The AMD/ATI Bar & Grill

Vi0L0 · 2013-11-24 09:10:11

ringtail wrote:

Has anybody an idea for my problem?
Is it OK that fglrx_dri.so opened with vi starts with "^?ELF@@@@" in /usr/lib/dri location and with "ELF@@@@" in /usr/lib/xorg/modules/dri/ ?

$ ls -l /usr/lib/dri/fglrx_dri.so
lrwxrwxrwx 1 root root 38 11-23 14:58 /usr/lib/dri/fglrx_dri.so -> /usr/lib/xorg/modules/dri/fglrx_dri.so

since it's just a symlink the answer for your question is no. You should be looking at the same file, there should be no differences.

Last edited by Vi0L0 (2013-11-24 09:10:44)

Kuci · 2013-11-24 11:05:51

People, how could I get rid of the second screen ? I used:

$ aticonfig --adapter=all --initial

because X doesn't start without configuration produced by this command. Here is the xorg.conf:

Section "ServerLayout"
	Identifier     "aticonfig Layout"
	Screen      0  "aticonfig-Screen[0]-0" 0 0
	Screen         "aticonfig-Screen[1]-0" RightOf "aticonfig-Screen[0]-0"
EndSection

Section "Module"
EndSection

Section "Monitor"
	Identifier   "aticonfig-Monitor[0]-0"
	Option	    "VendorName" "ATI Proprietary Driver"
	Option	    "ModelName" "Generic Autodetecting Monitor"
	Option	    "DPMS" "true"
EndSection

Section "Monitor"
	Identifier   "aticonfig-Monitor[1]-0"
	Option	    "VendorName" "ATI Proprietary Driver"
	Option	    "ModelName" "Generic Autodetecting Monitor"
	Option	    "DPMS" "true"
EndSection

Section "Device"
	Identifier  "aticonfig-Device[0]-0"
	Driver      "fglrx"
	BusID       "PCI:0:1:0"
EndSection

Section "Device"
	Identifier  "aticonfig-Device[1]-0"
	Driver      "fglrx"
	BusID       "PCI:1:0:0"
EndSection

Section "Screen"
	Identifier "aticonfig-Screen[0]-0"
	Device     "aticonfig-Device[0]-0"
	Monitor    "aticonfig-Monitor[0]-0"
	DefaultDepth     24
	SubSection "Display"
		Viewport   0 0
		Depth     24
	EndSubSection
EndSection

Section "Screen"
	Identifier "aticonfig-Screen[1]-0"
	Device     "aticonfig-Device[1]-0"
	Monitor    "aticonfig-Monitor[1]-0"
	DefaultDepth     24
	SubSection "Display"
		Viewport   0 0
		Depth     24
	EndSubSection
EndSection

I don't have attached any secondary monitor, that's why I want to get rid of it. However, X doesn't start when I remove line 4. Any solution ?

OramahMaalhur · 2013-11-24 11:12:37

Hello.
I'm using catalst-test-pxp (integrated intel card + HD7730M).
Since some version of catalyst my laptop cannot wake up after suspend and sometimes it even cannot suspend. This problem is actual for the current version of catalyst, 13.11. With blacklisted fglrx and switched on integrated card the laptop hibernates and wakes up well.
I cannot use nomodeset, 'cause in this case X crashes during start (integrated card cannot be initialized with disabled KMS). Anyway previously it worked with KMS. atieventsd is run.
What could be the cause?
p.s.
I guess this question was asked some time ago, just cannot find it in the topic.

