Package Details: catalyst-generator 15.9-12

Git Clone URL: https://aur.archlinux.org/catalyst-generator.git (read-only)
Package Base: catalyst-generator
Description: AMD/ATI drivers. Generator of catalyst-{kernver} packages with fglrx module inside.
Upstream URL: http://www.amd.com
Licenses: custom
Conflicts: catalyst, catalyst-daemon, catalyst-dkms, catalyst-hook, catalyst-test
Provides: catalyst=15.9
Submitter: Vi0L0
Maintainer: Vi0L0
Last Packager: Vi0L0
Votes: 25
Popularity: 0.000003
First Submitted: 2010-09-22 22:52
Last Updated: 2016-08-15 15:39

Dependencies (10)

Required by (3)

Sources (19)

  • 4.2-amd-from_crimson_15.11.patch
  • 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-v2.patch
  • ati_make.sh
  • catalyst-PKGBUILD
  • catalyst.install
  • catalyst_build_module
  • completion-bash
  • completion-zsh
  • crimson_i686_xg.patch
  • fglrx_gpl_symbol.patch
  • grsec_arch.patch
  • http://www2.ati.com/drivers/linux/amd-catalyst-15.9-linux-installer-15.201.1151-x86.x86_64.zip
  • lano1106_fglrx_intel_iommu.patch
  • lano1106_kcl_agp_13_4.patch
  • makefile_compat.patch

Latest Comments

Vi0L0 commented on 2016-01-20 14:27

15.9-9:
- added 4.4 kernel support

Vi0L0 commented on 2015-12-13 20:00

15.9-8:
just like commit says:
- adding new patch from crimson
- removing gcc49 dependency
- adding initial grsec kernel support

should be safe to remove gcc49 pkg :)

Vi0L0 commented on 2015-09-29 20:29

updated to 15.9
4.2 kernel is in [core] and we still don't have support for it.
So for now please use linux-lts kernel.

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 2015-06-04 13:23

updated to 15.5-1 - AKA "nothing interesting":
- it doesn't support xserver 1.17
- looks older than ubuntu vivid driver which we are using ie in catalyst-test@AUR
- i've added initial support for 4.1 kernel (using kolasa's patch, tested on 4.1rc5)
- rel notes: http://support.amd.com/en-us/kb-articles/Pages/AMDCatalyst15-5LINReleaseNotes.aspx

Vi0L0 commented on 2015-05-04 19:23

14.12-3:
- 4.0 kernel support (thanks to kolasa for patch)

Vi0L0 commented on 2015-03-08 13:37

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

Vi0L0 commented on 2014-10-03 20:54

updated to 14.9-1:
- official, stable release
http://support.amd.com/en-us/kb-articles/Pages/AMDCatalyst14-9LINReleaseNotes.aspx

- it doesn't support xserver 1.16

- 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 :P
2. im unable to test resume/suspend
3. im not the developer, just pressing random keys :P so the patch could be less/more noobish :-)

Vi0L0 commented on 2014-08-18 20:17

14.4-21: added dirty GPL-licensing solution - fglrx_gpl_symbol.patch - workaround for compilation error for i686 on 3.16.1 kernel

Vi0L0 commented on 2014-07-26 17:03

Thank You.
Yes, I also think it's great, many people worked on this before it get current shape.
I did promoted it many times, but it never get as much people's attention as the -hook. Well...I never presented it on a TV - that could be the reason of why my promotion was unsuccessful ;)

NoSuck commented on 2014-07-24 21:32

This packages is currently the most sensible way to manually install and upgrade Catalyst drivers, and the wiki seems to agree with me (https://wiki.archlinux.org/index.php/AMD_Catalyst#Catalyst-generator). This package should be promoted more.

You've done a great job with this. Thank you.

Vi0L0 commented on 2014-06-09 20:41

should work fine with linux 3.15

Vi0L0 commented on 2014-05-25 11:34

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

Vi0L0 commented on 2014-05-01 19:23

There's a bug in linux 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

Vi0L0 commented on 2014-04-28 20:33

updated to 14.4-10, catalyst 14.4 stable:
- xserver 1.15 support
(gotjazz wrote: little warning though - if you use xbmc, get sdl-nobackingstore from the AUR or you'll have tearing during video playback with xorg 1.15)
- many fixes here and there

Vi0L0 commented on 2014-04-05 11:25

13.12-2:
added Cold's 3.14 patch

Montague commented on 2014-03-09 16:57

Is it normal that this package works with xorg-server 1.15 ?
Sorry for the noob question, I know I am supposed to be using xorg-server 1.14 (which I was until yesterday), but I upgraded to xorg-server 1.15 “by accident”, and now Catalyst still works. Should I downgrade my xorg-server to version 1.14, or is it safe to keep using xorg-server version 1.15 ?

