Package Details: catalyst-test 15.12-8

Git Clone URL: https://aur.archlinux.org/catalyst-test.git (read-only)
Package Base: catalyst-test
Description: AMD/ATI Catalyst drivers for linux AKA Crimson. catalyst-dkms + catalyst-utils + lib32-catalyst-utils + experimental powerXpress suppport. PRE-GCN Radeons ARE NOT SUPPORTED
Upstream URL: http://www.amd.com
Licenses: custom
Conflicts: catalyst, catalyst-daemon, catalyst-dkms, catalyst-generator, catalyst-hook, catalyst-utils, libcl, libgl, mesa-libgl, mesa-libgl-git
Provides: catalyst=15.12, catalyst-hook=15.12, catalyst-libgl=15.12, catalyst-utils=15.12, dri, libatical=15.12, libcl, libgl, libtxc_dxtn, mesa-libgl, mesa-libgl-git, opencl-catalyst=15.12
Submitter: Vi0L0
Maintainer: Vi0L0
Last Packager: Vi0L0
Votes: 167
Popularity: 0.282200
First Submitted: 2010-02-17 20:49
Last Updated: 2016-07-25 20:37

Dependencies (22)

Required by (688)

Sources (26)

  • 4.3-gentoo-mtrr.patch
  • 4.3-kolasa-seq_printf.patch
  • 4.4-manjaro-xstate.patch
  • 4.6-arch-get_user_pages-page_cache_release.patch
  • 4.7-arch-cpu_has_pge.patch
  • arch-fglrx-authatieventsd_new.patch
  • ati-powermode.sh
  • atieventsd.service
  • atieventsd.sh
  • catalyst.conf
  • catalyst.sh
  • crimson_i686_xg.patch
  • dkms.conf
  • fglrx_gpl_symbol.patch
  • grsec_arch.patch
  • http://www2.ati.com/drivers/linux/radeon-crimson-15.12-15.302-151217a-297685e.zip
  • lano1106_fglrx_intel_iommu.patch
  • lano1106_kcl_agp_13_4.patch
  • lib32-catalyst.sh
  • makefile_compat.patch
  • makesh-dont-check-gcc-version.patch
  • pxp_switch_catalyst
  • switchlibGL
  • switchlibglx
  • temp-links-catalyst.service
  • temp_links_catalyst

Latest Comments

Vi0L0 commented on 2016-07-29 07:08

In amdcccle I'm always disabling v-sync as well, it's not needed when using tear free.
There's also simple tear free switcher which allows switching via console (it comes with desktop entry): https://aur.archlinux.org/packages/amdtfset-git/

Also keep in mind that working composite has negative impact on performance of fullscreen rendering apps. I'm always disabling this as well before running app/game - in KDE its easy with Alt+Shift+F12 shortcut

Vi0L0 commented on 2016-07-25 20:45

@irton: ah yes, will fix it later silently, thanks

for now I added 4.7 kernel support, it seems to work fine with my amateur patch.

errors that were generated without 4.7 patch were like (could be useful for searchers):
-----8<-------
/var/lib/dkms/fglrx/15.12/build/2.6.x/firegl_public.c: In function ?kcl_mem_pat_setup?:
./arch/x86/include/asm/bitops.h:337:29: error: passing argument 2 of ?constant_test_bit? from incompatible pointer type [-Werror=incompatible-pointer-types]
? constant_test_bit((nr), (addr)) \
^
/var/lib/dkms/fglrx/15.12/build/2.6.x/firegl_public.h:654:21: note: in expansion of macro ?test_bit?
#define cpu_has_pge test_bit(X86_FEATURE_PGE, &boot_cpu_data.x86_capability)
^~~~~~~~
/var/lib/dkms/fglrx/15.12/build/2.6.x/firegl_public.c:4547:9: note: in expansion of macro ?cpu_has_pge?
if (cpu_has_pge)
^~~~~~~~~~~
./arch/x86/include/asm/bitops.h:308:28: note: expected ?const volatile long unsigned int *? but argument is of type ?__u32 (*)[19] {aka unsigned int (*)[19]}?
static __always_inline int constant_test_bit(long nr, const volatile unsigned long *addr)
^~~~~~~~~~~~~~~~~
./arch/x86/include/asm/bitops.h:338:29: error: passing argument 2 of ?variable_test_bit? from incompatible pointer type [-Werror=incompatible-pointer-types]
: variable_test_bit((nr), (addr)))
^
/var/lib/dkms/fglrx/15.12/build/2.6.x/firegl_public.h:654:21: note: in expansion of macro ?test_bit?
#define cpu_has_pge test_bit(X86_FEATURE_PGE, &boot_cpu_data.x86_capability)
^~~~~~~~
(...)
-----8<-------

irton commented on 2016-07-04 05:29

script make.sh may be with execute permission? eq. 755

Vi0L0 commented on 2016-05-30 18:42

15.12-7: fixed dkms versioning
Also noticed:
- no more wh problems on cs:go on 4.5 and 4.6 kernels - probably valve fixed it;
- for some reason glxgears32 segfaults on my pc (no matter of what catalyst version, didnt check oss amdgpu/radeon):
'glxgears32' terminated by signal SIGSEGV (Address boundary error)
will check it later;
- tomorrow I should be able to update repos, at least [catalyst]

Vi0L0 commented on 2016-05-13 20:55

thanks again :) so it stores modules in two different places... oukaaaay 8)
Arch/dkms wiki should definitelly have more details

runnytu commented on 2016-05-13 20:26

@Vi0L0, no problem we are all humans after all ;P, the best way to clean is 'rm -fr /var/lib/dkms/fglrx' and to install the module for all the installed kernels with 'dkms autoinstall -m fglrx-15.12 -v 6' later with 'dkms status' should see a output like this for all installed kernels, 2 in my case:
fglrx-15.12, 6, 4.5.4-1-ARCH, x86_64: installed
fglrx-15.12, 6, 4.5.4-1-zen, x86_64: installed

Vi0L0 commented on 2016-05-13 19:48

you got the point runnytu, nice finding

I have removed dkms commands from install script, should be fine now.
Previous version was taken from original catalyst-dkms version and it didn't fit arch's newer dkms.
It was creating fglrx.ko module for booted kernel (with "dkms -v fglrx"), then it was overwritten (because defined module name is the same) by dkms itself (with "dkms -m fglrx-15.12").

I'm just not sure how is dkms going to find out about overwritten module (as it was generated using different version, could generate some trash from it that will have to be removed manually. Sorry for the inconvinience.
Possibly the best way to clean after -5 version is to:
1. use `dkms remove -m fglrx-15.12 -v 5 --all` or `dkms remove -m fglrx-15.12 -v 6 --all` after update to -6
2. use `find /usr/lib/modules/ -iname '*fglrx*'` to trace trash and use rm to remove it manually
3. use `dkms install -m fglrx-15.12 -v 6` to build module for booted kernel or reinstall catalyst-test if you got more kernels - I don't know simple way to build modules for all kernels. One can use some loop like `for` for it, but there should be some easier way available.
But even without cleaning it should work fine, eventually there will be some fglrx module trash left after kernel's version bump ;P.

Thanks again runnytu.
And sorry for missing it.

runnytu commented on 2016-05-12 18:04

Hi, @Vi0L0, you should remove the 'dkms install -m fglrx -v $1' in post_install() and post_upgrade() and the 'dkms remove -m fglrx -v $1 --all' in pre_remove() in the file catalyst-test.install because pacman installs and removes the DKMS modules automatically at the beginning, to remove and at the end to install, with this commands installs the DKMS modules twice, one with your scipt, named fglrx that fails sometimes and the automatic pacman, named fglrx-15.12 that works fine, could you check it?

rg_arc commented on 2016-05-12 16:48

@Vi0L0 Good deal. As always keep up the good work, I'll stick around.

Vi0L0 commented on 2016-05-12 16:22

definitelly not too old, phoronix said that it might be soon supported by amdgpu: http://www.phoronix.com/scan.php?page=news_item&px=AMD-SI-Might-Soon-Get-AMDGPU
For now use radeon, it should be good and stable.

rg_arc commented on 2016-05-12 15:02

@Vi0L0 I've used the open sourced one before as well and I think there was something minor that annoyed me, I can't remember at the moment. This is my current GPU: https://ptpb.pw/9tri is this under the "too old" category?

Vi0L0 commented on 2016-05-12 12:54

I keep switching from one to another. It's both safe and easy :). In fact I think that everyone should check out both.
Cs:go should work well on open driver, thought on mine 290x it's more fluid on fglrx.
For older cards pick up radeon, for newer amdgpu, it's all in the wiki.
For open I personally preffer lordheavy's [mesa-git] repo which comes with the newest git drivers + linux-mainline kernel (also availabe on miffe's repo).
BTW: just a note, you can always have linux-lts installed next to linux and switch between them, don't have to downgrade
Edit: and no, I heven't seen such wallhack glitches on open drivers.

rg_arc commented on 2016-05-12 12:17

@Vi0L0 I hate to ask, but would it be safe to say switching over to the open sourced driver would fix this issue? I'm not downgrading my kernel for CS:GO.

Vi0L0 commented on 2016-05-12 11:50

it doesn't solve this issue at all. It's still there for both 4.5 and 4.6. For now I don't know how to fix it.
The only workaround for now is to use kernel < 4.5, in example linux-lts (now in 4.4).

rg_arc commented on 2016-05-12 11:38

@ViOLO I am on Linux Kernel 4.5.3-1 and playing CS:GO certainly still has the wallhack glitches. However, does this new catalyst-test release solve that issue only for 4.6 or for 4.5 as well? CS:GO is completely unplayable with this ridiculous glitch.

Vi0L0 commented on 2016-05-11 21:18

updated to 15.12-5:
- initial 4.6 kernel support (tested on rc6 and rc7) thanks to trivial patch, like on 4.5 kernel cs:go still generates wallhack glitches
- switched from hook to dkms, dkms seems to work really well since pacman's hooks so this seems to be a good idea
- used acpi script (ati-powermode.sh) from aslmaswd

enjoy, test and let me know if it works or not


@herycp:
ad 1. turn off v-sync or/and tear-free

herycp commented on 2016-04-02 12:31

why my glgears :
1. max just about 60 FPS
2. fatal IO error 11 (Resource temporarily unavailable) on X server ":1.0"
after 997 requests (997 known processed) with 0 events remaining.


catalyst 15.12-3
vga 0=HD 7640G
1=HD 7470M

Xserver= 1.17.4-2 (repo extra)

afulm commented on 2016-03-26 09:02

Sorry, the fault lies with my system, not the package.

afulm commented on 2016-03-25 23:09

Ah, it also says:

==> WARNING: lib32-catalyst-utils will NOT be added to the package
and
:: steam: installing catalyst-test (15.12-3) breaks dependency 'lib32-libgl'

Vi0L0 commented on 2016-03-25 08:33

This warning has always been there, it's more like a note.
The update only bumped the package number and a kernel version, should not cause any regression.
Which deps are missing?

afulm commented on 2016-03-25 02:38

Strangely, trying to buld the latest update, it says:

==> WARNING: [multilib] repository must be uncommented in /etc/pacman.conf to add lib32-catalyst-utils into the package

though multilib is uncommented. Makes me unable to satisfy dependencies.

Vi0L0 commented on 2016-03-23 21:24

been using 4.5 kernel since rc5 on catalyst, but decided to make some game testing before commiting 4.5 support.
There was no problems except for cs:go which generates some glitches - missing walls :), it could be used as a wallhack, literally, not sure if it is bug or a feature :P. Occures on catalyst 15.9 and 15.12, catalyst-firepro works well.
So if you dont like this bug (could be that im the only one touched by problem) use older linux 4.4 or linux-lts or mentioned catalyst-firepro

awerush commented on 2016-03-18 22:48

This just in: http://support.amd.com/en-us/kb-articles/Pages/AMDGPU-PRO-Beta-Driver-for-Vulkan-Release-Notes.aspx

niqingliang2003 commented on 2016-02-13 06:58

my card is 6870, using hdmi interface with xf86-video-ati.
lcd is LG 2560x1080, but can't get 60HZ.
current mode is 2560x1080x30hz.
on windows 2560x1080x60hz.

any idea?

zer0def commented on 2016-02-02 23:52

@Vi0L0: has AMD worked out their issues with i686 in catalyst>15.9?

Vi0L0 commented on 2016-01-20 10:49

xserver support is hidden inside binary part of the driver.
Bringing such support could be done by the community but it involves skilled asm hacker and a lot of his free time - till now it never happened.
In other words we are patiently waiting for AMD to move their letters and bring it. I believe they already have such support for months and testing it on their secret beta channel.

JeansenVaars commented on 2016-01-18 22:27

I am curious on why xorg 1.18 is not supported, appreciated if anyone can shed some light

Vi0L0 commented on 2016-01-15 07:16

I would start with looking at logs: Xorg.0.log or journalctl + eventually ~/.xsession-errors if black screen happens when logging to DE

dualcorex2 commented on 2016-01-13 15:30

I seem to have the same problem as @horsemanoffaith, but with KDE Plasma 5. Not sure what to do.

Vi0L0 commented on 2016-01-04 07:41

Maybe installing patched cogl will help you:
i686: http://catalyst.wirephire.com/repo/catalyst/i686/cogl-1.22.0-2.5-i686.pkg.tar.xz
x86_64: http://catalyst.wirephire.com/repo/catalyst/x86_64/cogl-1.22.0-2.5-x86_64.pkg.tar.xz

horsemanoffaith commented on 2016-01-04 07:24

I just installed Catalyst-test on my laptop, which is running a HD6450M card. Seems to work fine. My only issue is that when I would start Cinnamon, all I would get is a black screen. I used R4v3nPr0's terminal commands in fluxbox, then logged into Cinnamon, and it works. Is there a fix for this, possibly a patch or a script that I can have my system run on start-up? I'd surely prefer not to have to enter those commands every time I want to use Cinnamon.

eggz commented on 2015-12-29 22:01

I am running this on a kabini (AM1) athlon 5150 APU with radeon r3 graphics (I know, it's a mouthful) And I get sporadic drivercrashes (like every few days)

I am using this APU fairly intensively (imo) on a TV and another screen, The TV has kodi (using VAAPI AND VDPAU acceleration -- I actually got both working afaik, Dont know which one kodi uses) and the other one has another X desktop with browsers and terminals. So its kind of a busy enviroment. Kodi is always the program making the driver crash (ofc while playing); last time I clearly spotted a segfault there. Probably in reaction of the drivercrash though.

I do have to say, this driver tackles the tearing much more handsomely than the OS driver.

Well, if someone is intrested, here is my crashdump: http://pastebin.com/fH7Ae54z

In any case, keep up the good work. My small APU is very dependent of good driver optimalisation and this here is the closest thing currently available.

Cheers.

Rikudark commented on 2015-12-29 13:15

I have installed latest version (15.12) using an HD5500 and it seems to work properly. Previous versions worked properly too on my system. Will edit this post if any error is found.

Pentium320 commented on 2015-12-23 02:44

That could be weird. I know, this package doesn't support pre-GCN cards now, but I wanted to test before switching to "total"... and it seems to be working, at least for my card.
Previous version (15.11) indeed didn't want to work - aticonfig said: no supported adapters detected. But on this version (15.12) it works again.
I'm using Mobility HD5650.
Can somebody confirm that, or am I the only one?

Vi0L0 commented on 2015-12-19 21:18

ah, i plan [catalyst] repo update tomorrow
edit: nope, sorry didn't make it, will update tomorrow (Monday)

Vi0L0 commented on 2015-12-19 21:16

updated to 15.12:
- not much changes to see on a naked eye, many files has been recompiled, so maybe you will notice some fixes, portal 2 should work better with new profile
- didn't play Dreamfall Chapter before on Linux, but darn how nice, fluid port is is
- still no xserver 1.18 support
- added initial 4.4 kernel support - thanks to Philip Müller @ Manjaro
- added testing grsec support

at least on kernel side it looks good:
# catalyst_build_module all
+ building fglrx module for 4.1.13-rt15-1-rt kernel ...
ok.
+ building fglrx module for 4.1.15-1-lts kernel ...
ok.
+ building fglrx module for 4.2.5-1-ARCH kernel ...
ok.
+ building fglrx module for 4.3.3.201512162141-1-grsec kernel ...
ok.
+ building fglrx module for 4.4.0-rc5-mainline kernel ...
ok.
:)

Vi0L0 commented on 2015-12-14 16:51

wait, whaa? :)
this package, catalyst-test, is a complex solution, if you will uncomment multilib repo and build this package it will have "built-in" lib32-catalyst* packages, or in other words it will contain the same same data as lib32-catalyst* packages.
This package is in version 15.11 and here AMD brought fixes so we are now able to build fglrx modules on gcc5.

Now when it comes to 15.9 - I borrowed those 15.11 fixes and yesterday implemented it to 15.9 packages, that's why there's no gcc49 dependency :)
(btw 15.9 has nothing to do with this catalyst-test 15.11 package)

oljik commented on 2015-12-14 01:34

oh, sorry to bother about it. I just realised that I can use lib32-catalyst-libgl 15.9 with it. After realising its and catalyst-libgl's dependencies. The number 15.9 was confusing to me.
And, when did you remove gcc49 dependecy?:)) I was waiting for it.
thank you for your work.

oljik commented on 2015-12-14 00:42

lib32-catalyst-libgl

Vi0L0 commented on 2015-12-12 20:28

@oljik: 32 bit libs? be more precize please

@nevax: no need to compilate things like this, it should be as simple as adding [xorg117] repo above other repos in /etc/pacman.conf and performing pacman -Syuu

@mnovick1988: I never seen community patches which brought ABI's compatibility for newer xserver, sounds like a hardcore low-level job

oljik commented on 2015-12-12 13:49

what about 32 bit libs? Is there any success?

nevax commented on 2015-12-11 23:47

Check here : https://wiki.archlinux.org/index.php/AMD_Catalyst#xorg117

I have also made some note for my personal use : https://drive.google.com/file/d/0BwwtZc6nm4zzTWRqOVJGX0NCZDQ/view?usp=sharing

mnovick1988 commented on 2015-12-11 23:19

now there is a real issue... since arch went to xserver(1.18x), the only way to use xserver>=1.17 is to either use a mirror or compile it yourself...

Thanks AMD... - any patches for 1.18 yet?

alpargyenge commented on 2015-12-09 09:59

Thank you for your work, first i have installed the stable catalyst and i have artefacts, now with catalyst-test artefacts gone ,everithyng work great.

Anonymous comment on 2015-11-27 19:22

Thank you, but i don't think this could help me. I installed driver by this package and driver works without problems. When i switch card to Radeon, restart X and show informations about graphics by glxinfo, it shows me vendor AMD, and OpenGL version 4.5(in Mesa drivers i have only 3.0). I can switch card by amdcccle(as i mention before, aticonfig not working). Driver works, but it performace is very bad. In some games i have worse perfomance than in Intel card. So this is problem: I can install driver and driver work, but i can't get full performance. In one game i have 10FPS in Intel card and above 30FPS in open source Radeon Driver. In fglrx Crimson i have only 6-7 FPS. Problem isn't exist in installation, so i don't think i can solve this by installing driver from AMD page.

R4v3nPr0 commented on 2015-11-27 18:54

Hi, as I promised this is the guide that I mentioned that lets you install Crimson even when it says your video card is unsupported.

If your video card is supported by Catalyst, hence, your video card is supported by Crimson. The only thing you have to do is extract the .run file and inside all extracted files, there one that are named detect_gpu.sh, open that file and modify the last line from exit $status to exit 0, with this you force Crimson to install even when it says your video card is unsupported.

Note that after installation you can’t use aticonfig because it says that your video card is unsupported, so you need to use your xorg.conf generated by aticonfig from Catalyst 15.9.

Anonymous comment on 2015-11-27 16:28

I don't know how i can make this work. I've removed fglrx driver and installed it again(by this package). Again, aticonfig shows me error: No supported graphics adapters found. I've configured xorg.conf manually and driver work, but again, perfomance is very bad(before update to Crimson i had above 1000FPS in fgl_glxgear, now i have max.100FPS, games works slower than before too). I don't know what i have to do to make fglrx work with good performance. I'm tired of this, i only want to play games with good performance. Now, what i should do?

R4v3nPr0 commented on 2015-11-26 18:07

@spr1n

If you install the package is in AUR, video cards that are supposedly not supported, they will work properly, but the guide I'll post is intended for all users of all distributions in general who need to use the installer that comes with the driver.

spr1n commented on 2015-11-26 17:41

Nice to see tty bug finally being addressed.

Also Radeon HD7670M is working just fine with Crimson driver without any additional tweaking.

Anonymous comment on 2015-11-26 16:34

@R4v3nPr0
Thank you, i will check your guide.

R4v3nPr0 commented on 2015-11-26 16:10

Hello @kamils096, your video card is supported by Crimson, in a few hours I will upload a guide of what to do to install Crimson for a video card it says it does not support.

Anonymous comment on 2015-11-26 15:38

@nevax
Thank you for answer, but your advice didn't help. I think i found a problem.

It seems Radeon HD7670M isn't supported by Crimson. When i navigate to Linux download page and show list of supported product there is:
"AMD Radeon™ HD 7000M Series"

but on Windows download page there is:
"AMD Radeon™ HD 7700M+ Series​"

I think AMD didn't update list for Linux. If HD7670M isn't supported in Windows, why AMD would still support it on Linux? Windows has higher priority than Linux for AMD.

R4v3nPr0 commented on 2015-11-26 05:31

Hello everyone, today I managed to install the Crimson AMD driver for a video card "HD 6770". To no longer, and more complicated this comment, I discovered that the driver module supports all video cards driver 15.9 support, but to achieve install the Crimson driver for a video card that supposedly does not support you have to change a text file. Tomorrow I will publish a guide on how I did it, so that users of video cards, AMD 5000/6000 have an opportunity to install the driver Crimson. And all users of video cards that supposedly are not supported by Crimson, but they are supported by Catalyst 15.9.

nevax commented on 2015-11-25 21:55

@kamils096
Try this :
https://wiki.archlinux.org/index.php/AMD_Catalyst#.22aticonfig:_No_supported_adapters_detected.22

-----

I have a problem on my GNOME with this package on 2 systems (one freshly installed).

Just here :
http://image.noelshack.com/fichiers/2015/48/1448493038-sans-titre.png
&
http://image.noelshack.com/fichiers/2015/48/1448493565-selection-001.png
Not evrywhere.

The "t" letter is missing.

Thanks for this package @Vi0L0

Anonymous comment on 2015-11-25 21:27

Hi everyone!

When i installed new drivers(Radeon Crimson) i have problems:
1. aticonfig not working. When i try to do something it shows me:
aticonfig: no supported graphics adapters found
before this upgrade everything was fine.

I have Radeon HD7670M, and this card is supported i think. I checked this on AMD site. I have hybrid graphics, so i can't change graphics card. I can do this from amdcccle, but there is another problem.

2. Performance is very bad. In fgl_glxgears and glxgears i have less FPS than in Intel card. For example: Before upgrade in glxgears i had more than 2000FPS. Now i have max. 250FPS. Games work slower too - in Kerbal Space Program i have lower FPS than in Intel card(with Open Source drivers and with previous catalyst driver i had a lot more FPS).

Did i miss something? English isn't my first language, so please excuse any mistakes.

Thank you in advance for answer.

Vi0L0 commented on 2015-11-25 19:48

thanks mate, 15.11-3 brings updated info about this fact.
I will soon updated catalyst-total @ AUR to the latest 15.9 version with linux 4.2 and 4.3 support.

Also my plan for repos is:
1. move [catalyst] to [catalyst-stable] repo - the only difference at this time is that [catalyst] supports 4.2 and 4.3 kernels and is using gcc49
2. place crimson inside [catalyst] repo so it could stay there for testing purposes for some time
...
3. move [catalyst-stable] to the new legacy repo called [catalyst-hd56k]
4. if there will be no problems with crimson move it to [catalyst-stable]

K5HV commented on 2015-11-25 19:13

Since update to 15.11-2 from 15.9-6 my 58XX cards are not OpenCL capable anymore.
It's my OCL ring so I don't have any GUI to check (Xorg only because AMD cards need it).
It seems than new version of Catalyst/Crimson dropped support for 5xxx and 6xxx cards for real.

Vi0L0 commented on 2015-11-25 18:16

I can't guarantee that it will support xserver 1.18

Dea993 commented on 2015-11-25 18:15

@ViOLO
so we must wait next update of crimson for xorg 1.18 support?

Vi0L0 commented on 2015-11-25 16:55

I can now confirm that it segfaults on xserver 1.18

Vi0L0 commented on 2015-11-25 16:31

Didn't miss, just haven't seen user with such a problems.
Release notes claims it does support 5 and 6 k, but it could be total BS.

So is it really not working on those cards?

R4v3nPr0 commented on 2015-11-25 15:37

@Vi0L0
You missed add to the package description, that the 5000 and 6000 series are not supported, and apparently some or all of the APU Mobile 8000G series are not supported.

I propose that a legacy Catalyst 15.9 package is created, similar to catalyst-total-hd234k.

Vi0L0 commented on 2015-11-25 09:08

will confirm no xserver 1.18 support later, didnt have time yesterday for that, and wanted to release it asap.
as for real crimson, dunno but even tests on windows shows no or tiny improvement.
It doesn't support 4.3 kernel, but at least have 4.2 + compiles on gcc 5 (well on i686 only 4.1 because AMD clearly forgot about this architecture) and fixed xserver 1.17 (no tty bug).

Also hey it fixed some legendary bugs like stuttering in glxgears and mouse cursor corruptions - maybe they decided it deserves new naming and all this marketing hype? xD

Dea993 commented on 2015-11-25 08:32

no support for xorg 1.18?
i hope that the real crimson for linux is not yet arrived (old interface, poor improvments of performance, no supporto for last xorg, no support for last kernel)

Vi0L0 commented on 2015-11-25 06:24

It needs 2 additional patches to compile on i686 - fglrx_gpl_symbol.patch and crimson_i686_xg.patch.
I am not able to perform tests on i686, could some user confirm if its working fine?

x86_64 testing shows:
- official support for up to 4.2 kernel, and we don't need gcc-4.9, plus thanks to community patches it's working on linux 4.3
- tty bug is fixed
- we still need patched cogl for GNOME/GDM - but it seems pretty normal...
- i can still see glitches when running video through vaapi on plasma desktop
- opengl performance is similar to 15.9, but ie. Steam's Bioshock Infinite was working poorly and now on 15.11 it's pretty good

Vi0L0 commented on 2015-11-24 21:43

@francoism90:
take a look at https://bbs.archlinux.org/viewtopic.php?pid=1581133#p1581133
new 15.11 is not that bad, except for i686...

Vi0L0 commented on 2015-11-24 21:41

Almost updated it, but after tests on x86_64 found that module compilation is failing on i686 :( Also on kernels that were mentioned by AMD as supported... like 4.2
Had no time to fix i686, maybe will do this tomorrow, so for now source for x86_64:
http://mirror.rts-informatique.fr/archlinux-catalyst/tarball/15.11/catalyst-test-15.11-1.src.tar.gz

francoism90 commented on 2015-11-24 21:41

@R4v3nPr0: Yeah, you're better of with an Intel GPU, unfortunately they are not that powerful. :/

Also Crimson is really a joke (e.g. still no Linux 4.* support); it's just a rewrite of the CP/GUI, no other fixes done.

I really like all the devs like Vi0L0, patching stuff, etc. - it's a shame they don't get any support from AMD. :(
AMD should release their sources, Christmas gift maybe? :P

R4v3nPr0 commented on 2015-11-24 17:14

Crimson is a piece of shit, I have an AMD 8510G, which is a video card with at least 2 years after being released, nor is supported by AMD Crimson.

Hot4ruh1me commented on 2015-11-24 14:15

Hey Vi0L0,
This just got out. http://support.amd.com/en-us/download/desktop?os=Linux+x86_64
To notice:
1 - "this release is finally building against the latest Linux 4.x kernels. (...)However, no word if this release supports the latest Linux 4.3 kernel nor if it can handle the fresh X.Org Server 1.18."
2 - "Crimson on both Linux and Windows has gone away with pre-GCN GPU support. The Radeon HD 5000/6000 series are no longer supported by the mainline Radeon Software (Catalyst) driver, leaving you with Catalyst 15.9 driver as the last official update."

Source: http://www.phoronix.com/scan.php?page=news_item&px=AMD-Crimson-Pre-GCN

Vi0L0 commented on 2015-11-23 10:10

I didn't hear about possibility to patch fglrx so it could support amdgpu.
I think it have to be done on AMD side.

Also how is kernel module related with xserver compatibility?

Dea993 commented on 2015-11-23 09:13

xf86-video-amdgpu has been released on repository.
is possible patch the catalyst to use the new kernel driver (amdgpu) bypassing problems with xorg 1.18 and kernel version?

Katsutoshi commented on 2015-11-17 22:05

Until now, no. Its very strange. When i reload cinnamon, i can see the desktop. Maybe this is a misconfiguration of xorg.conf/20-radeon.conf?

But now i installed Plasma Desktop and now everything is working until now. Later i will try to install Cinnamon again.

Thank you Vi0L0 for the help!

EDIT: Tested the R4v3nPr0 method and worked 100%. Thank you too R4v3nPr0 !

Vi0L0 commented on 2015-11-17 22:00

Did you install patched cogl?
Or maybe tried R4v3nPr0's method (can't confirm if it's working, didn't test it)?

Katsutoshi commented on 2015-11-17 21:33

There's any problem with LXDM and Cinnamon? I know this is not a support topic, but i managed to install the drivers but cinnamon keeps in a black screen (but LXDM shows normally).

Katsutoshi commented on 2015-11-17 19:34

Oh, sorry. Since i'm a newbie, i ended reinstalling the entire the system because i messed up to return to opensource.

But what i remember it was a lot of errors about methods and wrong types of data trying (i think) to compile the module with fpu-internal.h/internal.h.

If it not help in anything, i can try to install again and paste the log here.

Vi0L0 commented on 2015-11-17 18:56

what's inside /var/log/catalyst-install.log?

Katsutoshi commented on 2015-11-17 18:25

Here i'm getting errors even if i put a link of internal.h to asm/fpu-internal.h
And the catalyst-test never builds the module.

Someone knows how to fix this? (Kernel 4.2.5-1)

R4v3nPr0 commented on 2015-11-17 04:16

For those who have problems with GDM, just run the following in a terminal with root privileges.

echo "export COGL_DRIVER=gl" >> /etc/environment
echo "export COGL_OVERRIDE_GL_VERSION=1.4" >> /etc/environment
echo "export COGL_RENDERER=GLX" >> /etc/environment
echo "export LD_PRELOAD=/usr/lib/fglrx/fglrx-libGL.so.1.2" >> /etc/environment

And that's all I needed to run GDM.

ni3k commented on 2015-11-16 14:19

@bosyi

Thanks for your reply. I already use lightdm, gdm is not installed :-)

Before installing "mutter-catalyst", gnome did crash, too.
I have this problem since upgrade to Gnome 3.18, but I also did a fresh
install of Arch Linux. My ohter system with GTX980 runs great with Gnome
3.18. I will stick with xfce, for now.

bosyi commented on 2015-11-15 22:21

@ni3k
Problem may be in GDM. Try to install LightDM for example. Disable gdm, enable lightdm, and possibly delete mutter-catalyst.

ni3k commented on 2015-11-15 21:40

I can only run xfce (which I like, but I would prefer gnome for HIDPI support), gnome or cinnamon crashes (HD7950). I tried "mutter-catalyst" and have installed "cogl 1.22.0-2.5", but still no luck? I also have tried open source "radeon" driver, not working. My screen resolution is 3840x2160, maybe this is the problem? If somebody has any advice, I would highly appreciate.

Vi0L0 commented on 2015-11-14 07:28

I know nothing about xserver 1.18 support.
If 1.18 will hit arch repo please use [xorg117] repo (it's already available) .

securitybreach commented on 2015-11-14 00:53

So is support coming for xorg-server 1.18.0 or do we have to use a repo to hold it back for now?

