with 8.8 I have problems with mouse movements on q3, and with 8.9 now the mouse don't works into the game, and X crash.
I'm using Debian Lenny realase in an Acer Aspire 5672 with ati raden mobility X1400.
Now I ordered a new pc with nvidia card, but I hope AMD/ATI will come the best choice for linux users with the future opensource driver.
Bye

[Heiko]

I saw that the Windows 8.9 version already supports some extensions for OpenGL 3.0 support (they call it `OpenGL 3.0 support - Phase 1'). Lets hope these new features will soon also arrive for Linux!

[StefanHamminga]

Well, if I have to choose between xserver 1.5 support or opengl 3 I'd take xserver 1.5 support any day!

[TheWind]

this release does have improvement against some issue's that came along but an new one of course came up.

In this case the issue I'm having with this new release is in compiz fusion the alpha blur effect is become the equivalent of the saturation effect.

So if i have alpha blur enabled just about all my menu's become gray instead of having the blur effect like i had set them to have.

Another issue is I was being logged out after 3 seconds from these drivers, but that was only after i had logged in and out a few times which in most cases i never do.

Those are the only ones that are new for me this release.

Running an IGP 1200 Xpress ( rs690 ) [512 MiB memory shared to this IGP], the rest i don't think matters but AMD64x2 cpu, 2 GiB RAM, and Gigabyte Motherboard on Ubuntu 8.04. If you need more info just say so.

Actually it's xorg thats crashing nor me being logged out it has the same effect as Ctrl+Alt+Backspace though once it happens once it continuously happens afterwards. This is not limited to repeated logging in and out (thats the easy way to cause this fault) it can happen at random times too.

[NeoBrain]

Wow, just read the phoronix article.
I'm really looking forward to the upcomming 8.10 driver, hoping I'll get XServer 1.5 support on my birthday :D

EDIT: @TheWind:
I'm experiencing the same problem. This doesn't occur with focus blur though...

[Pedric]

Interesting things about the changelog/new features:

I first learned of the ATI Proprietary Linux Driver Releases news feed, which states:

Driver version: 8.532

NEW FEATURES:
(1) Production support for Red Flag DT6SP1 and OpenSUSE 11

ISSUES RESOLVED: (1) Desktop corruption when using full screen Wine application. (2) Compilation issues against Linux kernel 2.6.26. (3) HDMI Audio support on Radeon HD 3100/3200 products. (4) X fails after driver install for Radeon HD 4870 and FireGL V7700. (5) Choppy playback during video playback. (6) Fatal error while restarting X, after resolution change in amdcccle. (7) Segmentation fault when running textured video.

https://a248.e.akamai.net/f/674/9206/0/www2.ati.com/drivers/linux/catalyst_89_linux.html


However, on the page that the feed links to, there is no mention of these issues...

[Heiko]

Well, if I have to choose between xserver 1.5 support or opengl 3 I'd take xserver 1.5 support any day!

You are absolutely right, for most people xserver 1.5 is far more important then OpenGL 3.0. I am using Ubuntu 8.04 with the default 8.6.24-19, so currently I don't have any specific wishes for other kernel/xserver/xorg support (although I might upgrade to Ubuntu 8.10 when it arrives, so that would mean I also need support for all these things).

Wow, just read the phoronix article.
I'm really looking forward to the upcomming 8.10 driver, hoping I'll get XServer 1.5 support on my birthday :D

EDIT: @TheWind:
I'm experiencing the same problem. This doesn't occur with focus blur though...

Hah, if you present it like this: who knows I get a nice birthday present too (14th October, that is probably a bit to early for the next release).

[NeoBrain]

Hah, if you present it like this: who knows I get a nice birthday present too (14th October, that is probably a bit to early for the next release).

15th October for me, still way too early for a feature release ;-) I expect it about the 23rd.

Apart from that, I don't care much about OGL 3.0 either as long as there are no apps using it.

[Fran]

The only difference I've noticed wrt 8.8 is that now in amdcccle I cannot override the application settings for Antialiasing and Anisotropic filtering (everything is disabled). I can change Adaptive Antialiasing, Mipmap detail and "More Settings", though.

Anyone else? HD3850 here.

[NeoBrain]

The only difference I've noticed wrt 8.8 is that now in amdcccle I cannot override the application settings for Antialiasing and Anisotropic filtering (everything is disabled). I can change Adaptive Antialiasing, Mipmap detail and "More Settings", though.

Anyone else? HD3850 here.

amdcccle segfaults for me... I'm not using it too often anyways

[Extreme Coder]

While it's great to see such a large bugfix list, it's sad to see that new problems have been introduced, or that they haven't been solved completely. For example, when I started reading the article, I was like 'This is it! This is the driver release that will end all problems!', and then I get bummed by 3D corruption on the Unigine Tropics engine :/ Why is having normal 3D acceleration too hard? :(

[liels]

Is xv output (eg with mythtv) still tearing? While the changelog is impressive I won't be able to ditch my nvidia card until I can get tear-free output on mythtv.

[Vighy]

while everybody complains about small missing features :P
I do complain about a bad regression that segfaults my Xorg on start-up.

It already happened with Catalyst 8.7...


[...snip...]
(II) fglrx(0): Direct rendering enabled
(II) fglrx(0): Interrupt handler installed at IRQ 16.
(II) fglrx(0): Exposed events to the /proc interface
(II) fglrx(0): Finished Initialize PPLIB!
(II) fglrx(1): === [atiddxPreInit] === begin
(EE) fglrx(1): Quitting secondary screen -- no monitor specified.
(EE) fglrx(1): PreInit failed
(II) fglrx(1): === [atiddxPreInit] === end

Backtrace:
0: /usr/bin/X(xf86SigHandler+0x6a) [0x49263a]
1: /lib/libc.so.6 [0x7f9e61f5dda0]
2: /usr/lib64/xorg/modules/drivers//fglrx_drv.so(atiddxFreeScreen+0x2f4) [0x7f9e5fcfa434]
3: /usr/bin/X(xf86DeleteScreen+0x7e) [0x4a5ade]
4: /usr/bin/X(InitOutput+0xa1f) [0x46c0bf]
5: /usr/bin/X(main+0x2c6) [0x437256]
6: /lib/libc.so.6(__libc_start_main+0xe6) [0x7f9e61f491d6]
7: /usr/bin/X(FontFileCompleteXLFD+0x289) [0x4366f9]

Fatal server error:
Caught signal 11. Server aborting


exactly the same backtrace!! (info about my architecture are in that thread, too)
http://www.phoronix.com/forums/showthread.php?t=11570


and other people experienced it too:
http://www.phoronix.com/forums/showthread.php?t=11794
with other chipsets...