What I mean by “upraded by accident” is I was having an issue with mesa yesterday, so I switched to the opensource xf86-video-ati for a short while, and when I came back to catalyst-generator, I did all the steps from the wiki (uncomment the [xorg114] repo in my pacman.conf, etc.) but forgot to downgrade xorg-server right after that. I just noticed now my error, but that it works regardless.

Vi0L0 commented on 2013-12-19 21:45

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

Vi0L0 commented on 2013-11-18 19:37

silent update, no pkgrel bump:
- thanks to zoopp and his referer we no longer have to download installer manually :)

Vi0L0 commented on 2013-11-17 19:55

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

Vi0L0 commented on 2013-11-08 19:13

13.4-16:
- added workaround to bring support of kernel 3.12, by Krzysztof Kolasa

Vi0L0 commented on 2013-10-27 18:35

upgraded to 13.4-15:
- added staticlibs and libtool to the options and it fixed the whole libfglrx_ip.a bug
- the package is a lot bigger because I added !strip to the options, it's leaving symbols in the libraries so the package (and libs themselves) are bigger - but it could be useful, especially in some debugging

Vi0L0 commented on 2013-08-03 13:52

updated to 13.4-13 :
- switching to 3.10 kernel patch of Emil Karlson from gentoo which is genereting correct proc entries

Vi0L0 commented on 2013-06-01 12:48

updated to 13.4-5:
- thanks to lano1106's patch we don't need AGP support in kernel anymore

Vi0L0 commented on 2013-05-21 16:03

updated to 13.4-4:
- fixes support of intel's iommu (thanks to lano1106)

Vi0L0 commented on 2013-05-20 15:33

updated to 13.4-3:
- brings kernel 3.10 (tested on rc1) support (via procfs workaround);
- fixes ttys bug

BIG THANKS for these goes to Krzysztof Kolasa :: https://github.com/kolasa

Vi0L0 commented on 2013-05-01 09:42

updated to 13.4-2:
Thanks to Mikkel Oscar Lyderik catalyst_build_module script now has got autocompletion on bash and zsh! :D
Autocompletion is showing kernels listed in /usr/lib/modules and others options like all, remove, etc.

Vi0L0 commented on 2013-04-27 21:21

updated to 13.4-1:
http://support.amd.com/us/kbarticles/Pages/AMDCatalyst13-4LINReleaseNotes.aspx

lano1106:
:/ I will look at it tomorrow

lano1106 commented on 2013-04-18 04:45

weird. it is on but still doesn't work...
lano1106@whippet2 ~/dev/linux-ck/src/linux-3.8 $ grep CONFIG_DECOMPRESS_GZIP .config
CONFIG_DECOMPRESS_GZIP=y

Vi0L0 commented on 2013-04-13 08:57

guess that it's the kernel option:
CONFIG_DECOMPRESS_GZIP=y

lano1106 commented on 2013-04-13 04:41

Vi0L0, the generator works super well except for one detail. the generator output is a .ko.gz file and under this form my system doesn't load the module.

I gunzip it and all is fine. Is there a config switch I must turn on when compiling the kernel in order to be able to load generated fglrx.ko.gz file?

Vi0L0 commented on 2013-03-25 19:08

package has been once again updated, but silently, no pkgrel bump
because AMD has changed files that don't belong to this package, only catalyst-utils/total*

Vi0L0 commented on 2013-03-21 18:51

package has been updated, but silently, no pkgrel
because AMD has changed files that don't belong to this package, only catalyst-utils/total*

Vi0L0 commented on 2013-03-16 15:24

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 2013-01-18 14:50

updated to 13.1-1:
- changelog is here: http://support.amd.com/us/kbarticles/Pages/AMDCatalyst131ProprietaryLinuxGraphicsDriverReleaseNotes.aspx

Roken commented on 2012-12-29 21:33

That was it. despite manually updating to be sure (packer -S catalyst-generator) it didn't update.

I manually downloaded the tarball and extracted to /usr/share/ati/build_mod then ran catalyst-build-module. This re-downloaded and buitl the driver but errored out at "/tmp/catalyst.randomechars/catalyst-3.8-git....... not found. CD's to the /tmp folder and pacman -U catalyst-generator.tar.gz, then re-issued catalyst-build-module and it's successfully built and installed.

I have no idea why catalyst-generator didn't update. though :/

Thanks for pointing me in the right direction.

Roken commented on 2012-12-29 21:33

That was it. despite manually updating to be sure (packer -S catalyst-generator) it didn't update.

I manually downloaded the tarball and extracted to /usr/share/ati/build_mod then ran catalyst-build-module. This re-downloaded and buitl the driver but errored out at "/tmp/catalyst.randomechars/catalyst-3.8-git....... not found. CD's to the /tmp folder and pacman -U catalyst-generator.tar.gz, then re-issued catalyst-build-module and it's successfully built and installed.