Vi0L0 commented on 2015-11-11 12:09

maybe you were touched by:
-------------------------- TTY BUG -----------------------------
There's a bug in switching to VT/TTY
Using uvesafb is a workaround:
https://wiki.archlinux.org/index.php/uvesafb
----------------------------------------------------------------

herycp commented on 2015-11-10 10:20

@Vi0L0 sorry very late

my problem is fix after install mutter-catalyst


now my ploblem is :
when start gdm why transition from text mode to gui its very slow and its like blank (very dark like death)

its also happens when im change from gui to tty

Dea993 commented on 2015-11-10 08:11

@juan_kklzone
yes this work also on new kernel 4.3.
don't worry, i've used kernel 4.1 lts some weeks ago, now i'm using kernel 4.2, and i've also tried on testing the kernel 4.3

juan_kklzone commented on 2015-11-10 06:33

Someone knows if is it working on Xorg-server 1.17 w/ 4.1 LTS kernel? I want to give a try, but better if this time it'll really works for me.

Thanks in advice.

kainlite commented on 2015-11-06 22:04

Got it working, seems that i was doing something wrong in my config :/

Vi0L0 commented on 2015-11-06 19:37

here's mine: http://mirror.rts-informatique.fr/archlinux-catalyst/files/4.3/
original config from linux-mainline 4.3rc7, so I believe also from stable
feel free to check out

kainlite commented on 2015-11-06 18:51

I've a fx-8150 CPU, and I'm using the .config from the latest stable kernel 4.2.5-1, I've changed just 2 options:

Processor type and features -> Processor family -> Opteron/...
Processor type and features -> Timer frequency -> 1000 HZ

and then disabled all modules that I'm not using:

make localmodconfig

Vi0L0 commented on 2015-11-06 18:13

strange, it works fine here :/
is there anything exotic in your software or hardware configuration? [testing] repo, unusual 4.3 kernel config options, Skylake CPU maybe, something like that?
You can try to go deeper in debugging with:
coredumpctl list
coredumpctl gdb <pid>

kainlite commented on 2015-11-06 16:59

This is what I found not to be working, I mean the module gets compiled, loaded, but when it reaches graphical target which is gnome in my case and it just explodes, black screen over and over the only way to get in is using an older kernel for me... I'm using your patch for cogl which works great in 4.2

Nov 06 12:07:39 skynet-pc kernel: gnome-session-c[969]: segfault at 0 ip 00007f68fe187a0f sp 00007ffcbd7d6208 error 4 in libc-2.22.so[7f68fe0fc000+19b000]
Nov 06 12:07:39 skynet-pc systemd-coredump[970]: Process 969 (gnome-session-c) of user 120 dumped core.
Nov 06 12:07:39 skynet-pc kernel: gnome-session-c[971]: segfault at 0 ip 00007f1cc73ffa0f sp 00007ffd361a6bf8 error 4 in libc-2.22.so[7f1cc7374000+19b000]
Nov 06 12:07:39 skynet-pc gnome-session[966]: gnome-session-binary[966]: WARNING: software acceleration check failed: Child process killed by signal 11
Nov 06 12:07:39 skynet-pc systemd-coredump[972]: Process 971 (gnome-session-c) of user 120 dumped core.
Nov 06 12:07:39 skynet-pc kernel: gnome-session-f[973]: segfault at 0 ip 00007f8d32127a0f sp 00007fff9a2d3c08 error 4 in libc-2.22.so[7f8d3209c000+19b000]

Thanks.

Vi0L0 commented on 2015-11-06 07:42

@herycp:
can you paste journalctl output from a broken session on some pastebin site?
btw did you try to run it on a freshly created user?

@kainlite:
i tested on vanilla 4.3 kernel @ x86_64, didn't test i686.
what is journalctl saying?

kainlite commented on 2015-11-06 07:01

Hey @Vi0L0,

I've tried to run the driver in a custom 4.3 kernel with no success same deal as before with 4.2: black screen. have you been able to test it with a 4.3 kernel?

Thanks for your work on this package ;-)

herycp commented on 2015-11-03 21:40

@ViOLO still fail

cant enter gnome shell


i try gnome-shell --replace :

X Error of failed request: BadValue (integer parameter out of range for operation)

Major opcode of failed request: 141 (RANDR)
Minor opcode of failed request: 13 (RRChangeOutputProperty)
Value in failed request: 0x77
Serial number of failed request: 277
Current serial number in output stream: 278

Vi0L0 commented on 2015-11-03 18:54

updated to 15.9-6:
added initial 4.3 kernel support - thanks to kolasa' and gentoo' patches :)


hope that cogl will be fixed upstream soon as there's workaround for this problem: https://bugzilla.gnome.org/show_bug.cgi?id=756306

@herycp:
you can also try this version of cogl (with fix from upstream):
https://github.com/manjaro/packages-extra/tree/master/cogl

herycp commented on 2015-11-03 16:02

work in 4.2.5-1

but

i must change gdm to lightdm and i cant login to gnome shell

still waiting to fix that

fyi: i installed http://mirror.rts-informatique.fr/archlinux-catalyst/tarball/gnome318-problems/cogl-1.22.0-2.src.tar.gz

securitybreach commented on 2015-10-26 21:31

Works beautifully here on 4.2.4-1.

Thanks Vi0L0

Vi0L0 commented on 2015-10-26 20:22

Yes, gcc compiles for some time - it's a compilers pack after all, thats why I put it to the repo (well x86_64 for now).
It's not an update thought, it's additional package, but it's required.

I'm considering putting good old catalyst package with precompiled for current ARCH kernel fglrx module inside, it would save you a lot of time and memory if you are using ARCH kernel. Will think about it, but guess its a good idea

securitybreach commented on 2015-10-26 20:17

Yeah, I noticed that the gcc49 was added as a dependency automatically for me when updating. Seems like a big update as it has been compiling for about 10 minutes now on an i5/24gb ram

Thymo commented on 2015-10-26 20:13

Yaourt seems to be complaining that catalyst-test requires linux < 4.2. The PKGBUILD says linux < 4.3, a simple yaourt -Syuad should fix that.

Vi0L0 commented on 2015-10-26 20:07

No, as long as you have gcc-4.9 binary which is serverd by gcc49 package - then it will work.
gcc49 install itself next to current gcc, without conflict or something.

securitybreach commented on 2015-10-26 19:50

So it will now work on 4.2.x as long as gcc is up to date?

Vi0L0 commented on 2015-10-26 19:45

updated to 15.9-5:
- with workaround to bring linux kernel 4.2 support

based on http://ati.cchtml.com/show_bug.cgi?id=1189
fglrx module needs to be build on gcc 4.9.x (or < 5) to work on linux 4.2

install gcc49@AUR : https://aur.archlinux.org/packages/gcc49/
or use a package compiled by me for x86_64 : http://mirror.rts-informatique.fr/archlinux-catalyst/repo/catalyst/x86_64/gcc49-4.9.3-1-x86_64.pkg.tar.xz
(soon, maybe tomorrow, I will prepare this package for i686, and update repos as well)

harry218 commented on 2015-10-18 15:46

@Vi0L0, thanks man, the patch worked fine here!

Vi0L0 commented on 2015-10-18 09:36

GNOME 3.18 fix
In [catalyst] and [catalyst-stable] you may find a new package cogl 1.22.0-2. It brings fix for recent problems with new GNOME 3.18 and new GDM (it baheves strange with plasma though, after choosing plasma and entering password I have to switch to tty and then back to X to see plasma desktop, probably another bug).
Source and a patch is here: http://mirror.rts-informatique.fr/archlinux-catalyst/tarball/gnome318-problems/cogl-1.22.0-2.src.tar.gz

rajder commented on 2015-10-17 17:45

I can confirm that this driver (15.9-4) with latest gnome and linux-lts does not work on my system (R9 290). No errors in journalctl for latest boot but gnome is black, only the mouse cursor is visible plus some white blinking pixels.

straby187 commented on 2015-10-14 22:39

Its working in Debian Sid for me. I tried downloading from AMD and applying patches but it didnt not work but fglrx from the sid repo installed like a champ on 4.2.3-1

pic of uname & fglrxinfo
http://imgur.com/ifwumjQ

Xaero_Vincent commented on 2015-10-12 22:46

@Vi0L0: Heh. Funny pic.

For now, I've resorted to r600 drivers. If I had a newer Radeon card, I would probably never use Catalyst because RadeonSI has OpenGL 4.1 support now and with the SI scheduler patch even beats Catalyst performance in most cases. Sadly, the r600g driver is further behind and stuck at OpenGL 3.3, limiting support for some new AAA Linux games.

http://www.linuxsystems.it/2015/08/radeonsi-with-si-scheduler-humiliates-catalyst-in-all-tests/
http://openbenchmarking.org/result/1508257-DARK-150823638

Vi0L0 commented on 2015-10-12 19:14

today GNOME 3.18 is even more "interesting":
http://mirror.rts-informatique.fr/archlinux-catalyst/files/gnom_3.18.jpg


@francoism90:
well, at least AMD is doing really good on open source driver side :)

francoism90 commented on 2015-10-12 11:24

@Vi0L0: Thanks for all the updates. :)
Nothing you can do about the bugs in the Catalyst driver, AMD needs to fix issues, instead of the community (or make AMD should make it open-source).

To be honest I switched to radeon instead, and so far I'm happy with the performance and no problems with Gnome 3.18 and the latest Linux 4.3 RC4 kernel.

harry218 commented on 2015-10-12 07:49

My Gnome 3.18 and Cinnamon are also broken right now with this driver.

Vi0L0 commented on 2015-10-11 08:17

same here, thought atm i don't have time to look at this

Xaero_Vincent commented on 2015-10-10 22:23

It looks like Gnome 3.18 is broken with Catalyst 15.9, at least for me. I just upgraded and gnome shell boots to a black screen with just a few white dots but the mouse cursor is visible and moves. Has anyone tried 15.9 with Gnome 3.18?

Amarildo commented on 2015-10-10 16:55

So what does the future look like? Are we safe, at least until linux 5 comes if we use linux-lts?
And what about Xorg 18?

Thanks

Vi0L0 commented on 2015-10-06 18:56

I rather think that AMD needs to brings 4.2 kernel support - probably throught updated/new catalyst driver, as I believe that problem lays somewhere on binary side.
Hopefully it should happen soon as Ubuntu 15.10 should be released with 4.2 kernel.

Raansu commented on 2015-10-06 16:06

So I'm guessing more testing is needed before this package is updated for the linux kernel 4.2.2 that's in the repo?


catalyst-test: requires linux<4.2

Amarildo commented on 2015-10-04 23:30

My GPU is a R9 270X.

I can't tell which versions I was using on Ubuntu because I left it a while ago, around a week. But I installed the latest UbuntuMATE snapshot (15.10) at the time and it was completely up to date.

I can't past Xorg of Arch as well, because as soon as the system starts loading my screen will display funny colors and the whole system freezes.

I'm on Debian Jessie (Stable) now, they have Catalyst 15.9 backported and the Kernel 4.1 backported as well. But I can guarantee you that Catalyst 15.9 works on Debian Sid with linux 4.2, Plasma 5 and sddm. Well, at least with my card :D

Dea993 commented on 2015-10-04 19:43

what about catalyst and amdgpu on kernel 4.2 for carrizo and toga??

Vi0L0 commented on 2015-10-04 19:24

Just installed ubuntu 15.10 beta2 x86_64 and updated it. After switching to fglrx-update and reboot the result is exactly the same as on Arch - RIP on init...

securitybreach commented on 2015-10-04 16:59

Hehe, right

Vi0L0 commented on 2015-10-04 16:48

Same here, we are on Arch, we preffer new things!
I never even liked this quote "Better is the enemy of good." :P

securitybreach commented on 2015-10-04 16:38

Yup, lts works fine here but I would prefer 4.2.x..

Vi0L0 commented on 2015-10-04 16:22

So it looks more like its working for some ubuntu users and some of them are having very similar results to mine. Wondering what's causing this. Can you write whats your GPU, architecture, and gcc version you are using on Ubuntu?

As for lts - never had problems with it, other arch users confirmed its fine. Can you paste Xorg.0.log and dmesg on some pastebin portal?

Amarildo commented on 2015-10-04 15:42

Actually, Catalyst 15.9 works on both for me o.O

It works on Debian Sid with Linux 4.2
And it works on Ubuntu 15.10 with Linux 4.2

I installed Catalyst from your repo along with linux-lts, and after reboot my system freezes to a point where I can't do anything, I had to re-install Arch. Do you have a time-frame to make 15.9 (from your repo) compatible with linux-lts?

Thanks

Vi0L0 commented on 2015-10-04 10:12

Checked Debian, over us they only added one patch from ubuntu. It's not working :>.
In minutes I should update this package with mentioned patch (without release bump, as it's not that important).

I'm pretty sure fglrx is not working on Debian, as it's not working on Ubuntu:
https://bugs.launchpad.net/ubuntu/+source/fglrx-installer/+bug/1493888
https://bugs.launchpad.net/ubuntu/+source/fglrx-installer/+bug/1501276

Amarildo commented on 2015-09-30 23:10

Debian sid now has linux 4.2 and it also has Catalyst 15.9. Have you guys checked their source code to see if their packages are useful in Arch?

Just don't import their packages directly because Catalyst on Debian has a 2-month-old bug in 'amd-opencl-icd' that makes your GPU unusable for OpenCL. The Ubuntu package and openSUSE's work fine, though.

Vi0L0 commented on 2015-09-30 06:30

I'm also looking with hope at ubuntu 15.10

Xaero_Vincent commented on 2015-09-30 05:58

Thanks for the work Vi0L0. Everything works good with Linux LTS.

My guess is fully working kernel 4.2+ support probably needs a patch in the closed-source part of the Catalyst driver, so we'll just have to wait for AMD. AMD does track Ubuntu releases and it looks like 15.10 will have a 4.2 kernel, so that should be good news for Arch users in the near future.

Amarildo commented on 2015-09-29 21:27

Thank you for your work, Vi0l0.

If I install Catalyst from your repo, will it work (build it's modules) on linux-lts? The last time I tried it didn't work.

securitybreach commented on 2015-09-29 19:31

BTW keep up the great work man

securitybreach commented on 2015-09-29 19:31

Yeah, I knew about how to get back in but it is useful for many others.

Vi0L0 commented on 2015-09-29 19:29

yes, this tip is more general and important.
Some users will probably curse me if they will notice this tip too late... :>
Or miss it completelly.

Vi0L0 commented on 2015-09-29 19:24

yes, it can be all done with mentioned script, you just have to manually switch between kernels (I don't want script to mess with bootloader)...

securitybreach commented on 2015-09-29 19:24

Nice tip

securitybreach commented on 2015-09-29 19:23

Neat. Is there a script to do the opposite? Switching to radeon from catalyst? Just out of curiosity...

Vi0L0 commented on 2015-09-29 19:22

one more important thing:
be aware that I removed linux from dependency array and so it will let linux to update (while it's not supported tongue), if you will not configure bootloader to use kernel older than 4.2, like linux-lts, it will probably boot linux and you will probably gonna see a freeze or blank screen. System should work though, you can ssh to it and config bootloader from here, or use single mode if you can't ssh, or well there's possibility that linux-lts was added before, boot it and config as default kernel to boot.

Vi0L0 commented on 2015-09-29 19:21

once configured it should serve you well ;)

but that's the reason of why i'm not blocking linux update with this pkg - you can update linux to 4.2 and use radeon oss on it, and then easily switch to linux-lts and catalyst in a no time with pxp_switch_catalyst script

securitybreach commented on 2015-09-29 19:13

Vi0L0: thanks. I decided to go to LTS instead of switching to radeon. My four monitor setup takes a bit of configuring with radeon.

securitybreach commented on 2015-09-29 19:13

Vi0L0: thanks. I am on LTS now

Vi0L0 commented on 2015-09-29 09:44

Looks like I need to change linux to linux-lts in the dependencies to fix this permanently...

Pyromaniac commented on 2015-09-28 21:27

I'm not sure, but with this package from AUR it would probably be the same.

Pyromaniac commented on 2015-09-28 21:26

Sorry for bothering you again, but there's an issue with lts kernel. After switching to it, I can't update catalyst-generator. The problem is that <4.2 dependency and the fact that lts package provides "kernel26-lts", not "linux".

Vi0L0 commented on 2015-09-28 20:47

Also as I noted before please use linux-lts, it's in version 4.1 now and it will get updates, so it's much better than using downgraded linux all the time.

Vi0L0 commented on 2015-09-28 20:45

Thanks Pyromaniac. catalyst-generator never had linux dependency because, well, linux pkg is not a "must-have", but in this 4.2 circumstances it seems legit, so it has linux<4.2 now. As for other libgl and utils - these pkgs have nothing to do with the kernel itself so no changes on them


+ catalyst-test updated to 15.9-3:
i've rewritten acpi scripts, had no laptop hardware to check but it should work in theory :>
Those who are interested should also enable acpid.service

Pyromaniac commented on 2015-09-28 18:57

Hi ViOLO!

Just wanted to mention that there's no <4.2 restriction on catalyst-generator, catalyst-libgl and catalyst-utils in "catalyst" repo. So I had to rollback the kernel after updating my system. Can you do something about it?

And thank you for the great work you do!

securitybreach commented on 2015-09-28 18:04

Oh well, I guess its back to the Radeon driver as we have been waiting on 4.2 support for quite a long while. Oh well, thanks for the support Vi0L0

Vi0L0 commented on 2015-09-28 17:55

nope

securitybreach commented on 2015-09-28 17:54

So still no support for 4.2.1?

Vi0L0 commented on 2015-09-28 17:53

Also after installing linux-lts don't forget to add it to your bootloader

Vi0L0 commented on 2015-09-28 17:52

linux 4.2.1 is now in [core], the best way is to switch to linux-lts, you can then update linux to 4.2 with pacman -Syud or add linux to /etc/pacman.conf's IgnorePkg. This package should build module for linux-lts on its own, just don't forget to also install linux-lts-headers. After installing you can call module build with `catalyst_build_module all`

@untitaker:
which ubuntu patches are you talking about? here we are using some 4.2 ubuntu patches, but you could find newer ones

untitaker commented on 2015-09-28 16:23

That almost certainly doesn't work.

It seems that Ubuntu patched their dist of the catalyst drivers to be compatible with linux==4.2. Is it feasible to apply those patches to this package too?

archerous commented on 2015-09-28 12:52

@Dea993

Try to delete entry in depend section "linux<4.2" in Pkgbuild.Maybe it works

Dea993 commented on 2015-09-28 10:23

how to install it on kernel linux-lts 4.1.8?? when i try to install, pacman says
"errore: il seguente pacchetto non è stato trovato: linux<4.2"

Vi0L0 commented on 2015-09-26 09:30

@Raansu: lol, didn't know its on the wiki, will fix it soon, just have to rewrite acpi scripts which also became buggy latelly

@zer0def:
what do you mean with avoiding collisions with mesa-libgl package? (btw theres no collision with mesa, mesa is in dependencies of this pkg)
mesa-libgl doesn't actually brought libgl library, it broughts symlinks to mesa's libgl. We don't need it as catalyst-libgl is able to create such symlinks to mesa's libgl on it's own - and this is the way of how we are performing pxp switch

zer0def commented on 2015-09-25 12:15

Vi0l0, since you claim that you're introducing PXP switching to regular packages, wouldn't it be a bit more useful to avoid collisions with mesa and libgl?

Raansu commented on 2015-09-25 09:36

@francoism AND @ViOLO

Why is this still a thing with this package? I have to edit atieventsd.service every single time there is an update here. https://wiki.archlinux.org/index.php/AMD_Catalyst#Failed_to_start_atieventsd.service

galixigen commented on 2015-09-24 07:57

So, I tried to use this catalyst with gnome on Lenovo's IdeaCenter with integrated HD8330 and it didn't worked, KDE still works fine. So, as I can say, there's a problem with integrated gfx?

Vi0L0 commented on 2015-09-23 19:35

From Wiki:
Note: With every Catalyst update you should remove amdpcsdb file in this way: kill X, remove /etc/ati/amdpcsdb, start X and then run amdcccle - otherwise the version of Catalyst may display wrongly in amdcccle.
also:
cat /etc/ati/amdpcsdb.default | grep -i Release

Package comes with a lot of patches, but they are not used and stayed here only because it could be helpful for some users. They are comment in and not applied. Take a look at the PKGBUILD, we only use these patches:

patch -Np1 -i ../makefile_compat.patch
patch -Np1 -i ../lano1106_fglrx_intel_iommu.patch
patch -Np1 -i ../lano1106_kcl_agp_13_4.patch
patch -Np1 -i ../fglrx_gpl_symbol.patch
patch -Np1 -i ../ubuntu_buildfix_kernel_4.0.patch
patch -Np1 -i ../ubuntu_buildfix_kernel_4.1.patch
patch -Np1 -i ../ubuntu_buildfix_kernel_4.2.patch #three 4.2 patches left for testing purposes, use simultaneously - V
patch -Np1 -i ../4.2-fglrx-has_fpu.patch
patch -Np1 -i ../4.2-kolasa-fpu_save_init.patch

we are mainly using packages which are providing support for newer than 3.19 kernels. AMD didn't change much things in 15.9 when it comes in kernel support - they added like 4 lines which fixed module compilation problem on oldie ubuntu - this brought partial, but only partial, compatibility with 4.0 kernel (yes, we had it months ago).
Plus there is fglrx_gpl_symbol.patch that's changing license to "GPL\0Proprietary ..." because otherwise I'm unable to compile patch without optimizations (-O0, for debugging purposes, btw -O0 works only with additional -ftree-ter flag).
Also we are using two good old patches from lano1106, because they are very sane and because after all these years AMD still didn't take it's piiii together:
lano1106_fglrx_intel_iommu.patch: https://bbs.archlinux.org/viewtopic.php?pid=1275707#p1275707
lano1106_kcl_agp_13_4.patch: https://bbs.archlinux.org/viewtopic.php?pid=1277645#p1277645

If you will compare our patched files and files from kolasa you won't notice much differences... They work in most places in same logic, same way, the only real difference is that kolasa is using some strange patch for 3.19 kernel (I also checked it out, don't see difference).
I think we need to make 4.2 kernel work, before we jump into 4.3rc.

The good thing about AUR is that you can just take the source and change it, feel free to remove/comment_in or add some patches and check if it's working for you.
I already did it and still no success on 4.2 kernel. I also checked only files from kolasa repo, without success. I only saw couple of guys that were able to run 4.2 succesfully, or at least they said so. Month ago phoronix said that ubuntu fixed 4.2 support, yet I still see comments from ubuntu users that it's not working.
Maybe it's hardware related? Maybe it's something in kernel config. Dunno. I'm no expert but it's rather hard to debug: http://ati.cchtml.com/show_bug.cgi?id=1189
I really wish AMD fix it. We paid for their hardware, they should provide us decent support. I sent direct request for help weeks ago, still no answer...

Vi0L0 commented on 2015-09-23 19:34

From Wiki:
Note: With every Catalyst update you should remove amdpcsdb file in this way: kill X, remove /etc/ati/amdpcsdb, start X and then run amdcccle - otherwise the version of Catalyst may display wrongly in amdcccle.
also:
cat /etc/ati/amdpcsdb.default | grep -i Release

Package comes with a lot of patches, but they are not used and stayed here only because it could be helpful for some users. But they are comment in and not applied. Take a look at the PKGBUILD, we only use these patches:

patch -Np1 -i ../makefile_compat.patch
patch -Np1 -i ../lano1106_fglrx_intel_iommu.patch
patch -Np1 -i ../lano1106_kcl_agp_13_4.patch
patch -Np1 -i ../fglrx_gpl_symbol.patch
patch -Np1 -i ../ubuntu_buildfix_kernel_4.0.patch
patch -Np1 -i ../ubuntu_buildfix_kernel_4.1.patch
patch -Np1 -i ../ubuntu_buildfix_kernel_4.2.patch #three 4.2 patches left for testing purposes, use simultaneously - V
patch -Np1 -i ../4.2-fglrx-has_fpu.patch
patch -Np1 -i ../4.2-kolasa-fpu_save_init.patch

we are mainly using packages which are providing support for newer than 3.19 kernels. AMD didn't change much things in 15.9 when it comes in kernel support - they added like 4 lines which fixed module compilation problem on oldie ubuntu - this brought partial, but only partial, compatibility with 4.0 kernel (yes, we had it months ago).
Plus there is fglrx_gpl_symbol.patch that's changing license to "GPL\0Proprietary ..." because otherwise I'm unable to compile patch without optimizations (-O0, for debugging purposes, btw -O0 works only with additional -ftree-ter flag).
Also we are using two good old patches from lano1163, because they are very sane and because after all these years AMD still didn't take it's piiii together:
lano1106_fglrx_intel_iommu.patch: https://bbs.archlinux.org/viewtopic.php?pid=1275707#p1275707
lano1106_kcl_agp_13_4.patch: https://bbs.archlinux.org/viewtopic.php?pid=1277645#p1277645

If you will compare our patched files and files from kolasa you won't notice much differences... They work in most places in same logic, same way, the only real difference is that kolasa is using some strange patch for 3.19 kernel (I also checked it out, don't see difference).
I think we need to make 4.2 kernel work, before we jump into 4.3rc.

The good thing about AUR is that you can just take the source and change it, feel free to remove/comment_in or add some patches and check if it's working for you.
I already did it and still no success on 4.2 kernel. I also checked only files from kolasa repo, without success. I only saw couple of guys that were able to run 4.2 succesfully, or at least they said so. Month ago phoronix said that ubuntu fixed 4.2 support, yet I still see comments from ubuntu users that it's not working.
Maybe it's hardware related? Maybe it's something in kernel config. Dunno. I'm no expert but it's rather hard to debug: http://ati.cchtml.com/show_bug.cgi?id=1189
I really wish AMD fix it. We paid for their hardware, they should provide us decent support. I sent direct request for help weeks ago, still no answer...

Xaero_Vincent commented on 2015-09-23 10:47

@Vi0L0, thanks for updating the repo with 15.9. I'm just curious (maybe it's nothing) but I upgraded to 15.9 from the repo and looked at glxinfo and it says my GL string is 15.20.1046 but doing a google search reveals that is the same thing for 15.7? I restarted and did aticonfig --initial again too be sure. Shouldn't my string say 15.201.1151 or something after upgrading to 15.9?

I'm no expert but it looks like the Catalyst packages here have a lot of patches that been used with previous Catalyst releases too. Could some of these old patches be causing issues with 15.9? The page that mentioned kolasa's updates gave instructions to start with a vanilla Catalyst package and apply only his GIT patches for 4.2 and 4.3rc support.

galixigen commented on 2015-09-21 12:09

Vi0L0, I forgot to tell that I'm on HP laptop with "switchable graphic", integrated: 8610G, discreet: r7 m260.

Amarildo commented on 2015-09-21 08:26

OK, I assume linux-lts 4.1 will be supported by both 15.7 and 15.9. I'll install it now.

Vi0L0 commented on 2015-09-21 08:06

@Amarildo:
hard to tell. I don't want to be a pesymist, but it really doesn't look good. The good thing is that linux-lts 4.1 hit [testing].

@galixigen:
what vga do you have?

Amarildo commented on 2015-09-21 07:23

Will 15.9 or 15.7.1 be ready when linux-4.2 comes to Arch?

galixigen commented on 2015-09-20 21:23

Nope, i'm on freshly installed arch and yes, it crashing for all users.

Vi0L0 commented on 2015-09-20 20:54

Same here. Did update cause this? Is it crashing for all users?

galixigen commented on 2015-09-20 20:49

Vi0L0, x86_64 and 4.1.6-1

Vi0L0 commented on 2015-09-20 20:40

which architecture and what kernel?
GNOME works well here, ran from gdm (no fix needed) or sddm.

galixigen commented on 2015-09-20 20:05

I'm getting gnome-shell crash, not gdm, but gnome-shell, cause I use lightdm instead.
Here's the log after loggining in: http://pastebin.com/tetwnQN8
Can I fix this? KDE works fine, so this is gnome-specific error.

Vi0L0 commented on 2015-09-20 17:25

15.9-2: added mentioned kolasa's change
also uncommitted 4.2 kernel patching so you may easier test it.

It crashes on 4.2 here though, still linux<4.2 in deps

securitybreach commented on 2015-09-20 17:15

Thanks Vi0L0!! Hopefully we can finally update the kernel soon...

securitybreach commented on 2015-09-20 17:15

Thanks Vi0L0

Vi0L0 commented on 2015-09-20 17:13

I've tested kolasa's patches and have no luck with it. Still crashes on init.
This package is missing only one change that kolasa added latelly

I'm working on catalyst repo atm.

Xaero_Vincent commented on 2015-09-20 17:02

@francoism: I could be wrong but from my understanding, AMD doesn't yet officially support anything beyond Linux 3.19 with their Catalyst 15.9 driver. Linux 4.x kernels are all 3rd party community supported. It might be worth trying Kolasa's latest GIT patches to see how they work.

@Vi0L0: Will we see 15.9 land in the catalyst.wirephire repo?

francoism commented on 2015-09-20 09:33

@Raansu: Please see the catalyst wiki (https://wiki.archlinux.org/index.php/AMD_Catalyst#Failed_to_start_atieventsd.service)

@Xaero_Vincent: I haven't found the time to test the new catalyst-drivers, but as far I've found, there is no confirmation from AMD that it is working on Linux 4.1>. Yes, it stated that there is support for the latest kernel(s), but not given any numbers will not help devs like Vi0L0.

Xaero_Vincent commented on 2015-09-20 00:29

Vi0L0, does this incorporate kolasa's latest patches? Supposidly there some now that support 4.2.0 and even 4.3 rc1

https://www.reddit.com/r/linux_gaming/comments/3l786m/patch_your_catalyst_159_for_support_kernel_40_41/

https://github.com/kolasa/fglrx-core-15.201

Raansu commented on 2015-09-18 06:25

Hi,

I'm getting this following error, any ideas?

[raansu@Y40-80 ~]$ sudo systemctl start atieventsd.service
Failed to start atieventsd.service: Unit acpid.socket failed to load: No such file or directory.
[raansu@Y40-80 ~]$


securitybreach commented on 2015-09-18 00:26

rherbert: excellent and thanks for your contributions

rherbert commented on 2015-09-18 00:25

I'm not an Arch Linux user, but I do use the AMD/ATI Catalyst drivers for linux. I just want to thank you Arch Linux users for your contributions to the Linux community. Thanks to the patches included in the catalyst-test 15.9-1 package, I was able to successfully compile the AMD Catalyst 15.9 proprietary driver under Linux kernel 4.2. Thanks much!

securitybreach commented on 2015-09-17 22:41

Restarted X and now getting an odd little box under my cursor. Any ideas?

securitybreach commented on 2015-09-16 20:48

Well crap, ok. Thanks

Vi0L0 commented on 2015-09-16 20:47

because it doesnt support 4.2
I've added what we've got for now (so these "beta" 4.2 patches) for testing purposes, patches seems uncomplete. Maybe someone could take it and make something funny with it :>

securitybreach commented on 2015-09-16 20:44

The reason I ask is because this new release still says it needs a kernel below 4.2

securitybreach commented on 2015-09-16 20:38

Nice and thanks man. So we are good to go to upgrade the kernel now?

Vi0L0 commented on 2015-09-16 20:35

aah, CS:GO fans should check this release out

Vi0L0 commented on 2015-09-16 20:35

updated to 15.9:
- it's like 15.7 with game fixes

Problems with tty switch, glitches on chrome/vaapi, 4.2 kernel support persists
I left 4.2 patches for your testing purposes, just need to uncomment both simultaneously


@securitybreach:
too bad that he's not showing the cards...

securitybreach commented on 2015-09-16 12:11

Comments on third page says it works on 4.2:
http://www.phoronix.com/forums/forum/phoronix/latest-phoronix-articles/823500-catalyst-15-9-is-imminent-with-many-steam-linux-game-fixes?p=823604#post823604

francoism90 commented on 2015-09-15 20:35

Good news?
http://phoronix.com/scan.php?page=news_item&px=Catalyst-15.9-Soon

securitybreach commented on 2015-09-10 15:40

Any progress?

Vi0L0 commented on 2015-09-04 19:50

nope, last time I checked it failed on build, I promissed to look at it but had no time, maybe in this week I will find more time somehow :>

SummerRainbowz commented on 2015-09-04 19:46

Is there a way to use this with the grsec kernel? I'd use the OSS driver with my 7870 XT but it just makes a black screen with Xorg.