Catalyst 8.8 and 8.6 didn't suffer for the hang.
There may be a workaround?

[carpex]

Its funny that WINE fixes are supposed to be an important feature of this release, but there is no mention of WINE in the release notes. :confused:

[Kano]

That driver is fully crap for several reasons:

a) Using Debian etch + Xorg 7.1.1

OpenGL accelleration only works as root! The xorg.conf entry

Section "DRI"
Mode 0666
EndSection

Seems to be fully ignored. Absolutely unusable!

b) Absolutely no rendering improvements - even as root pointsprites are missing

Hey mtippet didn't you tell me at the end of JUNE that after 2 month all would work - like gl2benchmark and your example with flightgear?

c) Not even the most trivial fixes for 2.6.27 support - load the kernel module only smp_call_function must be fixed - for 2.6.27 and newer the 3rd option has to be removed, wow thats really hard that you can not make it.

Maybe AMD/ATI wants to wins the worst driver of the year award with that piece of shit. I guess the Linux devs there play at home with Nvidia cards in their pcs.

[Pfanne]

That driver is fully crap for several reasons:

a) Using Debian etch + Xorg 7.1.1

OpenGL accelleration only works as root! The xorg.conf entry

Section "DRI"
Mode 0666
EndSection

Seems to be fully ignored. Absolutely unusable!

b) Absolutely no rendering improvements - even as root pointsprites are missing

Hey mtippet didn't you tell me at the end of JUNE that after 2 month all would work - like gl2benchmark and your example with flightgear?

c) Not even the most trivial fixes for 2.6.27 support - load the kernel module only smp_call_function must be fixed - for 2.6.27 and newer the 3rd option has to be removed, wow thats really hard that you can not make it.

Maybe AMD/ATI wants to wins the worst driver of the year award with that piece of shit. I guess the Linux devs there play at home with Nvidia cards in their pcs.

well thats one hell of a great post...
why don't you try to rant a little bit less and at least tell us what gpu you are using...

[NeoBrain]

Having tested it a bit further now (esp. regarding Wine), I noticed that (apart from the usual screen corruption) virtual desktop switching works again and Wine doesn't segfault anymore when started twice on the same terminal.

By the way, @Michael: the official spelling is Wine, not WINE.

[BlackStar]

I couldn't get this driver to work: a few seconds after X started, it would (soft) crash. It's the first time a driver didn't work here and no amount of tinkering could get it in shape, so I'm back to 8.8 for now.

RV580 (X1950 Pro) on Ubuntu 8.04 x86_64 with 6GB of memory.

At least 8.8 works quite well, so it's not a problem. Looking forward to Ubuntu 8.10 with the open-source drivers now - already tested and they work flawlessly here!

[rbmorse]

Works fine, here (Ubuntu 8.04) with my HD 3870.

But, I'm not doing anything interesting.

[flami]

Well now "OffscreenRenderingMode fbo" works in wine again, so no stuff thats not supposed to on the screen appears in the bottom left, but I cant minimize or switch desktop anymore . I can interact with the desktop ( e.g. open something in the taskbar ) , but i cant see anything . It flickers for a second where i see the desktop , then wine's all over the place again .

( debian lenny , mobility x700 )

[Kano]

@Pfanne

The errors I describe are independent of the GPU used. The PointSprites are missing since driver release 8.41.7, that's nothing new, I mentioned that 13 month ago already. And when the only change from one driver to the next is a extremely bad new error then you would not get angry, right? I am responsible for lots of ATI users who use Kanotix and I create patches for ATI drivers on my own if possible. But if the base is so bad then this is completely lost time. If a huge company can not even fix issues introduced over a year ago they lost completely. But to make you happy: the card i used is X700 Pro (RV410) chipset - one of the chipsets with the worst open source driver experience!

[storma]

That driver is fully crap for several reasons:

a) Using Debian etch + Xorg 7.1.1

OpenGL accelleration only works as root! The xorg.conf entry

Section "DRI"
Mode 0666
EndSection

Seems to be fully ignored. Absolutely unusable!

Check your permissions on /dev/dri/card0

[Kano]

Good idea to check this:

ls -l /dev/dri/card0
crw-r----- 1 root root 226, 0 2008-09-18 18:01 /dev/dri/card0

[flami]

hum I got another nice "feature" ^^:
flikering lines on the top right part of my screen when I run anything that involves 3d
http://flami.net/Junk/8.9_screen_corrpution.jpg
Im going back 8.8 , Ill be back soon (tm)

[dokers]

Anyone try this driver with the AGP 3850 card yet? I haven't had success since the 8.05 driver? I'm on Hardy Heron. If I get a chance I'll try it in the next day or 2.

[flami]

Well now "OffscreenRenderingMode fbo" works in wine again, so no stuff thats not supposed to on the screen appears in the bottom left, but I cant minimize or switch desktop anymore . I can interact with the desktop ( e.g. open something in the taskbar ) , but i cant see anything . It flickers for a second where i see the desktop , then wine's all over the place again .

( debian lenny , mobility x700 )

hum I got another nice "feature" ^^:
flikering lines on the top right part of my screen when I run anything that involves 3d
...
Im going back 8.8 , Ill be back soon (tm)


back to 8.8, fbo doesn't work anymore , but i can switch to the desktop from wine.
those flickering lines ... I don't know , there still are some in ppracer, but theres less and they are not as big anymore. in wine they disappeared
But before i upgraded to 8.9 I never had them.

[Odur]

Works fine, here (Ubuntu 8.04) with my HD 3870.

But, I'm not doing anything interesting.
I too got a HD3870, and I can't switch users without getting a hard lock. Would you mind testing that?


I wrote this about 20 minutes ago in bug report #112518 at Launchpad:

I too got this bug and haven't found a solution yet.