AuHop · 2013-11-24 14:27:17

@OramahMaalhur
I encountered just about the exact same problem while trying to get around the backlight issues. With switchable graphics if I activated suspend and then immediately switched to a tty, it came back successfully. I don't know why X was crashing when trying to return to itself directly, but switching there from the tty worked fine and I was able to continue.
Besides that, all I can see available is to try to get a dedicated iGPU or dedicated dGPU mode working. Catalyst seems to be pretty good about supporting this in at least nominative sense, but much of the time it simply doesn't work. Given that even my backlight control hasn't worked, despite my hardware mux having been set explicitly to activate only the dGPU, I wouldn't probably get your hopes up (at all?) regarding a switchable configuration. The only reason it seemed to support backlight control there is because the display is still wired to the iGPU, where (Intel's) drivers have yet to languish and were providing the proper ACPI interface handling.
Tldr - Catalyst can barely support its own devices, let alone a dynamic configuration with Intel. Try switching to a tty simultaneously if you must attempt this.
Afaik Intel's driver will indeed fail without KMS.

Last edited by AuHop (2013-11-24 14:38:59)

ivdok · 2013-11-24 16:32:13

@Vi0L0 Wikipedia article claims this:

That means what I can buy Fluendo's batch of codecs, and use GStreamer-capable player like Totem to play with XvBA? Or GStreamer with -bad and -ugly will be sufficient to play? Or I am not right at all, and I must use recompiled MPlayer?

mojtabazn · 2013-11-24 17:03:09

hey i had installed latest catalyst test because of som fixed bug for my laptop from repositories in wiki
and now have some questions :

1: when i was installing i had removed open source driver and pacman asked me to remove libcl because of conflicting also libgl so is it natural? in wiki there is just it asks fo removing libgl

2: i used aticonfig --initial so do i need remove files in xorg.conf.d or just leave theme be there
because i have 20-radeon.conf for my open source driver ?

3 :here is the problem why every time i install catalyst from aur or repositories i have this error ?

(EE) AIGLX error: failed to open /usr/X11R6/lib64/modules/dri/fglrx_dri.so, error[/usr/X11R6/lib64/modules/dri/fglrx_dri.so: cannot open shared object file: No such file or directory]

with locate this file is in 3 dir

locate fglrx_dri.so 

/usr/lib/dri/fglrx_dri.so
/usr/lib/xorg/modules/dri/fglrx_dri.so
/usr/lib32/xorg/modules/dri/fglrx_dri.so

i know i should simlink but with which files or directories please give me exact command

and sorry for my bad english

ivdok · 2013-11-24 17:46:31

mojtabazn wrote:

hey i had installed latest catalyst test because of som fixed bug for my laptop from repositories in wiki
and now have some questions :
1: when i was installing i had removed open source driver and pacman asked me to remove libcl because of conflicting also libgl so is it natural? in wiki there is just it asks fo removing libgl
2: i used aticonfig --initial so do i need remove files in xorg.conf.d or just leave theme be there
because i have 20-radeon.conf for my open source driver ?
3 :here is the problem why every time i install catalyst from aur or repositories i have this error ?
(EE) AIGLX error: failed to open /usr/X11R6/lib64/modules/dri/fglrx_dri.so, error[/usr/X11R6/lib64/modules/dri/fglrx_dri.so: cannot open shared object file: No such file or directory]
with locate this file is in 3 dir
locate fglrx_dri.so 

/usr/lib/dri/fglrx_dri.so
/usr/lib/xorg/modules/dri/fglrx_dri.so
/usr/lib32/xorg/modules/dri/fglrx_dri.so
i know i should simlink but with which files or directories please give me exact command
and sorry for my bad english

Please, try to use Pacman-driven repo way of installing fglrx - it's the most error-proof way right now. With AUR you are most probably going to suffer from random errors, which occur from time to time, especially when not on stable package.

mojtabazn · 2013-11-24 18:25:55

i used exact one of the repositories in wiki

ivdok · 2013-11-24 18:33:21

mojtabazn wrote:

i used exact one of the repositories in wiki

Try to symlink 2nd or 1st one. As I remember, they all are the same. If they are not, you are doing it wrong, or didn't installed lib32-catalyst-utils.