Vi0L0 commented on 2015-09-04 18:39

Yes. It could take some time, thought

securitybreach commented on 2015-09-04 18:13

Ok, so we should hold back the kernel until a fix comes down?

Vi0L0 commented on 2015-09-04 15:14

@securitybreach:
I haven't seen a patch that worked for me on linux 4.2 :(
also debugging is pretty hard:
http://ati.cchtml.com/show_bug.cgi?id=1189

securitybreach commented on 2015-09-04 12:53

Please update as the 4.2 kernel hit testing last night and this package breaks updating. I know I could simply change the dependency line but I figured I would tell you.

error: failed to prepare transaction (could not satisfy dependencies)
:: catalyst-test: requires linux<4.2

securitybreach commented on 2015-09-04 12:53

error: failed to prepare transaction (could not satisfy dependencies)
:: catalyst-test: requires linux<4.2

securitybreach commented on 2015-09-04 12:52

Please update as the 4.2 kernel hit testing last night and this package breaks updating. I know I could simply change the dependency line but I figured I would tell you.

Chreymo commented on 2015-08-26 07:31

PxP still doesn't seem to work with xorg-server 1.17, having to revert to older xorg-server as before. Is this expected?

Log shows fglrx looks for integrated gpu even when discrete is enabled by aticonfig and nomodeset enabled on kernel, so fails on "fail to probe Intel vga device".

In other words, instead of going by the xorg conf file and detecting my discrete card, fglrx is deciding on its own to detect and load integrated gpu driver.

Chreymo commented on 2015-08-26 07:19

PxP still doesn't seem to work with xorg-server 1.17, having to revert to older xorg-server as before. Is this expected?

Vi0L0 commented on 2015-08-12 20:18

here: http://ati.cchtml.com/
don't forget to paste atigetsysteminfo.sh's fruit.

tuxayo commented on 2015-08-12 19:11

I can crash my X.Org by:
- suspending
- turning my monitors off
- waking up with the monitors still off
Where should I report this?

rajder commented on 2015-07-28 16:37

A tip for future newcomers:

>>> Enable multilib BEFORE installing <<<

(as instructed in the wiki) Or opengl will not be available for 32-bit applications.

rajder commented on 2015-07-27 00:39

@ViOLO: Thank you for this package, keep up the awesome work!

I had a problem with the nomodeset kernel parameter. With nomodeset I got this in journalctl:

Jul 27 01:59:00 xxx kernel: [drm] VGACON disable radeon kernel modesetting.
Jul 27 01:59:00 xxx kernel: [drm:radeon_init [radeon]] *ERROR* No UMS support in radeon module!

I removed the nomodeset parameter and the error did not appear during the next boot.

Gnome 3 is working flawlessly. Chrome seems to work well with Chrome WebGL Experiments, have not tested any games yet.

francoism commented on 2015-07-20 11:44

@Vi0L0: no issues so far on Xorg 1.17.2-4 and Gnome-shell :)

Vi0L0 commented on 2015-07-13 18:29

here's bug report of chromium glitch : http://ati.cchtml.com/show_bug.cgi?id=1182
feel free to add comment - this should make it more "visible"

dfanz0r commented on 2015-07-09 19:05

Awesome, ill give this a try when i get home. (Currently on the the opensource drivers.)

francoism commented on 2015-07-09 19:03

@Vi0L0: Thanks for the update! :)

@dfanz0r: Official support; devs found workarounds for the latest packages, but reading the feedback given by Vi0L0, it's still bad. Better using an Intel HD instead. ;p

Vi0L0 commented on 2015-07-09 18:53

updated to 15.7:
OpenGL is now in 4.5 version, it brings fixes for some titles like borderlands 2 which now works well, still looks like crap smile but at least its fluid. Even warsow seems snappier, but it's not mindblowing. Dota 2 is kinda laggy in some situations, plus it crashed on exit (sic!) and I wasn't able to kill its process (sadface, at least it wasnt eating resources, only ate some ram and sit)...

They fixed vaapi which latelly wasn't working with -va vaapi option.

It does support xserver 1.17, but tty - just like with ubuntu 15.4 version - is all black - another sadpanda.

And ofcourse glitching (on plasma DE) chromium - cherry on the top.


I modified kolasa's patch for 4.1 kernel support a bit

dfanz0r commented on 2015-07-09 18:05

@beta990 There should be patches to support 4.0 how else would the current package support it :)

But even so i guess it would be nice if they had official support...

francoism commented on 2015-07-09 16:02

@dfanz0r: a few changes, but still no Linux 4~ support. :(
OpenGL support seems to be improved.

@Vi0L0: thanks for the info. :)
Seems AMD releases their drivers in .deb, but thought the Ubuntu driver was different than the tar.gz one. So I assume there are no differences with performance, etc.

dfanz0r commented on 2015-07-08 21:51

15.7 is out
http://support.amd.com/en-us/download/desktop?os=Linux+x86

Vi0L0 commented on 2015-06-21 08:44

this is ported to Arch Ubuntu driver, just like in the description.
It doesn't differ much, binaries are the same, kernel patches could differ but Ubuntu was in the past taking patches from here. Surelly AMD is aiming to ubuntu with this driver.
What kind of differences you are talking about? Performance, functionality, stability? Desktop or laptop?

francoism commented on 2015-06-10 10:44

@Vi0L0: Thanks for updating! :)

I understand the Ubuntu AMD driver differs from the official one. Is it possible to 'port' the Ubuntu driver to Arch?
What makes it so much different? The code base or patches?

Thanks!

Vi0L0 commented on 2015-06-04 15:30

catalyst 15.5 doesn't even support xorg-server 1.17...
You may download it from stable AUR packages or [catalyst-stable] repo.

I will unflag this package and wait with its update till proper/nicer catalyst release arive.

Vi0L0 commented on 2015-06-03 10:55

I will get back late today, should be able to update it tomorrow.
Heard it's older than ubuntu version, wondering if it supports xserver 1.17

TheNumb commented on 2015-06-03 08:30

15.5 Beta is out now!
http://www2.ati.com/drivers/linux/amd-catalyst-omega-15.5-linux-run-installers.zip

I wasn't able to find a changelog :(

awerush commented on 2015-06-02 19:00

Having VAAPI issues with catalyst-test 15.3-1, xorg-server 1.17.1-5 and mpv 0.9.2-1.

When I try to play a file using hardware decoding I get this:

$ mpv --hwdec vaapi some_file.mkv
Playing: some_file.mkv
(+) Video --vid=1 (h264)
(+) Audio --aid=1 (*) (ac3)
libva info: VA-API version 0.37.0
libva info: va_getDriverName() returns 0
libva info: User requested driver 'fglrx'
libva info: Trying to open /usr/lib/dri/fglrx_drv_video.so
libva info: Found init function __vaDriverInit_0_33
libva info: va_openDriver() returns 0
Using hardware decoding.
AO: [pulse] 48000Hz stereo 2ch float
// hangs for about 3 to 4 seconds here ...
[vaapi] vaCreateContext(): operation failed
[ffmpeg/video] h264: decode_slice_header error
[ffmpeg/video] h264: no frame!
Error while decoding frame!
Error using hardware decoding, falling back to software decoding.
VO: [opengl] 1280x720 yuv420p
AV: 00:00:01 / 00:56:50 (0%) A-V: 0.000

Anyone else with similar VAAPI problems?

P.S.: VLC also falls back to software decoding:

[00007fd7ac0660d8] vaapi_drm generic error: Could not access rendering device: No such file or directory

awerush commented on 2015-06-02 18:59

Having VAAPI issues with catalyst-test 15.3-1, xorg-server 1.17.1-5 and mpv 0.9.2-1.

When I try to play a file using hardware decoding I get this:

$ mpv --hwdec vaapi some_file.mkv
Playing: some_file.mkv
(+) Video --vid=1 (h264)
(+) Audio --aid=1 (*) (ac3)
libva info: VA-API version 0.37.0
libva info: va_getDriverName() returns 0
libva info: User requested driver 'fglrx'
libva info: Trying to open /usr/lib/dri/fglrx_drv_video.so
libva info: Found init function __vaDriverInit_0_33
libva info: va_openDriver() returns 0
Using hardware decoding.
AO: [pulse] 48000Hz stereo 2ch float
// hangs for about 3 to 4 seconds here ...
[vaapi] vaCreateContext(): operation failed
[ffmpeg/video] h264: decode_slice_header error
[ffmpeg/video] h264: no frame!
Error while decoding frame!
Error using hardware decoding, falling back to software decoding.
VO: [opengl] 1280x720 yuv420p
AV: 00:00:01 / 00:56:50 (0%) A-V: 0.000

Anyone else with similar VAAPI issues?

P.S.: VLC also falls back to software decoding:

[00007fd7ac0660d8] vaapi_drm generic error: Could not access rendering device: No such file or directory

awerush commented on 2015-06-02 18:58

Having VAAPI issues with catalyst-test 15.3-1, xorg-server 1.17.1-5 and mpv 0.9.2-1.

When I try to play a file using hardware decoding I get this:

$ mpv --hwdec vaapi some_file.mkv
Playing: some_file.mkv
(+) Video --vid=1 (h264)
(+) Audio --aid=1 (*) (ac3)
libva info: VA-API version 0.37.0
libva info: va_getDriverName() returns 0
libva info: User requested driver 'fglrx'
libva info: Trying to open /usr/lib/dri/fglrx_drv_video.so
libva info: Found init function __vaDriverInit_0_33
libva info: va_openDriver() returns 0
Using hardware decoding.
AO: [pulse] 48000Hz stereo 2ch float
// hangs for about 3 to 4 seconds here ...
[vaapi] vaCreateContext(): operation failed
[ffmpeg/video] h264: decode_slice_header error
[ffmpeg/video] h264: no frame!
Error while decoding frame!
Error using hardware decoding, falling back to software decoding.
VO: [opengl] 1280x720 yuv420p
AV: 00:00:01 / 00:56:50 (0%) A-V: 0.000

Anyone else with similar VAAPI issues?

P.S.: VLC also fall back to software decoding:

[00007fd7ac0660d8] vaapi_drm generic error: Could not access rendering device: No such file or directory

awerush commented on 2015-06-02 18:57

Having VAAPI issues with catalyst-test 15.3-1, xorg-server 1.17.1-5 and mpv 0.9.2-1.

When I try to play a file using hardware decoding I'll get this:

$ mpv --hwdec vaapi some_file.mkv
Playing: some_file.mkv
(+) Video --vid=1 (h264)
(+) Audio --aid=1 (*) (ac3)
libva info: VA-API version 0.37.0
libva info: va_getDriverName() returns 0
libva info: User requested driver 'fglrx'
libva info: Trying to open /usr/lib/dri/fglrx_drv_video.so
libva info: Found init function __vaDriverInit_0_33
libva info: va_openDriver() returns 0
Using hardware decoding.
AO: [pulse] 48000Hz stereo 2ch float
// hangs for about 3 to 4 seconds here ...
[vaapi] vaCreateContext(): operation failed
[ffmpeg/video] h264: decode_slice_header error
[ffmpeg/video] h264: no frame!
Error while decoding frame!
Error using hardware decoding, falling back to software decoding.
VO: [opengl] 1280x720 yuv420p
AV: 00:00:01 / 00:56:50 (0%) A-V: 0.000

Anyone else with similar VAAPI issues?

P.S.: VLC also fall back to software rendering:

[00007fd7ac0660d8] vaapi_drm generic error: Could not access rendering device: No such file or directo

niqingliang2003 commented on 2015-05-11 13:59

after changed to xf86-video-ati, the memory occupied by xorg always 2x MB, looks like no memory leak anymore.

securitybreach commented on 2015-05-10 18:59

All TTYs work perfect for me. Running xorg-server 1.17.1-5 and catalyst-test 15.3-1

Vi0L0 commented on 2015-05-10 18:58

*repo ☺

Vi0L0 commented on 2015-05-10 18:57

the only solution I know about is to use xorg-server 1.16, ie. from [xorg116] tepo

ThePierrezou commented on 2015-05-10 18:54

When i start xorg, i can't use Tty with ctrl+alt+F*
Do you know can i resolve this problem ?

ThePierrezou commented on 2015-05-03 02:59

Thank you verry muc, it work verry well with my saphrire(amd) radeon hd 6450 2GO :)

niqingliang2003 commented on 2015-04-29 12:19

besides glitch in chromium, I found that the Xorg will use more and more memory, it looks like memory leak. and after some time, the border of all kde windows missed.

Vi0L0 commented on 2015-04-28 20:10

yeah, chromium generates really bad bad glitches on plasma
I've tried many setting, including disabling direct rendering of kde - by putting
export KWIN_DIRECT_GL=0
in profile (it seems to improve performance), nut nothing helped for good, it could work good for some time, longer time, but it gets back, always.
I thought that Opera is not touched by this bug, but it also glitches, just later.
darn...
Seems like only solution is to disable kwin.

Have you seen some bug report about this? I remember that kde in time before plasma was working very well with catalyst.
I guess it's time to post some bug report, maybe on both amd and kde sites?

niqingliang2003 commented on 2015-04-26 03:20

@ScOrian, I have, mainly in chromium under KDE

Hot4ruh1me commented on 2015-04-09 17:33

Hey Vi0L0, this came up in the news:
http://www.phoronix.com/scan.php?page=news_item&px=Ubuntu-14.04-Catalyst-15.3

my regars!

securitybreach commented on 2015-03-31 13:06

@Sc0rian, yes I have

bcc commented on 2015-03-31 11:22

http://i.imgur.com/IAPwpOs.jpg anyone seen this? I seem to get terrible tearing all over the place

francoism commented on 2015-03-16 17:26

Thanks Vi0L0 for all your time and updating this package. :)

@nuxlee: good times will come back. From what I read, AMD is busy adding patching the kernel, mesa, etc.
It all sounds promising and the work they do now, will solves most problems/issues they (AMD/devvers) facing now.

And more (positive) news: (AMD) Wayland support looks promising, so we can ditch X soon!

nuxlee commented on 2015-03-14 10:23

Thanks for your effort dude. It's really tough times for AMD users, I have issues with X and free driver because X crashes every time I try to run libreoffice or I'm working in emacs, and now I'm reading people comments about proprietary driver and all the problems they're having with it. I have a cheapend laptop with apu, radeon hd 8330, and I can't make my mind about what to use. Any suggestions?
Thanks

Vi0L0 commented on 2015-03-13 21:14

updated to 15.3-1:
- borrowed from ubuntu.
- brings xserver 1.17 support
- its buggy :P

I was thinking for long time about uploading this driver because I think it sucks :P. On xserver 1.17 xv video output doesn't work at all, as well as tty terminal, which is sad :(. After switching to xserver 1.16 it gets back to normal. I remember that in the past AMD did release ubuntu driver which had same bugs.
So it sucks, steam games are working fine, maybe better, but still.
And since this package is called catalyst-test i decided to make this update, you can switch to catalyst-total in no-time.

I also dropped many patches and left only most neccessary, so 1) for intel iommu, 2) for agp 3) for 4.0 kernel support. You can still uncomment dropped one if you want.

securitybreach commented on 2015-03-13 13:58

@AID Thanks

AlD commented on 2015-03-13 07:23

Yea, I had Xorg go up to several Gig. Seems to get worse after a few suspend, resume cycles. Definitely didn't happen with previous versions.

securitybreach commented on 2015-03-13 02:57

Has anyone noticed high memory usage with the xorg116 package and the last update? Xorg is using 428mb with i3 and 4 monitors. Luckily I have 24gb of ram but it seems like it is using way more ram than it should.

TheNumb commented on 2015-03-12 13:29

Looks like the gdm bug still hasn't been fixed. Oh well...

TheNumb commented on 2015-03-12 13:27

Hey, it indeed works with xorg 1.17 :)

$ fglrxinfo
display: :0 screen: 0
OpenGL vendor string: Advanced Micro Devices, Inc.
OpenGL renderer string: AMD Radeon HD 7700M Series
OpenGL version string: 4.4.13374 Compatibility Profile Context 15.20.1013

Now to test if they did fix the gdm segfault.

securitybreach commented on 2015-03-11 20:55

So does that mean we will be able to use the xorg 117 when the package is updated?

Vi0L0 commented on 2015-03-11 20:13

Thanks for the news! :)
Should be easy to convert ubuntu driver, but im afk and have lot of work these times, maybe I will be able to update on Friday, maybe on Sunday, dunno.
Please be patient.

Vi0L0 commented on 2015-03-11 20:08

Thanks for the news!

Vi0L0 commented on 2015-03-11 20:07

Thanks for the news!

Hot4ruh1me commented on 2015-03-11 20:01

Hey Vi0Lo,
News -> http://www.phoronix.com/scan.php?page=news_item&px=Ubuntu-15.04-15.3-Beta

I don't know if you can do anything with the package in the ubuntu repos, but at least is a new driver coming our way soon =D

Keep the great work!

Vi0L0 commented on 2015-03-09 19:57

Xserver 1.17 causes segfault here ☺

francoism commented on 2015-03-09 19:40

Hi Vi0L0,

Still haven't upgrade X (1.16.4-1), what are you experience with upgrading X?
Does it even work?

Thanks! :)

securitybreach commented on 2015-03-08 14:47

Thanks man

Vi0L0 commented on 2015-03-08 14:44

No news, im pretty sure that (like always) amd has fglrx with xserver 1.17 support and tests it internally. But I have no idea when will they release it.

securitybreach commented on 2015-03-08 14:41

Any xorg-server 1.17.1 support yet?

Vi0L0 commented on 2015-03-08 13:25

14.12-3:
- using kolasa's patch to cover 4.0 kernel support

francoism commented on 2015-02-13 10:51

Thanks Vi0L0!
Hope AMD will release a new version soon. :)

Vi0L0 commented on 2015-02-12 21:02

14.12-2:
- using kolasa's patch to cover 3.19 kernel support

Would be a lot harder with xserver 1.17, it renders nasty segfault.
For now you can use [xorg116] repo:
[xorg116]
#Server = http://catalyst.wirephire.com/repo/xorg116/$arch
Server = http://mirror.rts-informatique.fr/archlinux-catalyst/repo/xorg116/$arch
#Server = http://mirror.hactar.bz/Vi0L0/xorg116/$arch

securitybreach commented on 2015-02-12 18:51

@beta990: I ended up reverting a few packages and disabled testing until this gets fixed.

francoism commented on 2015-02-12 12:11

@securitybreach: This is the problem with the catalyst-drivers.. we need AMD to release a new version.
From what I read support for AMD will be better on Linux 3.20.
Atm. there are not much you can do than block the packages/disable testing.

securitybreach commented on 2015-02-11 17:12

And now xorg-server is uninstallable due to the catalyst-total package:

error: failed to prepare transaction (could not satisfy dependencies)
:: catalyst-test: requires linux<3.19
:: catalyst-test: requires xorg-server<1.17.0

securitybreach commented on 2015-02-11 06:03

Amazing enough, this is the first issue in over a year of running the Testing repos. I know it is far from your fault but I just thought I would mention it.

securitybreach commented on 2015-02-11 04:32

Amazing enough, this is the first issue in over a year of running the Testing repos. I know it if far from your fault but just thought I would mention it.

securitybreach commented on 2015-02-11 04:13

I had to add both the kernel(linux) and catalyst-test to the IgnorePkg section in pacman.conf until catalyst-test is fixed.

On an unrelated noted, I also had to disable virtualbox-host-modules as well to be able to update.

securitybreach commented on 2015-02-11 04:12

I had to add both the kernel(linux) and catalyst-test to the IgnorePkg section in pacman.conf until catalyst-test is fixed.

securitybreach commented on 2015-02-11 04:07

I was completely wrong. I started noticing that stuff was launching slow and other slowdowns so I started amdcccle and noticed that all of my monitors were grayed out even though they were functioning. Then I launched glxinfo and there was an error. So I reverted to last month's clonezilla clone. Now I am stuck and I am not able to update anything without removing catalyst-test first. Any news on when the fix will be available?

error: failed to prepare transaction (could not satisfy dependencies)
:: catalyst-test: requires linux<3.19

securitybreach commented on 2015-02-10 18:11

It must have.

Vi0L0 commented on 2015-02-10 17:56

maybe something changed from the last RC then

securitybreach commented on 2015-02-10 17:39

comhack@Cerberus ~ % pacman -Qi linux | grep Version
Version : 3.19-1

comhack@Cerberus ~ % pacman -Qi catalyst-test | grep Version
Version : 14.12-1

securitybreach commented on 2015-02-10 17:38

It worked just fine even after rebooting.

Vi0L0 commented on 2015-02-10 17:34

I doubt it would work, i needed a patch to work on 3.19 RC.
Im afk, unable to update atm, but if you want you can use 3.19 patch from this pkg:
http://catalyst.wirephire.com/tarball/catalyst-test/catalyst-test-14.9-12.src.tar.gz
You can also use catalyst-hook package from [catalyst] repo - it already contains mentioned patch.

securitybreach commented on 2015-02-10 17:11

I fixed it by editing the PKGBUILD and changing 3.19 to 3.20

securitybreach commented on 2015-02-10 17:09

I am getting the following error (I am on testing):

catalyst-test: requires linux<3.19

francoism commented on 2015-02-01 13:54

@Hot4ruh1me: Sorry! Indeed installing and enabling acpid.service fixes the error! Thanks. :)

Hot4ruh1me commented on 2015-01-30 14:46

@beta990 Do you have ACPID installed and enabled in your system? You can find it in community repo with that name. Having the service running should fix your problem. Regards

francoism commented on 2015-01-30 14:41

Sorry for double posting:
[ 11.031] (WW) Open ACPI failed (/var/run/acpid.socket) (No such file or directory)

Think there is another path in a service wrong. Will try to find it, but if you know, please tell me. :)

francoism commented on 2015-01-30 14:24

Hi Vi0L0,

Thanks for providing this package!
Could you please fix this issue: https://wiki.archlinux.org/index.php/AMD_Catalyst#Failed_to_start_atieventsd.service

mgd commented on 2015-01-28 12:21

Not catalyst driver related in particular:
I recently updated my local pacman keyring as recommended on the ArchLinux webpage under https://www.archlinux.org/news/gnupg-21-and-the-pacman-keyring/

Since then I get an error for the catalyst repo from 'pacman -Syu' complaining about the signing key.

I tried to add the key as of the AUR wiki by 'pacman-key -r 653C3094' but that did not work.

Is the catalyst repo key not yet updated or am I doing something stupid ?

Vi0L0 commented on 2015-01-10 16:32

pacman -Qo /usr/lib/dri/fglrx_drv_video.so
+
as described below:
IMPORTANT
thing is that from now on theres this new symlink:
ln -snf /usr/lib/libXvBAW.so.1.0 ${pkgdir}/usr/lib/dri/fglrx_drv_video.so
which will conflict with xvba-video packages. Thats amd's will, maybe they will do something with it. Feel free to remove this file(symlink) from your system before updating (if you will be touched by a conflict).

untitaker commented on 2015-01-10 15:34

I get conflicting files:

catalyst-test: /usr/lib/dri/fglrx_drv_video.so exists in filesystem

AlexTsi commented on 2015-01-02 15:27

@Vi0L0 some update about xf86-video-intel thing. It seems fixed for me.
I have latest catalyst-test package.
latest xf86-video-intel package (local/xf86-video-intel 2.99.917-1 (xorg-drivers xorg))
latest xorg-server (local/xorg-server 1.16.3-2 (xorg))
and also installed catalyst-fix-gdm.

Then did "pxp_switch_catalyst amd", rebooted, and gdm started normally, using fglrx (checked with fglrxinfo). Great work!

Vi0L0 commented on 2014-12-20 19:41

quote from bar and grill:
I can see that git version already have this patch included (as well as other fixes). Can some pxp user check if xf86-video-intel-git works fine on their systems with xserver 1.16 installed? You can get this pkg from AUR or from the [mesa-git] repo or you can use mine if you want[1], i built it minutes ago for x86_64.

[1]http://mirror.rts-informatique.fr/archlinux-catalyst/files/xf86-video-intel-git/

Vi0L0 commented on 2014-12-18 21:41

silent update, no pkg rel bump: added catalyst-libgl, opencl-catalyst, lib32-catalyst-libgl and lib32-opencl-catalyst into provides array to satisfy libva-xvba-driver (and maybe other that will come in the future) deps

@AlexTsi: we got more confirmations for this intel bug, take a look at grill and bar

AlexTsi commented on 2014-12-13 19:10

@Vi0l0 I confirm that this package catalyst-test works on my pxp system. I have intel/amd laptop. So perhaps catalyst-pxp-test is no longer needed.

However, I need to mention, that latest fglrx still seems not to work with xf86-video-intel 2.99.916-3.. At least not for me. It would be great if someone else could confirm that too.

dfanz0r commented on 2014-12-12 05:16

So far im seeing a decent amount of fixes. Maybe a minor performance improvement not really sure.

Vi0L0 commented on 2014-12-11 22:02

updated to 14.12-1, "famous" catalyst omega
http://support.amd.com/en-us/kb-articles/Pages/AMDCatalystOmegaLINReleaseNotes.aspx

quick tests showed that this driver is pretty good.
2d rendering was never that fast on 290X (its almost as fluid as on oss radeon)
cs:go works like a charm now ;P.

gdm still segfaults

vaapi does work (but it has problems with osd, including subtitles) here without additional xvba-video packages
if you want to check it just add
export LIBVA_DRIVER_NAME=fglrx
into ~/.bashrc, and if that won't work for all your players you can create something like /etc/profile.d/fglrx-vaapi.sh with lets say:
LIBVA_DRIVER_NAME=fglrx
export LIBVA_DRIVER_NAME
LIBVA_DRIVERS_PATH=/usr/lib/dri/
export LIBVA_DRIVERS_PATH


IMPORTANT
thing is that from now on theres this new symlink:
ln -snf /usr/lib/libXvBAW.so.1.0 ${pkgdir}/usr/lib/dri/fglrx_drv_video.so
which will conflict with xvba-video packages. Thats amd's will, maybe they will do something with it. Feel free to remove this file(symlink) from your system before updating (if you will be touched by a conflict).

Janma8 commented on 2014-12-09 14:45

New (awesome) Driver Released :) http://www.phoronix.com/scan.php?page=news_item&px=MTg1NzU

Janma8 commented on 2014-12-09 14:25

New (awesome) Driver released :)

harry218 commented on 2014-10-25 10:04

Hi Vi0L0, thanks for this nice package. I'm quite new about this Catalyst stuff so I would like to ask which one of your packages based on the current beta catalyst? http://support.amd.com/en-us/kb-articles/Pages/latest-linux-beta-driver.aspx The AMD numbering is messy, this beta version is 14.6 while the stable one is already at 14.9.

klenamenis commented on 2014-10-04 17:23

Thanks! I wonder if it's possible to patch the propietary driver to support framerates > 60Hz. Using the radeon driver in conjuction with xrandr it was no problem to achieve this, but catalyst produces errors when trying to change it via xrandr.

Vi0L0 commented on 2014-10-04 06:48

depends on your kernel package. If it's standard and `catalyst_build_module all` works for it then systemd service will also work

klenamenis commented on 2014-10-04 02:24

The package built fine with xorg 1.16.1 and my own custom kernel, thanks. Can I still use the catalyst-hook systemd-service while using a custom kernel or do I have to rebuild after every kernel update?

securitybreach commented on 2014-10-03 20:24

Thanks. I do not know how I missed that.

Vi0L0 commented on 2014-10-03 20:13

ofcourse, take a look at the dep 'xorg-server<1.17.0' and older comment here:

Comment by Vi0L0 2014-09-12 18:26
updated to 14.9-10:
this release was prepared for Ubuntu 14.10 and is versioned by AMD as 14.201

Sorry for late update, I just get back from work's travel, ufff.
Works fine on xserver 1.16.

securitybreach commented on 2014-10-03 18:12

So does that mean that the xorg115 repo is no longer required?

Vi0L0 commented on 2014-10-03 18:08

Yes, it does support xserver 1.16.
Instead of stable 14.9 which lacks such support.

klenamenis commented on 2014-10-03 14:58

So this supports xorg-server 1.16.x? Why using ubuntu's package when http://www2.ati.com/drivers/linux/amd-catalyst-14-9-linux-x86-x86-64.zip is available?

Ginkgo commented on 2014-09-27 16:30

Just letting you know that the OpenCL 2.0 driver from http://catalyst.wirephire.com/tarball/14.9-opencl2/ works. resume/suspend works and OpenCL 2.0 seems to be there. Haven't tried any real 2.0 features yet.

It actually seems a bit more stable than 14.6.

Vi0L0 commented on 2014-09-26 20:26

updated to 14.9-11:
- as mentioned earlier I added my fresh patch for kernel 3.17 rc6.
Notes about the patch:
1. it probably wont work on 3.17 kernels older than rc6 because I don't know how to tell preprocessor about rc version :PP
2. im unable to test resume/suspend
3. im not the developer, just pressing random keys tongue so the patch could be less/more noobish :-)

Vi0L0 commented on 2014-09-26 20:25

updated to 14.9-11:
- as mentioned earlier I added my fresh patch for kernel 3.17 rc6.
Notes about the patch:
1. it probably wont work on 3.17 kernels older than rc6 because I don't know how to tell preprocessor about rc version tongue
2. im unable to test resume/suspend
3. im not the developer, just pressing random keys tongue so the patch could be less/more noobish :-)

Vi0L0 commented on 2014-09-26 19:56

Yes, I've seen, as noted on grill:
"I can see that phoronix posted catalyst 14.50 ... but do I really want to check some closed source kernel module coming from an unknown source? not sure :P"

Not to mention: would I like to share this "closed source kernel module coming from an unknown sourc"? And the answer is: No, in fact I think it would be unethical (to share unsecure, unconfirmed soft which can harm your(and mine) system).

So I think that I would like to wait for official 14.50 :)

Hot4ruh1me commented on 2014-09-26 18:58

Don't know if you've seen this article : http://www.phoronix.com/scan.php?page=news_item&px=MTc5NzI

14.50 it's comming soon =)

Vi0L0 commented on 2014-09-26 18:35

let me copy-paste from the grill:
Looks like this AMD OpenCL 2.0 Driver, versioned as 14.41 doesn't support xserver 1.16.

Too bad, I wanted to replace catalyst-test@AUR and [catalyst] repo with it, but I believe that most of you preffer xserver 1.16 support over OpenCL 2.0.

So for now those of you who wants to check openCL 2.0 can download package src from here: http://catalyst.wirephire.com/tarball/14.9-opencl2/
There you will find modified catalyst-test package with my fresh patch for kernel 3.17 rc6. Notes about the patch:
1. it probably wont work on 3.17 kernels older than rc6 because I don't know how to tell preprocessor about rc version tongue
2. im unable to test resume/suspend
3. im not the developer, just pressing random keys tongue so the patch could be less/more noobish :-)
Next to catalyst-test src pkg you will find opencl-headers2 src package (modified opencl-headers12 src)

I can see that there's a new version of codeXL, soon I will check it and update amdapp-codexl pkg (cant promise that it will happen today)

I'm also going to import 3.17rc6 patch into current catalyst-test@AUR and [catalyst] packages

Vi0L0 commented on 2014-09-25 19:27

uuuuuuh :) very nice finding! :D
Thanks Hot4ruh1me, I will look at it tomorrow.


Meanwhile I just get back to my keyboard :P compiled 3.17rc6 and can see this error when compiling fglrx:

cl_acpi.c: In function ‘KCL_ACPI_Slot_No_Hotplug’:
kcl_acpi.c:845:21: error: ‘struct acpi_device_flags’ has no member named ‘no_hotplug’

I will look at it probably tomorrow or the day after (gotta find time)
3.17rc5 works fine

Hot4ruh1me commented on 2014-09-25 17:30

Hey Vi0L0,

Just gonna leave this here in case is useful:

http://support.amd.com/en-us/kb-articles/Pages/OpenCL2-Driver.aspx

abandonedaccount commented on 2014-09-13 13:32

ok, I should've read the previous comments - I have no excuse.

abandonedaccount commented on 2014-09-13 13:07

Hi. It currently says "for Ubuntu 14.10" and it's downloading the archive from ubuntu dot com and other related changes to reflect that.
Was this intended?
Also could you use https where possible? not just because md5sums are breakable.
Please take these constructively. I appreciate your work!