Dist: Kubuntu 8.04.1 Fully upgraded
Graphics card: PowerColor (ATI) HD3870 SCS3
Tested drivers: fglrx 8.3 through 8.9

Symptoms: When switching user the screen goes black and the system is totally unresponsive, including the keyboard. Not even caps-lock trigges the caps-lock light. The only thing that works is the reset button and to hold the powerbutton until it shuts down. Shutdown or restart works without problem (I have fixed the problem with atieventsd, see bug 118605).
If i use the vesa or radeonhd drivers everything is ok. But then I got no 3D acceleration, which I think would be nice...

This bug is really annoying to me, as we are switching users a lot on this computer. It started when I bought this card. I got a ATI X600 before, and that didn't suffer from this bug at all.

Anything else I could provide to help solving this bug, or at least get it triaged?

[Melcar]

hum I got another nice "feature" ^^:
flikering lines on the top right part of my screen when I run anything that involves 3d
http://flami.net/Junk/8.9_screen_corrpution.jpg
Im going back 8.8 , Ill be back soon (tm)


Distortion you mean, as in horizontal lines across the screen? That's nothing new. Even with vsync. on you can notice some level of distortion.
It also happens with video playback, but that I think is another separate issue.

[chrisr]

OpenGL screensavers are still not exiting to a command line when you quit. (Run /usr/libexec/xscreensaver/rubik, and then press ESC.) So this presumably means that I still can't use OpenGL screensavers.

xine is still crashing the Xserver when it uses Textured Video and reaches the end of the movie. So I still can't use XVideo for video playback.

These are both long-standing issues.

I am also seeing this in my Xorg.0.log file:

(II) Loading sub module "vbe"
(II) LoadModule: "vbe"
(II) Reloading /usr/lib/xorg/modules//libvbe.so
(EE) fglrx(0): vm86() syscall generated signal 11.
(II) fglrx(0): EAX=0x0000004f, EBX=0x00000000, ECX=0x00000000, EDX=0x00000000
(II) fglrx(0): ESP=0x00001000, EBP=0x00000000, ESI=0x00000000, EDI=0x00002000
(II) fglrx(0): CS=0x0000, SS=0x0100, DS=0x0040, ES=0x0000, FS=0x0000, GS=0x0000
(II) fglrx(0): EIP=0x00000600, EFLAGS=0x00033202
(II) fglrx(0): code at 0x00000600:
f4 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
(II) fglrx(0): VESA BIOS detected
(II) fglrx(0): VESA VBE Version 3.0
(II) fglrx(0): VESA VBE Total Mem: 16384 kB
(II) fglrx(0): VESA VBE OEM: ATI ATOMBIOS
(II) fglrx(0): VESA VBE OEM Software Rev: 9.12
(II) fglrx(0): VESA VBE OEM Vendor: (C) 1988-2005, ATI Technologies Inc.
(II) fglrx(0): VESA VBE OEM Product: M66GL
(II) fglrx(0): VESA VBE OEM Product Rev: 01.00
(EE) fglrx(0): vm86() syscall generated signal 11.
(II) fglrx(0): EAX=0x0000004f, EBX=0x00004116, ECX=0x00000000, EDX=0x00000000
(II) fglrx(0): ESP=0x00001000, EBP=0x00000000, ESI=0x00000000, EDI=0x00002000
(II) fglrx(0): CS=0x0000, SS=0x0100, DS=0x0040, ES=0x0000, FS=0x0000, GS=0x0000
(II) fglrx(0): EIP=0x00000600, EFLAGS=0x00033202
(II) fglrx(0): code at 0x00000600:
f4 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
(II) fglrx(0): ATI Video BIOS revision 9 or later detected
...
(II) fglrx(0): Using adapter: 1:0.0.
(II) fglrx(0): [FB] Find the MC FB aperturs range(MCFBBase = 0xc0000000, MCFBSize = 0x10000000)
(--) fglrx(0): Video RAM: 262144 kByte, Type: DDR3
(II) fglrx(0): PCIE card detected
(WW) fglrx(0): board is an unknown third party board, chipset is supported
(EE) fglrx(0): vm86() syscall generated signal 11.
(II) fglrx(0): EAX=0x0000004f, EBX=0x00000003, ECX=0x00000000, EDX=0x00000000
(II) fglrx(0): ESP=0x00001000, EBP=0x00000000, ESI=0x00000000, EDI=0x00002000
(II) fglrx(0): CS=0x0000, SS=0x0100, DS=0x0040, ES=0x0000, FS=0x0000, GS=0x0000
(II) fglrx(0): EIP=0x00000600, EFLAGS=0x00033202
(II) fglrx(0): code at 0x00000600:
f4 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
(II) fglrx(0): RandR 1.2 support is enabled!
(II) fglrx(0): RandR 1.2 rotation support is enabled!

I don't know what's going on there, but it doesn't look good.

Enabling RandR 1.2 support has reduced the FPS of glxgears from 3700 to ~2214.

[raulromania]

OS: Ubuntu (stable, latest)
Graphicscard: PowerColor HD3850 PCS AGP

First, very big thanks to the hard working developers.

Anyone try this driver with the AGP 3850 card yet? ...

Well, I have a PowerColor 3850 PCS for the very new AGP slot ^^

The driver works almost perfectly with it. I am very glad to see that.
Flash videos now working without lags, even with xv enabled. (see the following 'xorg.conf')

Section "InputDevice"
Identifier "CherryEvolutionFlat"
Driver "evdev"
Option "Device" "/dev/input/by-id/usb-046a_0023-event-kbd"
EndSection

Section "InputDevice"
Identifier "RazerDiamondback"
Driver "evdev"
Option "Device" "/dev/input/by-id/usb-Razer_Razer_1600dpi_Mouse-event-mouse"
EndSection

Section "InputDevice"
Identifier "KeySonicACK540RF"
Driver "evdev"
Option "Device" "/dev/input/by-id/usb-05af_2.4G_USB_RF_KeyBoard-event-kbd"
EndSection

Section "Device"
Identifier "VideoOut0"
BusID "PCI:1:0:0"
Driver "fglrx"
Screen 0
Option "UseInternalAGPGART" "true"
Option "VideoOverlay" "true"
Option "OpenGLOverlay" "false"
Option "BackingStore" "true"
Option "BlockSignalsOnLock" "true"
Option "TexturedVideo" "true"
Option "TexturedVideoSync" "true"
Option "Textured2D" "true"
Option "TexturedXRender" "true"
EndSection