I have no idea why catalyst-generator didn't update. though :/

Thanks for pointing me in the right direction.

Roken commented on 2012-12-29 20:43

That was it. despite manually updating to be sure (packer -S catalyst-generator) it didn't update.

I manually downloaded the tarball and extracted to /usr/share/ati/build_mod then ran catalyst-build-module. This re-downloaded and buitl the driver but errored out at "/tmp/catalyst.randomechars/catalyst-3.8-git....... not found. CD's to the /tmp folder and pacman -U catalyst-generator.tar.gz, then re-issued catalyst-build-module and it's successfully built and installed.

I have no idea why catalyst-generator didn't update. though :/

Thanks for pointing me in the right direction.

Vi0L0 commented on 2012-12-29 18:46

also are you sure that generator was updated successfully?
ie is this file present: /usr/share/ati/build_mod/arch-fglrx-3.8.patch

Maybe you have changed something after 3.7rc3, like the way that /tmp is mounted?

Vi0L0 commented on 2012-12-29 18:39

Strange...
I've build this kernel using only your files (i only removed lines with steve :> and used directly your config).
i didn't try to run this kernel because i've got other hardware, but i was able to build fglrx module for your kernel, everything went smooth:

# catalyst_build_module 3.8.0-rc1+
Building fglrx module for 3.8.0-rc1+ kernel ...
Ok.


the module is right where it should be:
# ls -hl /lib/modules/extramodules-3.8-git/
-rw-r--r-- 1 root root 1,8M 12-29 19:13 fglrx.ko.gz
-rw-r--r-- 1 root root 11 12-29 19:10 version
# cat /lib/modules/extramodules-3.8-git/version
3.8.0-rc1+

umm... so i guess it's not your config, the fact that module was build is telling that also headers are fine.
Have you tried building module when other kernel, like arch' [core] kernel, was booted?

Roken commented on 2012-12-28 20:27

Oh, and the install file is:

pkgname=linux-git
kernver=3.8.0-rc1+

post_install () {
echo ">"
echo "> Updating module dependencies. Please wait ..."
depmod $kernver

echo ">"
echo "> Generating initramfs, using mkinitcpio. Please wait..."
echo ">"
mkinitcpio -p $pkgname
}

post_upgrade() {
post_install
}

post_remove() {
rm -f -- "boot/initramfs-$pkgname.img"
}

Roken commented on 2012-12-28 20:26

PKGBUILD is at http://pastebin.com/vQXYCvVx

And config is at http://pastebin.com/N20f4ukp

The PKGBUILD is the AUR PKGBUILD for linux-git, slightly modifed to get the config of the currently running kernel.

Vi0L0 commented on 2012-12-28 19:48

sure,
i would also like to look at pkgbuild/install files of your kernel

Roken commented on 2012-12-26 19:39

Yes - it's the same error. I definitely have the uapi directory in kernel headers, and it does work for the core kernel, which is why I think I may be missing a kernel config option in more recent builds. I'm building with the same config, but given the changes to the kernel build, I'm not entirely sure what I may be missing.

Would it help if I posted my kernel config - maybe you can see what I'm missing?

Vi0L0 commented on 2012-12-25 18:40

is it returning same error you mentioned before?
this would be very strange.
btw are you including uapi directory into kernel' headers?
it's needed since 3.7
also: is generator working for core kernel?

ps. i dont have access to pc atm

Roken commented on 2012-12-25 12:42

No - I still can't build it with anything later than kernel 3.7-rc3. I'm obviously missing something but I have no idea what :(

Vi0L0 commented on 2012-12-22 21:41

@Roken, oh and as you said - there were no changes in kernel' drivers (like radeon) but there were changes in kernel's acpi section.
I've used my config from 3.7 kernel to build 3.8 one, no changes.

Vi0L0 commented on 2012-12-22 21:37

updated to 12.11-2:
- brings patch for 3.8 kernel

i really had not much time to work on it - less than 1 hour... it's compiling and looks like working fine, but i'm unable to perform more tests over this patch because i have to go to my family, will be back after a week.
Merry Christmas! ;)

Roken commented on 2012-12-22 16:59

I should add I'm using the main catalyst now (not hd234k) - I upgraded my gfx card :)

Roken commented on 2012-12-21 19:26

Trying to build with the latest git-kernel fails:

/tmp/catalyst.EwaasD/src/kcl_acpi.c:779:58: error: ‘struct dev_archdata’ has no member named ‘acpi_handle’

Now, acpi kernel options appear to have changed, and whilst I'm building everything that seems to be necessary, it's very possible I'm missing a kernel option (in which case please help). I can't see any significant changes to the graphics driver section.

