The AMD/ATI Bar & Grill

LadFromWales · 2011-06-15 17:00:03

Vi0L0 wrote:

@LadFromWales:
looks like a good material for a bug report at http://ati.cchtml.com
but first check what will happen if you will remove everything from xorg.conf's Section "Module"

Thanks Vi0L0. Given removing everything from xorg.conf "Module" section a go...same behaviour
Have submitted a bug report as suggested... http://ati.cchtml.com/show_bug.cgi?id=167

EDIT: Trying 11.6 now!

Yep, still happening with 11.6!

Last edited by LadFromWales (2011-06-15 18:56:28)

Vi0L0 · 2011-06-15 18:06:26

Looks like OpenGL acceleration for flash works fine even when Tear-Free Desktop is enabled.
To enable it as root:

mkdir /etc/adobe && touch /etc/adobe/mms.cfg && echo "OverrideGPUValidation=true" > /etc/adobe/mms.cfg

DarksideEE7 · 2011-06-15 18:39:21

Vi0L0 wrote:

Looks like OpenGL acceleration for flash works fine even when Tear-Free Desktop is enabled.
To enable it as root:
mkdir /etc/adobe && touch /etc/adobe/mms.cfg && echo "OverrideGPUValidation=true" > /etc/adobe/mms.cfg

Have you been having problems with fullscreen flash using kwin as well? The last flashplugin release helped a lot, but I still get flickering on Justin.tv with fullscreen (not youtube), and sometimes have to switch to a tty console using ctrl+alt+f{2-6}, then back to F{7,8}.

Updating now via ssh.....but I'm not at home so I can't test yet.

Last edited by DarksideEE7 (2011-06-15 18:47:05)

Vi0L0 · 2011-06-15 18:54:26

DarksideEE7 wrote:

Have you been having problems with fullscreen flash using kwin as well?

KDE is my main DE and my kwin is always on.

DarksideEE7 wrote:

The last flashplugin release helped a lot, but I still get flickering on Justin.tv with fullscreen (not youtube), and sometimes have to switch to a tty console using ctrl+alt+f{2-6}, then back to F{7,8}.

I had such problems even on youtube but only when using OpenGL acceleration for flash. Now with 11.6 everything is ok and Justin.tv is also working well.

Aarin · 2011-06-15 19:11:57

@Vi0L0

- nomodeset was already at the kernel line.
- fglrx was put correctly by the pxp_switch_catalyst at MODULES (so the problem shouldn't be there)
- since 11.6 was out, I decided to try it out. No changes. Same segmentation fault at xorg.
- I have removed from mkinitcpio.conf the i915 and intel_agp modules I had there, and done a mkinitcpio -p kernel26
- "modprobe.blacklist=i915" added to the kernel line. no changes... both drivers are loaded.

the funny thing is that Xorg says:

[ 23.231] (II) Loading /usr/lib/xorg/modules/drivers/fglrx_drv.so
[ 23.231] (II) Loading /usr/lib/xorg/modules/linux/libfglrxdrm.so
[ 23.231] (II) fglrx(0): pEnt->device->identifier=0xf72330
[ 23.233] (II) Loading /usr/lib/xorg/modules/drivers/intel_drv.so
[ 23.234] (II) intel(1): pEnt->device->identifier=(nil)
[ 23.234] (EE) Screen 1 deleted because of no matching config section.
[ 23.234] (II) UnloadModule: "intel"
[ 23.234] (II) Unloading intel

so the intel driver shouldn't be unloaded? Oo

After all this I decided to try to enter with the intel graphic card, so I've done a "pxp_switch_catalyst intel" (which had no errors) and I rebooted. This time xorg didn't loaded fglrx, but Xorg says that there isn't no screen available to the card (perhaps because pxp_switch_catalyst "deletes" xorg.conf when switching to the intel card?)

sumski · 2011-06-15 20:22:44

It seems catalyst 11.6 borked HDMI output, but "only" with X. tty is fine

agapito · 2011-06-15 20:39:02

phoronix user wrote:

alright, I just installed 11.6......honestly I have never been more livid with a release, last release there was a bug that causes everything to freeze up for a second or two and revs up the cpu all the way every single time u make the incredible mistake of moving ur f###ing mouse near the lower right corner of ur screen.....now such a glaring and frustrating bug in one release is annoying, but no ones perfect. However to have the same glaring bug which basically ruins my computing experience in the next release is intolerable. It just goes to show how much time and effort they really put into these releases.