Vi0L0 commented on 2014-09-12 18:28

Thanks for the head up scjet :)

Vi0L0 commented on 2014-09-12 18:26

updated to 14.9-10:
this release was prepared for Ubuntu 14.10 and is versioned by AMD as 14.201

Sorry for late update, I just get back from work's travel, ufff.
Works fine on xserver 1.16.
I tested:
games: dota2, l4d2, serious sam 2, civilization v, tf2
video: mplayer (vaapi), xbmc (xvba - latest xbmc version before xvba was dropped)
opengl benchs: heaven, valley
basic opencl apps

Vi0L0 commented on 2014-09-12 18:24

updated to 14.9-10:
ubuntu's 14.10 release, versioned by amd as 14.201

Sorry for late update, I just get back from work's travel, ufff.
Works fine on xserver 1.16.
I tested:
games: dota2, l4d2, serious sam 2, civilization v, tf2
video: mplayer (vaapi), xbmc (xvba - latest xbmc version before xvba was dropped)
opengl benchs: heaven, valley
basic opencl apps

Vi0L0 commented on 2014-09-12 18:22

Sorry for the lagg, I just get back from work and tested this.
Works fine on xserver 1.16.
I tested:
games: dota2, l4d2, serious sam 2, civilization v, tf2
video: mplayer (vaapi), xbmc (xvba - latest xbmc version before xvba was dropped)
opengl benchs: heaven, valley
basic opencl apps

scjet commented on 2014-09-12 06:37

http://www.phoronix.com/scan.php?page=news_item&px=MTc4NTQ
There this now ?

DottorGab commented on 2014-09-04 14:42

Thanks

Vi0L0 commented on 2014-09-03 19:18

About the new catalyst 14.30:

Once again Phoronix missed the truth. Phoronix wrote:
AMD put out the "Catalyst Software Suite for AMD Radeon R9 285" for Windows and Linux. While it's aimed for the R9 285, the driver will work with any supported Radeon GPU.

But it's not designed to work with cards other than 285.
It will work if you are updating, but it won't generate xorg.conf on systems with cards other than 285. aticonfig will say that they are not supported. So it does sucks for newcommers.

Look at the size of the control file:
$ ls -l 14.30/common/etc/ati/control
-rw-r--r-- 1 user users 312 08-29 15:54 14.30/common/etc/ati/control
$ ls -l 14.20/common/etc/ati/control
-rw-r--r-- 1 user users 89984 07-11 11:46 14.20/common/etc/ati/control

It has just one line...

That's why I won't update neither this nor catalyst-test-pxp nor the repo.

If you have 285 you can download sourceball and build the package by yourself here:
-test:
http://catalyst.wirephire.com/tarball/14.30-285/catalyst-test-14.9-1.src.tar.gz
-test-pxp:
http://catalyst.wirephire.com/tarball/14.30-285/catalyst-test-pxp-14.9-1.src.tar.gz

I versioned it as 14.9 so it would be easier to (auto-)update it with cower/yaourt/aur


And yes... there was no repo update for some time... we are still waiting for the new driver (which is suppose to fix opencl)

Vi0L0 commented on 2014-09-03 19:15

About the new catalyst 14.30:

Once again Phoronix missed the truth. Phoronix wrote:
AMD put out the "Catalyst Software Suite for AMD Radeon R9 285" for Windows and Linux. While it's aimed for the R9 285, the driver will work with any supported Radeon GPU.

But it's not designed to work with cards other than 285.
It will work if you are updating, but it won't generate xorg.conf on systems with cards other than 285. aticonfig will say that they are not supported. So it does sucks for newcommers.

Look at the size of the control file:
$ ls -l 14.30/common/etc/ati/control
-rw-r--r-- 1 user users 312 08-29 15:54 14.30/common/etc/ati/control
$ ls -l 14.20/common/etc/ati/control
-rw-r--r-- 1 user users 89984 07-11 11:46 14.20/common/etc/ati/control

It has just one line...

That's why I won't update neither this nor catalyst-test nor the repo.

If you have 285 you can download sourceball and build the package by yourself here:
-test-pxp:
http://catalyst.wirephire.com/tarball/14.30-285/catalyst-test-pxp-14.9-1.src.tar.gz
-test:
http://catalyst.wirephire.com/tarball/14.30-285/catalyst-test-14.9-1.src.tar.gz

I versioned it as 14.9 so it would be easier to (auto-)update it with cower/yaourt/aur

DottorGab commented on 2014-08-26 15:40

For the repo Server = http://catalyst.wirephire.com/repo/catalyst/$arch, because there was no update?

theandy94 commented on 2014-08-24 12:56

Oh... well, I've completely ignored the "<"...
Thank you for pointing that out, now everything works fine.
-theandy

DottorGab commented on 2014-08-24 08:37

On 1.16 doesn't work

"xorg-server<1.16.0"

you must use repo [xorg115]

theandy94 commented on 2014-08-22 23:17

Hi, I'm fairly new to Arch Linux and I always run into a problem, when I try to install catalyst-test.
Terminal gives me xorg-server 1.16.0 not found.
I already checked which xorg version I'm using, I'm on version 1.16.0-6.
I really appreciate any help.
Thanks in advance.
-theandy94

DottorGab commented on 2014-08-20 07:48

there are news for Catalyst and GDM? thank ViOLO

Vi0L0 commented on 2014-08-18 19:57

-12: typo

Vi0L0 commented on 2014-08-18 19:50

14.6-11: added dirty GPL-licensing solution - fglrx_gpl_symbol.patch - workaround for compilation error for i686 on 3.16.1 kernel

mid-kid commented on 2014-08-17 20:46

Oh, now I understand.
Thanks for explaining.

Vi0L0 commented on 2014-08-17 10:15

It's a driver for embedded gpus:
http://www.amd.com/en-us/products/embedded/graphics

AMD is releasing it from time to time:
http://support.amd.com/en-us/download/embedded/previous?os=Linux

It's build on a top of "normal" catalyst, thats why it's often[1] working with pcs gpus.

It's control file is different/crippled, I'm pretty sure that aticonfig will tell you that your adapter is not supported once you will remove xorg.conf - it means problems for newcommers.

If it would be any better than this 14.6 beta then I would think about maybe creating some additional package or something... but on my 290x its no good - it's dropping/missing frames on heaven 4.0 (the only test i made) while having same fps. Its easy to see with unarmed eye - maybe its touched by famous "frame pacing" bug(?)


I know that Phoronix is a good source of news, Im also reading it, but I think that Michael should look at and wrote about the details. It's not the first time he's missing "embedded" magic word...


[1] once I did update to embedded and get comments that its not working at all on some gpus.

niqingliang2003 commented on 2014-08-17 00:56

14.8 is out:
http://support.amd.com/en-us/download/embedded?os=Linux+x86&rev=13.151

mid-kid commented on 2014-08-16 19:40

http://www.phoronix.com/scan.php?page=news_item&px=MTc2MzM

Don't know why it's called "embedded", but I just installed it, and it works.
Couldn't find a way to download it with curl, tho.

Vi0L0 commented on 2014-08-15 21:11

Heard about it, it's an embedded graphics driver, useless for us. But keep the head up, heard that new driver should be released soon, hope before my travel

Hot4ruh1me commented on 2014-08-15 15:21

Hey Vi0L0, AMD left a silent update of the driver last week. In case you didn't see it, here is the link: http://support.amd.com/en-us/download/embedded?os=Linux+x86&rev=13.151

Keep up the amazing work! =)

SmilingWolf commented on 2014-08-05 07:05

Not working. Installed
xorg-server-1.16
xorg-server-common-1.16
xf86-input-evdev-2.9.0-2
along to this package and the fglrx module simply refused to load. Had to roll back to xorg-server-1.15.2-1, xorg-server-common-1.15.2-1 and xf86-input-evdev-2.9.0-1.

SmilingWolf commented on 2014-08-05 07:03

Not working. Installed
xorg-server-1.16
xorg-server-common-1.16
xf86-input-evdev-2.9.0-2
along to this package and the fglrx module simply refused to load.

Ryp commented on 2014-07-30 12:51

Not working with 1.16 for me neither:
xorg-server: 1.15.2-1
catalyst-test-pxp: 14.6-10

Ryp commented on 2014-07-30 12:50

Not working for me neither:
xorg-server: 1.15.2-1
catalyst-test-pxp: 14.6-10

Ryp commented on 2014-07-30 12:47

Not working for me neither =)

0x017 commented on 2014-07-29 19:48

Don't working :'(
http://pastebin.com/WDB4B1Dn
xorg crash when /usr/lib/xorg/modules/extensions/libglx.so try loading.

Vi0L0 commented on 2014-07-29 18:49

check if it's working with xorg-server 1.16 for you (feel free to change dependencies), if it is - let us know.
If not - let us know :) and use [xorg115] repo:
https://wiki.archlinux.org/index.php/AMD_Catalyst#Xorg_repositories

0x017 commented on 2014-07-29 16:42

error failed to commit transaction (conflicting files)
:: catalyst-test: required xorg-server<1.16.0

$ pacman -Qe | grep catalyst-test
catalyst-test 14.6-10

$ X -version
X.Org X Server 1.15.2


Vi0L0 commented on 2014-07-20 11:00

maybe pgp.mit server catched some temp problem, it works atm.
You can also run without specifying server ie:

pacman-key --recv-keys 0xabed422d653c3094
pacman-key --lsign-key 0xabed422d653c3094

(it used pool.sks-keyservers.net server here)

securitybreach commented on 2014-07-20 05:31

Vi0L0, I am having problems adding your key. Getting the error: 0xabed422d653c3094 could not be locally signed

I followed the directions on you user page:
https://wiki.archlinux.org/index.php/User:Vi0L0

securitybreach commented on 2014-07-20 05:31

Vi0L0, I am having problems adding your key. Getting the error: 0xabed422d653c3094 could not be locally signed

I followed the directions on you user page:
https://wiki.archlinux.org/index.php/User_talk:Vi0L0

securitybreach commented on 2014-07-20 04:19

BTW your key is not working right now:
root@Cerberus /home/comhack # pacman-key --keyserver pgp.mit.edu --recv-keys 0xabed422d653c3094
gpg: requesting key 653C3094 from hkp server pgp.mit.edu
gpg: keyserver timed out
gpg: keyserver receive failed: Keyserver error
==> ERROR: Remote key not fetched correctly from keyserver.

Vi0L0 commented on 2014-07-19 19:24

If 1.16 is working for you then you can leave it.
History shows that catalysts which gains fresh xserv support often suffers for a childhood disease. It may be more or less noticable, depending on systems config, but occurs. So it would be probably better to use older xserv versions for two or three months from initial support release, till things will be fixed. But its comming from history, and could change (though i see this every year).

I'm using newest catalyst because generally (key word) its the best release.

You can use xorg repo with any other repo, just put it atop of the others

securitybreach commented on 2014-07-19 17:57

Most of the reason I used the catalyst-test was that it worked with current xorg versions. I guess I will continue with this version anyway.

securitybreach commented on 2014-07-19 17:56

Ok I do have another machine with a 6790 in it that I have not updated yet. What method (repo) do you think I should use and will this work with the testing branch(conflicts)?

securitybreach commented on 2014-07-19 17:55

comhack@Cerberus ~ % pacman -Qi xorg-server | grep Version
Version : 1.16.0-2
comhack@Cerberus ~ % pacman -Qi catalyst-test | grep Version
Version : 14.6-10
comhack@Cerberus ~ %

Vi0L0 commented on 2014-07-19 17:52

<thinking> hmm... looks like I misunderstood you, because I thought that you are running xserver 1.16, but your xorg.0.log is telling otherwise (?).

Umm... catalyst had problems with supporting fresh versions of xorg-server. During 5 years only like two or three times it was otherwise - because there was no changes in xservers ABI or there were minor changes. The cure is often to use this xorg repo (which is just a simple copy-paste of packages from arch extra repo) or block update to newest xserver version.

But lets get it straight - are you running catalyst on xserv 1.16?
I had no luck, I didn't hear any news about 1.16 support, I heard that xserv devs changed abi in a late stage of developement and it pissed off nvidia :) because they were supporting 1.16 rc2 (or something) and then bam - support was broken and they had to compile and test it once again - and its taking time.
It pissed off nvidia but they made it on time. Now lets take AMD and their loong testing procedures... I would be really surprised if they would do this on time. Seriously.
But everything is possible, and there could be something on my side, some missconfig or something. That's why I'm asking you so many questions :) and thats why I asked for confirmation

securitybreach commented on 2014-07-19 17:16

You mention building xorg114 and xorg115 but will it work with the current versions (1.16.0.2) like the previous versions did. I know mine is working but I am referring to how catalyst-test didn't require a custom xorg repo(downgraded version).

securitybreach commented on 2014-07-19 17:15

You mention building xorg114 and xorg115 but will it work with the current builds (1.16.0.2) like the previous versions did. I know mine is working but you know what I mean.

securitybreach commented on 2014-07-19 17:12

xorg-conf: http://ix.io/dto
Xorg.0.log: http://ix.io/dtp

Vi0L0 commented on 2014-07-19 17:01

I checked also my 7850, same error. Everything works well though with testing + xorg114 or xorg115 (im installing only 3 packages from xorg repo).
Maybe you changed some system conf file recently? Could you post your xorg.conf and xorg.0.log somewhere?

securitybreach commented on 2014-07-19 14:08

I have a XFX Radeon HD 7870 and no, I am not using any special hack or option.

Vi0L0 commented on 2014-07-19 10:07

I'm not that lucky, getting "symbol lookup error <path_here>/libglx.so: undefined symbol: LoadExtension"
IgnoreABI didn't help :/
Did you use some hack or special option? :> btw which gpu?

securitybreach commented on 2014-07-19 01:22

I'm on the testing Branch and the new xorg-server works beautifully with this package.

Vi0L0 commented on 2014-07-18 22:14

Yep, xorg-server 1.16 is in [testing], tomorrow I should be able to create [xorg115] repo

securitybreach commented on 2014-07-18 16:16

I just got this error:
error: failed to prepare transaction (could not satisfy dependencies)
:: catalyst-test: requires xorg-server<1.16.0

Vi0L0 commented on 2014-07-16 17:45

Thank you ChemBro :* :>

Vi0L0 commented on 2014-07-16 17:39

updated to 14.6 beta 1.0 jul11:
- changes? what changes? :>
- couple of libs and confs differs
- steam games are smooth
- gdm still segfaults

ChemBro commented on 2014-07-15 21:53

New version is something like 14.6beta-jul11, though I don't know what is new, because AMD did not change the changelog.

dequis commented on 2014-06-13 08:43

Oh dammit sorry, misfire, didn't want to flag out of date.

mid-kid commented on 2014-06-12 19:16

I understand, but I still think it shouldn't be like that.
I'll just stick to patching the PKGBUILD.

Vi0L0 commented on 2014-06-11 22:17

catalyst-hook originaly came without dependency on any kernel, because I was (am) not using linux package, and seen no sense in such a dep. But... people were still posting about problems, even such simple like no headers. optdep was not enought, there was no cure, thanks to explicit dep its all gone. Im not gonna walk without it.

You can check out catalyst-generator + utils, you wont find such dep there, and its closer to ancient arch way

mid-kid commented on 2014-06-11 20:51

Please remove the explicit dependency on linux, it makes it impossible to remove it when switching to another kernel.

AlD commented on 2014-06-10 09:58

Is none of you affected by the DPMS crash bug? http://ati.cchtml.com/show_bug.cgi?id=1026
Any ideas what might be causing this? I have an HD 6800.

clfarron4 commented on 2014-06-09 18:22

Ooh, catalyst-test works with the main 3.15 kernel (post rc release) quite nicely.

mgd commented on 2014-05-30 19:39

Just for the record:
This version seems to have issues with sgminer. At least the settings that used to work since 13.11 and up to 14.4-20 no longer work. Therefor I reverted back to 14.4-20 for the time being.

Vi0L0 commented on 2014-05-29 19:59

gdm still seg faults here...

marmotta commented on 2014-05-29 16:49

Any news for gdm?

Vi0L0 commented on 2014-05-28 19:19

good point mid-kid, removed this historical entry from optdeps (no pkrel bump)

mid-kid commented on 2014-05-28 19:01

linux-headers is in both depends and optdepends... Derp?

securitybreach commented on 2014-05-28 17:57

Thanks a lot Vi0L0

Vi0L0 commented on 2014-05-28 17:39

btw i like this new aur :)

Vi0L0 commented on 2014-05-28 17:38

updated to 14.6 beta 1:
- http://support.amd.com/en-us/kb-articles/Pages/latest-linux-beta-driver.aspx
- for me it means better performance in serious sam 3
- phoronix forums user Espionage724 said: "I have a under/overscan option for my laptop's built-in screen now"

Vi0L0 commented on 2014-05-25 11:35

updated to 14.4-20 (catalyst 14.4 rev2):
- which brings some fixes for Witcher 2

Thanks for the head up, Hot4ruh1me

Vi0L0 commented on 2014-05-25 11:29

updated to 14.4-20 (catalyst 14.4 rev2):
- which brings some fixed for Witcher 2


Thanks for the head up, Hot4ruh1me

Hot4ruh1me commented on 2014-05-23 20:34

Hey Vi0L0,
I gotta say, you do an outstanding job mantaining this package compatible to arch. Kudos on that from me and my bf.
So, beside the congrats, here ar my two cents in case you didn't know:
AMD released Catalyst™ 14.4 for Linux. Here are the RN: http://support.amd.com/en-us/kb-articles/Pages/AMDCatalyst14-4LINReleaseNotes.aspx

Thanks for the hard work!

Vi0L0 commented on 2014-05-20 10:12

Should be very easy.
Just install this package and generate xorg.conf with aticonfig (with one monitor should be `aticonfig --initial`, for more info read here:https://wiki.archlinux.org/index.php/AMD_Catalyst#Configuring_the_driver)

While installing this package pacman should ask you about removing mesa-libgl, do it.
You could remove oss ati packages but you don't have to :P, just be sure that you don't have xorg.conf(.d) files with radeon inside.
You can also add nomodeset or radeon.modeset=0, but it should not be a "must have".



Now let say about what's fun with this package :).

Actually if you didn't mess with your system you should be able to have both oss radeon and catalyst installed and switch between them with my pxp_switch_catalyst script, only reboot is required. It's working well on my system, so should in yours. Only if you want to get it to work with your custom xorg.conf radeon file, you will have too look at mentioned script.

pxp_switch_catalyst in a basic way
( so if you don't want to change anything in you system, which means no graphic bootloader modifies [NO nomodeset NOR radeon.modeset=0 flags added, also NO radeon.dpm=1,no vga, no video nor any gfx boot option], and don't have xorg.conf(.d) with radeon inside )
AND if you only have one graphics card AND it is radeon, should work as follows:
- pxp_switch_catalyst amd - this command will switch to catalyst, xorg.conf is set to use fglrx, fglrx is added to modules load, reboot and you are working on catalyst.
Short story: at boot it will load fglrx because its inside modules-load, it will also load radeon module becase we didn't add nomodeset, but Xserver will use fglrx because it's inside xorg.conf, and it should work well.
- pxp_switch_catalyst intel - this command will switch to oss driver, so in our case to radeon (command could misslead, yes, but its because its designed to pxp switch), script removes fglrx from modules-load, and removes xorg.conf, reboot and you are working on oss radeon.
Short story: script removes fglrx from modules-load so it's not loaded, radeon is loaded, xorg.conf doesn't exist so by default xserver is using radeon.

Vi0L0 commented on 2014-05-20 10:04

Should be very easy.
Just install this package and generate xorg.conf with aticonfig (with one monitor should be `aticonfig --initial`, for more read here:https://wiki.archlinux.org/index.php/AMD_Catalyst#Configuring_the_driver)

While installing this package pacman should ask you about removing mesa-libgl.
You could remove oss ati packages but you don't have to, just be sure that you don't have xorg.conf(.d) files with radeon inside.
You can also add nomodeset or radeon.modeset=0, but it should not be a "must have".

Now let say about what's fun with this package :).

Actually if you didn't mess with your system you should be able to have both oss radeon and catalyst installed and switch between them with my pxp_switch_catalyst script, only reboot is required. It's working well on my system, so should in yours. Only if you want to get it to work with your custom xorg.conf radeon file, then you have too look at mentioned script.

pxp_switch_catalyst in a basic way
( so if you don't want to change anything in you system, which means no bootloader modifies [NO nomodeset NOR radeon.modeset=0 flags added], and don't have xorg.conf(.d) wiith radeon inside )
AND if you only have one graphics card and its radeon, should work as follows:
- pxp_switch_catalyst amd - this command will switch to catalyst, xorg.conf is set to use fglrx, fglrx is added to modules load, reboot and you are working on catalyst.
Short story: at boot it will load fglrx because its inside modules-load, it will also load radeon module becase we didn't add nomodeset, but Xserver will use fglrx because it's inside xorg.conf, and it should work well.
- pxp_switch_catalyst intel - this command will switch to oss driver, so in our case to radeon (command could misslead, yes, but its because its designed to pxp switch), script removes fglrx from modules-load, and removes xorg.conf, reboot and you are working on oss radeon.
Short story: script removes fglrx from modules-load so it's not loaded, radeon is loaded, xorg.conf doesn't exist so by default xserver is using radeon.

moggers commented on 2014-05-18 09:28

Is there an idiot proof guide to switching from xf86-video-ati to catalyst-test? There's a slew of cross dependancies and crap that makes the whole thing a nightmare, like libEGL/libGLES conflicts.

Vi0L0 commented on 2014-05-01 19:21

There's a bug in 3.14.2:
https://bugs.archlinux.org/index.php?do=details&action=details.addvote&task_id=40171

Temporary workaround is to downgrade linux and linux-headers to 3.14.1 or use gcc from [testing]. The first solution should be imho safer. You can find 3.14.1 at arm:

x86_64:
ftp://seblu.net/archlinux/arm/2014/04/25/core/os/x86_64/linux-3.14.1-1-x86_64.pkg.tar.xz
ftp://seblu.net/archlinux/arm/2014/04/25/core/os/x86_64/linux-headers-3.14.1-1-x86_64.pkg.tar.xz

i686:
ftp://seblu.net/archlinux/arm/2014/04/25/core/os/i686/linux-3.14.1-1-i686.pkg.tar.xz
ftp://seblu.net/archlinux/arm/2014/04/25/core/os/i686/linux-headers-3.14.1-1-i686.pkg.tar.xz

Synch commented on 2014-05-01 10:54

Getting an error while trying to install, GCC supports ‘-fstack-protector-strong’ from version 4.9, but the core provides 4.8, installing the gcc from testing solves the issue.

make: Entering directory '/usr/lib/modules/3.14.2-1-ARCH/build'
Makefile:615: Cannot use CONFIG_CC_STACKPROTECTOR_STRONG: -fstack-protector-strong not supported by compiler
CC [M] /tmp/catalyst.3bC0Fa/firegl_public.o
gcc: error: unrecognized command line option ‘-fstack-protector-strong’

kamelie1706 commented on 2014-05-01 08:18

found answered in forum
https://bbs.archlinux.org/viewtopic.php?id=57084&p=261

kamelie1706 commented on 2014-05-01 07:57

lib32-catalyst-utils : /usr/lib32/libAMDXvBA.so.1 est déjà présent dans le système de fichiers
lib32-catalyst-utils : /usr/lib32/libXvBAW.so.1 est déjà présent dans le système de fichiers
lib32-catalyst-utils : /usr/lib32/libatiuki.so.1 est déjà présent dans le système de fichiers

in english "library already in file system"

Any idea how to proceed?

retr0virus commented on 2014-04-29 19:14

Ah, sorry... your answer was too fast. ^^
I tried again and this time it was 14.4 already... maybe I used my old 14.3 tarball (same directory...)
Thanks anyway. :-)

Vi0L0 commented on 2014-04-29 19:08

you mean the tarball of this, catalyst-test, package?
downloads fine here

retr0virus commented on 2014-04-29 19:00

The tarball seems to be still at 14.3.

anadon commented on 2014-04-29 00:16

Despite more hours, I still can't get this to work with the a10-7850k.

Vi0L0 commented on 2014-04-27 11:44

14.4-11:
- fixing symlinks, making them more fit to mesa's one

Vi0L0 commented on 2014-04-26 14:02

updated to 14.4-10, catalyst 14.4 stable

it's hard to find difference between stable version and rc :P

experimental pxp support got new functionality. new catalyst-test release is conflicting and providing mesa-libgl and mesa-libgl-git - yes providing both could be messy if mesa-libgl-git would work on other libs versions than mesa-libgl, but I didn't notice version change for months, so lets check how it will work.
It also doesn't conflict with any oss driver (either intel or ati), so with it you should be able to perform a switch from one to another driver without installing/removing packages, simply run pxp_switch_catalyst to link libs and xorg.conf. Tested on 290x and 3.14 kernel with catalyst and oss ati, with [mesa-git], it worked - without 3d acceleration because afaik it's not available for 290x on 3.14 kernel, but kms and 2d accel were working just fine (could be that dpm also since fan was working quietelly).

securitybreach commented on 2014-04-25 02:10

@AID Thanks a lot

AlD commented on 2014-04-24 18:58

@securitybreach: http://pkgbuild.com/git/aur-mirror.git/tree/catalyst-test?id=414a82e5df916a79bd4811fab405cd0cf1d17bc9

securitybreach commented on 2014-04-24 18:31

Can someone give me the previous package version so I can downgrade?

C8483 commented on 2014-04-23 22:04

This build finally fixed flickering in chrome and spotify-linux for me. Among other problems! Wooo!

securitybreach commented on 2014-04-23 18:13

Also, I tried running the suggested command:
aticonfig --set-pcs-val=MCIL,DigitalHDTVDefaultUnderscan,0
per the wiki, but it just made it too big. Basically the top of my terminal is cut off now.

securitybreach commented on 2014-04-23 18:09

Also, I tried running the suggested command: aticonfig --set-pcs-val=MCIL,DigitalHDTVDefaultUnderscan,0 per the wiki for my monitor and number and all it made it too big. Basically the top of my terminal is cut off now.

securitybreach commented on 2014-04-23 17:38

With the latest build, the Scaling Options panel in Adjustments is missing so I have a black bar around one of my hdmi monitors? Is there another way of fixing this or should I roll back the version? Thanks

securitybreach commented on 2014-04-23 17:33

With the latest build, the Scaling Options panel un Adjustments is missing so I have a black bar around one of my hdmi monitors? Is there another way of fixing this or should I roll back the version? Thank

securitybreach commented on 2014-04-23 16:58

With the latest build, the Scaling Options panel is missing so I have a black bar around one of my hdmi monitors? Is there another way of fixing this or should I roll back the version? Thanks

securitybreach commented on 2014-04-23 16:57

With the latest build, the Scaling Options panel is missing so I have a black bar around one of my monitors? Is there another way of fixing this or should I roll back the version? Thanks

Vi0L0 commented on 2014-04-22 21:06

btw i will update repo tomorrow

Vi0L0 commented on 2014-04-22 20:59

updated to 14.4-1, catalyst 14.4 rc 1.0:
- opengl 4.4 support
- i'm finally having good 2D performance on 290x with this one

@rockjock:
this is a known bug related with gdm, you may temporary downgrade gdm or use other dm

rockjock commented on 2014-04-20 22:40

Using Gnome 3.12. Xorg fails to start on boot; system hangs on message "Reached target graphical interface". Alt-F2 followed by startx works though.

springah commented on 2014-04-18 13:35

new install here - unable to wake from suspend, nomodeset and vga=0 set, radeon 6870. otherwise, working great!

Ganitzsh commented on 2014-04-16 16:46

Hi,

I updated my kernel to 3.14.1 unfortunately I havent been careful enough, and now my display is not working anymore :/ I removed catalyst to install open source drivers instead. Is it the same for everyone?

Vi0L0 commented on 2014-03-30 11:56

Place of where mesa's libs are stored is different.
Also nvidia started to provide egl libs, so egl symlinks were moved from mesa to mesa-libgl and because we're in conflict with mesa-libgl we had to provides egl symlinks on our own.

mcxplode commented on 2014-03-29 20:47

Violo you have done a peachy job mantaining this driver. ----====:)====----
Catalyst Beta 14.3 fixes longterm bug with Planetary Annihilation. Up until now PA would only run with software-ui because of gpu_buffer_intialize_bug which occurs with coherent-ui. Coherent-ui uses chromium v28 and thats were the bug occurs so it also fixes AMD GPU acceleration with certain versions of chromium. PA works great now with this driver.
I am suprised I didn't break my system updating to mesa 10.1.o-4, haven't updated my Arch system in over a year, but beta drivers work great. What changed with new version of mesa now that properiatory drivers depend on specific version? Is it the symbolic linking thats changed because catalyst doesn't use mesa libs it symlinks to its own libs.

Vi0L0 commented on 2014-03-22 17:09

thanks retr0virus.
I updated package one more time, just without pkgrel bump (there was no such a need, previous warning was harmless).
Yes it was related with recent pxp changes, I changed method a bit, now there's no warning.

retr0virus commented on 2014-03-22 16:32

I get the following output when installing the new 14.3-1 (german language):
:: Installation fortsetzen? [J/n] j
(1/1) Prüfe Schlüssel im Schlüsselring [######################] 100%
(1/1) Überprüfe Paket-Integrität [######################] 100%
(1/1) Lade Paket-Dateien [######################] 100%
(1/1) Prüfe auf Dateikonflikte [######################] 100%
(1/1) Überprüfe verfügbaren Festplattenspeicher [######################] 100%
/tmp/alpm_I7j3pu/.INSTALL: Zeile 50: usr/lib/fglrx/switchlibGL: Datei oder Verzeichnis nicht gefunden
/tmp/alpm_I7j3pu/.INSTALL: Zeile 51: usr/lib/fglrx/switchlibglx: Datei oder Verzeichnis nicht gefunden
(1/1) Aktualisiere catalyst-test [######################] 100%
- /usr/lib/modules/3.10.33-1-lts looks like unused, maybe remove it manualy?
- removing fglrx module from /usr/lib/modules/3.13.6-1-ARCH
+ building fglrx module for 3.13.6-1-ARCH kernel ...
ok.
(And so on...)
I am not sure if the output about the two missing files switchlibGL and switchlibglx already existed in 14.2. Maybe it has something to do with your added check_switch?

Vi0L0 commented on 2014-03-22 16:30

14.3-1:
- xv video output works fine now on xserver 1.15
- some opengl fixes

- I added 'experimental powerXpress support'. Idea is descibed here [1]. I decided to add check_switch into .install file. Before extracting files it's checking if your pxp is set to intel, and if so it's switching to intel after extraction. check_switch implementation is n00bish (because Im creating file to store data) and I don't like it, maybe someone knows some hack-trick which will allow to transfer data from pre_upgrade to post_upgrade?

[1] https://bbs.archlinux.org/viewtopic.php?pid=1391639#p1391639

Vi0L0 commented on 2014-03-22 16:28

14.3-1:
- xv video output works fine now on xserver 1.15
- some opengl fixes

- I added 'experimental powerXpress support'. Idea is descibed here [1]. I decided to add check_switch into .install file. Before extracting files it's checking if your pxp is set to intel, and if so it's switching to intel after extraction. check_switch implementation is n00bish and I don't like it, maybe someone knows some hack-trick which will allow to transfer data from pre_upgrade to post_upgrade?

[1] https://bbs.archlinux.org/viewtopic.php?pid=1391639#p1391639

Vi0L0 commented on 2014-03-18 18:24

guys,
i dont have access to my pc and so i won't be able to update before late friday/saturday

doomguard88 commented on 2014-03-18 10:50

14.3 is out.

Ganitzsh commented on 2014-03-16 23:39

Thx! It works well with GLX (XCB).

kamelie1706 commented on 2014-03-15 16:58

new package fix my issue, Thx!

kamelie1706 commented on 2014-03-15 16:55

ok then it explains that
https://bbs.archlinux.org/viewtopic.php?id=178563

I should hold on when I see mesa update :-(

kamelie1706 commented on 2014-03-15 16:54

ok then explain that
https://bbs.archlinux.org/viewtopic.php?id=178563

I should hold on when I see mesa update :-(

Vi0L0 commented on 2014-03-15 16:37

rebuild because of mesa 10.1.0-4 changes

okubax commented on 2014-03-15 08:56

Ganitzsh: change the video output in vlc or mplayer or downgrade xorg to 1.14 using the xorg repo

kamelie1706 commented on 2014-03-15 08:31

Try vlc and change video output. Works for me with video output GLX(XCB).