Vi0L0 commented on 2012-10-16 18:38

12.9-2:
- added patch to support 3.7rc1 kernel, it's my own so i don't trust it ;P pls test (yup, seems to work fine here)

Vi0L0 commented on 2012-10-13 18:35

updated to 12.9:
- now compatible with xserv 1.13 :>
- it is a catalyst driver prepared especially for ubuntu

Vi0L0 commented on 2012-10-10 20:52

1.13 is only in [testing] now, i guess i will pull -test when 1.13 will hit [extra]

hd234k unfortunatelly doesn't support 1.13 atm :(

Roken commented on 2012-10-10 19:55

Personally, I'll trust your judgement. I'm holding back the xorg upgrade ATM until you decide catalyst is worth it. I have to ask, though, will it work with catalyst hd234k (xorg 1.13 I mean)

Vi0L0 commented on 2012-10-10 14:42

there's catalyst-test with xserver 1.13 support
i'm not moving to it (maybe i should?) since it's not well tested atm

Roken commented on 2012-10-09 20:39

Is there a chance of a build for xorg 1.13?

Roken commented on 2012-10-09 20:36

Is there a chance of a build for xorg 1.13?

Vi0L0 commented on 2012-10-06 16:01

updated to 12.8-2:
fixed to work with fresh patch (with patch --follow-symlinks)

Roken commented on 2012-09-22 19:22

Scratch my last message.

patch-2.7-1 breaks catalyst-generator

Reverting to patch-2.6.1-3 fixes things.

Roken commented on 2012-09-22 12:43

Ahh - it seems to be a filesystem problem here. Booted into my backup installation, installed the new kernel and ran catalyst-generator on that without a problem. Now restoring the backup to my main drive.

Vi0L0 commented on 2012-09-21 07:43

So it's not so obvious...
What if you will boot into stable kernel, like -ARCH and then use this cmd:

catalyst_build_module all

this should build modules for all installed kernels, should also bypass -git kernel' error which i assume exist.

Roken commented on 2012-09-20 21:27

results as follows:

cd /tmp/catalyst.fQnedX/
(22:19)/tmp/catalyst.fQnedX$ diff Makefile /usr/share/ati/build_mod/Makefile
(22:20)/tmp/catalyst.fQnedX$

So they are identical, but the failure remains :/

Roken commented on 2012-09-20 21:20

results as follows:

cd /tmp/catalyst.fQnedX/
(22:19)/tmp/catalyst.fQnedX$ diff Makefile /usr/share/ati/build_mod/Makefile
(22:20)/tmp/catalyst.fQnedX$

So they are identical, but the failure remains :/

Roken commented on 2012-09-19 22:53

I'll check tomorrow and post back.

Vi0L0 commented on 2012-09-19 21:33

I'm not able to download -git kernel atm, but for me this error does seems to be related with kernel, not generator package...
Maybe something with filesystem handling etc?

Do the test : after catalyst-generator failure enter /tmp/catalyst.<some_random_chars_here> , open Makefile and compare it's content with /usr/share/ati/build_mod/Makefile, should be the same

Roken commented on 2012-09-19 19:16

Sure: - build error (with 3.6.0-rc6+ using teh linux-git PKGBUILD and the same .config as my previous builds)

==> Making package: catalyst- 12.6-4 (Wed Sep 19 20:07:01 BST 2012)
==> Checking runtime dependencies...
==> Checking buildtime dependencies...
==> Retrieving Sources...
==> Extracting Sources...
==> Starting build()...
==> Unpacking archive ...
==> Applying patch(es) ...
File Makefile is not a regular file -- can't patch
1 out of 1 hunk ignored -- saving rejects to file Makefile.rej
==> ERROR: A failure occurred in build().
Aborting...