WTF.

FUCK AMD.

DarksideEE7 · 2011-06-16 02:07:22

Vi0L0 wrote:

DarksideEE7 wrote:
Have you been having problems with fullscreen flash using kwin as well?
KDE is my main DE and my kwin is always on.
DarksideEE7 wrote:
The last flashplugin release helped a lot, but I still get flickering on Justin.tv with fullscreen (not youtube), and sometimes have to switch to a tty console using ctrl+alt+f{2-6}, then back to F{7,8}.
I had such problems even on youtube but only when using OpenGL acceleration for flash. Now with 11.6 everything is ok and Justin.tv is also working well.

I tested with Justin.tv and youtube, and I'm still experiencing some flickering from the Justin.tv player's overlay after going fullscreen (KDE with kwin, opengl). Youtube seems to work well.

I'm going to add this line to my mms.cfg that you mentioned:
OverrideGPUValidation=true

So should this line be uncommented as well?
EnableLinuxHWVideoDecode=1

The file was already present for me, so I didn't have to create it by touching it.

EDIT:

Excellent, adding this line:
OverrideGPUValidation=true
fixed the flickering problem with Justin.tv's overlay. Thanks for that, I didn't even notice that the adobe directory existed.....how embarrassing.

Last edited by DarksideEE7 (2011-06-16 02:17:12)

sumski · 2011-06-16 05:49:14

sumski wrote:

It seems catalyst 11.6 borked HDMI output, but "only" with X. tty is fine

Can anyone reproduce this?

yimm · 2011-06-16 08:09:14

sumski wrote:

sumski wrote:
It seems catalyst 11.6 borked HDMI output, but "only" with X. tty is fine
Can anyone reproduce this?

I will try

Edit : HDMI output isn't detected when I connect hdmi cable. I have to switch to tty, and switch again to X in order to detect hdmi cable.

Last edited by yimm (2011-06-16 08:28:18)

Gillian00 · 2011-06-16 09:41:43

so, no change at all for gnome 3 ??

Vi0L0 · 2011-06-16 10:40:37

Aarin wrote:

@Vi0L0
- nomodeset was already at the kernel line.
- fglrx was put correctly by the pxp_switch_catalyst at MODULES (so the problem shouldn't be there)
- since 11.6 was out, I decided to try it out. No changes. Same segmentation fault at xorg.
- I have removed from mkinitcpio.conf the i915 and intel_agp modules I had there, and done a mkinitcpio -p kernel26
- "modprobe.blacklist=i915" added to the kernel line. no changes... both drivers are loaded.
the funny thing is that Xorg says:
[ 23.231] (II) Loading /usr/lib/xorg/modules/drivers/fglrx_drv.so
[ 23.231] (II) Loading /usr/lib/xorg/modules/linux/libfglrxdrm.so
[ 23.231] (II) fglrx(0): pEnt->device->identifier=0xf72330
[ 23.233] (II) Loading /usr/lib/xorg/modules/drivers/intel_drv.so
[ 23.234] (II) intel(1): pEnt->device->identifier=(nil)
[ 23.234] (EE) Screen 1 deleted because of no matching config section.
[ 23.234] (II) UnloadModule: "intel"
[ 23.234] (II) Unloading intel
so the intel driver shouldn't be unloaded? Oo

Strange that xserver is even trying to load intel, it shouldn't be loaded at all.
It is unloaded (line 130) and loaded again (line 159).
Maybe you've got some conf file in /etc/X11/xorg.conf.d dir, like 20-intel.conf?
Or maybe intel needs to be blacklisted in xorg.conf (i dunno if it's even possible)?