Ganitzsh commented on 2014-03-14 16:08

Hi, I'm having trouble watching a video since the update. When I open a video my session crash and I have to reboot. Any ideas?

jbernardo commented on 2014-03-14 13:11

I am still getting target not found when I run "pacman -Syu catalyst-utils catalyst-libgl opencl-catalyst lib32-catalyst-utils lib32-catalyst-libgl lib32-opencl-catalyst ":

$ sudo pacman -Syu catalyst-utils catalyst-libgl opencl-catalyst lib32-catalyst-utils lib32-catalyst-libgl lib32-opencl-catalyst
[sudo] password for user:
:: A sincronizar a base de dados de pacotes...
core está actualizado
extra está actualizado
community está actualizado
multilib está actualizado
archlinuxfr está actualizado
erro: alvo não encontrado: catalyst-utils
erro: alvo não encontrado: catalyst-libgl
erro: alvo não encontrado: opencl-catalyst
erro: alvo não encontrado: lib32-catalyst-utils
erro: alvo não encontrado: lib32-catalyst-libgl
erro: alvo não encontrado: lib32-opencl-catalyst

:(

kamelie1706 commented on 2014-03-11 05:03

Thx. Works for me.

Vi0L0 commented on 2014-03-10 22:23

should be fine now

kamelie1706 commented on 2014-03-10 21:45

Still some challenges with some packages

x86_64:
pacman -Syu catalyst-utils catalyst-libgl opencl-catalyst lib32-catalyst-utils lib32-catalyst-libgl lib32-opencl-catalyst

erreur : échec de récupération du fichier « lib32-catalyst-utils-14.2-2-x86_64.pkg.tar.xz » depuis catalyst.wirephire.com : The requested URL returned error: 404 Not Found
avertissement : erreur lors de la récupération de certains fichiers
erreur : échec de récupération du fichier « lib32-catalyst-libgl-14.2-2-x86_64.pkg.tar.xz » depuis catalyst.wirephire.com : The requested URL returned error: 404 Not Found
avertissement : erreur lors de la récupération de certains fichiers
erreur : échec de récupération du fichier « lib32-opencl-catalyst-14.2-2-x86_64.pkg.tar.xz » depuis catalyst.wirephire.com : The requested URL returned error: 404 Not Found
avertissement : erreur lors de la récupération de certains fichiers
erreur : la validation de la transaction a échoué (erreur de la bibliothèque de téléchargement)

Vi0L0 commented on 2014-03-10 21:30

it's because of https://bbs.archlinux.org/viewtopic.php?pid=1390853#p1390853 :

Repo users should note that:
- catalyst-utils is now splitted into:
catalyst-utils
catalyst-libgl
opencl-catalyst
- lib32-catalyst-utils is now splitted into:
lib32-catalyst-utils
lib32-catalyst-libgl
lib32-opencl-catalyst

You MOST probably want to also install catalyst-libgl and (if using x86_64) lib32-catalyst-libgl


In other words to install all packages:
i686:
pacman -Syu catalyst-utils catalyst-libgl opencl-catalyst

x86_64:
pacman -Syu catalyst-utils catalyst-libgl opencl-catalyst lib32-catalyst-utils lib32-catalyst-libgl lib32-opencl-catalyst

kamelie1706 commented on 2014-03-10 20:41

Tried your package but got a conflict

I choose the first option
1) catalyst-libgl

I get
:: imagemagick : requiert libcl
:: lib32-cairo : requiert lib32-libgl
:: lib32-glu : requiert lib32-libgl
:: lib32-qt4 : requiert lib32-libgl
:: steam : requiert lib32-libgl

kamelie1706 commented on 2014-03-10 08:43

no hurry, I can hold update for a while. Highly appreciate your support!

Vi0L0 commented on 2014-03-10 08:36

I'm planning to update repo today (I will get back home in 8 hours). I want to rebuild packages and provide catalyst-libgl, so it will take me some time

kamelie1706 commented on 2014-03-10 06:59

Thx Vi0L0. Any plan to update your package repo?

waders commented on 2014-03-10 00:45

Ty once again Vi0L0.

Vi0L0 commented on 2014-03-09 22:25

updated:
- added mesa and lib32-mesa to the dependencies lists
- added symlinks to egl and gles libs of mesa and lib32-mesa

@noobfromby:
modprobe fglrx
command is loading fglrx module to the kernel, check if there's no errors

after this update downgrade of mesa isn't necessary

noobfromby commented on 2014-03-09 18:58

@Vi0L0
What i can do? What is modprobe fglrx?

I reinstall catalyst after this issue and downgrade linux but issue not solved.

I delete symlink, error with downgrades mesa also not solved(

Pls help, how to solve problem with exist file in file system.

noobfromby commented on 2014-03-09 15:38

@Vi0L0
What i can do? What is modprobe fglrx?

I reinstall catalyst after this issue and downgrade linux but issue not solved.

I delete symlink, error with downgrades mesa also not solved(

Vi0L0 commented on 2014-03-09 14:44

@noobfromby:
are you sure that fglrx module was rebuilt correctly?
Is:
modprobe fglrx
working?

Downgrade is returning an error because you created a symlink.

kamelie1706 commented on 2014-03-09 13:59

Thx ... will wait the package to test

noobfromby commented on 2014-03-09 12:48

@VI0L0
Your method with symlinc doesn't help me(

noobfromby commented on 2014-03-09 12:01

@kamelie1706
I want to downgrade mesa, but when i do sudo pacman -U /var/cache/pacman/...
I see the following:
mesa-libEGL.so.1.0.0 libEGL.so:exist in filesystem
ect...(many files)
what is mean??? My mesa don't downgrade(

Vi0L0 commented on 2014-03-09 10:42

cd /usr/lib
ln -s mesa-libEGL.so.1.0.0 libEGL.so
ln -s mesa-libEGL.so.1.0.0 libEGL.so.1
ln -s mesa-libEGL.so.1.0.0 libEGL.so.1.0.0
ln -s mesa-libGLESv1_CM.so.1.1.0 libGLESv1_CM.so
ln -s mesa-libGLESv1_CM.so.1.1.0 libGLESv1_CM.so.1
ln -s mesa-libGLESv1_CM.so.1.1.0 libGLESv1_CM.so.1.1.0
ln -s mesa-libGLESv2.so.2.0.0 libGLESv2.so
ln -s mesa-libGLESv2.so.2.0.0 libGLESv2.so.2
ln -s mesa-libGLESv2.so.2.0.0 libGLESv2.so.2.0.0

later today I will update package(s)

kamelie1706 commented on 2014-03-09 10:36

Seems to be same than for nvidia and it looks like indeed we need to have symlink updated in catalyst package
https://bbs.archlinux.org/viewtopic.php?id=178189

@BoTuLox
Thx for the tip, not sure if this is the only symlink we need to fix

@noobfromby
Go back to mesa-10.0.3-1 lib32-mesa-10.0.3-1
**
# go to your pacman cache
cd /var/cache/pacman/pkg/
# search for mesa package
ls | grep mesa
# install the previous package found in my case
sudo pacman -U mesa-10.0.3-1-x86_64.pkg.tar.xz

Worked for me ...

kamelie1706 commented on 2014-03-09 10:34

Seems to be same than for nvidia and it looks like indeed we need to have simlink updated in catalyst package
https://bbs.archlinux.org/viewtopic.php?id=178189

@BoTuLox
Thx for the tip, not sure if this is the only symlink we need to fix

@noobfromby
Go back to mesa-10.0.3-1 lib32-mesa-10.0.3-1
**
# go to your pacman cache
cd /var/cache/pacman/pkg/
# search for mesa package
ls | grep mesa
# install the previous package found in my case
sudo pacman -U mesa-10.0.3-1-x86_64.pkg.tar.xz

Worked for me ...

noobfromby commented on 2014-03-09 07:14

Hello people, after upgrade mesa(10.1-2) and linux kernel(3.13.6) i have crashed system, after this upgrade i reboot system. Then, when my system was loading I saw a flickering screen, then I got into console. X-org not working, say me that he crashed. I try this method "sudo ln -s /lib/mesa-libEGL.so.1.0.0 /lib/libEGL.so.1" but it's not helped me. Pls who know how restore this bug write please. Sorry for my English.

BoTuLoX commented on 2014-03-08 23:13

I had to do a symlink today since my apps were complaining about not finding libEGL.so.1 after updating mesa.

Turns out latest mesa renames it to mesa-libEGL.so.1.0.0, so I just ran "sudo ln -s /lib/mesa-libEGL.so.1.0.0 /lib/libEGL.so.1" and problem solved.

If I run into any issues I'll update.

kamelie1706 commented on 2014-03-08 22:08

Problem not related to kernel but to mesa lib mesa-10.1.0-2-x86_64.

I went back to mesa-10.0.3-1-x86_64

kamelie1706 commented on 2014-03-08 21:45

Latest kernel make the system unstable.
I tried # catalyst_build_module all
Did not improve.

I can log but Cairo-Dock, chromium crash. I can launch dolphin, konsole but cannot move windows ( stuck on top left). Do not recommend to update to latest kernel.

kamelie1706 commented on 2014-03-08 21:44

Latest kernel make the system unstable.
I tried # catalyst_build_module all
Did not improve.

I can log but Cairo-dick, chromium crash. I can launch dolphin, konsole but cannot move windows ( stuck on top left). Do not recommend to update to latest kernel.

Delinquenz commented on 2014-03-04 23:57

I have the same issue like the others. VLC (only when a video is played) crashes my X, same with MPlayer. Totem works fine.
I am also not able to play a level in Portal 2. I don't know if this is related to my linux-ck kernel or to the AMD drivers.

kamelie1706 commented on 2014-03-02 15:47

Actually ... I can access the menu in vlc, that is jus a bug when vlc fullscreen.

kamelie1706 commented on 2014-03-02 15:43

It freezes also vlc interface ... cannot even access menu
[ 2904.888]
X.Org X Server 1.15.0
Release Date: 2013-12-27
[ 2904.888] X Protocol Version 11, Revision 0
[ 2904.888] Build Operating System: Linux 3.12.5-1-ARCH x86_64
[ 2904.888] Current Operating System: Linux BIGONE 3.13.5-1-ARCH #1 SMP PREEMPT Sun Feb 23 00:25:24 CET 2014 x86_64
[ 2904.888] Kernel command line: BOOT_IMAGE=/vmlinuz-linux root=UUID=a2077c26-9485-4361-b683-087602c30c8c rw quiet nomodeset radeon.dpm=0

And I confirm I get video/X crashes with any video tool
[ 5925.987] (EE) Backtrace:
[ 5925.988] (EE) 0: /usr/bin/X (xorg_backtrace+0x48) [0x584ae8]
[ 5925.988] (EE) 1: /usr/bin/X (0x400000+0x1887c9) [0x5887c9]
[ 5925.988] (EE) 2: /usr/lib/libpthread.so.0 (0x7fd4a993c000+0xf880) [0x7fd4a994b880]
[ 5925.988] (EE) 3: /usr/lib/xorg/modules/drivers/fglrx_drv.so (xdl_xs115_atiddxPixmapIsTypeOf+0x25) [0x7fd4a66c7af5]
[ 5925.989] (EE) 4: /usr/lib/xorg/modules/drivers/fglrx_drv.so (atiddxPixmapIsTypeOf+0xd) [0x7fd4a64e146d]
[ 5925.989] (EE) 5: /usr/lib/xorg/modules/amdxmm.so (X740XvPutImage+0x197) [0x7fd4a3663957]
[ 5925.989] (EE) 6: /usr/bin/X (0x400000+0x8b9c2) [0x48b9c2]
[ 5925.989] (EE) 7: /usr/bin/X (0x400000+0xd31c2) [0x4d31c2]
[ 5925.989] (EE) 8: /usr/bin/X (0x400000+0x35d1e) [0x435d1e]
[ 5925.989] (EE) 9: /usr/bin/X (0x400000+0x39b2a) [0x439b2a]
[ 5925.989] (EE) 10: /usr/lib/libc.so.6 (__libc_start_main+0xf5) [0x7fd4a85b3b05]
[ 5925.989] (EE) 11: /usr/bin/X (0x400000+0x250ce) [0x4250ce]
[ 5925.990] (EE)
[ 5925.990] (EE) Segmentation fault at address 0x47
[ 5925.990] (EE)
Fatal server error:
[ 5925.990] (EE) Caught signal 11 (Segmentation fault). Server aborting
[ 5925.990] (EE)
[ 5925.990] (EE)

clfarron4 commented on 2014-02-27 10:53

It seems to be working fine with 3.13.5-lqx without any patches.

Vi0L0 commented on 2014-02-26 21:07

14.2-1:
it's 14.1 with two fixes:
- Minecraft: missing textures have been resolved
- Resolves corruption issues seen in X-plane

It's _not_ solving problems with xv video output on xserver 1.15

I should be able to update [catalyst] repo to this version within 2 days

http://support.amd.com/en-us/kb-articles/Pages/Latest-LINUX-Beta-Driver.aspx

kamelie1706 commented on 2014-02-26 03:03

Beta 14.2 available
http://www2.ati.com/drivers/beta/amd-catalyst-14.2-betav1.3-linux-x86.x86_64.zip

kamelie1706 commented on 2014-02-26 02:54

Beta 14.2 available from AMD

Vi0L0 commented on 2014-02-23 19:11

14.1-2:
- added initial 3.14 kernel support
(turned out that Cold's patch for lqx kernel is working perfectly fine for 3.14 kernel, thanks Cold :))

Vamp898 commented on 2014-02-23 11:52

Using VLC with OpenGL Interface doesnt cause any crash...

theOptimizedCode commented on 2014-02-23 00:29

Same issue as others, running VLC/Mplayer crashes X

Vi0L0 commented on 2014-02-16 09:20

thanks for 'lib32-libxinerama' dep goes to sas

Vi0L0 commented on 2014-02-16 09:10

silent update - added 'lib32-libxinerama' to the deps

Vi0L0 commented on 2014-02-07 20:34

Got it, the are 2 ways to solve the problem:
1. use GL, OpenGL or vaapi video output - but not all software got this option, like skype
2. use xorg-server 1.14 package (ie from [xorg114] repo)- this one solved the problem completely here

Vi0L0 commented on 2014-02-04 23:04

Unfortunatelly since Monday Im AFK. Should get back on Saturday.
I was only testing video playback on mpalyer-vaapi and its vaapi video output, but it was working very well. I will look at the problem when I get back home, but I can't promise a thing.

tomasini commented on 2014-02-04 19:55

Hi Vi0L0,

thanks for the support for the catalyst-test!

I have the same experience with VLC crash and X. Temporary solution for the VLC player is to use driver X11 video output (XCB) instead of the default settings. The video quality is not the best, but it does not crash X. I believe that Vi0L0 find a solution soon :).

zoopp commented on 2014-02-04 10:29

@C84843: and there I though I was the only one. I managed to avoid making VLC crash X by setting video output to 'OpenGL GLX video output (XCB)'. This setting can be found in 'Tools > Preferences > Video'.

In case someone is interested in a X crash log file: http://codepad.org/rthlnJfF.

kalle commented on 2014-02-04 10:27

@C8483 Same here. Mplayer makes X crash and wine makes problems too.

idecasso commented on 2014-02-04 06:55

@cifarron4 I use Chrome with Catalyst 13.111 and I see the bug. In Catalyst 13.11 Beta and 13.!2, the bug not exist.

C8483 commented on 2014-02-04 01:12

Is it just me that VLC/Dragon Player/Skype Videocall makes X instantly crash with this version?

Perberos commented on 2014-02-03 13:21

Thankyou Vi0L0

mid-kid commented on 2014-02-03 11:51

Holyshit. Magic lamp animation in KDE works now without breaking.
/me is happy.

mid-kid commented on 2014-02-03 11:02

About time they supported xorg 15.
Actually, I'm pretty surprised they archived kernel 3.13 support this fast.

Vi0L0 commented on 2014-02-02 15:00

updated to 14.1 beta 1.3:
http://support.amd.com/en-us/kb-articles/Pages/latest-linux-beta-driver.aspx

Official features:
- Xserver 1.15 support
- Kernel 3.13 support

clfarron4 commented on 2014-01-29 19:29

@idecasso: I only have this problem with Chrome and Chromium.

idecasso commented on 2014-01-29 09:06

With 13,111 driver my computer has a bug. Slowly begins to consume RAM even if not running any programs, only OS and desktop. I tried it with Gnome and KDE. The same problem occurs in both. Having only opened Chrome, the system consumes 4.8GB after a few hours, and stays close consuming 3.7GB, I have 8GB of ram. This problem did not occur with the 13.11 Beta, I try 13.12 stable to confirm if there also have this problem.

AlD commented on 2014-01-23 21:37

I'm fine with whatever works. :-)

Vi0L0 commented on 2014-01-23 20:30

None of the AUR maintainers asked for it, ever, and I could do this for you, but... when looking at:
- official packages (like xf86-video-ati), I can see that they depend/conflict with both xorg-server and X-ABI-EXTENSION_VERSION - basing on this I should also use both;
- xorg-server* aur packages, I can see that they do provide both X-ABI-VIDEODRV_VERSION and xorg-server - once again, both.

Basing on this I can tell that your package also should provide them both.
As well as mine should depend on both.
But is it worth it? Does it make much sense?
If everybody would switch to the one *-ABI-EXTENSION_VERSION conception and reffer to it then it would make sense.
Now its only "triumph of the form over the content" or "multiplication of unneeded entities" (as I would say in polish).

But you almost convinced me, feel free to throw something into this topic.

As for the versioning:
1. AMD won't release 13.13 (there's no 13th month), I heard that they should release 14.1 in the end of january
2. there's no way of telling to pacman that 13.11 is higher version than 13.12 (at least I don't know about any). 13.111 is as good (or as bad) as any other number higher than 13.12.
3. If anyone care that much about versioning (as we can see AMD doesn't care ;P) then description of the package is always showing the right version.

AlD commented on 2014-01-23 20:00

I had put together an xorg114 PKGBUILD that wasn't split, so I didn't have an xorg-server package installed. Of course it was easily fixed by Provides: xorg-server=1.14.0, I just thought referring the ABI version was cleaner.

Versioning this with 13.111 is a bit messy, especially if AMD releases a 13.13.

Vi0L0 commented on 2014-01-23 18:42

@AID:
imho X-ABI-EXTENSION_VERSION variable makes any sense only in a theory, in practice every new release of xorg-server comes with a new ABI.
Why in your opinion should I use it? :)

@mid-kid:
it was not my idea to release 13.11 after 13.12, blame AMD :P.
As for me I added another "1" at the end to let your client know that this is a newer version. Pretty simple.

mid-kid commented on 2014-01-23 17:17

I don't get the versioning of this package. Somehow 13.11 is newer than 13.12, and for a new version: "Hey, let's slap another 1 at the end"

AlD commented on 2014-01-22 12:40

Instead of depending on certain versions of xorg-server it would probably make more sense to depend on the ABI version, i. e. X-ABI-EXTENSION_VERSION=7.0.

Vi0L0 commented on 2014-01-16 16:36

Either use [xorg114] which is compiled against libgcrypt 1.6 or recompile xorg-server 1.14 on your own. You can also try libgcrypt15 package

zoopp commented on 2014-01-16 10:27

Are there other people stuck with a broken X since the libgcrypt update?

X won't start due to missing 'libgcrypt.so.11'. I suppose if get rid of catalyst and update X the issue would be gone.

Vi0L0 commented on 2014-01-15 18:08

There's catalyst 13.30 rc3 released yesterday but I don't know what to do with it, I don't have a place for it...

It's not "stable", and for me it's definitely worse than 13.11 beta V9.95.
It's like 13.12 with "Supported Products:
AMD A10-7850K Desktop APU with AMD Radeon R7 Series graphics
AMD A10-7700K Desktop APU with AMD Radeon R5 Series graphics"

You may find source files here [1] and build it on your own

[1] http://catalyst.wirephire.com/tarball/13.30-rc3/

Vi0L0 commented on 2014-01-14 11:36

Yes, segfault here.

David-UPCT commented on 2014-01-14 01:10

xorg-server has been updated to 1.15 and catalyst is marked to be only compatible with xorg-server<1.15. Has someone tested catalyst with this new version? Thanks

mid-kid commented on 2014-01-11 12:27

These drivers work great on my Laptop. With the oss drivers I always got ~70C core temperature. With these, it stays around 60C.
I've got only one problem though: Where are the opengl headers located? (more specifically: GL/glu.h)
I'm using catalyst-hook from your repo.

clfarron4 commented on 2014-01-10 20:04

Lets try that comment again (bleedin' internet):

13.111-1 builds the LQX module for fglrx perfectly fine without the patch required for the last version(s) here.

Vi0L0 commented on 2014-01-10 19:17

They really do fixed some opengl bugs.
For example:
- gwenview on fullscreen in opengl mode works correctly
- there's no "blue screen" when watching html5 yt videos on fullscreen

clfarron4 commented on 2014-01-10 19:15

13.111-1 builds the module for fglrx perfectly fine without the patch required for the last version(s) here.

clfarron4 commented on 2014-01-10 19:15

13.111-1 builds the module for fglrx perfectly fine without the patch required for the last version(s) here.

clfarron4 commented on 2014-01-10 19:14

13.111-1 builds the LQX kernel fine this time.

Vi0L0 commented on 2014-01-10 18:38

Updated to 13.111-1, better known as 13.11 beta V9.95.
Forgive the versioning, it was the easiest way.

I had to remove /etc/ati/amdpcsdb to get kwins opengl back to work.

Probably tomorrow I will update [catalyst] repo with this.

C8483 commented on 2014-01-09 23:28

Thank you for the package. Installed on a fully updated arch machine and KDE with Compositing type OpenGl 3.1, Raster. No problems detected. Didn't test xrender (should I have this on?)

Vi0L0 commented on 2014-01-09 22:33

It seems so, not sure what does it changing though (http://support.amd.com/en-us/kb-articles/Pages/latest-linux-beta-driver.aspx).
Except for that I'm unable to run kdes kwin in opengl mode, xrender is working fine. gnome is also fine, as well as opengl games/apps.

This kwins opengl is blocking me from the update, i don't have time for further investigation now. Maybe tomorrow i will dig in it, if you will find some workaround please share.

For now you can use these sources to build and test 13.11 beta V9.95:
http://mirror.rts-informatique.fr/archlinux-catalyst/tarball/13.11-beta995/catalyst-test-13.111-1.src.tar.gz

C8483 commented on 2014-01-09 16:38

It looks like "13.11 Beta V9.95" is newer than 13.12-1

Vi0L0 commented on 2014-01-08 09:25

@bakgwailo:
+ I mean pacman.conf file

Vi0L0 commented on 2014-01-08 09:23

@bakgwailo:
you need to put [xorg114-testing] higher than arch's repos, so also above extra and testing. It's important

AlexTsi commented on 2014-01-07 22:43

Anyone else having issues with suspend in KDE with latest catalyst?
Specifically, when I resume after suspend screen freezes.

bakgwailo commented on 2014-01-07 17:52

So I searched a bit to figure out how to get around this:

resolving dependencies...
looking for inter-conflicts...
:: xorg-server and xf86-input-evdev are in conflict (X-ABI-XINPUT_VERSION). Remove xf86-input-evdev? [y/N] y
error: failed to prepare transaction (could not satisfy dependencies)
:: xorg-server: requires xf86-input-evdev
:: catalyst-test: requires xorg-server<1.15.0

I am using this package and:

[xorg114-testing]
Server = http://mirror.rts-informatique.fr/archlinux/repo/xorg114/$arch

Should I just hold this back, or is it an issue with the xorg114-testing repo?

lembang commented on 2014-01-03 20:41

@Cold

Thanks for the patch, yes it works very well
I'm happy with my liquorix kernel + the latest AMD driver.

AlD commented on 2014-01-03 17:40

If you're encountering the (Konsole) resize crash (http://ati.cchtml.com/show_bug.cgi?id=657), use the latest beta release for the time being: http://pkgbuild.com/git/aur-mirror.git/plain/catalyst-test/PKGBUILD?id=d4ec783

clfarron4 commented on 2013-12-31 12:08

@Cold

The patch seems to work pretty well for the LQX kernel for me.

Cold commented on 2013-12-30 03:53

@lembang

modified pkg for lqx kernel ->
https://mega.co.nz/#!pNp3xQRD!KSVZQTJ1aoZ9RmtU9PoZgElB2xhiukR7PoqnwIs57jU

patch :

--- ./common/lib/modules/fglrx/build_mod/firegl_public.c
+++ ./common/lib/modules/fglrx/build_mod/firegl_public.c
@@ -1755,7 +1755,7 @@
KCL_TYPE_Uid ATI_API_CALL KCL_GetEffectiveUid(void)
{
#ifdef current_euid
- return current_euid();
+ return __kuid_val(current_euid());
#else
return current->euid;
#endif

Vi0L0 commented on 2013-12-21 22:17

General note:
I'm AFK and will so be till next saturday.
In case of problems please write on bar and grill thread. If no one will help and there's some serious bug - please write the bug report on bugzilla.

Meanwhile you can still use archive repos like ones for 13.4 and 13.11-beta94:
[catalyst-stable-13.4] and [catalyst-13.11-beta94]
Server's URLs are like on the wiki for [catalyst].

Or build pkgs on your own basing on archive tarballs (use newest ver):
http://catalyst.wirephire.com/tarball/13.4/
http://catalyst.wirephire.com/tarball/13.11-beta94/

Thanks and merry christmass ;)

lembang commented on 2013-12-21 06:31

I tried using this package, I have no problem with compilation, but after I did "pacman -U catalyst-test"
i got
====
Building fglrx module for 3.12.5-1-lqx kernel ...
Failed!!! Check out log: /var/log/catalyst-install.log
====

and the content:
======
633 make: Entering directory '/usr/src/linux-3.12.5-1-lqx'
634 CC [M] /tmp/catalyst.gTQHG7/firegl_public.o
635 /tmp/catalyst.gTQHG7/firegl_public.c: In function 'KCL_GetEffectiveUid':
636 /tmp/catalyst.gTQHG7/firegl_public.c:1759:5: error: incompatible types when returning type 'kuid_t' but 'KCL_TYPE_Uid' was expected
637 return current_euid();
======

anyone have an idea what should i do to fix this?
Thank you.

Vi0L0 commented on 2013-12-19 20:41

no

zoopp commented on 2013-12-19 20:16

patching file etc/ati/authatieventsd.sh
patching file etc/ati/authatieventsd.sh
patch unexpectedly ends in middle of line

Should I be worried..?

Vi0L0 commented on 2013-12-19 19:53

13.12-1:
- AMD marked it as stable, so am I...
- http://support.amd.com/en-us/kb-articles/Pages/AMDCatalyst13-12LINReleaseNotes.aspx

Don't worry about kernels ;)

yousry commented on 2013-12-19 15:35

The release notes exclude the 3.12 Linux kernel:
System Requirements - Linux kernel 2.6 or above (up to 3.11)

cguenther commented on 2013-12-19 09:46

catalyst 13.12 is out now. See http://support.amd.com/en-us/kb-articles/Pages/AMDCatalyst13-12LINReleaseNotes.aspx

Toost_Inc commented on 2013-12-18 20:42

Is this just me or is the tarball 0 bytes?

Vi0L0 commented on 2013-12-17 20:43

13.11-11 now supports 3.13 kernel (tested on < rc5)

tangram commented on 2013-11-27 12:08

Hello Vi0L0
Two days ago I installed this driver, but I have issues to start x-server.
I posted detailed description here:
https://bbs.archlinux.org/viewtopic.php?id=173486
Could you please take a look there.
Thank you in advance

roukoswarf commented on 2013-11-24 20:54

Wow, years later they finally fixed the xrender transparency explosion (kde resize was the most common case of it hanging)

Thanks for the update Vi0l0.

Vi0L0 commented on 2013-11-23 16:21

updated to 13.11 beta 9.4 (9.4? why not 10? also why there's no consequence in file naming and archive naming schema?
It looks like someone is archiving and then someone else is publishing drivers, then when its time to publish the newer driver people are switched with completely different people who had no idea how their predecessors did this before. And it's month after month, release after release, nobody knows what's happening... c'mon AMD!)
anyway:
- phoronix said that: fixed are a crash when resizing KDE's Konsole, brightness adjustment issues, kernel module compile issues, avoids using a new GPL symbol, and resolving a system hang to resume from S4 with an OpenGL screen-saver running
- i've modified remove and remove_all (ra) in catalyst_build_module script, it should now remove modules better.

Vi0L0 commented on 2013-11-18 18:57

@zoopp: you were right! thanks! :)
decided to use curl (makepkg.conf's default) and 2 lines:

url_ref="http://support.amd.com/en-us/kb-articles/Pages/latest-linux-beta-driver.aspx"
DLAGENTS="http::/usr/bin/curl --referer ${url_ref} -o %o %u"

volle commented on 2013-11-18 09:11

@zoopp:

thanks, using makepkg worked (aura is a package manager like pacman but for AUR)

zoopp commented on 2013-11-17 22:17

@Vi0l0: you can still download from AMD if you set the referrer. For example:

curl --referer http://support.amd.com/en-us/kb-articles/Pages/latest-linux-beta-driver.aspx http://www2.ati.com/drivers/beta/amd-catalyst-13.11-beta6-linux-x86.x86_64.zip
wget --referer=http://support.amd.com/en-us/kb-articles/Pages/latest-linux-beta-driver.aspx http://www2.ati.com/drivers/beta/amd-catalyst-13.10-beta2-linux-x86.x86_64.zip
..downloads the archive without any issues.

@volle: I've never used 'aura' however for the time being you could just 'makepkg'.

zoopp commented on 2013-11-17 22:06

@Vi0l0: you can still download from AMD if you set the referrer. For example:

curl --referer http://support.amd.com/en-us/kb-articles/Pages/latest-linux-beta-driver.aspx http://www2.ati.com/drivers/beta/amd-catalyst-13.11-beta6-linux-x86.x86_64.zip

..downloads the archive without any issues. I really doubt this has any legal implications.

@volle: I've never used 'aura' however for the time being you could just 'makepkg'.

volle commented on 2013-11-17 21:08

sorry for this dumb question but i'm totally new to this.
where do i have to put the zip file if i'm using aura?

i tried to copy it to
/var/cache/aura/pkgbuilds
~
/var/chache/aura/src

and aura always tries to download the file (and doesn't succeed).

Vi0L0 commented on 2013-11-17 20:45

doesn't make sense at all, aspecially when it comes to archive files - we simply be not able to perform downgrade...

that's why i'm testing cloud disks, trying to find the best free place for archive amd installers, check out, and feel free to ask for other archive zips:

https://mega.co.nz/#F!J4E2zIRD!ODSaaCP1cP9FRp-RLMm_Ig

+ md5/sha sums: https://docs.google.com/document/d/1MdkB0lOa40UTCxNl70hpB22lvcQmKRdp59HNZbGEnow/edit?usp=sharing

bakgwailo commented on 2013-11-17 20:34

Ahhh, yeah, I kind of figured that. Sorry, should have read the top of the pkgbuild (kind of funny as I was just editing/trying different URLs in there, lol). Kind of sucks that AMD did that - doesn't really make sense to me.

Vi0L0 commented on 2013-11-17 19:53

it's because amd's site is not allowing direct download anymore...
read the top of pkgbuild for instructions.

i also added symlink as sugested by lano1106:
/usr/lib32/libXvBAW.so -> libXvBAW.so.1.0

bakgwailo commented on 2013-11-17 19:49

So, maybe I am missing something, but I no longer can download the file directly from the package build/yaourt. It basically downloads an HTML file instead which is along the lines of

"Download Not Complete
​Review the guide below for solutions to download your file


We’re sorry, but we were unable to complete your download. There could be several reasons for this:

You were linked to this download from a site outside of AMD.com
- See more at: file:///tmp/yaourt-tmp-joe/aur-catalyst-test/amd-catalyst-13.11-beta6-linux-x86.x86_64.html#sthash.bv8z2vbF.dpuf"

etc.

xenos commented on 2013-11-05 21:45

Thanks Violo for all the work you put into maintaining this.