I can't find Makefile.rej

#lspci
00:00.0 RAM memory: NVIDIA Corporation MCP78S [GeForce 8200] Memory Controller (rev a2)
00:01.0 ISA bridge: NVIDIA Corporation MCP78S [GeForce 8200] LPC Bridge (rev a2)
00:01.1 SMBus: NVIDIA Corporation MCP78S [GeForce 8200] SMBus (rev a1)
00:01.2 RAM memory: NVIDIA Corporation MCP78S [GeForce 8200] Memory Controller (rev a1)
00:01.4 RAM memory: NVIDIA Corporation MCP78S [GeForce 8200] Memory Controller (rev a1)
00:02.0 USB controller: NVIDIA Corporation MCP78S [GeForce 8200] OHCI USB 1.1 Controller (rev a1)
00:02.1 USB controller: NVIDIA Corporation MCP78S [GeForce 8200] EHCI USB 2.0 Controller (rev a1)
00:04.0 USB controller: NVIDIA Corporation MCP78S [GeForce 8200] OHCI USB 1.1 Controller (rev a1)
00:04.1 USB controller: NVIDIA Corporation MCP78S [GeForce 8200] EHCI USB 2.0 Controller (rev a1)
00:06.0 IDE interface: NVIDIA Corporation MCP78S [GeForce 8200] IDE (rev a1)
00:07.0 Audio device: NVIDIA Corporation MCP72XE/MCP72P/MCP78U/MCP78S High Definition Audio (rev a1)
00:08.0 PCI bridge: NVIDIA Corporation MCP78S [GeForce 8200] PCI Bridge (rev a1)
00:09.0 SATA controller: NVIDIA Corporation MCP78S [GeForce 8200] AHCI Controller (rev a2)
00:10.0 PCI bridge: NVIDIA Corporation MCP78S [GeForce 8200] PCI Express Bridge (rev a1)
00:18.0 Host bridge: Advanced Micro Devices [AMD] Family 10h Processor HyperTransport Configuration
00:18.1 Host bridge: Advanced Micro Devices [AMD] Family 10h Processor Address Map
00:18.2 Host bridge: Advanced Micro Devices [AMD] Family 10h Processor DRAM Controller
00:18.3 Host bridge: Advanced Micro Devices [AMD] Family 10h Processor Miscellaneous Control
00:18.4 Host bridge: Advanced Micro Devices [AMD] Family 10h Processor Link Control
01:08.0 Multimedia audio controller: Creative Labs CA0106 Soundblaster
01:09.0 Network controller: Ralink corp. RT2561/RT61 802.11g PCI
02:00.0 VGA compatible controller: Advanced Micro Devices [AMD] nee ATI RV730XT [Radeon HD 4670]
02:00.1 Audio device: Advanced Micro Devices [AMD] nee ATI RV710/730 HDMI Audio [Radeon HD 4000 series]

Not sure what else you will need. catalyst-generator works just fine with kernels pre rc6+

Vi0L0 commented on 2012-09-19 18:33

@Roken: can you provide more details about the error and your system?

Vi0L0 commented on 2012-09-19 18:18

hmm... works fine on my custom 3.6 rc6 x86_64, non-testing system

Roken commented on 2012-09-18 21:59

This was working fine right up to Linux 3.6.0-rc6+, then it fails to build complaining that the Makefile is non-standard. Any chance of a patch to fix it?

Vi0L0 commented on 2012-08-06 22:10

What for? I still beleieve that there are people who aren't using linux package at all
ie we also got linux-lts package in da core...

I can add info about headers into postinstall echo, though

Radioactiveman commented on 2012-07-21 20:19

linux-headers should be a dependency.

Vi0L0 commented on 2012-07-15 21:49

12.6-3:
fixes in .install and "catalyst_build_module all"

Vi0L0 commented on 2012-07-14 19:57

update:
12.6-2 should work fine with glibc 2.16.0-2 and kernel 3.5

Vi0L0 commented on 2012-06-16 12:52

updated to 12.4-7:
got now nice clean patch for 3.4 support on both architectures :)

thx to 3V0LU710N for finding it, and Sebastian Siebert from OpenSUSE for a patch

3V0LU710N commented on 2012-06-10 23:14

@Vi0L0: I've accidentally removed the link of Ubuntu's patch... Here it is available again: https://bugs.launchpad.net/ubuntu/+source/fglrx-installer/+bug/998486

Cheers

Vi0L0 commented on 2012-06-10 20:42

@3V0LU710N:
i've used other, dirty, patch which working is confirmed

The patch you have found is more clean though, i think that i have to find some tester before applying it...
anyway thanks :)

Vi0L0 commented on 2012-06-10 20:42

@3V0LU710N:
i've used other, dirty, patch which working is confirmed

The patch you have found is more clean though, i think that i have to find some tester before applying it...
anyway thanks :)

Vi0L0 commented on 2012-06-10 19:37

@3V0LU710N:
i've used other, dirty, patch which working is confirmed

The patch you have found is more clean though, i think that i have to find some tester before applying it...
anyway thanks :)

Vi0L0 commented on 2012-06-10 19:29

should now work also on i686 with kernel 3.4

3V0LU710N commented on 2012-06-05 12:40

Greetings Vi0L0,

I found this patch on Ubuntu's Launchpad, but I'm not sure if it's working on x86 (32 bit) architectures and kernel 3.4.x:

https://bugs.launchpad.net/ubuntu/+source/fglrx-installer/+bug/998486
The patch: https://bugs.launchpad.net/ubuntu/+source/fglrx-installer/+bug/998486/+attachment/3143181/+files/doit.patch

Cheers

Vi0L0 commented on 2012-06-03 21:58

updated to 12.4-4:
- 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 21:40

updated to 12.4-4:
- 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 2012-06-02 00:34

Just thought that i can get no of cpus using:
cat /proc/cpuinfo | grep processor | wc -l
So you won't have to type it manually when using 3.4 kernel, i will change it in next release