Last edited by ivdok (2013-11-24 18:34:27)

mojtabazn · 2013-11-24 18:50:56

i had installed this package also and then simlinked this way

#ln -s /usr/lib64/xorg/modules /usr/X11R6/lib64

but now have error and black screen

cat /var/log/X0rg.0.log |grep WW

[   141.667] (WW) The directory "/usr/share/fonts/OTF/" does not exist.
[   141.708] (WW) `fonts.dir' not found (or not valid) in "/usr/share/fonts/100dpi/".
[   141.709] (WW) `fonts.dir' not found (or not valid) in "/usr/share/fonts/75dpi/".
[   142.958] (WW) Falling back to old probe method for fglrx
[   143.321] (WW) fglrx: No matching Device section for instance (BusID PCI:0@1:0:1) found
[   143.439] (WW) fglrx(0): board is an unknown third party board, chipset is supported
[   146.719] (WW) evdev: USB Optical Mouse: ignoring absolute axes.
[   180.123] (WW) xf86CloseConsole: KDSETMODE failed: Input/output error
[   180.123] (WW) xf86CloseConsole: VT_GETMODE failed: Input/output error

cat /var/log/X0rg.0.log |grep EE

[   180.123] (EE) 
[   180.123] (EE) xf86CloseConsole: VT_ACTIVATE failed: Input/output error
[   180.123] (EE) 
[   180.123] (EE) 
[   180.123] (EE) Please also check the log file at "/var/log/Xorg.0.log" for additional information.
[   180.123] (EE) 
[   180.123] (EE) 
[   180.123] (EE) Backtrace:

note: system doesnt crash i can turn off with power button just black screen

Last edited by mojtabazn (2013-11-24 18:54:36)

ivdok · 2013-11-24 18:59:28

mojtabazn wrote:

i had installed this package also and then simlinked this way

#ln -s /usr/lib64/xorg/modules /usr/X11R6/lib64

but now have error and black screen

cat /var/log/X0rg.0.log |grep WW

[   141.667] (WW) The directory "/usr/share/fonts/OTF/" does not exist.
[   141.708] (WW) `fonts.dir' not found (or not valid) in "/usr/share/fonts/100dpi/".
[   141.709] (WW) `fonts.dir' not found (or not valid) in "/usr/share/fonts/75dpi/".
[   142.958] (WW) Falling back to old probe method for fglrx
[   143.321] (WW) fglrx: No matching Device section for instance (BusID PCI:0@1:0:1) found
[   143.439] (WW) fglrx(0): board is an unknown third party board, chipset is supported
[   146.719] (WW) evdev: USB Optical Mouse: ignoring absolute axes.
[   180.123] (WW) xf86CloseConsole: KDSETMODE failed: Input/output error
[   180.123] (WW) xf86CloseConsole: VT_GETMODE failed: Input/output error

cat /var/log/X0rg.0.log |grep EE

[   180.123] (EE) 
[   180.123] (EE) xf86CloseConsole: VT_ACTIVATE failed: Input/output error
[   180.123] (EE) 
[   180.123] (EE) 
[   180.123] (EE) Please also check the log file at "/var/log/Xorg.0.log" for additional information.
[   180.123] (EE) 
[   180.123] (EE) 
[   180.123] (EE) Backtrace:

note: system doesnt crash i can turn off with power button just black screen

> ln -s /usr/lib64/xorg/modules /usr/X11R6/lib64
Oh god, why entire folder, and not fglrx_dri.so?

Vi0L0 · 2013-11-24 20:28:30

ivdok wrote:

@Vi0L0 Wikipedia article claims this:
http://i5.5cm.ru/i/vlOH.png
That means what I can buy Fluendo's batch of codecs, and use GStreamer-capable player like Totem to play with XvBA? Or GStreamer with -bad and -ugly will be sufficient to play? Or I am not right at all, and I must use recompiled MPlayer?