Vi0L0 commented on 2013-10-28 18:27

13.11-8: added lib32-libcl to conflicts/provides

PelPix commented on 2013-10-28 15:22

Thanks so much for the fix. There's one more thing, Vi0L0. Sorry for the trouble:
Please add lib32-libcl to provides.

coco42 commented on 2013-10-28 03:30

You do not get enough of this. Thank you for your efferts.

Vi0L0 commented on 2013-10-27 09:06

fixed by adding staticlibs and libtool options to the package - i still don't get why they were removed...
Thanks Hetivn!
I also added !strip to the options, it's leaving symbols in the package so the package (and libs themselves) will be bigger, but it could be useful

Vi0L0 commented on 2013-10-27 09:04

fixed by adding options table to the package.
I also added !strip to the options, it's leaving symbols in the package so the package (and libs themselves) will be bigger, but it could be useful

rabinhood commented on 2013-10-27 07:54

Than you. It helped. I had to modify makekg.conf according to your instructions.

Hetivn commented on 2013-10-27 05:40

With regard to the following problem:

...
make[1]: *** No rule to make target '/tmp/catalyst.077UAR/libfglrx_ip.a', needed by '/tmp/catalyst.077UAR/fglrx.o'. Pare.
...

At least for me, this was caused by a recent change in /etc/makepkg.conf. There, the global package options were defined as follows:
OPTIONS=(strip docs !libtool !staticlibs emptydirs zipman purge !upx !debug)

However, for building this (all?) catalyst packages I needed:
OPTIONS=(strip docs libtool staticlibs emptydirs zipman purge !upx !debug)

PelPix commented on 2013-10-26 22:43

All your previous packages are retroactively broken. I expect there's been some sort of incompatible update to a build tool.

PelPix commented on 2013-10-26 22:35

Same as below.

olace commented on 2013-10-26 16:52

I miss something?

[2013-10-26 14:40] [PACMAN] Running '/usr/bin/pacman -U /tmp/catalyst-test/catalyst-test-13.11-6-x86_64.pkg.tar.xz'
[2013-10-26 14:40] [ALPM-SCRIPTLET] + removing fglrx module from /usr/lib/modules/3.11.6-1-ARCH
[2013-10-26 14:40] [ALPM-SCRIPTLET] Building fglrx module for 3.11.6-1-ARCH kernel ...
[2013-10-26 14:40] [ALPM-SCRIPTLET] Failed!!! Check out log: /var/log/catalyst-install.log


2013-10-26 14:50:56
Building fglrx module for 3.11.6-1-ARCH kernel ...
--------
OsVersion says: SMP=1
file /proc/kallsyms says: SMP=1
file /usr/lib/modules/3.11.6-1-ARCH/build/include/generated/autoconf.h says: SMP=1
file /usr/lib/modules/3.11.6-1-ARCH/build/include/generated/autoconf.h says: MODVERSIONS=1
file /usr/lib/modules/3.11.6-1-ARCH/build/arch/x86/include/asm/compat.h says: COMPAT_ALLOC_USER_SPACE=arch_compat_alloc_user_space
make: Entering directory '/usr/src/linux-3.11.6-1-ARCH'
CC [M] /tmp/catalyst.077UAR/firegl_public.o
CC [M] /tmp/catalyst.077UAR/kcl_acpi.o
CC [M] /tmp/catalyst.077UAR/kcl_agp.o
CC [M] /tmp/catalyst.077UAR/kcl_debug.o
CC [M] /tmp/catalyst.077UAR/kcl_ioctl.o
CC [M] /tmp/catalyst.077UAR/kcl_io.o
CC [M] /tmp/catalyst.077UAR/kcl_pci.o
CC [M] /tmp/catalyst.077UAR/kcl_str.o
CC [M] /tmp/catalyst.077UAR/kcl_iommu.o
CC [M] /tmp/catalyst.077UAR/kcl.o
CC [M] /tmp/catalyst.077UAR/kcl_wait.o
make[1]: *** No rule to make target '/tmp/catalyst.077UAR/libfglrx_ip.a', needed by '/tmp/catalyst.077UAR/fglrx.o'. Pare.
Makefile:1225: recipe for target '_module_/tmp/catalyst.077UAR' failed
make: *** [_module_/tmp/catalyst.077UAR] Error 2
make: Leaving directory '/usr/src/linux-3.11.6-1-ARCH'

Vi0L0 commented on 2013-10-26 12:48

Thanks David-UPCT.
Updated to 13.11-6 (beta 6):
- support for R9 290(X)
- fixed so called 30 FPS / Tear-Free / V-Sync bug
- a new huge (40MB) libamdocl library is inside /usr/lib/hsa
- amdcccle supports more languauges: cs, da, hu, ja, ko

David-UPCT commented on 2013-10-25 21:12

I've marked it out of date because there's a new 13.11 beta (beta 6)

r0ottm commented on 2013-10-25 15:37

I've been unable to build the catalyst module after updating my arch yesterday, I also tried installing an older kernel,nothing changed, I'm however able to install the *.zip from ati.com, though, with the latest kernel, the powerplay menu cannot be accessed from ccc. Anyone else experiencing this ?

Vi0L0 commented on 2013-10-18 20:41

You can simply change package version and md5sum or use archive build files located here http://catalyst.wirephire.com/tarball/catalyst-test/

PelPix commented on 2013-10-17 22:55

Or do I have to tell it not to apply certain patches?

PelPix commented on 2013-10-17 22:55

Any way we can downgrade? This broke Enlightenment for me. Do I just uncomment the appropriate parts of the PKGBUILD and change the package version?

Vi0L0 commented on 2013-10-12 16:26

updated to 13.11 beta 1 (13.11-1):
- allows new amd's gpus to use up to 6 monitors

deags commented on 2013-10-11 09:25

@ SysGhost:
How are you setting up multi mon? I've been using catalyst-test with a 7950 driving 3 monitors in latest xorg for several months without issue, using xrandr.

mizi commented on 2013-10-08 16:51

Today I installed the new version 13.11 beta
http://support.amd.com/us/kbarticles/Pages/AMDCatalyst13-11LINBetaV1Driver.aspx
but the problem remains

SysGhost commented on 2013-10-08 09:18

Any catalyst version beyond 13.4 won't work too well with multiple monitors.
It'll hang/freeze the system when additional monitors are connected.
It will also crash/freeze the system if more than one monitor is connected when Xorg loads/initializes.

Single monitor setups seems to be working however.

mizi commented on 2013-10-07 18:07

Same problem of @pinoteres
APU AMD A4-4355M
Trinity [Radeon HD 7400G]
on Samsung Notebook Series5 535U3C A1


dmesg | grep fglrx
[ 11.013098] fglrx: module license 'Proprietary. (C) 2002 - ATI Technologies, Starnberg, GERMANY' taints kernel.
[ 11.020328] fglrx: module verification failed: signature and/or required key missing - tainting kernel
[ 11.029304] <6>[fglrx] Maximum main memory to use for locked dma buffers: 3208 MBytes.
[ 11.029545] <6>[fglrx] vendor: 1002 device: 9994 count: 1
[ 11.030190] <6>[fglrx] ioport: bar 1, base 0xf000, size: 0x100
[ 11.030830] <6>[fglrx] Kernel PAT support is enabled
[ 11.030858] <6>[fglrx] module loaded - fglrx 13.20.5 [Sep 21 2013] with 1 minors
[ 25.779918] fglrx_pci 0000:00:01.0: irq 46 for MSI/MSI-X
[ 25.780964] <6>[fglrx] Firegl kernel thread PID: 1860
[ 25.781071] <6>[fglrx] Firegl kernel thread PID: 1861
[ 25.781171] <6>[fglrx] Firegl kernel thread PID: 1862
[ 25.781293] <6>[fglrx] IRQ 46 Enabled
[ 25.786714] <6>[fglrx] Reserved FB block: Shared offset:0, size:1000000
[ 25.786718] <6>[fglrx] Reserved FB block: Unshared offset:fc49000, size:3000
[ 25.786721] <6>[fglrx] Reserved FB block: Unshared offset:fc4c000, size:3b4000
[ 25.786723] <6>[fglrx] Reserved FB block: Unshared offset:1ffee000, size:12000
[ 3930.423611] <6>[fglrx] IRQ 46 Disabled
[ 3930.423709] <6>[fglrx] Preparing suspend fglrx in kernel.
[ 3930.698724] <6>[fglrx] Suspending fglrx in kernel completed.
[ 3930.698726] <6>[fglrx] Power down the ASIC .
[ 3936.488155] <6>[fglrx] Power up the ASIC
[ 3936.496333] <6>[fglrx] Preparing resume fglrx in kernel.
[ 3936.503989] <6>[fglrx] Resuming fglrx in kernel completed.
[ 3936.504057] <6>[fglrx] IRQ 46 Enabled

pinoteres commented on 2013-10-05 12:27

In my case
(APU A6-3410MX with Radeon HD 6520G, linux 3.11.2-1, acpi 1.6-1, acpid 2.0.20-1)
release 13.10 broke backlight.
I could not adjust brightness in any way.

# echo 1 > /sys/class/backlight/acpi_video0/brightness
# cat /sys/class/backlight/acpi_video0/max_brightness
20
# vbetool vbefp setbrightness 1
Real mode call failed
# dmesg | grep fglrx
[ 7.652215] <6>[fglrx] Maximum main memory to use for locked dma buffers: 15153 MBytes.
[ 7.652340] <6>[fglrx] vendor: 1002 device: 9647 count: 1
[ 7.652881] <6>[fglrx] ioport: bar 1, base 0x6000, size: 0x100
[ 7.653551] <6>[fglrx] Kernel PAT support is enabled
[ 7.653570] <6>[fglrx] module loaded - fglrx 13.20.5 [Sep 21 2013] with 1 minors
[ 22.395840] fglrx_pci 0000:00:01.0: irq 43 for MSI/MSI-X
[ 22.397086] <6>[fglrx] Firegl kernel thread PID: 1362
[ 22.397605] <6>[fglrx] Firegl kernel thread PID: 1363
[ 22.398053] <6>[fglrx] Firegl kernel thread PID: 1364
[ 22.398315] <6>[fglrx] IRQ 43 Enabled
[ 22.407985] <6>[fglrx] Reserved FB block: Shared offset:0, size:1000000
[ 22.407987] <6>[fglrx] Reserved FB block: Unshared offset:fb78000, size:4000
[ 22.407989] <6>[fglrx] Reserved FB block: Unshared offset:fb7c000, size:484000
[ 22.407991] <6>[fglrx] Reserved FB block: Unshared offset:1fff4000, size:c000

Downgrade to catalyst-test 13.8-7 resolved issue.

Squarex commented on 2013-10-04 11:58

Last update fixed some issues with KWin.

zoopp commented on 2013-10-01 18:50

This update fixes the backlight issues I had.

Vi0L0 commented on 2013-10-01 15:54

updated to 13.10-1:
- http://support.amd.com/us/kbarticles/Pages/AMDCatalyst13-10LINBetaDriver.aspx
- amd used jacob's patch, i added Olivier's patch also

quick testing didn't show any other than typical behavior

Vi0L0 commented on 2013-10-01 06:21

Thanks David,
I don't have time today, but tomorrow it should be updated.

David-UPCT commented on 2013-09-30 21:18

@Vi0l0: amd has uploaded a 13.10 beta

Vi0L0 commented on 2013-09-30 17:39

updated to 13.8-7:
- added Olivier's (lano1106) patch to improve jacob's patch
- added controversial GPL\0 workaround to bring 3.12 kernel support as suggested by kolasa

Vi0L0 commented on 2013-09-30 09:38

edit: by "remove catalyst package" i mean removing it from the repos only...

Vi0L0 commented on 2013-09-30 09:36

I was waiting for 3.12 rc3 to see what will happen with acpi's GPL check.
I'm almost sure that it stayed, but i'm unable to check it now, will do this later today.
I think that this new GPL check is pretty controversial and if it will stay I think that I won't provide fglrx modules built for 3.12 kernel. It means that I will remove catalyst package (only this one, hook or generator will stay) until it will be claryfied or fixed upstream.

Anyhow soon I will add Olivier's patch and 3.12 support to this package.

synergy commented on 2013-09-29 00:31

@Vi0L0,
Olivier's patch is a correct solution: http://ati.cchtml.com/show_bug.cgi?id=866#c19

I recommend pulling jacob's patch in favor of his.

Would there be an interest in adding an additional patch to allow LK 3.12 build completion and functionality?

Thank you.

synergy commented on 2013-09-29 00:29

@Vi0L0,
Olivier's patch is a correct solution: http://ati.cchtml.com/show_bug.cgi?id=866#c19

I recommend pulling jacob's patch in favor of his.

Thank you.

Vi0L0 commented on 2013-09-24 15:12

updated to 13.8-4:
- removed flim_fglrx_13.8_proc_perm.patch
- added jacob_access_right.patch

This should get a rid of memory leaks! Thanks goes to jacob from amd!
http://ati.cchtml.com/show_bug.cgi?id=866

Vi0L0 commented on 2013-09-21 08:35

thanks zoopp!
Downgraded to 13.8-3

zoopp commented on 2013-09-20 14:58

According to this: http://devgurus.amd.com/thread/167491, 13.9 is an older version of 13.8 that was posted by mistake (see post #3).

Koshon commented on 2013-09-19 13:44

ahh no, glx not work

Koshon commented on 2013-09-19 12:51

Vi0L0, hm.. yeah, with this folder all work fine. But why dkms package worked without this directory? But.. it does not matter. thx

zoopp commented on 2013-09-19 12:08

Backlight is still broken for me no matter the combination of acpi_osi/acpi_backlight. God damn you AMD!

zoopp commented on 2013-09-19 12:07

Backlight is still broken for me no matter the combination of apci_osi and apci_backlight..god damn you AMD.

Vi0L0 commented on 2013-09-18 21:58

updated to 13.9-1:
- looks like 13.6 but leaks for me...

Vi0L0 commented on 2013-09-17 20:24

is `uname -r` same as `cat /usr/lib/modules/extramodules-<your-zen-kernel-big-version-here>/version` ?

Koshon commented on 2013-09-17 14:44

install: missing file operand

Vi0L0 commented on 2013-09-16 16:01

what is log saying?

Koshon commented on 2013-09-16 11:57

not work with zen kernel. Error on a install, when dkms build. 3.10.10.
only catalyst-dkms-test work fine, but deprecated.

Koshon commented on 2013-09-16 11:55

not work with zen kernel. Error on a install, when dkms build.

hellnest commented on 2013-09-11 03:55

Failed to install due package conflict

Vi0L0 commented on 2013-09-08 17:45

@ttux:
to avoid all those problems in future please read more about catalyst-hook:
https://wiki.archlinux.org/index.php/AMD_Catalyst#Catalyst-hook
:)

Vi0L0 commented on 2013-09-08 17:44

@abhisek:
shouldn't you use -pxp package / or comment elsewhere?
If you are sure that you are writing in right place then which radeon do you have?

ttux commented on 2013-09-08 12:36

I re-installed the catalyst-test package and it works now. Sorry for disturbing with my comments.

ttux commented on 2013-09-08 12:11

I created the folder /usr/lib/modules/extramodules-3.11.0 and the file version in it containing 3.11.0 so now it worked and it put fglrx.ko.gz in the extramodules folder but I can't load the module. I get Module fglrx not found.

ttux commented on 2013-09-08 11:57

I get this error when I run catalyst_build_module:
install: missing destination file operand after 'fglrx.ko.gz'

Any idea?

Thanks

Anonymous comment on 2013-09-08 05:36

After updating to 13.8-beta-2 my system can not load kdm while switched to AMD GPU, while Intel one is working fine. Please fix this.

gran commented on 2013-09-07 18:02

remember to add

sudo systemctl enable catalyst-hook
sudo systemctl start catalyst-hook
sudo systemctl enable temp-links-catalyst
sudo systemctl start temp-links-catalyst

:) if not it keeps reconfiguring to default on boot
On ubuntu i didn't know about that and i was always keeping pc in "sleep" mode :D

zoopp commented on 2013-09-07 09:45

Backlight control breaks after I install this on my Dell Inspiron N7010 no matter in which way I try to change levels and no matter what I set acpi_osi and acpi_backlight to.

The only way to change backlight levels is during boot time.

Anyone else experiencing something similar?

MindLikeWater commented on 2013-09-06 13:41

Yesterday's kernel update left me with this in my /var/log/catalyst-install.log:
Kernel header files are absent: directory /usr/lib/modules/3.10.10-1-ARCH/build doesn't exist! Game over

:-(

LeetShiva commented on 2013-09-03 18:59

Kernel i updated on 3.11 so should be also pkgbuild :)
Anyway great work Vi0lo for maintaining this package

Anonymous comment on 2013-08-26 16:56

Hello Violo,

since i updated to this catalyst KDE crashes quite immediately. Btw the desktop effects did not work with 13.8-1.

greets

MindLikeWater commented on 2013-08-26 14:49

Fix for the current download issue:

Open PKGBUILD with an editor, search "source=(" and change the http to https. https://www2.ati.com/drivers/beta/amd-catalyst-13.8-beta2-linux-x86.x86_64.zip is the right URL to download the catalyst driver.

(Credits to Tonno87 for this solution)

tuxayo commented on 2013-08-25 12:08

13.8-2 is still getting me short freezes before a complete freeze on Team Fortress 2
@Vi0L0 Thanks for the 13.6 repo, 13.6 works well for me :-)

MindLikeWater commented on 2013-08-25 08:46

Same as with "catalyst-test-pxp": the file http://www2.ati.com/drivers/beta/amd-catalyst-13.8-beta2-linux-x86.x86_64.zip cannot be downloaded.

idecasso commented on 2013-08-25 06:23

I have found the following problem in 13.8.1 and 13.8.2 drivers. In Steam games belonging to the Source engine, the system never releases the memory allocated ram once the game ends. I tried the 13.6 beta driver and open source driver, do not exhibit this problem. I have a AMD HD6950 2GB, Phenom II x4 955, 8GB RAM, Gnome 3.10, and probe with stock kernels and ck.

idecasso commented on 2013-08-25 06:22

I have found the following problem in 13.8.1 and 13.8.2 drivers. In Steam games belonging to the Source engine, the system never releases the memory allocated ram once the game ends. I tried the 13.6 beta driver and open source driver, do not exhibit this problem. I have a 2GB HD6950 AMD Phenom II x4 955, 8GB RAM, Gnome 3.10, and probe with stock kernels and ck.

idecasso commented on 2013-08-25 04:27

I have found the following problem in 13.8.1 and 13.8.2 drivers. In Steam games belonging to the Source engine, the system never releases the memory allocated ram once the game ends. I tried the 13.6 beta driver free driver, do not exhibit this problem. I have a 2GB HD6950 AMD Phenom II x4 955, 8GB RAM, Gnome 3.10, and probe with stock kernels and ck.

Vi0L0 commented on 2013-08-24 18:43

updated to 13.8 beta 2:
personally I don't see any positive difference when comparing to beta 1, for me this update is useless, but maybe you will find some pros.
After updating to kde 4.11 and this beta 2 (I was updating from catalyst 13.1) I noticed that "Magic Lamp" compositing effect is very well turning my desktop into garbage, with only glitches on it - it was working well on catalyst 13.1

And1G commented on 2013-08-21 21:11

The Beta2 link (http://www2.ati.com/drivers/beta/amd-catalyst-13.8-beta2-linux-x86.x86_64.zip) just downloads a 0 byte file...

Vi0L0 commented on 2013-08-20 17:23

Thanks for the head up, but i'm away from home and my arch machine, i'm unable to update, i should be back on saturday

PelPix commented on 2013-08-20 16:01

@Vi0L0

http://www2.ati.com/drivers/beta/amd-catalyst-13.8-beta2-linux-x86.x86_64.zip

Vi0L0 commented on 2013-08-18 16:47

13.6
- tarballs: http://catalyst.wirephire.com/tarball/13.6-beta/
- repo:
[catalyst-13.6]
Server = http://catalyst.wirephire.com/repo/catalyst/$arch

Cold commented on 2013-08-18 14:42

Is someone still having the 13.6 ones ?

performance is horrible with 13.8...
With my HD6950, sometimes, I got less than 60fps in xonotic with medium setting @1280x1024.

Before that, I was running @120fps at least with normal graphics.

Cold commented on 2013-08-18 14:40

Is someone still having the 13.6 ones ?

performance is horrible with 13.8...
With my HD6950, sometimes, I got less than 60fps in xonotic with medium setting @1280x1024

It runs better with a GT220... great...

roukoswarf commented on 2013-08-08 19:13

Just reporting back that 13.8-1 still has full gpu lockups.

more details:
3 monitor kde compositioned.

Edit: removed amdpcsdb and andpcsdb.current, seems stable now, will report back when fully configured again. might be a settings conflict that was causing the crashes.

roukoswarf commented on 2013-08-08 19:05

Just reporting back that 13.8-1 still has full gpu lockups.

more details:
3 monitor kde compositioned.

Vi0L0 commented on 2013-08-04 19:50

im not using yaourt, though iirc it's building packages somewhere near /var or something, maybe check there, or simply use makepkg ;)

cguenther commented on 2013-08-04 13:01

I am running linux-ck-bobcat and want to compile this package by running yaourt. But i get the error "Not enough space left in . (0KB) to decompress ./amd-catalyst-13.8-beta1-linux-x86.x86_64.run (76 KB)"

full commandline log is shown here:
http://pastebin.com/3JYvqnQy

I have 3.2 GB free RAM and more than 50GB free space on root partition. So is there an issue with the rights for the decompression or something like that?

kind regards,
Christian Günther

cguenther commented on 2013-08-04 12:59

I am running linux-ck-bobcat and want to compile this package by running yaourt. But i get the error "Not enough space lieft in . (0KB) to decompress ./amd-catalyst-13.8-beta1-linux-x86.x86_64.run (76 KB)"

The full commandline log is shown in:
<script src="http://pastebin.com/embed_js.php?i=3JYvqnQy"></script>

or:
http://pastebin.com/3JYvqnQy

I have 3.2 GB free RAM and more than 50GB free space on root partition. So is there an issue with the rights for the decompression or something like that?

kind regards,
Christian Günther

Anonymous comment on 2013-08-04 09:15

to the previous comment deactivating one screen does the trick. still that seems a bit of an annoying solution.

Anonymous comment on 2013-08-04 08:59

Hello Violo,

hmm could it be a problem with dual Screen setup? my specs are

ARCH linux newest stuff catalyst 13.8
HD 6950
i7 3770K

What setup do you use. i really ont know what to do because it seems all source engine games i play suffer from the same movement seem laggy although i have constant fps rates. Maybe i' try deactivate one screen ...

Vi0L0 commented on 2013-08-03 15:00

there's a freshly baked catalyst-test-pxp package on the AUR:
https://aur.archlinux.org/packages/catalyst-test-pxp/

phindela: I can not share your experience, for me perf is pretty good.

Anonymous comment on 2013-08-03 08:25

Hello,

today i tried the new catalyst especcially for the point source engine game hang on high performance asic. its still horrible totally laggy gameplay. Dont know wether its amd or valves fault but it feels much worse than with catalyst 13.6

Cold commented on 2013-08-03 03:05

correct me if I'm wrong...

pacman can't grab package from the AUR, use yaourt instead.
"yaourt -Sy catalyst-test" will rebuild and install the new package ( so it will be up to date )

otherwise you can do (I think) a "yaourt -Syua" to update packages which are from the AUR.

Thanks Vi0L0

dlsolo commented on 2013-08-02 23:09

Will a pacman -Syu update my catalyst drivers or do I need to uninstall, revert back to xf86-video-ati, and then update?

Vi0L0 commented on 2013-08-02 18:44

updated to 13.8-1:
- for me it's working just like our last "13.7-1", though it has amd's proper 3.10 kernel support
- added flim's fix for proper /proc/ati handling

bakgwailo commented on 2013-07-31 21:40

OK, I dowgraded back to stable and the KWin issues are still happening (although not as badly/often). Most likely a 4.11 Kwin issue.

bakgwailo commented on 2013-07-29 20:36

13.7-1 seems to have fixed my bug/issue of broken hibernate/sleep in the previous beta. However, on KDE 4.11rc2, I am getting pretty funky graphical glitching. On startup of the desktop, I get a blank screen until everything loads, and then if I maximize/restore windows I end up losing the top two corners where the buttons are (they just become transparent). If I cause a repaint/refresh of the app they come back correctly, and if I mouse over where the buttons were supposed to be they do show up again (although the border around them is still transparent).

Anonymous comment on 2013-07-27 04:15

I am running KDE as well. I found that 13.7-1 worked fine with tear-free off. So that seems to be the problem.

roukoswarf commented on 2013-07-27 03:32

Currently running on KDE, but the problem occurs any time there is GPU acceleration in play.

I've never had lag on tear-free ect, on any version, just as a note.

roukoswarf commented on 2013-07-27 03:31

Currently running on KDE, but the problem occurs any time there is GPU acceleration in play.

Vi0L0 commented on 2013-07-26 21:46

I was wondering on which DE you are working?
13.7-1 seems to work pretty good on my 7850, except of course lagging tear-free

Anonymous comment on 2013-07-26 20:42

I upgraded to 13.7-1 today. I confirm roukoswarf's observation. fgl_glxgears hangs the computer. I can only hard reset. Radeon HD7950.

roukoswarf commented on 2013-07-26 06:18

@Vi0L0: Getting full GPU hangs after seeing some graphics corruption, only solution is a full system reset, X cannot be killed successfully. Only happens on 13.7-1, confirmed working as usual on 13.6-3. Radeon 7850.

Vi0L0 commented on 2013-07-24 19:39

@roukoswarf: removing ~/.Xauthority didn't help here :/ strange case, I don't have much time to look at this much closer, though

@AlexTsi: don't bother yourself with it, I'm pretty sure that removing /etc/ati/amdpcsdb will fix the numbers but it will remove your config if you were changing anything via aticonfig or amdcccle

Vi0L0 commented on 2013-07-24 19:33

updated to 13.7 beta with opengl 4.3 support, here's my experience on hd 7850 (in comparison to 13.6 beta):
- still same laggy DE behaviour when compositing and Tear-Free Desktop are enabled: http://ati.cchtml.com/show_bug.cgi?id=838
- games performance is better, serious sam 3 just a tiny bit, but left 4 dead 2 shines like on windows
- unigine haeven and valley works a bit better
- opencl performance is pretty same (tested on oclHashcat), amdapp-sdk samples are working fine except for GL ones, which aren't rendering correctly
- i added zoqaeski's patch for authatieventsd' lightdm

AlexTsi commented on 2013-07-24 13:22

Hi, I compined this package with catalyst-total-pxp to install 13.6 beta on my hybrid A+I laptop. Everything is fine.. However I noticed 2 things that I do not know if are normal. Perhaps you could help me?

1: fglrx info gives :
OpenGL version string: 4.2.12337 Compatibility Profile Context 12.104
"4.2.12337" is the right version I think, but why do I have "Compatibility Profile Context 12.104"? Shouldn't that be 13.101?

2: In amdcccle, under "information", under "software" I have:
Driver Packaging Version: 12.104-130328a-155999C-ATI
2D driver version : 13.10.10

2D driver version seems fine, but should packaging version be something like "13.101-......"??

BTW THANKS A LOT VI0L0! This is the only package that I manage to make it work on my pc! EXCELLENT WORK! KEEP UP!

AlexTsi commented on 2013-07-24 13:15

Hi, I compined this package with catalyst-total-pxp to install 13.6 beta on my hybrid A+I laptop. Everything is fine.. However I noticed 2 things that I do not know if are normal. Perhaps you could help me?

1: fglrx info gives :
OpenGL version string: 4.2.12337 Compatibility Profile Context 12.104
"4.2.12337" is the right version I think, but why do I have "Compatibility Profile Context 12.104"? Shouldn't that be 13.101?

2: In amdcccle, under "information", under "software" I have:
Driver Packaging Version: 12.104-130328a-155999C-ATI
2D driver version : 13.10.10

2D driver version seems fine, but should packaging version be something like "13.101-......"??

BTW THANKS A LOT VI0L0! This is the only package that I manage to make it work on my pc! EXCELLENT WORK! KEEP UP!

Vi0L0 commented on 2013-07-24 06:44

Sorry, I was afk, I will check out opengl 4.3 driver today after getting back from work

roukoswarf commented on 2013-07-10 18:33

@ViOLO It appears to come back every few updates, and appears to be solved by removing ~/.Xauthority if that gives you any leads.

Vi0L0 commented on 2013-07-07 19:51

updated to 13.6-4:
- uncommented patching, changed deps

@roukoswarf:
thanks for the head up! No, it wasn't known issue, I looked ad it shortly - the problem with xauth is strange because manually I'm able to do what atieventsd.sh has... maybe some problem with amd's bin? dunno for now, will look at it, and maybe make bug report, will let you know anyway

roukoswarf commented on 2013-06-30 04:58

atieventsd[374]: xauth: file /tmp/atievntX.vsNG8N does not exist
atieventsd[374]: xauth: (argv):1: bad "add" command line

These messages appear to pop up every 1 second in journal, has been happening for a long while with this catalyst package, the -total, and the catalyst repo, is this a known issue?

Anonymous comment on 2013-06-26 20:48

I have only 30 fps if Tear Free Desktop feature is enabled. The same situation with any version above 13.1.
Catalyst 13.1 worked normally in that case. Did anyone face the same problem?

bparker commented on 2013-06-24 15:29

Needs qt4 dependency added as I could not launch amdcccle without it:

amdcccle: error while loading shared libraries: libQtGui.so.4: cannot open shared object file: No such file or directory

bvad commented on 2013-06-15 14:00

I could only get lib32-catalyst-utils to work by installing it from [MultiLib]. I got direct draw errors without it.

Vi0L0 commented on 2013-06-04 11:27

module should land in /usr/lib/modules/extramodules-3.9(something) - i don't remember the whole path

you can also run:
systemctl enable catalyst-hook
systemctl start catalyst-hook

to enable automatic re-compilation of fglrx modules for all installed kernels whenever kernel gets an update

ttux commented on 2013-06-04 07:31

Yes, I was building from inside extracted catalyst files. That's usually how I do it. I tried the catalyst_build_module script now and it says:
Building fglrx module for 3.9.4 kernel ...
Ok.

But it doesn't seem it has built a fglrx.ko

Vi0L0 commented on 2013-06-04 06:51

@ttux:
how are you trying to build it? Is it make inside extracted catalyst files? Because it does look like so.

Have you tried using catalyst_build_module script ?

ttux commented on 2013-06-04 06:34

I used the amd installer directly and it worked, which I think is a first for me.

ttux commented on 2013-06-04 06:15

I can't build the driver for some reason. I'm getting the following:
...
CC [M] .../catalyst/src/archive_files/common/lib/modules/fglrx/build_mod/2.6.x/kcl_wait.o
<command-line>:0:0: warning: "COMPAT_ALLOC_USER_SPACE" redefined [enabled by default]
<command-line>:0:0: note: this is the location of the previous definition
make[2]: *** No rule to make target `.../catalyst/src/archive_files/common/lib/modules/fglrx/build_mod/2.6.x/libfglrx_ip.a', needed by `.../catalyst/src/archive_files/common/lib/modules/fglrx/build_mod/2.6.x/fglrx.o'. Stop.
make[1]: *** [_module_.../catalyst/src/archive_files/common/lib/modules/fglrx/build_mod/2.6.x] Error 2

I have a custom kernel 3.9.4

scjet commented on 2013-06-03 06:11

@Vi0L0
Hey, Thankyou very much, good work. :)

Anonymous comment on 2013-06-02 16:38

i love you Vi0L0.

Vi0L0 commented on 2013-06-01 12:14

updated to 13.6-2:
- all bins have been moved to /usr/bin
- added workaround for i/o bug

@bakgwailo:
#2: you should fill the bug report @ http://ati.cchtml.com (firstly check that someone already did it)

bakgwailo commented on 2013-06-01 03:25

Two issues:

1) Looks like it conflicts now due to the move out from /bin and /sbin (preventing upgrades)

2) The latest beta totally broke sleep for me on my laptop (KDE) - resume just goes to a blank back lit screen

#1 you probably already know about (heh), #2, not sure if I should file a bug report anywhere.... either way, thanks for the great work and package!