Section "Device"
Identifier "VideoOut1"
BusID "PCI:1:0:0"
Driver "fglrx"
Screen 1
Option "UseInternalAGPGART" "true"
Option "VideoOverlay" "true"
Option "OpenGLOverlay" "false"
Option "BackingStore" "true"
Option "BlockSignalsOnLock" "true"
Option "TexturedVideo" "true"
Option "TexturedVideoSync" "true"
Option "Textured2D" "true"
Option "TexturedXRender" "true"
EndSection

Section "Monitor"
Identifier "MonitorSettings"
Option "DPMS"
EndSection

Section "Screen"
Identifier "Screen0"
Monitor "MonitorSettings"
Device "VideoOut0"
DefaultDepth 24
SubSection "Display"
Depth 24
Modes "1280x1024"
EndSubSection
EndSection

Section "Screen"
Identifier "Screen1"
Monitor "MonitorSettings"
Device "VideoOut1"
DefaultDepth 24
SubSection "Display"
Depth 24
Modes "1920x1080"
EndSubSection
EndSection

Section "ServerLayout"
Identifier "DisplayLayout"
Screen "Screen0"
Screen "Screen1" RightOf "Screen0"
EndSection

Section "ServerFlags"
Option "AllowMouseOpenFail" "true"
EndSection

WINE works well, too (still hard work needed).

Well, there is a thing, i would be happy if they fix it soon.
I have an Samsung LE40A656 FULL HD LCD and I get this "black border" issue.
I think the developers are aware of the underscan issue.

See → http://www.phoronix.com/forums/showthread.php?t=10073&page=4#40

The workaround with ...

sudo aticonfig --set-pcs-val=MCIL,DigitalHDTVDefaultUnderscan,0

... works very good, but I'd like to adjust it easier, i.e. in the AMDCCCLE.

greetz.

[chrisr]

Fedora's package building scripts have been broken for ages, and so I've been trying to updates them manually. However, I'm not sure where the new libdri.so should go. At the moment, I've moved the libdri.so from xorg-x11-server-Xorg into hiding and copied fglrx's one in manually to replace it. But this violates RPM! Is this new libdri.so meant to go somewhere else, please?

[chrisr]

That's me done with 8.9 - if it doesn't suspend correctly then it's going straight into the dustbin. This is the last entry in Xorg.0.log:
Synaptics DeviceOff called
(II) AIGLX: Suspending AIGLX clients for VT switch
(EE) fglrx(0): vm86() syscall generated signal 11.
(II) fglrx(0): EAX=0x0000004f, EBX=0x00004116, ECX=0x00000000, EDX=0x00000000
(II) fglrx(0): ESP=0x00001000, EBP=0x00000000, ESI=0x00000000, EDI=0x00002000
(II) fglrx(0): CS=0x0000, SS=0x0100, DS=0x0040, ES=0x0000, FS=0x0000, GS=0x0000
(II) fglrx(0): EIP=0x00000600, EFLAGS=0x00033202
(II) fglrx(0): code at 0x00000600:
f4 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00

[Kano]

Next major fault with Debian Etch:

I made a hotfix as c tool:
#include <sys/types.h>
#include <sys/stat.h>
int main(int argc, char **argv) {
chmod("/dev/dri/card0", 0666);
return 0;
}

which could be set to root:root and chmod 4755 and put into autostart, but then BERYL is still not working with Xorg 7.1.1. It worked before (just without water effect). So not only permissions are broken but even more. Too funny that wine works with it ;)

[grantek]

Woohoo! BIG improvements for me on my HD4850:
- 32-bit ET:QW on 64-bit OS WORKS - for me, this is the turning point between the driver being "not good" and "good with caveats" :)
- Wine is fixed! Yay! The 3d games I've tried so far haven't responded well though, but that was in my 64-bit OS - will need to do some more solid testing.
- On a more minor note, looks like the debian packaging problems (http://www.phoronix.com/forums/showthread.php?t=12706) have been fixed. I haven't tried uninstalling yet, but 8.9 only sets up diverts for the libGL (and now libdri) files themselves, not the symlinks (which seemed to be the problem before).

[nstamoul0]

I have had no luck building the latest drivers,nor finding a solution on the net.Here is what I get:

dpkg-shlibdeps: warning: symbol _XFlush used by debian/fglrx-driver/usr/lib/libfglrx_gamma.so.1.0 found in none of the libraries.
dpkg-shlibdeps: warning: symbol XextAddDisplay used by debian/fglrx-driver/usr/lib/libfglrx_gamma.so.1.0 found in none of the libraries.
dpkg-shlibdeps: warning: symbol _XReply used by debian/fglrx-driver/usr/lib/libfglrx_gamma.so.1.0 found in none of the libraries.
dpkg-shlibdeps: warning: symbol XextRemoveDisplay used by debian/fglrx-driver/usr/lib/libfglrx_gamma.so.1.0 found in none of the libraries.
dpkg-shlibdeps: warning: symbol XextCreateExtension used by debian/fglrx-driver/usr/lib/libfglrx_gamma.so.1.0 found in none of the libraries.
dpkg-shlibdeps: warning: symbol XextFindDisplay used by debian/fglrx-driver/usr/lib/libfglrx_gamma.so.1.0 found in none of the libraries.
dpkg-shlibdeps: warning: symbol _XFlush used by debian/fglrx-driver/emul/ia32-linux/usr/lib/libfglrx_pp.so.1.0 found in none of the libraries.
dpkg-shlibdeps: warning: symbol XextAddDisplay used by debian/fglrx-driver/emul/ia32-linux/usr/lib/libfglrx_pp.so.1.0 found in none of the libraries.
dpkg-shlibdeps: warning: symbol _XReply used by debian/fglrx-driver/emul/ia32-linux/usr/lib/libfglrx_pp.so.1.0 found in none of the libraries.
dpkg-shlibdeps: warning: symbol XextRemoveDisplay used by debian/fglrx-driver/emul/ia32-linux/usr/lib/libfglrx_pp.so.1.0 found in none of the libraries.
dpkg-shlibdeps: warning: symbol XextCreateExtension used by debian/fglrx-driver/emul/ia32-linux/usr/lib/libfglrx_pp.so.1.0 found in none of the libraries.
dpkg-shlibdeps: warning: symbol XextFindDisplay used by debian/fglrx-driver/emul/ia32-linux/usr/lib/libfglrx_pp.so.1.0 found in none of the libraries.
dpkg-shlibdeps: failure: couldn't find library libstdc++.so.5 needed by debian/fglrx-driver/usr/lib/libAMDXvBA.so.1.0 (its RPATH is '').
Note: libraries are not searched in other binary packages that do not have any shlibs or symbols file.
To help dpkg-shlibdeps find private libraries, you might need to set LD_LIBRARY_PATH.
dh_shlibdeps: command returned error code 512
make: *** [binary] Error 1
dpkg-buildpackage: failure: debian/rules binary gave error exit status 2

Any suggestions?