Vi0L0 commented on 2012-06-02 00:13

updated to 12.4-2:
- added dirt dirty support of 3.4 kernel (basing on w41ter's patches http://www.gossamer-threads.com/lists/gentoo/user/251951)

Vi0L0 commented on 2011-12-07 18:00

@Deewiant:
thank you very much.
Fixed by changing sed's delimiter - | instead of /

Deewiant commented on 2011-12-07 13:44

There's a problem in the PKGBUILD. In check_ext() the 'destidir' variable can get set to "${kernver}/video". Then in install_module(), the command 'sed -i -e "s/_destidir=.*/_destidir=${destidir}/" PKGBUILD' gets run. This obviously doesn't work: '${destidir}' contains a "/", so the sed command gets borked and what ends up happening is the following error while building:

sed: -e expression #1, char 41: unknown option to `s'

And the install directory becomes, at least in my case, '/lib/modules/whatever' instead of '/lib/modules/${kernver}/video'.

This could be fixed by using a different character than '/' in the sed command(s). For example, an exclamation mark works well: 'sed -i -e "s!_destidir=.*!_destidir=${destidir}! PKGBUILD'. Of course, it's possible that somebody has an exclamation mark in their kernel version... the robust fix would be to not use sed (or at least its 's' command) at all.

Vi0L0 commented on 2011-05-12 23:27

# catalyst_build_module remove_all
isn't that helpful in packages other than catalyst-generator, because there this command can be - and is - used autmatically if you will remove main package (like catalyst-hook or ie. catalyst-total).
Here, in generator, it cannot be. generator is more manual, but it has got some pros like:
- you can build fglrx module as user in fakeroot, so it's safer;
- CFLAGS will be used;
- every fglrx module is packed into package so pacman knows about it - there are no files thrown here and there.
It fits to Arch Way better.

I picked catalyst-hook instead of generator in catalyst-total only because hook got more votes.

Anonymous comment on 2011-05-12 22:21

I didn't get the remove package - or wasn't aware of it in the process of installing catalyst-total - and the automatic deinstallation. The first time, I acknowledged it was today, after the first yaourt installation.
Anyway - thanks for telling me about
# catalyst_build_module remove_all

Vi0L0 commented on 2011-05-12 21:09

This behaviour is expected... i mean the point here was to show user that he need to remove those packages when he is removing catalyst-generator package.
Plus post remove should gave you info that you need to manually remove those packages...

Thats specification of pacman - it's unable to remove created by catalyst-generator catalyst-{kernver} packages because there could be only one instance of pacman ran in same time.

That's why it's easier to run:
# catalyst_build_module remove_all
before removing catalyst-generator package - it's all in the wiki.

Anyway solution here is to remove all those (catalyst-2.6.3*) packages manually.

If you got catalyst-total installed it's better to run:
# catalyst_build_module all
after whole removing operation, because it will propbably also remove all fglrx kernel modules from your system - pacman will probably think that they belong to those catalyst-2.6.3* packages

Anonymous comment on 2011-05-12 17:04

The remove process of this package seems to have a little flaw:
missing dependency for catalyst-2.6.32-lts : catalyst-generator
missing dependency for catalyst-2.6.38-ARCH : catalyst-generator
missing dependency for catalyst-2.6.38-lqx : catalyst-generator
missing dependency for catalyst-2.6.39-rc3-mainline : catalyst-generator
missing dependency for catalyst-2.6.39-rc5-mainline : catalyst-generator

Installing catalyst-total instead was the starting point for this. Now, every yaourt installation does have these messages at the end.

Vi0L0 commented on 2011-04-06 13:59

silent update:
2.6.39 patches should be now be applied in better way

Vi0L0 commented on 2011-03-30 02:50

silent update: 2.6.39 rc1 support

in 2.6.39 kernel BKL was removed completely, and thats why 2.6.39_bkl.patch appear (used in combination with no_bkl.patch)

Vi0L0 commented on 2011-02-17 14:09

another silent update:
no_bkl.patch will be added when kernel's # CONFIG_BKL is not set

got working 2.6.38 rc5 without Big Kernel Lock and even my dvd seems to work fine

so guess it's safe to remove BKL out now ;)

Vi0L0 commented on 2011-02-10 17:51

Silent update.
I didn't like the way how ati's code was cluttering PKGBUILDs or catalyst_build_module scripts, so i turned this code into ati_make.sh file and import it's content with . (source)
New way also fits to ati license better... ;P
Oh and now all AUR catalyst's packages have 2.6.38 patch so they support 38 rc3, rc4 and probably newer rcs (no rc1, rc2 support, but hey rc3 and rc4 are much better).
Change is mainly cosmetical, there's no need to update packages.