Dunno, never done such things with gstreamer, i have seen gstreamer-vaapi but never native. Also the info about native xvba in mplayer... i'm not sure if it's true, i never heard about this, wiki claims some old 5 yo phoronix art, while newer 2 yo art is telling otherwise: http://www.phoronix.com/scan.php?page=n … &px=OTM3MA,

phoronix wrote:

X-Video Bitstream Acceleration is the native format AMD exposes by their binary Linux blob for exposing the UVD2/UVD3 engine. XvBA resembles Microsoft's DXVA, so this new MPlayer support may help out anyone interested in a native XvBA implementation for MPlayer, but there's no word of anyone acting on that at the moment.

@mojtabazn:
i doubt that making such a symlink will help, because i doubt that fglrx_dri.so is the problem here.
/usr/X11R6/lib64/modules/dri/fglrx_dri.so is stupidly hardcoded (by AMD) path, but system should find this lib anyhow.
Can you paste whole Xorg.0.log somewhere?

Last edited by Vi0L0 (2013-11-24 20:29:44)

OramahMaalhur · 2013-11-25 07:11:05

AuHop wrote:

@OramahMaalhur
Try switching to a tty simultaneously if you must attempt this.

Good idea, thank you!

I've just tried to make systemd services to switch to another tty automatically before suspend, but have no luck. Switching back to the tty with X after resume works, here is it if anyone is interested in it:
/etc/systemd/system/catalyst-resume.service

[Unit]
Description=Catalyst suspend workaround
After=suspend.target

[Service]
Type=simple
ExecStart=/usr/bin/chvt 1

[Install]
WantedBy=suspend.target

chvt 1 - switches to the tty with X, tty1 in my case

mojtabazn · 2013-11-25 08:19:42

Vi0L0 wrote:

ivdok wrote:
@Vi0L0 Wikipedia article claims this:
http://i5.5cm.ru/i/vlOH.png
That means what I can buy Fluendo's batch of codecs, and use GStreamer-capable player like Totem to play with XvBA? Or GStreamer with -bad and -ugly will be sufficient to play? Or I am not right at all, and I must use recompiled MPlayer?
Dunno, never done such things with gstreamer, i have seen gstreamer-vaapi but never native. Also the info about native xvba in mplayer... i'm not sure if it's true, i never heard about this, wiki claims some old 5 yo phoronix art, while newer 2 yo art is telling otherwise: http://www.phoronix.com/scan.php?page=n … &px=OTM3MA,
phoronix wrote:
X-Video Bitstream Acceleration is the native format AMD exposes by their binary Linux blob for exposing the UVD2/UVD3 engine. XvBA resembles Microsoft's DXVA, so this new MPlayer support may help out anyone interested in a native XvBA implementation for MPlayer, but there's no word of anyone acting on that at the moment.
@mojtabazn:
i doubt that making such a symlink will help, because i doubt that fglrx_dri.so is the problem here.
/usr/X11R6/lib64/modules/dri/fglrx_dri.so is stupidly hardcoded (by AMD) path, but system should find this lib anyhow.
Can you paste whole Xorg.0.log somewhere?

hi

i installed fresh arch and after installing again this error exists without any simlink

i use startx to start xfce

and here is my whole Xorg.0.log

http://privatepaste.com/download/7f756dfbcc

Edit 1 : i tried to removing the .xinitrc and xserverrc from my home folder

to use default x windows i can see borders of windows hardly because of low brightness. fn+increasing brithness button do nothing and other VTs are blank screen after starting x

Last edited by mojtabazn (2013-11-25 11:17:22)

MRWITEK · 2013-11-25 16:09:33

Finally! The 13.11 beta v9.4 driver works properly in general. Last time AMD made this good driver for my laptop [Samsung laptop NP355V5X-S01RU AMD HD 7520G+7670M] was more than year ago! I hope there won't be huge bugs in this one.
P.S. Thanks A LOT Vi0L0. There is really small amount of people (I think) who can do the stuff you do and I am glad to see your work on catalyst being part of the community.