[Kano]

apt-get install ia32-libs libstdc++5

[nstamoul0]

I was missing the libstdc++5 package.Thanks a lot!

[rbmorse]

I too got a HD3870, and I can't switch users without getting a hard lock. Would you mind testing that?
User switch locks for me, too.

[Pulfer]

User switch locks for me, too.

Does it lock even when you boot without splash=silent (grub) and without loading atieventsd (/etc/rc5.d)? Haven't tried 8.9 driver yet but this workaround solved the problem for me with 8.8 and earlier driver versions.

[Moronix]

Does this still need 'Enable unused/obsoleted exported symbols' in the kernel?

[Odur]

Does it lock even when you boot without splash=silent (grub) and without loading atieventsd (/etc/rc5.d)? Haven't tried 8.9 driver yet but this workaround solved the problem for me with 8.8 and earlier driver versions.
You didn't quote me, but I'll answer anyway :)
I haven't tested without splash=silent, but I will when I get home from work.

The problem with atieventsd is not related to the hard lock when switching user. That bug locks the computer even on restart and power off, but it's not a hard lock. It's only when switching user, with or without atieventsd that the hard lock occur.

I fixed the atieventsd problem with:
sudo mkdir -p /var/lib/xdm/authdir
sudo ln -s /var/run/xauth /var/lib/xdm/authdir/authfiles
See bug #118605 at Launchpad for a discussion on the atieventsd problem.

[Pulfer]

You didn't quote me, but I'll answer anyway :)
I haven't tested without splash=silent, but I will when I get home from work.

For me the solution worked in Mandriva 2008.1 only when I booted without splash=silent and without atieventsd. Booting just without splash=silent or just without atieventsd didn't help.

I fixed the atieventsd problem with:
sudo mkdir -p /var/lib/xdm/authdir
sudo ln -s /var/run/xauth /var/lib/xdm/authdir/authfiles
See bug #118605 at Launchpad for a discussion on the atieventsd problem.

I'll try this solution.

[energyman]

The only difference I've noticed wrt 8.8 is that now in amdcccle I cannot override the application settings for Antialiasing and Anisotropic filtering (everything is disabled). I can change Adaptive Antialiasing, Mipmap detail and "More Settings", though.

Anyone else? HD3850 here.

works fine for me - when it finally pops up.
HD3870

[Fran]

works fine for me - when it finally pops up.
HD3870
Weird, I don't get the spam in xorg.0.log and amdcccle starts as quick as ever, but those two features are disabled.

Well, it's not like I play a lot in linux, so I don't care that much, but they were enabled in 8.8.

[Aostrich]

The only difference I've noticed wrt 8.8 is that now in amdcccle I cannot override the application settings for Antialiasing and Anisotropic filtering (everything is disabled). I can change Adaptive Antialiasing, Mipmap detail and "More Settings", though.
Try running amdcccle with root privileges. I have the same with 8.8.

8.9 don't work at all for me - hang up after starting X. I think (even hope) that the issue is in my xorg version, I have xorg-server-1.4.0.90 in my Samd64 12.1. Today I'll try to update it to 1.4.2, because I had no hang-ups with 8.9 in my experimental Sabayon linux installation, and it uses xorg-server 1.4.2.

[Fran]

Try running amdcccle with root privileges. I have the same with 8.8.
Nope, it's disabled even as root. I had that problem before 8.8, which seem to have fixed it, but now it's back.

[chrisr]

Some further analysis has revealed that my laptop's recent inability to suspend is actually due to the presence of a SATA drive in its docking station rather that fglrx.

However, the segmentation violation errors in Xorg.0.log persist.

[Peruna]

For me this new release is worst ever.

After installing (by purging old one and removing /etc/ati) new drivers on Ubuntu 8.04.1 mplayer fails to render video (some videos show only green, but with osd rendered and some just crash) with good old settings:
-vo gl:yuv=5:force-pbo:ati-hack:lscale=1

Those settings are fast enough to run most 1080p video, but now I only can watch video with -vo gl:yuv5:lscale=1 and it is terribly slow.

Xv is not an options because of tearing.

I wonder if something is wrong and somehow I am using Mesa somewhere so here is my Xorg.log: http://pastebin.com/m305a4aaf and xorg.conf: http://pastebin.com/m5a8eb1d3

I actually added "AIGLX" "off" after installation, but it was slow even without that setting (I think that it was not enabled with 8.8, but I might be wrong).

I have HD3200 (780g).

Any ideas?

[mirv]

Well, I have big issues with 8.9 - though I am using an "unofficial" gentoo ebuild. No opengl apps will run, dmesg spits out:

[fglrx:firegl_lock_free] *ERROR* lock was not held by 2! (*lock=0x80000001)
[fglrx:firegl_unlock] *ERROR* firegl_lock_free failed!

-- amd64, x1600, xorg-x11 7.3, xorg-server 1.4.2

-- Additionally, only happens when in dual head mode. Xorg log has the following:

(EE) fglrx(1): [DRI] Locking deadlock.
Already locked with context 22167732,
trying to lock with context 2.
(EE) fglrx(1): [DRI] Unlocking inconsistency:
Context 22167732 trying to unlock lock held by context 2

[dscharrer]

Well, I have big issues with 8.9 - though I am using an "unofficial" gentoo ebuild. No opengl apps will run, dmesg spits out:
The ati-drivers-8.522 with only the version and SRC_URI changed works for me. That is on Gentoo amd64 and a HD 4850. Which version did you base your're ebuild on? If I get to it I'll try it on a 32 bit Gentoo installation with a X1600.

Screen corruption is gone. :)

However, like other's mentioned this updates prevents ALT+TAB switching between from a Wine fullscreen app (and also from fullscreen Kaffeine for me). Like others mentioned, the desktop still reacts to mouse and keyboard input and the cursor changes appropriately when hovering over text fields, etc.