Vi0L0 commented on 2011-02-08 22:36

note: supports also 38 rc4

Vi0L0 commented on 2011-02-01 17:03

11.1-2 -> 11.1-3 update:
2.6.38_console.patch brings 2.6.38 rc3 (and maybe newer) support, i dropped rc2 and rc1 support :P

Vi0L0 commented on 2011-01-20 21:59

little update = no pkgrel bump

added 2.6.38 rc1 support (using same patches as on 2.6.37)

Vi0L0 commented on 2010-11-27 14:19

@dickeny: ok then, thanks

but instead of your 'sudo -ll -n /usr/bin/pacman' i used simple file existance check

it's because it wasn't work for me right: my user is in wheel and my sudoers line looks like:
%wheel ALL=(ALL) ALL
and then:
$ sudo -ll -n /usr/bin/pacman
sudo: sorry, a password is required to run sudo

Anonymous comment on 2010-11-27 12:49

@Vi0L0
for some scripts building and installing pkg (e.g. yaourt), they prefer to use sudo which may be more secure (i think).

catalyst-generator will NOT depend on sudo. I think the good way is that catalyst_build_module try to find and run sudo, or run su.

anyway i don't think su to root is a good way.

Vi0L0 commented on 2010-11-27 09:30

@dickeny: why? With sudo i will have to add sudo to dependencies and users will have to add their user to sudoers.
And whats the difference?
su: type root password
sudo: type user password

sudo can be more 'userfriendly' if you will add your user to some NOPASSWD group, but i think that it's not secure.

Anonymous comment on 2010-11-26 14:39

Suggest to use 'sudo' to install pkg. just like this:
http://aur.pastebin.com/eLsn2FGy

Vi0L0 commented on 2010-11-19 19:03

h5n1 here: http://www.phoronix.com/forums/showpost.php?p=157233&postcount=29
wrote that catalyst 10.11 (with patches that we are using for some time) works on .37 rc2-git5 kernel.

I just tested it, and can confirm - it's compiling and working fine.

Vi0L0 commented on 2010-11-16 23:24

today i tested it against .37 rc2 - compilation went without errors, but X is crashing while start :(
i've tested many patches with same problem, sigh, and because of no compilation error is rly hard for me to tell whats going on...
so please be patient for .37 rc2 support

Xorg's log most important lines looks like this:

[ 14.567] (==) fglrx(0): NoAccel = NO
[ 14.568] (==) fglrx(0): ATI 2D Acceleration Architecture enabled
[ 14.568] (--) fglrx(0): Chipset: "ATI Radeon HD 4800 Series " (Chipset = 0x9442)
[ 14.568] (--) fglrx(0): (PciSubVendor = 0x1002, PciSubDevice = 0x0502)
[ 14.568] (==) fglrx(0): board vendor info: original ATI graphics adapter
[ 14.568] (--) fglrx(0): Linear framebuffer (phys) at 0xd0000000
[ 14.568] (--) fglrx(0): MMIO registers at 0xfe9f0000
[ 14.568] (--) fglrx(0): I/O port at 0x0000c000
[ 14.568] (==) fglrx(0): ROM-BIOS at 0x000c0000
[ 14.568] (EE) fglrx(0): failed to map MMIO register space!(EE) fglrx(0): MapMMIO failed
[ 14.568] (EE) fglrx(0): PreInitAdapter failed
[ 14.568] (EE) fglrx(0): PreInit failed
[ 14.568] (II) fglrx(0): === [xdl_x760_atiddxPreInit] === end
[ 14.570] SetVBEMode failed
[ 14.570] (II) UnloadModule: "fglrx"
[ 14.571] (II) UnloadModule: "fglrxdrm"
[ 14.571] (II) UnloadModule: "vgahw"
[ 14.571] (II) Unloading /usr/lib/xorg/modules/libvgahw.so
[ 14.571] (II) UnloadModule: "fglrxdrm"
[ 14.571] (EE) Screen(s) found, but none have a usable configuration.
[ 14.571]
Fatal server error:
[ 14.571] no screens found

Vi0L0 commented on 2010-11-04 16:44

@agapito: thanks, md5sum is updated now

but unfortunatelly because of my slow connection im not even able to download kernel26-git, i'll try to provide support when .37 rc2 will arrive