Vi0L0 commented on 2013-05-29 19:36

updated to 13.6:
- xserver 1.14 support,
- many fixes: http://support.amd.com/us/kbarticles/Pages/AMDCatalyst13-6LINBetaDriver.aspx

For now I am disabling all patches, but if You want You can enable them by uncommenting patching lines under catalyst-hook section of PKGBUILD
Reason: https://bbs.archlinux.org/viewtopic.php?pid=1279795#p1279795

Anonymous comment on 2013-05-29 08:58

Looking forward to see gnome-shell behave with GLX_EXT_buffer_age

conan513 commented on 2013-05-28 21:55

AMD Catalyst 13.6 Linux Graphics Driver Beta Released
http://www2.ati.com/drivers/beta/amd-driver-installer-catalyst-13-6-beta-x86.x86_64.zip

Highlights of the AMD Catalyst™ 13.6 Beta Driver for Linux:
This release of AMD Catalyst™ Linux introduces support for the following new features:

Ubuntu 13.04 Production Support
Xserver 1.14 Support
GLX_EXT_buffer_age extension support
Resolved Issues:

[371937] Fix black screen while entering Team Fortress 2 under cinnamon desktop environment
[366812] Fix tearing for some OpenGL applications with VSYNC on or off
[373575] Fix taskbar missing after UVD playback on Ubuntu 12.04
[370955] Fix corruption in Team Fortress 2 when running Dustbowl map
[377199] Fix the error for APL initialization failure
[378086] Fix screen corruption when kill X on Ubuntu 13.04
[374657] Fix some WebGL conformance test failures
[377379] Fix the regression of performance drop with XBMC for UVD playback

derJupp commented on 2013-05-12 15:47

My fault, did not read the wiki https://wiki.archlinux.org/index.php/Catalyst :(
Now everything works ... Great!

derJupp commented on 2013-05-09 17:24

Installation fails with the following message :(

Fehler: Ziel nicht gefunden: xorg-server<1.14.0
==> Restart building catalyst-test ? [y/N]

Anonymous comment on 2013-04-29 11:43

Hey, thanks for keeping this up to date. I installed this driver and everything is working great except of the backlight control of my notebook... I think it is ATIs fault, just wanted to mention... Greetz

Vi0L0 commented on 2013-04-27 21:20

updated to 13.4-1:
http://support.amd.com/us/kbarticles/Pages/AMDCatalyst13-4LINReleaseNotes.aspx

Vi0L0 commented on 2013-04-21 17:36

little update:
removed Before from Service section of catalyst-hook.service

roukoswarf commented on 2013-04-17 01:41

13.3 beta 3 definitely fixed the long standing kwin xrender issues they claimed to have fixed.
Ever since 12.10 or so my xorg would crash on resizing a Xrender window (particularly konsole) and required me to use the segfault heaven opengl backend on the window.

roukoswarf commented on 2013-04-17 01:40

13.3 definitely fixed the long standing kwin xrender issues.
Ever since 12.10 or so my xorg would crash on resizing a Xrender window (particularly konsole) and required me to use the segfault heaven opengl backend on the window.

Vi0L0 commented on 2013-04-11 17:09

Forget my last sentence, phoronix is wrong :P

Vi0L0 commented on 2013-04-11 15:45

updated to 13.3-7:
fixing little "bug":
https://bbs.archlinux.org/viewtopic.php?pid=1257358#p1257358

Although if I were you I would not update this today - could be that I will update to the possibly newer 13.3 beta 3 (gotta check it out)

Vi0L0 commented on 2013-04-07 17:09

updated to 13.3-5:
SyncFirst was removed from the pacman 4.1 and so we can no longer force linux-headers to be installed in first place...

Instead of using old catalyst-hook method which was not working well with pacman 4.1 + linux kernel, I have added a new method - automatic re-compilation while system shutdown/reboot

To enable 'automatic re-compilation while system shutdown/reboot' testing feature run these commands as root:
systemctl enable catalyst-hook
systemctl start catalyst-hook

More info here:
https://bbs.archlinux.org/viewtopic.php?pid=1255575#p1255575

Vi0L0 commented on 2013-03-28 09:03

i wouldn't be that much optimistic about the time of releasing the driver with xserver 1.14 support
Ubuntu 13.04 will run on xserver 1.13, so for AMD there's no need to be hurry.
They will release it after deep, internal testing.

waaaaargh commented on 2013-03-28 00:05

still not compatible with xorg 1.14.0?

Vi0L0 commented on 2013-03-27 19:37

Thanks lkoos!

AMD has updated it with minor changes in amdpcsdb, control, profiles, etc

lkoos commented on 2013-03-27 01:12

I think the driver package (amd-driver-installer-catalyst-13.3-beta3-linux-x86.x86_64.zip) has been silently updated. The file I'm downloading has md5sum a8e2dd9384cfc3371c502eaebd542176.

I hope their repository hasn't been compromised.

Vi0L0 commented on 2013-03-23 11:43

+ Rel notes:
http://support.amd.com/us/kbarticles/Pages/AMDCatalyst13-3LINBetaDriver.aspx

Vi0L0 commented on 2013-03-23 11:37

updated to 13.3 beta 3:
- there are some little changes in amdpcsdb and profiles;
- latelly added catzilla profile and ARB_debug_group's methods was removed;
- opengl version is the same;
- it's not compatible with xorg-server 1.14

conan513 commented on 2013-03-23 10:51

and beta 2 no longer available

conan513 commented on 2013-03-23 10:44

and beta 2 no longer available

Anonymous comment on 2013-03-23 07:30

Beta 3 released.

Xorg 1.14 compatible?

Anonymous comment on 2013-03-20 11:34

Thank you so much Vi0L0.

Vi0L0 commented on 2013-03-19 18:04

@koex:
here's our bug report with workaround:
http://ati.cchtml.com/show_bug.cgi?id=718

@mityukov:
i believe that dependency tells it all: xorg-server<1.14.0.
currently there's no public amd driver with xserver 1.14 support

mityukov commented on 2013-03-19 16:20

This one doesn't support xorg-server 1.14 either (just like catalyst-dkms), does it?

Anonymous comment on 2013-03-19 14:38

VT switch (Alt+F1...F7) works for you? All i get is a black screen. VT is functional but cannot see what i'm typing.

Disable KMS and 'aticonfig --acpi-services=off' makes no difference.

Any workaround?

Anonymous comment on 2013-03-18 12:05

Awesome as always Vi0L0, thanks for the hard work :)

Vi0L0 commented on 2013-03-17 18:58

updated to 13.3 beta2, i can see that:
- amd has used my patches to bring 3.7 & 3.8 kernel support and so I don't have to use them on my own ;P
- a few profiles were modified, there's also new profile for catzilla :P
- there are some new ARB_debug_group's methods
- new variables in amdpcsdb

Vi0L0 commented on 2013-03-16 17:05

silent update, no pkgrel bump:
- i've cleaned up PKGBUILD and catalyst_build_module, from now on it supports only kernels build with proper extramodules support, tested on -ARCH, -ck, -lts and -mainline

Vi0L0 commented on 2013-03-09 17:48

AFAIK there are problems with brightness on 13.2 betas.
I've heard that 13.2 beta 7 fixed some of them, so i guess it's a known problem.
If you want/can you can always report a bug on ati.cchtml.com ;)

Vi0L0 commented on 2013-03-09 17:48

AFAIK there are problems with brightness on 13.2 betas.
I've heard that 13.2 beta 7 fixed some of them, so i guess it's a known problem.
If you want/can you can always report a bug on ati.cchtml.com ;)

Anonymous comment on 2013-03-07 20:02

Hi guys, i have a problem with this driver! It works better than catalyst 13.1 because i can use my 6620G without freeze at boot, but, i can't change display brightness. Is there a way to solve or have i just to wait?

Thanks for the work Vi0l0

eduardoeae commented on 2013-03-04 20:26

Does it compile for anyone with kernel 3.8?

conan513 commented on 2013-03-02 17:34

beta 7 is the newest
I don't know why flaged out of date

Vi0L0 commented on 2013-03-01 16:38

Why flagging out of date? Is there something newer than this?

Vi0L0 commented on 2013-02-28 16:48

updated to 13.2 beta 7

conan513 commented on 2013-02-23 23:37

Huge parformance improve in TF2 with 5670
now its running like on Windows with 4xAA

Vi0L0 commented on 2013-02-19 18:40

updated to 13.2 beta 6:
- afaik there's no changelog atm and it's hard to see the difference except that steam's big picture has been fixed
- as for the steam i can only test serious sam 3 and there's no difference on my hd 7850 in a matter of performance

Vi0L0 commented on 2013-02-19 17:51

guys, we have already reported this bug.
If you want feel free to add comment:
http://ati.cchtml.com/show_bug.cgi?id=718

shingo commented on 2013-02-18 18:35

I have the same problem! I installed catalyst-test and reboot, black screen. To return to the command line, you must add the parameter "3" before starting with syslinux. I reinstalled catalyst-total, still black screen or the screen sometimes funny color.

I installed the ATI driver free and it works. I try to find a solution. thank you

weirddan455 commented on 2013-02-13 02:57

OK I fixed the Serious Sam 3 issue by deleting the ShaderCache from inside the game directory. I think I've found a real bug with this driver though. Once I've started X, I can't switch back to the console by doing Ctrl+Alt+F2-F6 or by logging out of XFCE. I just get a black screen. And I don't have this issue with 13.1 from [community]. Hopefully someone finds a workaround on this thread:

https://bbs.archlinux.org/viewtopic.php?id=158001

weirddan455 commented on 2013-02-01 23:35

I can't get Serious Sam 3 to work with this package. Team Fortress 2 and Open Arena work so I know I have it installed right. Serious Sam 3 works with the 13.1 Catalyst from [community] and it even works with the open source driver (though it only gets about 3 FPS with that). When I try it with this package though, I click play, the screen flickers a little, then nothing. I tried reinstalling and then went back to 13.1 where it works.

Vi0L0 commented on 2013-01-30 18:52

updated to 13.2-1, 13.2 beta 3:
- Resolves a Team Fortress 2 performance issue with AMD Radeon™ HD 7000, 6000 Series GPU. Up to 300% performance improvement
- Resolves a Serious Sam 3 launching and corruption issue with AMD Radeon™ HD 7000, 6000 Series GPU

Vi0L0 commented on 2013-01-16 20:42

i could but i don't see sense of doing this.
12.12 is aimed to embedded devices, it's contol file is very small so it's not supporting many of the GPUs

m1ck3y commented on 2013-01-16 18:30

ViOLO, can you update catalyst-test package to 12.12?

Vi0L0 commented on 2012-12-15 13:35

updated to 12.9-5:
- added patch that brings 3.7 kernel support

jarda-wien commented on 2012-10-29 13:55

Ok ViOLO thanks. That's what I thought, just needed a confirmation.

Vi0L0 commented on 2012-10-26 17:51

This package REALLY is catalyst-hook + catalyst-utils + lib32-catalyst-utils(optional). Nothing more, nothing less. No special paths, tricks, hacks.
Looks like 12.9 ubu is a driver for you :)

jarda-wien commented on 2012-10-26 15:23

I was using this package during the last few weeks because I had problems with fullscreen flash video. I tried several methods like editing /etc/adobe/mms.conf and reverting to flash 10 or downgrading Xorg. Nothing really helped and I didn't find out which update caused this.
Anyway, trying this driver fixed it for me! There is a new catalyst version now (12.11) and the the problems are back. Fullscreen video is very slow and the X server doesn't even react to mouse movement if I go HD on Youtube. Windowed mode runs 1080p smoothly and fullscreen has problems even with 480p. I suspected the system downclocking the card for some reason, but a power meter tells a different story and tweaking amdoverdrivectrl so that all profiles are using the same fast clocks didn't help either...

I am using a HD 5850 and Core i3 540 with 8 GB of RAM so hardware wise, I am ok.

The question is: is this driver packages using a special patch or something? What exactly is different to catalyst-hook and catalyst-utils that could break fullscreen video?

Thanks in advance for trying to find a logical explanation.

Vi0L0 commented on 2012-10-10 18:16

update to 12.9-4:
- added atieventsd.service

Vi0L0 commented on 2012-10-08 19:44

... i mean for me, on my hd 7850 ofc

Vi0L0 commented on 2012-10-08 19:44

after jumping to [testing] it turned out it's working on xserv 1.13 :)

Vi0L0 commented on 2012-10-06 16:24

updated to 12.9-3:
this is ubuntu 12.10 quantal's beta driver - amd's "special" driver with xorg-server 1.13 support.

Doesn't work on my hd7850 but should work on other gpus.


If you got problems with it - here you can find sources to build 12.9-2 (amd's "normal" beta, without xserv 1.3 support) : http://catalyst.apocalypsus.net/tarball/12.9-test/catalyst-test-12.9-2.src.tar.gz

esclapion commented on 2012-09-27 11:37

Thanks a lot. It works fine now on my ATI 6870 :-)

Vi0L0 commented on 2012-09-26 18:08

updated to 12.9-2, AMD's catalyst 12.9 beta:
the only different between this new beta (version 9.00) and already released 12.9 for emb devs (also version 9.00, our catalyst-test) is in those three files located in /etc/ati:
- amdpcsdb.default
- control
- signature

Beta supports more gpus.

catalyst-test 12.9-1 users can download those files from:
http://catalyst.apocalypsus.net/files/control-12.9-beta/
and paste it into /etc/ati, perms 644, and in this way avoid update to -2 version

esclapion commented on 2012-09-24 21:53

Sorry, I missed the message Vi0L0 bottom of this page.
With control of 12.8, it works perfectly, thank you to him.
I now have two solutions :-)

Anonymous comment on 2012-09-24 17:01

Guys, be more specific, have you tried Vi0L0's solution?

"if it's not working with your hardware try to use /etc/ati/control file from catalyst 12.8
, you can download it from http://catalyst.apocalypsus.net/files/control-12.8/"

esclapion commented on 2012-09-20 21:14

Solved: The driver works well, but the aticonfig does not recognize my hardware.
Must manually create the .conf file in /etc/X11/xorg.conf.d

esclapion commented on 2012-09-20 12:33

Does not work with a HD 6870 : aticonfig: No supported adapters detected

ttux commented on 2012-09-18 17:43

I just intalled it, I have a HD6950 and I'm getting this watermark, is it normal?

tobias_de commented on 2012-09-17 18:09

Fix for the watermark: http://phoronix.com/forums/showthread.php?19875-Unsupported-Hardware-watermark

tobias_de commented on 2012-09-17 18:03

Works with HD 5830, but I get a stupid "AMD Unsupported hardware" logo at the bottom right.

antiuser commented on 2012-09-17 03:05

work fine with hd6310

antiuser commented on 2012-09-17 02:50

work fine with hd6310

Vi0L0 commented on 2012-09-16 17:58

pacman -Qi xorg-server

also: soon i will prepare [xorg112] repository, watch for info on ATI Catalyst @ Arch wiki

thirtythreeforty commented on 2012-09-16 17:37

Hmmm. How do I determine my version of X? If I'm fully up to date is it 1.13? If so, is there a way to downgrade?

Vi0L0 commented on 2012-09-16 17:05

updated to 12.9:
targeted for embedded devices, but seems to work pretty fine on my hd 7850.

- linux 3.5 support (works also on 3.6rc4, didn't test on 3.6rc5 yet)
- not working with xserver 1.13
- more info on phoronix: http://www.phoronix.com/scan.php?page=news_item&px=MTE4NDk

if it's not working with your hardware try to use /etc/ati/control file from catalyst 12.8
, you can download it from http://catalyst.apocalypsus.net/files/control-12.8/

Vi0L0 commented on 2012-07-15 20:36

12.6-7:
should work fine with glibc 2.16.0-2 and kernel 3.5

int0x80 commented on 2012-07-07 09:19

Downloading amd-driver-installer-8.98-x86.x86_64.zip...
% Total % Received % Xferd Average Speed Time Time Time Current
Dload Upload Total Spent Left Speed
100 178 100 178 0 0 8 0 0:00:22 0:00:20 0:00:02 1977
curl: (7) couldn't connect to host
==> ERROR: Failure while downloading amd-driver-installer-8.98-x86.x86_64.zip
Aborting...
==> ERROR: Makepkg was unable to build catalyst-test.

www.downforeveryoneorjustme.com is saying that the whole www2.ati.com domain is down.

Vi0L0 commented on 2012-07-01 17:56

yes, -test is for beta/not-stable drivers, unfortunatelly there's no 12.7 beta for linux so -test won't be updated atm

catalyst have been updated to stable 12.6 but there's afaik zero changes when comparing it to this beta driver...

ttux commented on 2012-06-29 16:26

There is catalyst 12.7 actually which is beta so I suppose catalyst-test will become 12.7 and catalyst will be 12.6

Enverex commented on 2012-06-29 15:12

I assume that if the official version is newer than the beta, that this would be updated to the official? (it doesn't make sense to have to switch between catalyst and catalyst-test every time a new official release comes out).

Therefore flagging this out of date (sorry if the above isn't true).

Anonymous comment on 2012-06-29 11:16

First of all, thanks for the package - works great with my 5770.
Are you going to upgrade it to the final version of Catalyst 12.6 released yesterday or wait for the next "beta" driver?

Anonymous comment on 2012-06-26 18:23

Same as PerisH and maxi_jac. Still with a HD6870.

maxi_jac commented on 2012-06-25 20:56

Same as Perish for me with 6870, startx freezes, no way back to vtty with Ctrl+alt+F1

PerisH commented on 2012-06-20 16:04

Hi! I'm using an Ati 6870 and this package doesn't work for me. GDM freezes and with startx command i can get the sessions load, but then freezes too.

Anonymous comment on 2012-06-19 12:35

For me works very good!! Thanks for your work.

Anonymous comment on 2012-06-19 11:39

@einseenai: You can install a microcode update 'amd-ucode' from the extra repo to get rid of the message.

einseenai commented on 2012-06-16 16:03

Upgraded kernel to 3.4.2, works fine, but on startup returns something like amd u-code bin not loaded.

Vi0L0 commented on 2012-06-16 12:23

updated to 12.6-6:
got now nice clean patch, 3.4 support for both architectures :)

thx to 3V0LU710N for finding it, and Sebastian Siebert from OpenSUSE for a patch

Vi0L0 commented on 2012-06-16 12:22

ttux: if you will use catalyst_build_module it will tell you where's log when fail will happen (/var/log/catalyst-install.log)

ttux commented on 2012-06-14 06:22

All good, I built the package again and it managed to build a fglrx.ko which it put in /lib/modules/3.4.2/video/

Thanks for all the hard work you put into the catalyst packages.

ttux commented on 2012-06-14 06:00

I put this patch and it works:
https://lkml.org/lkml/2012/4/10/485

Now I'm back to my previous error:
make[2]: *** No rule to make target `/root/catalyst/12.6/catalyst-test/src/archive_files/common/lib/modules/fglrx/build_mod/2.6.x/libfglrx_ip.a', needed by `/root/catalyst/12.6/catalyst-test/src/archive_files/common/lib/modules/fglrx/build_mod/2.6.x/fglrx.o'. Stop.

ttux commented on 2012-06-14 05:55

I've upgraded my kernel to 3.4.2, I now get the following when running make.sh
[root@arch build_mod]# ./make.sh
AMD kernel module generator version 2.1
./make.sh: line 390: [: =: unary operator expected
doing Makefile based build for kernel 2.6.x and higher
rm -rf *.c *.h *.o *.ko *.a .??* *.symvers
make -C /lib/modules/3.4.2/build SUBDIRS=/root/catalyst/12.6/catalyst-test/src/archive_files/common/lib/modules/fglrx/build_mod/2.6.x modules
make[1]: Entering directory `/usr/src/linux-3.4.2'
CC [M] /root/catalyst/12.6/catalyst-test/src/archive_files/common/lib/modules/fglrx/build_mod/2.6.x/firegl_public.o
/root/catalyst/12.6/catalyst-test/src/archive_files/common/lib/modules/fglrx/build_mod/2.6.x/firegl_public.c: In function ‘kasInitExecutionLevels’:
/root/catalyst/12.6/catalyst-test/src/archive_files/common/lib/modules/fglrx/build_mod/2.6.x/firegl_public.c:4159:5: error: ‘cpu_possible_map’ undeclared (first use in this function)

My guess is that I have to patch the kernel to put the cpu_possible_map function back?

Treens commented on 2012-06-12 20:09

@Vi0l0 : using #include "thefile.h" tells make/gcc to search the file in the compilation directory instead of searching the "normal" include folders. So we can maybe modify the firegl patch with #include "fpu-internal.h"... And then the script copies the fpu-internal.h from the /usr/src/... into the compilation directory, patches this file and compiles.

@ttux : i'll try to look at that tomorrow, but, ummm, i think you could post what you wrote in the forum

ttux commented on 2012-06-12 18:51

Sorry if I missed something but I have a problem getting fglrx.ko

I'm running make.sh from src/archive_files/common/lib/modules/fglrx/build_mod and I get the following:
[root@arch build_mod]# ./make.sh
AMD kernel module generator version 2.1
./make.sh: line 390: [: too many arguments
./make.sh: line 396: [: too many arguments
doing Makefile based build for kernel 2.6.x and higher
rm -rf *.c *.h *.o *.ko *.a .??* *.symvers
make -C /lib/modules/3.3.6/build SUBDIRS=/root/catalyst/12.6/catalyst-test/src/archive_files/common/lib/modules/fglrx/build_mod/2.6.x modules
make[1]: Entering directory `/usr/src/linux-3.3.6'
CC [M] /root/catalyst/12.6/catalyst-test/src/archive_files/common/lib/modules/fglrx/build_mod/2.6.x/firegl_public.o
CC [M] /root/catalyst/12.6/catalyst-test/src/archive_files/common/lib/modules/fglrx/build_mod/2.6.x/kcl_acpi.o
CC [M] /root/catalyst/12.6/catalyst-test/src/archive_files/common/lib/modules/fglrx/build_mod/2.6.x/kcl_agp.o
CC [M] /root/catalyst/12.6/catalyst-test/src/archive_files/common/lib/modules/fglrx/build_mod/2.6.x/kcl_debug.o
CC [M] /root/catalyst/12.6/catalyst-test/src/archive_files/common/lib/modules/fglrx/build_mod/2.6.x/kcl_ioctl.o
CC [M] /root/catalyst/12.6/catalyst-test/src/archive_files/common/lib/modules/fglrx/build_mod/2.6.x/kcl_io.o
CC [M] /root/catalyst/12.6/catalyst-test/src/archive_files/common/lib/modules/fglrx/build_mod/2.6.x/kcl_pci.o
CC [M] /root/catalyst/12.6/catalyst-test/src/archive_files/common/lib/modules/fglrx/build_mod/2.6.x/kcl_str.o
CC [M] /root/catalyst/12.6/catalyst-test/src/archive_files/common/lib/modules/fglrx/build_mod/2.6.x/kcl_iommu.o
CC [M] /root/catalyst/12.6/catalyst-test/src/archive_files/common/lib/modules/fglrx/build_mod/2.6.x/kcl.o
CC [M] /root/catalyst/12.6/catalyst-test/src/archive_files/common/lib/modules/fglrx/build_mod/2.6.x/kcl_wait.o
make[2]: *** No rule to make target `/root/catalyst/12.6/catalyst-test/src/archive_files/common/lib/modules/fglrx/build_mod/2.6.x/libfglrx_ip.a', needed by `/root/catalyst/12.6/catalyst-test/src/archive_files/common/lib/modules/fglrx/build_mod/2.6.x/fglrx.o'. Stop.
make[1]: *** [_module_/root/catalyst/12.6/catalyst-test/src/archive_files/common/lib/modules/fglrx/build_mod/2.6.x] Error 2
make[1]: Leaving directory `/usr/src/linux-3.3.6'
make: *** [kmod_build] Error 2
build failed with return value 2

It usually worked in previous catalyst.

I tried to use catalyst_build_module but there is no output and it looks like nothing happens. The fglrx.ko is not being built and the one for my kernel is not updated (/lib/modules/3.3.6/kernel/drivers/char/drm/fglrx.ko is still the 12.4 one).

Any idea?

Vi0L0 commented on 2012-06-12 15:29

umm... but how to tell make to use this header instead of the one from kernel src directory? <~~ this would rocks :)

copying all headers into tmp dir is not a way.

anyway your kinda dirty patch is not that bad actually since after module compilation we are restoring original header...

Treens commented on 2012-06-11 15:29

What do you think about creating a temporary fpu-internal.h in the build directory? It's maybe not as dirty as modifying kernel headers.

nob.

Treens commented on 2012-06-11 15:16

Ok, my bad. Patching commands are in catalyst_build_module file >< . But what do you think about creating a temporary fpu-internal.h in the build directory? It's maybe not as dirty as modifying kernel headers.

nob.

Treens commented on 2012-06-11 14:25

I quickly looked at the PKGBUILD and your modified version of the fpu-internal.h patch.

How will it be applied on kernel headers (which are in /usr/src/linux-3.4.XYZ-ARCH) ? I mean, you wrote the line "install -m644 "${srcdir}/fpu.3.4.i686.patch" "${pkgdir}/usr/share/ati/build_mod/" || return 1". If you take a copy of the file (adding some cp command in the PKGBUILD), maybe it will work without modfying dirty the kernel headers, but otherwise i don't see how it will get patched. Further, users which don't have a 3.4+ kernel will get a fail when patching (adding some test on kernel version).
I'm not really good with bash and PKGBUILDs language, so i'm maybe wrong.

nob.

Vi0L0 commented on 2012-06-10 19:17

@nob:
sorry for late update, i was afk
I can see that with your patches module is compiling fine on i686, im unable to test it but i believe it should work so i've added it into all pkgs (well except catalyst - because there i don't know how to change kernel's header, is it even possible in fakeroot ?- i don't think so)
anyway: BIG THX! :D

btw: in -ARCH kernel there is:
arch/x86/include/asm/fpu-internal.h
instead of:
arch/include/asm/fpu-internal.h

so i used one from -ARCH, why pae differ - dunno, should not i believe

Vi0L0 commented on 2012-06-10 18:34

@nob:
sorry for late update, i was afk
I can see that with your patches module is compiling fine on i686, im unable to test it but i believe it should work so i've added it into all pkgs (well except catalyst - because there i don't know how to change kernel's header, is it even possible in fakeroot ?- i don't think so)
anyway: BIG THX! :D

btw: in -ARCH kernel there is:
arch/x86/include/asm/fpu-internal.h
instead of:
arch/include/asm/fpu-internal.h

so i used one from -ARCH, why pae differ - dunno, should not i believe

Treens commented on 2012-06-06 14:18

P.S. : It functions, I tested ;) .

Treens commented on 2012-06-06 14:17

Vi0L0, it even didn't function for me :/ . I tried to find the error (cf. /var/log/catalyst-install.log) : i found an error at the end the firegl_public.c. It comes from a missing include file (#include <asm/fpu-ineternal.h>, which is broken too (missing #include <asm/i387.h>.

I corrected the both errors with those two patches :
-firegl_public.patch-

--- a/usr/share/ati/build_mod/firegl_public.c 2012-06-06 13:54:52.333289403 +0200
+++ b/usr/share/ati/build_mod/firegl_public.c 2012-06-06 13:55:54.446620112 +0200
@@ -34,6 +34,10 @@
#include <linux/autoconf.h>
#endif

+#if LINUX_VERSION_CODE >= KERNEL_VERSION(3,4,0)
+#include <asm/fpu-internal.h>
+#endif
+
#if !defined(CONFIG_X86)
#if !defined(CONFIG_X86_PC)
#if !defined(CONFIG_X86_XEN)

-fpu-internal.patch-

--- a/usr/src/linux-3.4.0-1-pae/arch/include/asm/fpu-internal.h 2012-06-06 13:54:20.599957408 +0200
+++ b/usr/src/linux-3.4.0-1-pae/arch/include/asm/fpu-internal.h 2012-06-06 13:55:29.156621180 +0200
@@ -14,6 +14,7 @@
#include <linux/regset.h>
#include <linux/slab.h>
#include <asm/asm.h>
+#include <asm/i387.h>
#include <asm/cpufeature.h>
#include <asm/processor.h>
#include <asm/sigcontext.h>

I'm under PAE-enabled kernel, but u can(?) remplace linux-3.4.0-1-pae in the second patch with linux-3.4.0-1-ARCH for the kernel headers.
Has someone else the same error?

nob.

Anonymous comment on 2012-06-06 06:16

Any chance of symlinking /usr/lib/libGL.so* and /usr/lib/xorg/modules/extensions/libglx.so in the .install? That way the package won't conflict with https://aur.archlinux.org/packages.php?ID=59794 and it'd be easy to write a script to flip the symlinks for hybrid graphics.

Anonymous comment on 2012-06-04 23:29

On the first attempt installation failed at "lib32-ati-dri: requires lib32-libgl=8.0.3". Adding "lib32-ati-dri" to conflicts in PKGBUILD helped.

Vi0L0 commented on 2012-06-04 15:21

yes, hd < 5 xxx are dropped
i decided to put this info into pkg description - should be more clear now

Anonymous comment on 2012-06-04 10:49

from 12.6 cards below HD 5xxx are dropped, aint it?

Anonymous comment on 2012-06-04 10:29

ok returning to xorg 1.11 and catalyst 11.12 it seems working.

z.

Anonymous comment on 2012-06-04 10:09

Hello. I tried to install 12.6 on my ATI RV710 Mobility Radeon HD 4300 on kernel 3.3.7 but i had these error:

[fglrx] Maximum main memory to use for locked dma buffers: 3661 MBytes
[fglrx:firegl_init_device_list] *ERROR* No supported display adapters were found
[fglrx:firegl_init_module] *ERROR* firegl_init_devices failed

and i had the same errors on kernel 3.4.

I do not know what i've to do because xorg 1.12 have no fglrx working package and it's marked as stable... so it's boring....

zeld

Vi0L0 commented on 2012-06-03 21:19

updated to 12.6-3:
- support of 3.4 kernel is now provided by other patch found here: http://devgurus.amd.com/thread/159241
clean and kinda official :)

Vi0L0 commented on 2012-06-03 21:10

updated to 12.6-3:
- support of 3.4 kernel is now provided by other patch found here: http://devgurus.amd.com/thread/159241
clean and kinda official :)

Vi0L0 commented on 2012-06-03 18:56

updated to 12.6-2:
- added dirt dirty support of 3.4 kernel (basing on w41ter's patches http://www.gossamer-threads.com/lists/gentoo/user/251951)

3V0LU710N commented on 2012-06-01 18:40

Yes, I tried to install 12.6 in my "old" cards... If you do a "sudo aticonfig --initial", it gives you a "No supported adapters found" error... And even hacking the "control" and "signature" files, your card is not detected...

I'm talking about the FOSS ATI/AMD drivers... On desktop, PM doesn't affect you too much, but you can only achieve a PM similar to Catalyst if you put your card in "low" profile / dynpm...

Unfortunately, at the moment, FOSS ATI drivers are still lacking in some features, but I hope/expect that the FOSS drivers will get some feature parity in the future (vs Catalyst).

Cheers :)

LeCrayonVert commented on 2012-06-01 17:56

3V0LU710N > ...and no proper power management ....
Have you tried to install 12.6 with your "old" cards ?

3V0LU710N commented on 2012-06-01 00:15

Yes, I've looked at the news on Phoronix. It seems there's a lot of outrage in the forums...

For me, it's the end of line for both my HD4650AGP and HD3870 PCI-E cards... Now, I have to choose between maintaining xorg 1.11 for a long time and freeze kernel updates, or use FOSS ATI drivers, which never worked very well for me (e.g. no proper H.264 video HW accel, significantly slower 3D performance and almost no OpenCL support ATM) =(

Cheers

Vi0L0 commented on 2012-05-31 23:29

@3V0LU710N:
also: have you seen sad sad news mentioned about Weegee here: https://bbs.archlinux.org/viewtopic.php?pid=1109069#p1109069 ???

Vi0L0 commented on 2012-05-31 23:17

@3V0LU710N:
thank you very much for this info!
I just get back from my little job journey, and used this 11.6 beta driver against my freshly bought hd 7850
Here's what it's offer:
- xserver 1.12 support;
- kde's direct rendering is finally working well, no errors, no flickering at all (!)

Vi0L0 commented on 2012-05-31 23:16

@3V0LU710N:
thank you very much for this info!
I just get back from my little job journey, and used this 11.6 beta driver against my freshly bought hd 7850
Here's what it's offer:
- xserver 1.12 support;
- kde's direct rendering is finally working well, no errors, no flickering at all (!)