Aarin wrote:

After all this I decided to try to enter with the intel graphic card, so I've done a "pxp_switch_catalyst intel" (which had no errors) and I rebooted. This time xorg didn't loaded fglrx, but Xorg says that there isn't no screen available to the card (perhaps because pxp_switch_catalyst "deletes" xorg.conf when switching to the intel card?)

Script do not deletes any xorg.conf, it's renaming it into:
- xorg.conf.cat if it will find fglrx (it will be linked later to xorg.conf);
- xorg.conf.oth if it will find intel or radeon (will be linked later);
- xorg.conf.wth if it will not find fglrx, intel or radeon (won't be linked or used, it's backup made because of unperfection of the script and my poor knowledge about intel).
So your xorg.conf for intel do not have "intel" word inside? If so and you've got ie. "i915" let me know, i will update script.

BTW: still it should load intel succesfully, even if xorg.conf is not present - at least it's working for radeon.

Vi0L0 · 2011-06-16 10:50:46

DarksideEE7 wrote:

So should this line be uncommented as well?
EnableLinuxHWVideoDecode=1

No, afaik it's used by nvidia drivers to accelerate video via vdpau.

DarksideEE7 wrote:

The file was already present for me, so I didn't have to create it by touching it.

Ah i didn't think about it, yes it could be present.
I'm using flashplugin-prerelease from AUR and there even /etc/adobe dir is not present.

Gillian00 wrote:

so, no change at all for gnome 3 ??

Indeed.

Last edited by Vi0L0 (2011-06-16 11:13:37)

Aarin · 2011-06-16 11:57:35

Vi0L0 wrote:

Aarin wrote:
@Vi0L0
- nomodeset was already at the kernel line.
- fglrx was put correctly by the pxp_switch_catalyst at MODULES (so the problem shouldn't be there)
- since 11.6 was out, I decided to try it out. No changes. Same segmentation fault at xorg.
- I have removed from mkinitcpio.conf the i915 and intel_agp modules I had there, and done a mkinitcpio -p kernel26
- "modprobe.blacklist=i915" added to the kernel line. no changes... both drivers are loaded.
the funny thing is that Xorg says:
[ 23.231] (II) Loading /usr/lib/xorg/modules/drivers/fglrx_drv.so
[ 23.231] (II) Loading /usr/lib/xorg/modules/linux/libfglrxdrm.so
[ 23.231] (II) fglrx(0): pEnt->device->identifier=0xf72330
[ 23.233] (II) Loading /usr/lib/xorg/modules/drivers/intel_drv.so
[ 23.234] (II) intel(1): pEnt->device->identifier=(nil)
[ 23.234] (EE) Screen 1 deleted because of no matching config section.
[ 23.234] (II) UnloadModule: "intel"
[ 23.234] (II) Unloading intel
so the intel driver shouldn't be unloaded? Oo
Strange that xserver is even trying to load intel, it shouldn't be loaded at all.
It is unloaded (line 130) and loaded again (line 159).
Maybe you've got some conf file in /etc/X11/xorg.conf.d dir, like 20-intel.conf?
Or maybe intel needs to be blacklisted in xorg.conf (i dunno if it's even possible)?
Aarin wrote:
After all this I decided to try to enter with the intel graphic card, so I've done a "pxp_switch_catalyst intel" (which had no errors) and I rebooted. This time xorg didn't loaded fglrx, but Xorg says that there isn't no screen available to the card (perhaps because pxp_switch_catalyst "deletes" xorg.conf when switching to the intel card?)
Script do not deletes any xorg.conf, it's renaming it into:
- xorg.conf.cat if it will find fglrx (it will be linked later to xorg.conf);
- xorg.conf.oth if it will find intel or radeon (will be linked later);
- xorg.conf.wth if it will not find fglrx, intel or radeon (won't be linked or used, it's backup made because of unperfection of the script and my poor knowledge about intel).
So your xorg.conf for intel do not have "intel" word inside? If so and you've got ie. "i915" let me know, i will update script.
BTW: still it should load intel succesfully, even if xorg.conf is not present - at least it's working for radeon.