Peace

Half-Shot · 2013-11-25 18:17:19

Latest update seems to have upset my graphics card. Gnome 3 dies after 10 or so seconds (when my programs start to execute on startup). Openbox is ok but there can be graphics corruption when a lot is going on. This didn't happen on the last beta i had installed (beta 6). I'm sure it will sort itself out eventually but any thoughts?

EDIT: 3D Gaming works fine. No corruption or problems of any kind so it can't be a power related issue. Tested on Team Fortress Classic and Team Fortress 2
EDIT2: Fixed by reinstalling the gnome package. Sorry i can't be more specific.

Last edited by Half-Shot (2013-11-25 19:47:43)

lano1106 · 2013-11-26 04:25:18

ivdok wrote:

@Vi0L0 Wikipedia article claims this:
http://i5.5cm.ru/i/vlOH.png
That means what I can buy Fluendo's batch of codecs, and use GStreamer-capable player like Totem to play with XvBA? Or GStreamer with -bad and -ugly will be sufficient to play? Or I am not right at all, and I must use recompiled MPlayer?

If hardware video decoding is very important to you, I would suggest to try out the OSS driver radeon as video decoding support is superior to what catalyst offer. Just double check what is supported. I think that I have read that UVD1 was supported and not UVD2 chips (or maybe it is the exact opposite) Anyway, be careful as I think that the freedesktop wiki is really not up to date regarding video decoding capability in their driver feature matrix. I have learn the video decoding breakthrough of the OSS driver on phoronix in an article that dates back to April 2013.

Please do not buy anything if Fluendo codecs aren't free. You should first check the other free alternatives. I can check how mplayer behaves with Xvba (if they still have Xvba support in recent mplayer releases...) but do not hold your breathe because

1. the vaapi-xvba driver is very thin. If this guy do not work, I do not see how a native Xvba app would do it.
2. The bug that I have filled highlights clearly that the problem is fundamentally on the Xvba side and happens right at the initialisation phase by doing something very basic and simple. The probability that mplayer or any other xvba app can workaround the issue are very low. but maybe a second confirmation that XvbA is broken isn't a bad thing.

tcdavis · 2013-11-26 06:35:41

I'm pretty close to giving up on the dGPU entirely, but here goes one last try...

My goals are pretty simple:

Have my laptop display work (Intel integrated)
Have the external monitor attached to my docking station (wired to Radeon 8750M) take over for the laptop when I'm at work. Or run in a different X session, I'm not picky.

The first item was easy. The second item is proving impossible. I started out with the OSS AMD drivers. I crafted an X config that worked to the extent that I could say

startx -- -screen External

which was setup to only use the BusID of the Radeon GPU. I had an external display, but the driver caused too many lock-ups/crashes of X. And at some point decided to start redrawing Emacs windows at 1FPS.

I then tried Catalyst drivers. First from the AUR with various variants. I eventually started fresh (reinstalled Arch) and used the repository instead. Now I'm stuck again. The closest I can get to a working external display is an extended desktop to a black screen and "X" cursor. This was the result of

aticonfig --initial

then using amdcccle to choose "High Performance - Rendering and Additional Displays" as the mode (this never persists past one restart). Unfortunately, because the "working" external display is never present in xrandr output, I can't even play with those settings in order to make it work.

Whenever I try to create an X config with the more advanced options (like desktop cloning, xinerama, etc.) to see what that might do, aticonfig complains that it's not supported when using RandR 1.2 (which I have tried to disable via /etc/ati/amdpcsdb and via the X config to no avail).

As a final note, I'll also point out that xrandr seems to be my only choice, because amdcccle doesn't even contain a section for "Displays", despite it being a very common reference in these discussions.