Vi0L0 commented on 2012-05-31 22:48

@3V0LU710N:
thank you very much for this info!
I just get back from my little job journey, and used this 11.6 beta driver against my freshly bought hd 7850
Here's what it's offer:
- xserver 1.12 support;
- kde's direct rendering is finally working well, no errors, no flickering at all (!)

3V0LU710N commented on 2012-05-31 20:49

Catalyst 12.6 Beta Released:

http://support.amd.com/us/kbarticles/Pages/AMDCatalyst126beta.aspx
http://www2.ati.com/drivers/hotfix/catalyst_12.6_hotfixes/amd-driver-installer-8.98-x86.x86_64.zip

p.s.: Doesn't support r600/r700 (Radeon HD <5xxx series) cards anymore... =(

Vi0L0 commented on 2011-05-12 23:44

this package supposed to be kind of a "bridge" between stable catalyst version which is not providing fresh xserver support on one side - and a newer one that will provide such support on the other side... so it won't be updated untill there will be such need.

phoronix said that in ubuntu 11.10 there will be xserver 1.10... cuz of proprietary drivers.
i guess that you know what does it mean...

Vi0L0 commented on 2011-04-29 11:40

Yes, it's better to switch to catalyst 11.4 which is in ver 8.841 so it's newer than this one

admiralspark commented on 2011-04-29 08:49

I can verify that catalyst-utils package now works excellent with xorg 1.10 and kde 4.6.x, though still experiencing the rainbow with gnome3. Thanks for the quick updates ViOLO!

Anonymous comment on 2011-04-29 04:42

Vi0L0,

First of all, thank you very very much for your quick and wonderful support to all ATI/ArchLinux users.

Second, now that 11.4 it's officially out... should I go back to standard catalyst / catalyst-utils aur package or it's fine stay on test?

niqingliang2003 commented on 2011-04-19 13:11

thanks!

Vi0L0 commented on 2011-04-17 13:58

Amdcccle is showing different catalyst's version:
https://wiki.archlinux.org/index.php/ATI_Catalyst#Amdcccle_is_showing_different_catalyst.27s_version

GPU/Mem frequency, Temperature, Fan speed, Overclocking utilities: https://wiki.archlinux.org/index.php/ATI_Catalyst#GPU.2FMem_frequency.2C_Temperature.2C_Fan_speed.2C_Overclocking_utilities

Linking fglrx to radeon won't make any difference, you can add vga= (like vga=794) parameter into kernel boot option to set higher resolution of CUI, but it won't be as high as with KMS.

niqingliang2003 commented on 2011-04-16 14:32

another problem, my CCC display:

Software:
Catalyst: 11.2

and no overlock option, and no fan speed, temperature, ...

It sounds like the fan is always full speed running.

niqingliang2003 commented on 2011-04-16 01:46

and if I link the fglrx to the radeon, does the CUI will work with high resolution?

niqingliang2003 commented on 2011-04-16 01:43

thanks!
the "nomodeset" work.

the problem is I never used it, and catalyst always working. so after upgrade to 2.6.38, I neglected the "note".

niqingliang2003 commented on 2011-04-16 01:28

but I never set "nomodeset", and always using catalyst, only 2.6.38 result problem.
I'll restore radeon and try the "nomodeset".

niqingliang2003 commented on 2011-04-16 01:27

but I never set "nomodeset", and always using catalyst, only 2.6.38 result problem.

Vi0L0 commented on 2011-04-15 13:54

there is important note in postinstall of this package:

Add nomodeset to your kernel line in /boot/grub/menu.lst , ie.:
kernel /boot/vmlinuz26 root=/dev/sda1 ro nomodeset

it should do wonders ;)

niqingliang2003 commented on 2011-04-15 12:00

and if I use radeon, after login console, use 'gdm' to start GUI, and then Alt+F1 switch back to console, only got a black screen.

so maybe catalyst can delete the radeon.ko.gz and substitute it with a soft link to fglrx.ko.gz

niqingliang2003 commented on 2011-04-15 11:27

maybe catalyst can delete the radeon.ko.gz and substitute it with a soft link to fglrx.ko.gz

niqingliang2003 commented on 2011-04-15 11:24

IMPORTANT!!!!!!!!!
hello, everybody, I finally found the reason.

the 2.6.38 has a radeon kernel module by itself, and it will be loaded automatically.
and result conflict with catalyst's fglrx.

after I mv /lib/modules/2.6.38-ARCH/kernel/drivers/gpu/drm/radeon/radeon.ko.gz to /lib/modules/2.6.38-ARCH/kernel/drivers/gpu/drm/radeon/radeon.ko.gz.bak
and reboot
I got the catalyst rework!!!!!!!!!!


thanks everybody.

BTW, I add "!radeon" in rc.conf's DAEMON, but no effect, so I must change the radeon's name.

I can't rmmod radeon, it said it is being used.

So I think this problem is still need catalyst's process.

niqingliang2003 commented on 2011-04-15 00:46

not mobile version, mine is desktop pc.

Vi0L0 commented on 2011-04-14 14:06

wait... is it mobile version? if yes - then i wouldn't be surprised - ati is contantly braking and fixing support for mobility gfx :(
The best way in this situation is to use last working driver, aspecially when catalyst-test isn't any better than 11.3
if you got mobility use 11.3 and [xorg19] repo

Vi0L0 commented on 2011-04-14 14:02

ahhh... it does not look good (the log) - for me it looks that ati has brake 6870 support
ask on Bar & Grill if there is some 6870 user - maybe someone has figure it out

Vi0L0 commented on 2011-04-14 13:57

strange... i can not start kdm without xorg.conf... was fglrx initialized correctly then? (you can check that in Xorg.0.log)
as for wnck - im not gnome expert - please write on Bar & Grill forum: https://bbs.archlinux.org/viewtopic.php?id=57084

niqingliang2003 commented on 2011-04-14 13:54

I have pasted the log in
http://pastebin.com/iPqA0PGJ
it said "Not enough video memory to allocate primary surface (frame buffer)."
but my 6870 has 1G video memory.

niqingliang2003 commented on 2011-04-14 13:25

if no /etc/X11/xorg.conf, I can runinto gnome, but one process named wnck will occupy 100% CPU, numerous windows are displayed in the bottom panel.

niqingliang2003 commented on 2011-04-14 13:19

if no /etc/X11/xorg.conf, I can runinto gnome, but one process named wnck will occupy 100% CPU, numerous windows are displayed in the bottom panel.

niqingliang2003 commented on 2011-04-14 13:16

the libgl work ok, and then I installed this package, executed catalyst_build_module
and no /etc/X11/xorg.conf
so I executed: aticonfig --initial --nobackup, but like that:

sudo aticonfig --initial --nobackup
Uninitialised file found, configuring.
Fail to link to fglrx-libglx.so, please check whether driver is installed correctly
Using /etc/X11/xorg.conf

Vi0L0 commented on 2011-04-14 11:36

as for resolution i guess that 1440 * 900 is only saved in your gnome account settings, not in /etc/X11/xorg.conf - so after login run amdcccle (it needs qt to run) as root and change resolution, this should overwrite xorg.conf and make 1440 * 900 your native resolution

catalyst 11.3 is not supporting xorg-server 1.10 - so the "no signal" could appear, although it should just put some errors in tty

use catalyst-test, dont forget to type:
catalyst_build_module
as unpriviliged user after installing it, then reboot, if black screen is still there - put log and dmesg on pestbin service - yes, the log should still be there after reboot

niqingliang2003 commented on 2011-04-14 06:38

the whole process is that:

before update kernel and x server, the catalyst 11.3 is OK, but when I change the original VGA cable(connect with 6870's first DVI through a adaptor) to a 5m VGA cable, my GDM login interface's resolution changed to 1024 * 768 (my display is 1440 * 900), and after login, I changed the resolution to 1440 * 900, but after reboot, the gdm login interface is also 1024 * 768, after login is 1440 * 900

after two days, I updated kernel to 2.6.38, and reinstalled catalyst and utils; when I start gdm, my display display "no signal", like no VGA input.

and then uninstall catalyst...

and then reboot, update x server to 1.10, install catalyst-test 11.4, and I got a black screen, and snowflake, 2.6.38 and 1.10 x server.

niqingliang2003 commented on 2011-04-14 06:38

the whole process is that:

before update kernel and x server, the catalyst 10.3 is OK, but when I change the original VGA cable(connect with 6870's first DVI through a adaptor) to a 5m VGA cable, my GDM login interface's resolution changed to 1024 * 768 (my display is 1440 * 900), and after login, I changed the resolution to 1440 * 900, but after reboot, the gdm login interface is also 1024 * 768, after login is 1440 * 900

after two days, I updated kernel to 2.6.38, and reinstalled catalyst and utils; when I start gdm, my display display "no signal", like no VGA input.

and then uninstall catalyst...

and then reboot, update x server to 1.10, install catalyst-test 11.4, and I got a black screen, and snowflake, 2.6.38 and 1.10 x server.

niqingliang2003 commented on 2011-04-14 06:34

tonight I will reproduce this problem, and then ... a question, after code reboot, does the log stay there?

Vi0L0 commented on 2011-04-13 14:56

@niqingliang2003: can you paste /var/log/Xorg.0.log and dmesg on some pastebin server?

niqingliang2003 commented on 2011-04-13 13:55

I got a black screen, and snowflake, 2.6.38 and 1.10 x server. video card is 6870

Vi0L0 commented on 2011-04-06 14:09

silent update:
2.6.39 patches should be now be applied in better way

Vi0L0 commented on 2011-03-30 15:02

updated to 11.4-111:
- official 2.6.38 support
- xorg-server 1.10 support
- vaapi isn't working (guess xvba-video needs to be updated)
- some artifacts in kde 4 when kwin is enabled - but it's not that bad

- initial 2.6.39 support (through patches)

Vi0L0 commented on 2010-11-27 23:05

Koshon: pretty easy, as root:
catalyst_build_module remove_all
pacman -Rd catalyst-test

+ additionally you can install libgl (which were deleted before this pkg install):
pacman -S libgl

It's all in the wiki: https://wiki.archlinux.org/index.php/ATI_Catalyst#catalyst-generator_-_a_tool_for_generating_catalyst-.7Bkernver.7D_packages

Koshon commented on 2010-11-27 20:53

fuuuuu! How i can delete this sh... So many dependens! >_<

Koshon commented on 2010-11-27 20:49

fuuuuu! How i can delete this sh... So many dependens! >_<

Vi0L0 commented on 2010-10-23 19:05

This catalyst-test is 10.10 RC1, so it should be older than new official catalyst 10.10.
i'm not sure will you notice any difference, but agapito said that with new catalyst 10.10 he's not experiencing memory leaks any more - so fresh catalyst-utils + catalyst (or -hook or -daemon or -generator) is probably better.

Anonymous comment on 2010-10-23 18:22

Is catalyst-utils 10.10 (official release) package same as this package (regarding pkgver)? I use catalyst-test in my Arch64.

Vi0L0 commented on 2010-10-15 12:37

@17.poison:
BusyBox - i never used it, but it looks it got problems...
From http://www.busybox.net/about.html:
"The utilities in BusyBox generally have fewer options"

Maybe remove it and install proper arch packages? :)

Vi0L0 commented on 2010-10-15 12:36

BusyBox - i never used it, but it looks it got problems...
From http://www.busybox.net/about.html:
"The utilities in BusyBox generally have fewer options"

Maybe remove it and install proper arch packages? :)

govindarajulu commented on 2010-10-15 06:14

after removing -u option and without changing anything else, this is what is shown
[root@poison3 ~]# catalyst_build_module
Please specify your (unprivileged) user name:
poison

Building catalyst-2.6.35-ARCH package for 2.6.35-ARCH kernel ...
--------
sed: can't read PKGBUILD: No such file or directory
==> ERROR: PKGBUILD does not exist.

govindarajulu commented on 2010-10-15 06:11

cp "/usr/share/ati/build_mod" "${workdir}" -R

i guess this 'll copy the directory build_mod to /tmp/catalyst.sdserf/ , not all files in /usr/share/ati/build_mod/ to /tmp/catalyst.sdserf
# ls /tmp/catalyst.sdserf
build_mod

so you have to cd /tmp/catalyst.sdserf/build_mod to find PKGBUILD file.

govindarajulu commented on 2010-10-15 06:04

[root@poison3 ~]# mktemp -du
mktemp: invalid option -- 'u'
BusyBox v1.17.2 (2010-09-04 15:52:21 MSD) multi-call binary.

Usage: mktemp [-dt] [-p DIR] [TEMPLATE]

Create a temporary file with name based on TEMPLATE and print its name.
TEMPLATE must end with XXXXXX (e.g. [/dir/]nameXXXXXX).

Options:
-d Make a directory instead of a file
-t Generate a path rooted in temporary directory
-p DIR Use DIR as a temporary directory (implies -t)

For -t or -p, directory is chosen as follows:
$TMPDIR if set, else -p DIR, else /tmp

Vi0L0 commented on 2010-10-14 20:50

@17.poison:
-u is standard mktemp option, from gnu.org:
"‘-u’, ‘--dry-run’ Generate a temporary name that does not name an existing file, without changing the file system contents."
i really dunno why it's telling you it's bad option.

And ${workdir}/build_mod? Actually directory called build_mod shouldn't exist in ${workdir}

Are you sure that everything is fine with your mktemp? Maybe you should reinstall coreutils package?

govindarajulu commented on 2010-10-14 17:29

ref: catalyst_build_module file

hey my system says there in no -u option in mktemp command (line 19)
and i guess
line 22 is -- cd "$workdir" should be -- cd "$workdir/build_mod"
line 30 is -- su - ${user} -c "cd ${workdir} && /usr/bin/makepkg -c" || return 1 , should be -- su - ${user} -c "cd ${workdir}/build_mod && /usr/bin/makepkg -c" || return 1
line 37 -s -- su - root -c "/usr/bin/pacman -Ud ${workdir}/catalyst-${kernver}-${catver}-$(uname -m).pkg" || return 1
should be su - root -c "/usr/bin/pacman -Ud ${workdir}/build_mod/catalyst-${kernver}-${catver}-$(uname -m).pkg" || return 1




Vi0L0 commented on 2010-10-13 02:06

@sati: check out agapito's and next posts: https://bbs.archlinux.org/viewtopic.php?pid=837440#p837440

sati commented on 2010-10-13 00:20

Anyone experiencing massive memory leaks with X using this driver? X seems to gobble a load of memory and doesn't appear to be releasing any of it..

Vi0L0 commented on 2010-10-11 01:54

@FallenWizard: thanks! updated! :)

It's slightly newer than ubuntu's version and was having "Testing use only" water sign, but old trick with signature once again did it's job ;)

Anonymous comment on 2010-10-10 18:26

I forgot the changelog :)

* Kernel module build fix for x86_64 kernels with security fix for CVE-2010-3081
* Generalized kernel module build fix for CVE-2010-3081 to work on RHEL and SLED/SUSE
* Ubuntu 10.10 early-look support (Xserver 1.9 support)
* Updated packaging scripts to support Ubuntu 10.10
* openSUSE 11.3 production support
* Fixed packaging scripts for SLED/SUSE
* Updated packaging scripts
* Checks for available video memory in CCC-LE before allowing to enable memory-intensive features
* Fixed Xserver segfaults and soft hangs when VT switching back to X
* Fixed missing CCC-LE icons in the System-Preferences menu
* Fixed corruption in Xinerama mode
* Fixed CCC-LE Gamma correction applied to wrong displays on openSUSE 11.3
* Fixed text corruption in Steam games
* Added missing maverick symlink in Ubuntu packaging script directory
* Fixed X version detection problems in installer, fixes installation on Fedora 11

Anonymous comment on 2010-10-10 18:22

New leaked beta released: http://www.ati-forum.de/files/Downloads/Driver/ati-driver-installer-8.783_RC1-x86.x86_64.rar

Anonymous comment on 2010-10-08 14:35

@ViOLO: Very good job, thx for the good work!

Vi0L0 commented on 2010-10-08 14:21

lib32-catalyst-utils-test is right here: http://aur.archlinux.org/packages.php?ID=41559

Vi0L0 commented on 2010-10-08 14:14

@agapito: i thought that building lib32-catalyst-utils-test is so easy that i don't need to upload it ;)
but ok, i will upload it in minutes

Vi0L0 commented on 2010-10-08 14:12

update: enriched module compilation options, more here: https://bbs.archlinux.org/viewtopic.php?pid=834227#p834227

@brunopereira81:
since i'm maintainer here (~1 year) nobody noticed that, but you are right - it should be useful for new users.
qt is now in optdeps.
Also if you don't want to install qt package you may uncomment lines 114 & 115 of PKGBUILD to install 2 qt libraries provided with catalyst by ati.

agapito commented on 2010-10-08 14:09

Why lib32-catalyst-utils-test is not in aur?

Anonymous comment on 2010-10-07 20:37

Qt is necessary to run amdcccle, shouldn't that count as a dependency?

Anonymous comment on 2010-10-04 15:28

This package make the window colors strange in my KDE, instead of the standard greyish look the color is similar to light violet blended with some gray.

10.9 work correctly.

I tried 10.10-888 package also from ViOLO repository and the same problem occours using xorg 1.9 of course.

Anonymous comment on 2010-10-04 15:28

This package make the window colors strange in my KDE, instead of the standard greyish look the color is similar to light violet blended with some gray.

10.9 work correctly.

I tried 10.10-888 package also from ViOLO repository and the same problem occours using xorg 1.9 of course.

Vi0L0 commented on 2010-10-01 21:19

@nemolivier:
lib32-catalyst.sh is needed if you want to build this pkg by yourself with makepkg - then you need PKGBUILD + lib32-catalyst.sh (+ optionally lib32-catalyst-utils-test.install) in same directory.

Anonymous comment on 2010-10-01 20:52

@ViOLO, great, but where am-I supposed to put lib32-catalyst.sh ? Is-it only for the AUR package, or for the [catalyst] one too ?
(I’ve installed the [catalyst] one…)
In fact I allready have a lib32-catalyst.sh in my /etc/profile.d !

Vi0L0 commented on 2010-09-30 23:52

@nemolivier:
download lib32-catalyst-utils-test from [catalyst] repo :)

or if you prefer aur here's PKGBUILD: http://pastie.org/1192409
don't forget to take lib32-catalyst.sh from http://aur.archlinux.org/packages.php?ID=24824

Anonymous comment on 2010-09-30 20:46

I can’t install it on my x86_64 because I need -lib32 packages and it seem’s that it’s not compatible with this package…

Vi0L0 commented on 2010-09-28 18:06

@b4283: it's because of bug in kernel26-headers, it should be fixed soon

Anonymous comment on 2010-09-28 18:00

Wow, at last celebrating to have working 3D games on my desktop and now this happens. :p

Vi0L0 commented on 2010-09-28 02:34

x86_64 users:
Temporary solution when updating kernel26 2.6.35.5 -> 2.6.35.6
Do NOT rebuild fglrx module after kernel's update (so don't use catalyst_build_module at all)

If you already rebuilt fglrx module and booted new kernel and you can see nice black tty login screen - then:
1. as root downgrade kernel:
pacman -U /var/cache/pacman/pkg/kernel26-headers-2.6.35.5-1-x86_64.pkg.tar.xz
pacman -U /var/cache/pacman/pkg/kernel26-2.6.35.5-1-x86_64.pkg.tar.xz
2. as unprivileged user type:
catalyst_build_module 2.6.35-ARCH
3. as root update kernel:
pacman -Su
4. reboot
It should work fine.

Vi0L0 commented on 2010-09-26 16:33

update: it now consists of catalyst-utils and catalyst-generator, so we are able to use catalyst_build_module

More info here: http://wiki.archlinux.org/index.php/ATI_Catalyst#catalyst-test_-_catalyst_with_xorg-server_1.9_support

Anonymous comment on 2010-09-26 10:26

It's starting to look like something playing with this update in VA-API hwaccel, evergreen (radeon hd 5650): http://ompldr.org/vNW44cA

Vi0L0 commented on 2010-09-24 03:06

little update: should now work with unpatched (with important security patches) x86_64 kernels

Vi0L0 commented on 2010-09-23 23:11

updated.

works ONLY with xorg-server 1.9, so atm you need to use [testing] repository

no auto-recompilation, no catalyst_build_module inside this package, just catalyst + catalyst-utils

have fun ;)

agapito commented on 2010-09-23 21:17

fglrx-installer (2:8.780-0ubuntu1) maverick; urgency=low

* New upstream release.
- Fix build issues with kernel fix for CVE-2010-3081 (LP: #642518).
- Add compatibility with 2.6.35 kernels (LP: #573748).
- Add compatibility with xserver 1.9 (LP: #630599).
* Make the driver Depend on the appropriate xserver-xorg-video-$ABI
(LP: #616215).

http://packages.ubuntu.com/maverick/fglrx

Vi0L0 commented on 2010-08-28 22:23

little update:
libstdc++5 has been removed from deps array, because it's not required by catalyst - gcc-libs package brings newer libstdc++ lib.
Feel free to remove libstdc++5 package from your system (ofcourse if it's not required by other package).

Vi0L0 commented on 2010-08-28 22:07

little update:
libstdc++5 has been removed from deps array, because it's not required by catalyst - gcc-libs package brings newer libstdc++ lib.
Feel free to remove libstdc++5 package from you system (ofcourse if it's not required by other package).

Vi0L0 commented on 2010-08-27 16:36

updated:
catalyst should now work with 2.6.36 rc2 kernel
patch wroten by Spasswolf, i found it here: http://www.phoronix.com/forums/showthread.php?t=25485

IMPORTANT NOTE in post_install/post_upgrade:
If you are using libva and xvba-video packages then you need to reinstall them, and type this as root:
rm /usr/lib/xorg/modules/dri/*_drv_video.so

WARNING:
__catalyst isn't supporting xorg-server 1.9___ which just landed
in [testing] repo and soon will be pushed to mainline
To avoid black screen use [xorg18] repository:
http://wiki.archlinux.org/index.php/ATI_Catalyst#.5Bxorg18.5D_repository

Vi0L0 commented on 2010-08-15 13:36

update: catalyst-test now fits to ATi Catalyst wiki's page section 'Stock and Custom Kernels'

Vi0L0 commented on 2010-08-14 14:55

Release notes of this firepro driver is telling that it's supporting OpenCL SDK 2.1, so i guess it's not supporting OpenCL 1.1.
Also it's not supporting kernel 34/35, although patch is working. If you are not using FirePro card (like me) you will see "Unsupported hardware" sign in righ bottom corner of the screen - but ive bypassed it with /etc/ati/control file of catalyst 10.7 ;)

I dont actually see any sense of replacing hotfix catalyst driver with this somehow older stuff.

But here is PKGBUILD and other needed files to build firepro driver: http://catalyst.apocalypsus.net/tarball/firepro-8.743.3.3.tar.gz
Important notes for this package:

NOTE: before building please download firepro_8.743.3.3_linux_103646.zip using this site:
http://support.amd.com/us/gpudownload/fire/Pages/fire_linux.aspx
and move it next to this PKGBUILD

NOTE: If you are using radeon card =/= Firepro - you can overwrite /etc/ati/control file with 'control' file provided by this package to make "Unsupported hardware" sign go away

Anonymous comment on 2010-08-13 15:50

I saw that AMD has posted a new version of the firpro driver for linux, version 8.743.3.3. I think this has the openCL 1.1 and newer OpenGL support as well. Could that be used in place of this hotfix version?

Vi0L0 commented on 2010-08-09 02:40

* Updated to 10.7 with OpenCL 1.1 support

+ NOTE: before building please download ati_catalyst_10.7b_opencl1.1_linux.zip using this site:
http://support.amd.com/us/kbarticles/Pages/OpenCL11ATICat107UpdateDriver.aspx
and move it next to PKGBUILD

+ PKGBUILD of older catalyst-test 10.4-666 is still available here:
http://catalyst.apocalypsus.net/tarball/catalyst-test-10.4-666.tar.gz

Vi0L0 commented on 2010-08-09 02:23

* Updated to 10.7 with OpenCL 1.1 support, more info here:
http://support.amd.com/us/kbarticles/Pages/OpenCL11ATICat107UpdateDriver.aspx

+ NOTE: before building please download ati_catalyst_10.7b_opencl1.1_linux.zip using this URL:
https://a248.e.akamai.net/f/674/20307/1m/www2.ati.com/drivers/hotfix/catalyst_10.7b_opencl1.1/ati_catalyst_10.7b_opencl1.1_linux.zip
and move it next to PKGBUILD

+ PKGBUILD of older catalyst-test 10.4-666 is still available here:
http://catalyst.apocalypsus.net/tarball/catalyst-test-10.4-666.tar.gz

Anonymous comment on 2010-05-19 12:53

Working fine with 2.6.34-1 kernel. <3

Vi0L0 commented on 2010-04-29 00:08

[xorg18] has been moved to [testing] :/

Vi0L0 commented on 2010-04-28 22:00

The addresses of [catalyst-test-xorg18] repos had been changed, here are the new ones:

For i686:
[catalyst-test-xorg18]
Server = http://dl.dropbox.com/u/4321633/repo/catalyst-xorg18/i686

For x86_64:
[catalyst-test-xorg18]
Server = http://dl.dropbox.com/u/4321633/repo/catalyst-xorg18/x86_64

Vi0L0 commented on 2010-04-28 20:11

catalyst 10.4 has been released, aur's catalyst has been updated. [catalyst] repo will be updated soon

Anonymous comment on 2010-04-28 07:42

@stromblade:
Have you updated your XServer, if you have done so you possible should rebuild your package.
Or perhaps you havent updated your xorg.conf file. running aticonfig --initial is always a good start.

Anonymous comment on 2010-04-28 01:22

cant start x with this package, other than what has already been posted, are there any other configuration changed that need to be made? All i get is a black screen

silvernode commented on 2010-04-20 00:49

Tried to download with yaourt and got this:
WARNING: certificate common name “a248.e.akamai.net” doesn’t match requested host name “launchpad.net”.
HTTP request sent, awaiting response... 400 Bad Request
2010-04-19 19:44:32 ERROR 400: Bad Request.

silvernode commented on 2010-04-20 00:48

Tried to download with yaourt and got this:
WARNING: certificate common name “a248.e.akamai.net” doesn’t match requested host name “launchpad.net”.
HTTP request sent, awaiting response... 400 Bad Request
2010-04-19 19:44:32 ERROR 400: Bad Request.

Anonymous comment on 2010-04-18 10:23

@chosig : You should have read the wiki : http://wiki.archlinux.org/index.php/ATI_Catalyst#Installation ;)

Anonymous comment on 2010-04-17 22:12

Bit if a merry-go-around when installing.

catalyst-testing is blocked by libgl, libgl is needed by xserver-xorg, catalyst-test definitely won't install without xserver-xorg, which pulls libgl back.

took a while for me to find the -d switch :P

perhaps force-uninstall libgl on catalyst-test install or give a hint on pacman -Rd libgl

+1 from me anyhow, works like a charm when installed, btw / nomodeset was unnessesary for me (just make sure to block ati/radeon/radeonhd from loading)

Vi0L0 commented on 2010-04-13 19:30

update - added support for 2.6.34-rc4 kernel

Vi0L0 commented on 2010-04-13 19:04

update - added support for 3.6.34-rc4 kernel

Vi0L0 commented on 2010-04-11 17:26

updated

agapito commented on 2010-04-11 15:57

https://launchpad.net/ubuntu/lucid/+source/fglrx-installer/2:8.723.1-0ubuntu1/+files/fglrx-installer_8.723.1.orig.tar.gz

Vi0L0 commented on 2010-04-10 14:07

Wanna test xserver 1.8? Read more here: http://bbs.archlinux.org/viewtopic.php?pid=740353#p740353

Vi0L0 commented on 2010-04-09 22:27

! Add nomodeset to your kernel line in /boot/grub/menu.lst , so it may looks like this:
kernel /boot/vmlinuz26 root=/dev/disk/by-uuid/1aa731c0-f626-4d01-a551-8eed3c2a35f3 ro nomodeset vga=794 splash quiet

Vi0L0 commented on 2010-04-08 19:16

kottin: "Remember, YOU MUST recompile the catalyst when you change kernel or you'll cry a lot."
Me: Good point! :)
kottin: "Later I will try with kernel26-bfs, but I think they're the same."
Me: With ck patchset you are getting more than just bfs scheduler

Anonymous comment on 2010-04-08 18:41

ok I did it with kernel26-ck:
Remember, YOU MUST recompile the catalyst when you change kernel or you'll cry a lot.
Later I will try with kernel26-bfs, but I think they're the same.

Vi0L0 commented on 2010-04-08 15:32

well... its always good to do post on forum http://bbs.archlinux.org/viewtopic.php?id=57084&p=66
write there info about your gpu and sys architecture + /var/log/Xorg.0.log
probably also /etc/X11/xorg.conf + ~/.xsession-errors may be helpful

Anonymous comment on 2010-04-08 10:43

I've just installed kernel26-bfs with this catalyst and the Xorg-maximized-fix from Vi0L0:
Xorg start, but it's impossible to enable composition, and whatever I do is really slow.
Maybe I must change Xorg.conf or try using the Xorg from core, I really don't know, some idea?
Sorry for the bad english.

Anonymous comment on 2010-04-05 17:23

With 2.6.32-ARCH kernel


glxinfo

name of display: :0.0
X Error of failed request: BadRequest (invalid request code or no such operation)
Major opcode of failed request: 157 (GLX)
Minor opcode of failed request: 19 (X_GLXQueryServerString)
Serial number of failed request: 20
Current serial number in output stream: 20

Anonymous comment on 2010-04-04 12:14

I get a segmentation fault on x86_64, normal kernel.

Vi0L0 commented on 2010-03-31 19:33

@maveric7911: its working fine with kernel26-ck, so i think it must work with kernel26-bfs also.
About rt - dunno.

maveric7911 commented on 2010-03-31 14:41

any word on getting this working with kernel26-bfs or rt?

maveric7911 commented on 2010-03-31 14:23

any word on getting this working with kernel26-bfs or rt?

Anonymous comment on 2010-03-29 18:48

Thanks. Works well on my Arch64 Box with RV620 LE [Radeon HD 3450] :-)

matkam commented on 2010-03-28 20:51

You guys manage to get AIGLX working for Compiz on the r800 chipset (Radeon HD 5000 Series)?

matkam commented on 2010-03-28 20:48

You guys manage to get Compiz working with this on the r800 chipset (HD 5000 series)? If so, pastebin your xorg.conf files ^,^

sovok_kpss commented on 2010-03-28 03:08

How can I install this with my rt-kernel? some patches e.t.c.

guisacouto commented on 2010-03-26 20:40

its working very well for now, however I cannot get slim to work with it.
when I boot and I try to login from Slim, the keyboard doesnt work (neither the ctrl+alt+fX). I have to ssh with my laptop, into the desktop, stop the slim daemon, kill X, and then startx..

did anyone had this kind of trouble with slim and keyboard(or another login manager)?

regards

adq commented on 2010-03-25 18:30

Hi, thanks for this package!

To get the atieventsd Xauth authentication working with KDM, I needed to modify /etc/ati/authatieventsd.sh line 37 so that it reads as follows:
XDM_AUTH_MASK=/var/run/xauth/A$1*

Anonymous comment on 2010-03-25 17:08

Thank you guys, I'll try that immediately

Vi0L0 commented on 2010-03-25 13:00

this lib32-catalyst-utils is better than mine: http://bbs.archlinux.org/viewtopic.php?pid=728763#p728763
also here is package builded for catalyst-test: http://dl.dropbox.com/u/4321633/lib32-catalyst-utils-10.4-333-x86_64.pkg.tar.xz

Anonymous comment on 2010-03-25 11:43

@Kariddi: I have used this PKGBULD http://bbs.archlinux.org/viewtopic.php?pid=728747#p728747

Anonymous comment on 2010-03-25 10:26

There is a way to use 32 bit opengl applications on 64 bit systems with this? Something like a lib32 package for the libraries? (or are them automatically installed?)

Anonymous comment on 2010-03-24 13:51

Work's great on compaq 615 RS780 (Radeon HD 3200) with kwin out-of-the-box.

Anonymous comment on 2010-03-24 10:31

ViOLO: yeah, i was using xv, cause opengl jst lagged, now with mplayer:vaapi everything is works. Thanks.