This is especially annoying, when wine hangs. Switching to VT1 still works, but killing the wine app causes the screen (int VT1, just plain text console) to be corrupted and switching back to X doesn't work. Again, the console stays reactive to keyboard input.

Also, when coming back from fullscreen mode in Kaffein sometimes the the screen stays black and things don't get redrawn until the applications request it (as in mose over, etc.). Note that switching from fullscreen to non-fullscreen in Kaffeine did cause that nice screen corruption to appear with Catalyst 8.6 through 8.8.

Thankfully, as with the Wine screen corruption, using the Mesa libGL insted of the Catalyst one makes all these problems go away for me, while retaining harware accelerated direct rendering. Of course, this disables all the 3D-related settings in the Catalyst Control Center.

[mirv]

The ati-drivers-8.522 with only the version and SRC_URI changed works for me. That is on Gentoo amd64 and a HD 4850. Which version did you base your're ebuild on? If I get to it I'll try it on a 32 bit Gentoo installation with a X1600.


Based it off the same version - just changed the SRC_URI as well. I rather suspect that I have something different to most people too - I've never been able to run amdcccle. Do you use a dual head system, or only single screen?
I use enlightenment if that makes a difference, though the 8.8's worked just fine.

[Kano]

First check:

ldd $(which amdcccle)

[dscharrer]

Do you use a dual head system, or only single screen?
single screen

[mirv]

First check:

ldd $(which amdcccle)

Output is:
linux-vdso.so.1 => (0x00007fffd25fe000)
libICE.so.6 => /usr/lib/libICE.so.6 (0x00007f53ca21a000)
libSM.so.6 => /usr/lib/libSM.so.6 (0x00007f53ca112000)
libdl.so.2 => /lib/libdl.so.2 (0x00007f53ca00e000)
libXext.so.6 => /usr/lib/libXext.so.6 (0x00007f53c9dfc000)
libX11.so.6 => /usr/lib/libX11.so.6 (0x00007f53c9af4000)
libfontconfig.so.1 => /usr/lib/libfontconfig.so.1 (0x00007f53c98c4000)
libXrender.so.1 => /usr/lib/libXrender.so.1 (0x00007f53c97bb000)
libXrandr.so.2 => /usr/lib/libXrandr.so.2 (0x00007f53c96b4000)
libfreetype.so.6 => /usr/lib/libfreetype.so.6 (0x00007f53c952a000)
libXcursor.so.1 => /usr/lib/libXcursor.so.1 (0x00007f53c9420000)
libpthread.so.0 => /lib/libpthread.so.0 (0x00007f53c9306000)
libXi.so.6 => /usr/lib/libXi.so.6 (0x00007f53c91fd000)
libGL.so.1 => //usr/lib64/opengl/ati/lib/libGL.so.1 (0x00007f53c906a000)
libstdc++.so.5 => /usr/lib64/libstdc++-v3/libstdc++.so.5 (0x00007f53c8e8d000)
libm.so.6 => /lib/libm.so.6 (0x00007f53c8d38000)
libgcc_s.so.1 => /lib/libgcc_s.so.1 (0x00007f53c8c2b000)
libc.so.6 => /lib/libc.so.6 (0x00007f53c89f0000)
/lib64/ld-linux-x86-64.so.2 (0x00007f53ca334000)
libXau.so.6 => /usr/lib/libXau.so.6 (0x00007f53c88ed000)
libXdmcp.so.6 => /usr/lib/libXdmcp.so.6 (0x00007f53c87e8000)
libz.so.1 => /lib/libz.so.1 (0x00007f53c86d4000)
libexpat.so.1 => /usr/lib/libexpat.so.1 (0x00007f53c84b0000)
libXfixes.so.3 => /usr/lib/libXfixes.so.3 (0x00007f53c83ab000)


As for my 8.9 problem, I suspect that may be something to do with the custom libdri that's apparently there now - I'll have to look into that a bit more.

-- As I suspected: /usr/lib64/xorg/modules/extensions/libdri.so had to be replaced by hand. amdcccle still doesn't work, but the rest of the driver does!

[wfeltmate]

Is anyone else having an issue running dual-head mode? I'm running on an HP dv5 laptop, which has an ati HD3200 chip and running my Samsung HDTV as a secondary monitor over an hdmi connection. Whenever I try to boot in to Ubuntu(8.04.1) fglrx crashes and BulletProofX loads.
I have no issue running in big desktop or clone mode, either booting with this setting or enabling it after my system has fully loaded.

[raulromania]

Is anyone else having an issue running dual-head mode? I'm running on an HP dv5 laptop, which has an ati HD3200 chip and running my Samsung HDTV as a secondary monitor over an hdmi connection. Whenever I try to boot in to Ubuntu(8.04.1) fglrx crashes and BulletProofX loads.
I have no issue running in big desktop or clone mode, either booting with this setting or enabling it after my system has fully loaded.

I've had the same issue with 8.8, but not with an notebook.
I have done the following steps and it worked.

[sudo] gedit /etc/X11/xorg.conf

Section "Device"
...
Driver "vesa"
...
EndSection

[sudo] apt-get remove --purge [all ati packages] && [sudo] rm -dfrv /etc/ati/
reboot

Then I reinstalled the ati packages and activate the fglrx driver but NO OTHER OPTION!!

Then reboot!
After reboot, I have activated the other fglrx options like ...

→ http://phoronix.com/forums/showthread.php?t=12727&page=5#50

The steps are not all logic, but it worked.

[balihb]

If you can't wait for debian to release the new package, than you can do this:
http://balihb.wik.is/Fglrx_8.9_on_Debian

[Heiko]

Is anyone else having an issue running dual-head mode? I'm running on an HP dv5 laptop, which has an ati HD3200 chip and running my Samsung HDTV as a secondary monitor over an hdmi connection. Whenever I try to boot in to Ubuntu(8.04.1) fglrx crashes and BulletProofX loads.
I have no issue running in big desktop or clone mode, either booting with this setting or enabling it after my system has fully loaded.

Well, on my dual-head system (AMD 3450: tv and tft monitor, two separate x-screens), OpenGL programs are not terminated. For example: glxinfo hangs (after outputting all info). An OpenGL program I wrote myself also hangs (after it was finished). Besides that, no really big problems.