Well for the first time in a while I finally got to enter X with success. But with the intel driver...
To put it to work I did:

- removed the i915 driver from the blacklist
- removed the "nomodeset" from the kernel line
- done a "pxp_switch_catalyst intel"

and it started without any problems.

with a "pxp_switch_catalyst amd" after and a Xserver restart gave the same error from yesterday
I do not have any file regarding intel or i915 in xorg.conf.d or anywhere else.

and that I know of, Its not possible to blacklist a driver in Xserver. But even if it was, it wouldn't be a very good solution, since I wouldn't be able to switch back to the intel driver when I'd like to.
why Is xserver pulling the intel driver again? A mistery indeed...

Vi0L0 · 2011-06-16 16:15:48

Aarin wrote:

and that I know of, Its not possible to blacklist a driver in Xserver.

Yes, a quick search return me no easy way of doing this...

Aarin wrote:

But even if it was, it wouldn't be a very good solution, since I wouldn't be able to switch back to the intel driver when I'd like to.

Actually we could throw it into pxp_switch script so it would be easy, but since "its not possible to blacklist a driver in Xserver" it's a no go.

Aarin wrote:

why Is xserver pulling the intel driver again? A mistery indeed...

I'm out of ideas
Well maybe except the one that you forgot to remove intel or i915 from MODULES array, but guess you didn't ;-)

I've checked pxp impementations in other distros, but found nothing interesting.
Later i will look at intel package, maybe there's something in it.

hal8000 · 2011-06-16 22:32:02

I hope this is the correct thread, but I cannot install the latest Catalyst 11.6 drivers.

I have an ATI HD5770 card previously working with catalyst 11.5 and 2.6.38.2 kernel.
Kernel is now the stock kernel 2.6.39.1 and have downloaded catalyst-utils-11.6-1
from the AUR.

Problem is lingl is in conflict and I cant remove it

[anc@orac catalyst-utils]$ sudo pacman -U catalyst-utils-11.6-1-i686.pkg.tar.xz
Password:
resolving dependencies...
looking for inter-conflicts...
:: catalyst-utils and libgl are in conflict. Remove libgl? [y/N] y

Remove (1): libgl-7.10.2-2

Total Removed Size: 3.66 MB

Targets (1): catalyst-utils-11.6-1

Total Download Size: 0.00 MB
Total Installed Size: 73.91 MB

Proceed with installation? [Y/n]
(1/1) checking package integrity [#####################################] 100%
(1/1) checking for file conflicts [#####################################] 100%
error: failed to commit transaction (conflicting files)
catalyst-utils: /etc/ati/amdpcsdb.default exists in filesystem
--snip--

When I reboot, I think I am using KMS settings as no fglrx module is loaded and the module isnt present for this
kernel:

[root@orac catalyst-utils]# modinfo fglrx
ERROR: modinfo: could not find module fglrx
[root@orac catalyst-utils]# modprobe fglrx
FATAL: Module fglrx not found.

If I try and remove libgl

[root@orac catalyst-utils]# pacman -Q libgl
libgl 7.10.2-2
[root@orac catalyst-utils]# pacman -R libgl
checking dependencies...
error: failed to prepare transaction (could not satisfy dependencies)
:: gnome-session: requires libgl
:: libreoffice: requires libgl
:: libva: requires libgl
:: mesa: requires libgl
:: mplayer: requires libgl
:: qt: requires libgl
:: wxgtk: requires libgl
:: xine-lib: requires libgl
:: xorg-xdriinfo: requires libgl

Not sure what to try next, thanks in advance for help

Vi0L0 · 2011-06-16 23:32:20

@hal8000:
it's all in the wiki: https://wiki.archlinux.org/index.php/ATI_Catalyst
pacman -Rdd libgl
+ don't forget to install catalyst package or one of it's substitutes: catalyst-daemon, catalyst-generator or catalyst-hook
+ add nomodeset into kernel options

afaik [catalyst] repo is working fine if you want to download less

Aarin · 2011-06-17 00:51:00

@Vi0L0

I've discovered a thing or two.

first, apparently nomodeset doens't affect intel graphic cards. Said here: Intel - Arch Wiki

"i915.modeset=0" is the intel equivalent to "nomodeset" for other video cards.