Can anybody help?

ivdok · 2013-11-26 16:49:37

lano1106 wrote:

If hardware video decoding is very important to you, I would suggest to try out the OSS driver radeon as video decoding support is superior to what catalyst offer. Just double check what is supported. I think that I have read that UVD1 was supported and not UVD2 chips (or maybe it is the exact opposite) Anyway, be careful as I think that the freedesktop wiki is really not up to date regarding video decoding capability in their driver feature matrix. I have learn the video decoding breakthrough of the OSS driver on phoronix in an article that dates back to April 2013.

I frequently watch videos in HD, so yes, I need it, but I can't yet switch to FOSS drivers yet - I also use wine frequently for old games, and performance is better on catalyst.

lano1106 wrote:

Please do not buy anything

Don't worry - I won't yet If Fluendo capabilities are provided via non-free GStreamer (-bad, -ugly), I shall use them.

lano1106 wrote:

...if Fluendo codecs aren't free. You should first check the other free alternatives.

In my country algorithms are not subject for copyright law, so I'm free to use, for example, libdvdcss(2) and all GStreamer packs. And I don't worry extremely severe like RMS about openness - if the tool/lib does it's job, and does right - I won't request sources or alike. That's why I use catalyst - it does it's job, and I'm yet satisfied with the result. And while it's better in 3D, I will prefer it over FOSS radeon(si) driver.

lano1106 wrote:

I can check how mplayer behaves with Xvba (if they still have Xvba support in recent mplayer releases...) but do not hold your breathe

Yes-yes, if it's not hard for you

lano1106 wrote:

1. the vaapi-xvba driver is very thin. If this guy do not work, I do not see how a native Xvba app would do it.

I would belive you, since media transcoding is not my area.

lano1106 wrote:

2. The bug that I have filled highlights clearly that the problem is fundamentally on the Xvba side and happens right at the initialisation phase by doing something very basic and simple. The probability that mplayer or any other xvba app can workaround the issue are very low. but maybe a second confirmation that XvbA is broken isn't a bad thing.

Fundamentally? Oh no. I don't want to wait for DxVA implementation (buggy, unnative for Linux code!) or switch to Windows (I have no plans to do so - my current state fully relieves me). I will bear with switching Xorgs (is it performance-hog?), but is it possible to have both Radeon Xorg (main, with KMS and XvBA-VAAPI) and Catalyst Xorg (external)? AFAIK, I had to blacklist radeon to be able to startx.

Last edited by ivdok (2013-11-26 16:49:59)

Vi0L0 · 2013-11-26 17:23:30

MRWITEK wrote:

Thanks A LOT Vi0L0. There is really small amount of people (I think) who can do the stuff you do and I am glad to see your work on catalyst being part of the community.

In fact it's pretty easy, everyone could do that. You only need willingness and time, but it's not eating much time after writing some scripts.

ivdok wrote:

I will bear with switching Xorgs (is it performance-hog?), but is it possible to have both Radeon Xorg (main, with KMS and XvBA-VAAPI) and Catalyst Xorg (external)? AFAIK, I had to blacklist radeon to be able to startx.

If you mean working on same xorg-server version it would be pretty easy. Easiest with -pxp and it's pxp-switch-catalyst script, its already adding/removing fglrx from modules, so only blacklist radeon or nomodeset is required. Switching between other versions of xorg-server will require additional scripting. No, there's no difference in performance between xorg-server versions.
But yes, it's stupid, let's just hope that AMD will fix xvba support... or jump quickly with all its hands and legs to the oss sea ;P.

Last edited by Vi0L0 (2013-11-26 17:34:15)

ivdok · 2013-11-26 18:37:38

Vi0L0 wrote:

If you mean working on same xorg-server version it would be pretty easy. Easiest with -pxp and it's pxp-switch-catalyst script, its already adding/removing fglrx from modules, so only blacklist radeon or nomodeset is required. Switching between other versions of xorg-server will require additional scripting. No, there's no difference in performance between xorg-server versions.
But yes, it's stupid, let's just hope that AMD will fix xvba support... or jump quickly with all its hands and legs to the oss sea ;P.