My other computer (single monitor) does not have these problems.

btw: what is the best location to report bugs? I did report some bugs to the unofficial bugzilla, but is that ever read by the devs?

[aniruddha]

I can't modprobe fglrx with *.deb package build from ATI's *.run file:

# modprobe fglrx
FATAL: Module fglrx not found.

Any ideas where I should start to look for a solution? I run Debian testing x86_64

Update:
The regular drivers don't work for me either. I need to issue a:
module-assistant auto-install fglrx-kernel-src
to get 'modprobe fglrx' working. problem is when I try to do that with the latest drivers (which i manually build from ATI's *.sh), it failed because it tried to update to the 'latest' fglrx driver.

Any idea how to resolve the problems with version conflicts?

Update2
I managed to build the kernel-src but now I get this error :confused::
# modprobe fglrx
FATAL: Error inserting fglrx (/lib/modules/2.6.26-1-amd64/misc/fglrx.ko): Unknown symbol in module, or unknown parameter (see dmesg)

[Kano]

You can try my script:

http://kanotix.com/files/install-fglrx-debian.sh

but you need a supported kernel. You need the CONFIG_PCI_LEGACY=y config option set and you should have got kernel headers installed correctly. Use

sh install-fglrx-debian.sh -v 8-9

to try 8-9 driver, 8-8 is currently default because of the stupid debian etch dri bug.

[Chris Clearwater]

It seems my half-life 2 issue was actually a problem with Wine/Alsa/PulseAudio, not Catalyst 8.9, sorry for the false alarm!

[Almindor]

One misinformation for this: "fixes ETQW slowness and artifacts", NOT. It's still slow and I still get the artifacts with 8.9.

Combined with the GL regressions, I guess they "deliver as expected", making more bugs along the way.

Good work AMD, as always! (never, ever, ever, not even in windows do their drivers work right)

[energyman]

which GL regression? The two games I play (vegastrike, ut2004) got better in low fps situations (vegastrike) or no changes at all (ut2004)

[Almindor]

which GL regression? The two games I play (vegastrike, ut2004) got better in low fps situations (vegastrike) or no changes at all (ut2004)

See article here, about unigine test regressions: http://www.phoronix.com/scan.php?page=article&item=amd_catalyst_89_linux&num=1

[energyman]

oh that? I have seen distortions like that with 8.8 too - when I run it for the second time.

[aniruddha]

You can try my script:

http://kanotix.com/files/install-fglrx-debian.sh

but you need a supported kernel. You need the CONFIG_PCI_LEGACY=y config option set and you should have got kernel headers installed correctly. Use

sh install-fglrx-debian.sh -v 8-9

to try 8-9 driver, 8-8 is currently default because of the stupid debian etch dri bug.

You script yields exactly the same results:

# modprobe fglrx
FATAL: Error inserting fglrx (/lib/modules/2.6.26-1-amd64/misc/fglrx.ko): Unknown symbol in module, or unknown parameter (see dmesg)

[ 35.598690] fglrx: module license 'Proprietary. (C) 2002 - ATI Technologies, Starnberg, GERMANY' taints kernel.
[ 35.612610] [fglrx] vendor: 1002 device: 9505 count: 1
[ 35.613093] [fglrx] ioport: bar 4, base 0xc800, size: 0x100
[ 35.613097] [fglrx] Maximum main memory to use for locked dma buffers: 3808 MBytes.
[ 35.614345] [fglrx] PAT is enabled successfully!
[ 35.614345] [fglrx] module loaded - fglrx 8.51.3 [Jul 3 2008] with 1 minors
[ 35.622723] ACPI: PCI Interrupt 0000:05:00.0[A] -> GSI 16 (level, low) -> IRQ 16
[ 35.772804] RPC: Registered udp transport module.
[ 35.772804] RPC: Registered tcp transport module.
[ 36.632763] [fglrx] Reserved FB block: Shared offset:0, size:1000000
[ 36.632767] [fglrx] Reserved FB block: Unshared offset:ff7f000, size:80000
[ 51.825335] eth0: no IPv6 routers present
[ 767.350895] atieventsd[3249]: segfault at 10 ip 7f74920375bf sp 7fff9a50dca0 error 4 in libX11.so.6.2.0[7f7491fe8000+107000]
[ 937.149217] mtrr: type mismatch for d0000000,10000000 old: write-back new: write-combining
[ 966.096958] [fglrx] Disable PAT
[ 966.096976] [fglrx] asyncIODestroy finished!
[ 966.097030] [fglrx] module unloaded - fglrx 8.51.3 [Jul 3 2008]
[ 968.926298] fglrx: Unknown symbol flush_tlb_page
[ 1011.044485] fglrx: Unknown symbol flush_tlb_page
[ 1046.575856] fglrx: Unknown symbol flush_tlb_page
[ 1048.566838] mtrr: type mismatch for d0000000,10000000 old: write-back new: write-combining
[ 1264.372089] fglrx: Unknown symbol flush_tlb_page
[ 1490.082945] fglrx: Unknown symbol flush_tlb_page
[ 6812.894912] fglrx: Unknown symbol flush_tlb_page
[ 6814.909401] mtrr: type mismatch for d0000000,10000000 old: write-back new: write-combining
[ 6958.821938] atieventsd[26673]: segfault at 10 ip 7f3f538a15bf sp 7fff5bd7a5e0 error 4 in libX11.so.6.2.0[7f3f53852000+107000]
[ 6959.432850] fglrx: Unknown symbol flush_tlb_page
[ 6961.421042] mtrr: type mismatch for d0000000,10000000 old: write-back new: write-combining
[ 7208.562554] fglrx: Unknown symbol flush_tlb_page
[ 7210.550816] mtrr: type mismatch for d0000000,10000000 old: write-back new: write-combining
[ 7255.927186] fglrx: Unknown symbol flush_tlb_page
[ 7279.044878] fglrx: Unknown symbol flush_tlb_page
[ 7281.031710] mtrr: type mismatch for d0000000,10000000 old: write-back new: write-combining
[ 7319.673089] fglrx: Unknown symbol flush_tlb_page

[Kano]

Did you use -v 8-9? Maybe you can try -v 8-7 too.

[nphoronix]

Anyone still having problems with Xvideo? I still get video flickering and video tearing (ATI 3870).

Any plans of actually FIXING those problems? I've been trying all sorts of xorg.conf I could find online but the problem's still there, so I doubt it's a xorg.conf misconfiguration.