agapito commented on 2010-11-04 15:22

Don't work with latest kernel26-git

catalyst_build_module

Building catalyst-2.6.37-rc1-00027-gff8b16d-dirty package for 2.6.37-rc1-00027-gff8b16d-dirty kernel ...
--------
==> Creando el paquete: catalyst-2.6.37-rc1-00027-gff8b16d-dirty 10.10-3 (jue nov 4 16:17:31 CET 2010)
==> Comprobando Dependencias en tiempo de Ejecución...
==> Comprobando Dependencias en tiempo de Compilación...
==> Obteniendo Fuentes...
==> Descomprimiendo fuentes...
==> Iniciando build()...
patching file Makefile
patching file firegl_public.c
patch unexpectedly ends in middle of line
Hunk #2 succeeded at 411 with fuzz 1.
patching file firegl_public.c
Hunk #1 succeeded at 5102 (offset 8 lines).
OsVersion says: SMP=1
file /proc/kallsyms says: SMP=1
file /lib/modules/2.6.37-rc1-00027-gff8b16d-dirty/build/include/generated/autoconf.h says: SMP=1
file /lib/modules/2.6.37-rc1-00027-gff8b16d-dirty/build/include/generated/autoconf.h says: MODVERSIONS=
file /lib/modules/2.6.37-rc1-00027-gff8b16d-dirty/build/arch/x86/include/asm/compat.h says: COMPAT_ALLOC_USER_SPACE=arch_compat_alloc_user_space
make: se ingresa al directorio `/usr/src/linux-2.6.37-rc1-00027-gff8b16d-dirty'
CC [M] /tmp/catalyst.ZU7hjv/src/firegl_public.o
CC [M] /tmp/catalyst.ZU7hjv/src/kcl_acpi.o
In file included from include/linux/io.h:22:0,
from include/linux/pci.h:54,
from /tmp/catalyst.ZU7hjv/src/firegl_public.c:114:
/usr/src/linux-2.6.37-rc1-00027-gff8b16d-dirty/arch/x86/include/asm/io.h:44:21: error fatal: xen/xen.h: No existe el fichero o el directorio
compilación terminada.
CC [M] /tmp/catalyst.ZU7hjv/src/kcl_agp.o
make[1]: *** [/tmp/catalyst.ZU7hjv/src/firegl_public.o] Error 1
make[1]: *** Se espera a que terminen otras tareas....
In file included from include/linux/io.h:22:0,
from include/linux/pci.h:54,
from /tmp/catalyst.ZU7hjv/src/kcl_acpi.c:24:
/usr/src/linux-2.6.37-rc1-00027-gff8b16d-dirty/arch/x86/include/asm/io.h:44:21: error fatal: xen/xen.h: No existe el fichero o el directorio
compilación terminada.
make[1]: *** [/tmp/catalyst.ZU7hjv/src/kcl_acpi.o] Error 1
In file included from include/linux/io.h:22:0,
from include/linux/pci.h:54,
from /tmp/catalyst.ZU7hjv/src/kcl_agp.c:39:
/usr/src/linux-2.6.37-rc1-00027-gff8b16d-dirty/arch/x86/include/asm/io.h:44:21: error fatal: xen/xen.h: No existe el fichero o el directorio
compilación terminada.
make[1]: *** [/tmp/catalyst.ZU7hjv/src/kcl_agp.o] Error 1
make[1]: *** wait: No hay ningún proceso hijo. Alto.
make: *** [_module_/tmp/catalyst.ZU7hjv/src] Error 2
make: se sale del directorio `/usr/src/linux-2.6.37-rc1-00027-gff8b16d-dirty'

agapito commented on 2010-11-04 14:56

catalyst-PKGBUILD md5sum is wrong

Vi0L0 commented on 2010-11-04 13:30

update: added testing support for 2.6.37 RC1 kernel (sema_init.patch)

thanks goes to DestroyFX from http://forums.gentoo.org/viewtopic-p-6473144.html?sid=0d087b0e5e3fd2c5fa9091ac993db4d0 site

Vi0L0 commented on 2010-10-24 13:37

thanks, updated

10.10-2 shall fix depmod issue

in catalyst-test i did it right, but i totally forgot about it in original catalyst-generator... sorry mates

Vi0L0 commented on 2010-10-24 13:35

thanks, updated

10.10-2 shall fix depmod issue

Anonymous comment on 2010-10-24 03:26

(1/1) checking for file conflicts [#########################################] 100%
(1/1) installing catalyst-2.6.34-ice [#########################################] 100%
/tmp/alpm_pfNlDW/.INSTALL: line 5: depmod: command not found
error: command failed to execute correctly
---
please use '/sbin/depmod' instead of 'depmod'.

Vi0L0 commented on 2010-09-26 18:27

More info about this package is here: http://wiki.archlinux.org/index.php/ATI_Catalyst#catalyst-generator_-_a_tool_for_generating_catalyst-.7Bkernver.7D_packages

Vi0L0 commented on 2010-09-25 21:27

update: cleaning
As for me this package is in good shape atm.

Vi0L0 commented on 2010-09-24 21:07

update! more here: https://bbs.archlinux.org/viewtopic.php?pid=830754#p830754

Vi0L0 commented on 2010-09-23 01:35

More info here: https://bbs.archlinux.org/viewtopic.php?pid=829880#p829880