Maybe I was not really clear, but I have two PCs - one laptop, which I don't really use, and desktop, which I prefer. Both have some kind of HD5xxx, but desktop has good powerful HD5770 (At least it was in 2010, when I bought it), and about laptop I don't know much more than even at Windows it has bad drivers. I'd like to use two Xorgs (not simultaneously is acceptable) - one with free drivers and perfect 2D, and "SUPAH GAEMING", which will NOT turn off first one. Second would be just a booster to run games, while first is for everyday work. I know what it's hard, but is it possible? I can disable KMS for both, if that's the case (plymouth sucks IMO), but blacklisting radeon? This is clearly bad move. I can understand unloading a kernel module, but disabling it? That means only catalyst Xorg, which I currently use. I can continue using 13.6, but how far can this continue? Holding packages is not an Arch Way. If I'd like oldstable packages - I'd be using Debian right now.

//Just thinking out loud, no offence or calling for actions meant.

Vi0L0 · 2013-11-26 19:48:11

ivdok wrote:

I have two PCs - one laptop, which I don't really use, and desktop, which I prefer. Both have some kind of HD5xxx, but desktop has good powerful HD5770 (At least it was in 2010, when I bought it), and about laptop I don't know much more than even at Windows it has bad drivers. I'd like to use two Xorgs (not simultaneously is acceptable) - one with free drivers and perfect 2D, and "SUPAH GAEMING", which will NOT turn off first one. Second would be just a booster to run games, while first is for everyday work. I know what it's hard, but is it possible?

The easy way, and i was talking about it when mentioning "scripting", is to use local repo with older xorg-server inside and switching between older and newer version with pacman, inside the script.
The hard way is to have both, older and newer, xorg-server (libs and drivers, like input-evdev) versions installed simultaneously. But to do this you will have to perform some hackish reconfiguration (look at for example of how -pxp works, you may have both arch's and catalyst's libgls installed simultaneously) and probably hackish recompilation of older xorg-server and xf86-input-* packages.

ivdok · 2013-11-27 06:04:28

Vi0L0 wrote:

The easy way, and i was talking about it when mentioning "scripting", is to use local repo with older xorg-server inside and switching between older and newer version with pacman, inside the script.
The hard way is to have both, older and newer, xorg-server (libs and drivers, like input-evdev) versions installed simultaneously. But to do this you will have to perform some hackish reconfiguration (look at for example of how -pxp works, you may have both arch's and catalyst's libgls installed simultaneously) and probably hackish recompilation of older xorg-server and xf86-input-* packages.

I agree, it's hackish way. But I will try to fix it later - when I reinstall Arch with x64 architecture - right now I use x86. Maybe fresh start would help me to fix XvBA-VDPAU/VAAPI connector via other way, as described here or here

Last edited by ivdok (2013-11-27 06:05:05)

Vi0L0 · 2013-11-27 09:58:16

xvba/vaapi is working fine with newest catalyst (13.11 beta 9.4) and xorg-server 1.13.4. I have used [xorg113] repo. Both cvlc and mplayer are working without an error here. Can someone confirm?

[xorg113]
Server = http://catalyst.wirephire.com/repo/xorg113/$arch

I'm typically downgrading 3 packages:
pacman -S xorg-server xorg-server-common xf86-input-evdev

visit · 2013-11-27 16:57:18

Hi guys,

is it normal, that fglrx is disabling my ATI cards PCI device on Xserver start?
It is not possible to switch to dgpu after that, as aticonfig --px-dgpu tells me "No supported adapters detected".