[Vighy]

Anyone still having problems with Xvideo? I still get video flickering and video tearing (ATI 3870).

Any plans of actually FIXING those problems? I've been trying all sorts of xorg.conf I could find online but the problem's still there, so I doubt it's a xorg.conf misconfiguration.

If i'm right, the best xorg.conf is the minimal one :D doing the most throught the driver db... so xorg.conf shouldn't be that much important right now.

[aniruddha]

Did you use -v 8-9? Maybe you can try -v 8-7 too.

Yep, I used this exact command:

./install-fglrx-debian.sh -v 8-9


8-7 works but I want 8-9 for the Wine fix ;)

[Kano]

With your kernel you can use -v 8-5 too. that Has got no wine problems.

[nphoronix]

If i'm right, the best xorg.conf is the minimal one :D doing the most throught the driver db... so xorg.conf shouldn't be that much important right now.

LOL, true, that's what I'm usually using anyway. Although I have the xv overlay enabled in xorg.conf. As it is, I CAN have flicker-free/tear-free playback using opengl and disabling compiz, however, I'd like the xv extension fixed and not having to disable compiz all the time.

[piquadrat]

Unfortunately, this driver butchered my dual screen setup. No problem, I thought, let's try out that fancy new RandR 1.2 support. Well, apparently, Firefox didn't like that:

http://img502.imageshack.us/img502/874/ffot7.th.png (http://img502.imageshack.us/my.php?image=ffot7.png)http://img502.imageshack.us/images/thpix.gif (http://g.imageshack.us/thpix.php)

All other applications don't have this problem, not even Thunderbird. Maybe it's related to the cairo rendering backend.

I hope AMD will fix this "little" rendering glitch before officially announcing randr1.2 support :)

[dscharrer]

I have now opened bug reports on ati.cchtml.com for the new bugs I get with Catalyst 8.9:
Alt+Tab from fullscreen Wine 3D app broken (http://ati.cchtml.com/show_bug.cgi?id=1278)
Killing fullscreen 3D app from VT-1 corrupts screen (http://ati.cchtml.com/show_bug.cgi?id=1279)
windows not redrawn when exiting fullscreen Kaffeine (OpenGL output) (http://ati.cchtml.com/show_bug.cgi?id=1280)

[Gordy]

Is anyone seeing random GCC crashes (other then me) when compiling wine-1.0 from source (or other large projects) with the new 8.9 catalyst driver installed? I can confirm this does not happen on my config when the 8.9 catalyst driver is removed and the fglrx 8.37.6 driver is installed. Here's my hw/sw config:

SAPPHIRE ATI 9600XT 128MB AGP (8x)
CRT 19" monitor: Philips 109B (109B10) with a resolution of 1600x1200 or 1280x1024
Motherboard: ASUS K8V SE Deluxe/ VIA K8T800 & VIA VT8237
Processor type: AMD64 3000+
Amount of memory: 1GB
Distro used: Ubuntu 7.10 Gutsy Gibbon (32-bit) with latest patches applied.

[schinkelm]

Hello,

2D remains slow with this version on 3870 and dual-head, X Server 1.4.2 (with AIGLX).

I tried all combinations of driver options in xorg.conf and even updated to kernel 2.6 27-rc7 (needed to patch firegl_public.c).

Especially Firefox is horribly slow. This affects (flawlessly) running video playback (via mplayer/opengl): The image stands still while scrolling. I observed very high cpu utilization on one core. Same problem with many other apps.

OpenGL works without problems.
Are there any helping global X server tweaks?
Is this a known problem? Is ATI working to solve it?

Best Regards,

Marco

[aniruddha]

With your kernel you can use -v 8-5 too. that Has got no wine problems.

I just tried installing 8-5 with your script, but I've got into all sorts of troubles:

1st Instead of Debian/testing it installed a ubuntu version :confused:
2nd I can't remove this version:

# apt-get remove xorg-driver-fglrx
Reading package lists... Done
Building dependency tree
Reading state information... Done
The following packages were automatically installed and are no longer required:
...
Use 'apt-get autoremove' to remove them.
The following packages will be REMOVED:
xorg-driver-fglrx
0 upgraded, 0 newly installed, 1 to remove and 35 not upgraded.
1 not fully installed or removed.
After this operation, 44.0MB disk space will be freed.
Do you want to continue [Y/n]?
(Reading database ... 243339 files and directories currently installed.)
Removing xorg-driver-fglrx ...
dpkg-divert: mismatch on divert-to
when removing `diversion of /usr/X11R6/lib32/libGL.so.1 to /usr/X11R6/lib32/fglrx/libGL.so.1.2.xlibmesa by xorg-driver-fglrx'
found `diversion of /usr/X11R6/lib32/libGL.so.1 to /usr/X11R6/lib32/fglrx/libGL.so.1.xlibmesa by xorg-driver-fglrx'
dpkg: error processing xorg-driver-fglrx (--remove):
subprocess post-removal script returned error exit status 2
Errors were encountered while processing:
xorg-driver-fglrx
Updating Debian Packages of System Configurations (dpsyco).
E: Sub-process /usr/bin/dpkg returned an error code (1)

[Kano]

Well you can remove it with the script as it has the hotfix integrated. If you want to do that manually do:

sed -i 'sX> /dev/nullX||trueX' /var/lib/dpkg/info/xorg-driver-fglrx.postrm

My script downloads dkms (updated url yesterday, before it might be broken for Debian) then the Ubuntu debs can be used. Latest script can use -v8-9 on 64 bit already. Working on the rest. dkms is nice when you switch kernels, then you always get the correct fglrx module - also you can trigger patches based on kernel version.

[aniruddha]

Well you can remove it with the script as it has the hotfix integrated. If you want to do that manually do:

sed -i 'sX> /dev/nullX||trueX' /var/lib/dpkg/info/xorg-driver-fglrx.postrm

My script downloads dkms (updated url yesterday, before it might be broken for Debian) then the Ubuntu debs can be used. Latest script can use -v8-9 on 64 bit already. Working on the rest. dkms is nice when you switch kernels, then you always get the correct fglrx module - also you can trigger patches based on kernel version.

How do I use your script to remove drivers? (the help function doesn't show this).

[Kano]

Well you can not directly remove only, but when you install another driver the old one is uninstalled.