Nov 27 12:04:10 quorra kernel: fglrx_pci 0000:06:00.0: irq 65 for MSI/MSI-X
Nov 27 12:04:10 quorra kernel: <6>[fglrx] Firegl kernel thread PID: 1127
Nov 27 12:04:10 quorra kernel: <6>[fglrx] Firegl kernel thread PID: 1128
Nov 27 12:04:10 quorra kernel: <6>[fglrx] Firegl kernel thread PID: 1129
Nov 27 12:04:10 quorra kernel: <6>[fglrx] IRQ 65 Enabled
Nov 27 12:04:10 quorra kernel: <6>[fglrx] Reserved FB block: Shared offset:0, size:1000000 
Nov 27 12:04:10 quorra kernel: <6>[fglrx] Reserved FB block: Unshared offset:f878000, size:4000 
Nov 27 12:04:10 quorra kernel: <6>[fglrx] Reserved FB block: Unshared offset:f87c000, size:484000 
Nov 27 12:04:10 quorra kernel: <6>[fglrx] Reserved FB block: Unshared offset:7fff4000, size:c000 
Nov 27 12:04:10 quorra kernel: vgaarb: this pci device is not a vga device
Nov 27 12:04:10 quorra kernel: <6>[fglrx] IRQ 65 Disabled
Nov 27 12:04:10 quorra acpid[906]: input device has been disconnected, fd 6
Nov 27 12:04:10 quorra kernel: <6>[fglrx] module unloaded - fglrx 13.25.5 [Nov 22 2013]
Nov 27 12:04:10 quorra kernel: vgaarb: invalid PCI address 0:6:0

catalyst-hook 13.11-10
catalyst-utils-pxp 13.11-10
xf86-video-intel 2.21.15-1
intel-dri 9.2.3-2
xorg-server 1.14.4-1

I'd appreciate any help. Thx.

Arch Linux

#5901 2013-11-24 09:10:11

Re: The AMD/ATI Bar & Grill

#5902 2013-11-24 11:05:51

Re: The AMD/ATI Bar & Grill

#5903 2013-11-24 11:12:37

Re: The AMD/ATI Bar & Grill

#5904 2013-11-24 14:27:17

Re: The AMD/ATI Bar & Grill

#5905 2013-11-24 16:32:13

Re: The AMD/ATI Bar & Grill

#5906 2013-11-24 17:03:09

Re: The AMD/ATI Bar & Grill

#5907 2013-11-24 17:46:31

Re: The AMD/ATI Bar & Grill

#5908 2013-11-24 18:25:55

Re: The AMD/ATI Bar & Grill

#5909 2013-11-24 18:33:21

Re: The AMD/ATI Bar & Grill

#5910 2013-11-24 18:50:56

Re: The AMD/ATI Bar & Grill

#5911 2013-11-24 18:59:28

Re: The AMD/ATI Bar & Grill

#5912 2013-11-24 20:28:30

Re: The AMD/ATI Bar & Grill

#5913 2013-11-25 07:11:05

Re: The AMD/ATI Bar & Grill

#5914 2013-11-25 08:19:42

Re: The AMD/ATI Bar & Grill

#5915 2013-11-25 16:09:33

Re: The AMD/ATI Bar & Grill

#5916 2013-11-25 18:17:19

Re: The AMD/ATI Bar & Grill

#5917 2013-11-26 04:25:18

Re: The AMD/ATI Bar & Grill

#5918 2013-11-26 06:35:41

Re: The AMD/ATI Bar & Grill

#5919 2013-11-26 16:49:37

Re: The AMD/ATI Bar & Grill

#5920 2013-11-26 17:23:30

Re: The AMD/ATI Bar & Grill

#5921 2013-11-26 18:37:38

Re: The AMD/ATI Bar & Grill

#5922 2013-11-26 19:48:11

Re: The AMD/ATI Bar & Grill

#5923 2013-11-27 06:04:28

Re: The AMD/ATI Bar & Grill

#5924 2013-11-27 09:58:16

Re: The AMD/ATI Bar & Grill

#5925 2013-11-27 16:57:18

Re: The AMD/ATI Bar & Grill

Board footer