But its also there that:

KMS is required in order to run X (Gnome, KDE, etc).

And indeed gdm does not open with Intel with the KMS off,

Also, when i915.modeset=0 was put in the kernel in order that KMS doesn't start, the intel drv was not loaded like before when starting gdm. But the segfault is still there.

Here's the Xorg.0.log

[    43.483] 
X.Org X Server 1.10.2
Release Date: 2011-05-28
[    43.483] X Protocol Version 11, Revision 0
[    43.483] Build Operating System: Linux 2.6.38-ARCH x86_64 
[    43.483] Current Operating System: Linux HP 2.6.39-ARCH #1 SMP PREEMPT Mon Jun 6 22:37:55 CEST 2011 x86_64
[    43.483] Kernel command line: initrd=/boot/kernel26.img root=/dev/disk/by-uuid/92375f13-aaa6-4a24-a515-10df921c1945 rootfstype=ext4 i915.modeset=0 BOOT_IMAGE=/boot/vmlinuz26 
[    43.483] Build Date: 30 May 2011  08:47:00AM
[    43.483]  
[    43.483] Current version of pixman: 0.22.0
[    43.483]     Before reporting problems, check http://wiki.x.org
    to make sure that you have the latest version.
[    43.483] Markers: (--) probed, (**) from config file, (==) default setting,
    (++) from command line, (!!) notice, (II) informational,
    (WW) warning, (EE) error, (NI) not implemented, (??) unknown.
[    43.483] (==) Log file: "/var/log/Xorg.0.log", Time: Fri Jun 17 01:22:17 2011
[    43.525] (==) Using config file: "/etc/X11/xorg.conf"
[    43.525] (==) Using config directory: "/etc/X11/xorg.conf.d"
[    43.584] (==) ServerLayout "aticonfig Layout"
[    43.584] (**) |-->Screen "aticonfig-Screen[0]-0" (0)
[    43.584] (**) |   |-->Monitor "aticonfig-Monitor[0]-0"
[    43.584] (**) |   |-->Device "aticonfig-Device[0]-0"
[    43.584] (==) Automatically adding devices
[    43.584] (==) Automatically enabling devices
[    43.622] (WW) The directory "/usr/share/fonts/OTF/" does not exist.
[    43.622]     Entry deleted from font path.
[    43.686] (==) FontPath set to:
    /usr/share/fonts/misc/,
    /usr/share/fonts/TTF/,
    /usr/share/fonts/Type1/,
    /usr/share/fonts/100dpi/,
    /usr/share/fonts/75dpi/
[    43.686] (**) ModulePath set to "/usr/lib/xorg/modules/updates,/usr/lib/xorg/modules"
[    43.686] (II) The server relies on udev to provide the list of input devices.
    If no devices become available, reconfigure udev or disable AutoAddDevices.
[    43.686] (II) Loader magic: 0x7d3440
[    43.686] (II) Module ABI versions:
[    43.686]     X.Org ANSI C Emulation: 0.4
[    43.686]     X.Org Video Driver: 10.0
[    43.686]     X.Org XInput driver : 12.2
[    43.686]     X.Org Server Extension : 5.0
[    43.687] (--) PCI:*(0:0:2:0) 8086:0046:103c:144a rev 2, Mem @ 0xc0000000/4194304, 0xb0000000/268435456, I/O @ 0x00005050/8
[    43.687] (--) PCI: (0:1:0:0) 1002:68c1:103c:144a rev 0, Mem @ 0xa0000000/268435456, 0xc4400000/131072, I/O @ 0x00004000/256, BIOS @ 0x????????/131072
[    43.688] (II) Open ACPI successful (/var/run/acpid.socket)
[    43.688] (II) "extmod" will be loaded by default.
[    43.688] (II) "dbe" will be loaded by default.
[    43.688] (II) "glx" will be loaded by default.
[    43.688] (II) "record" will be loaded by default.
[    43.688] (II) "dri" will be loaded by default.
[    43.688] (II) "dri2" will be loaded by default.
[    43.688] (II) LoadModule: "extmod"
[    43.812] (II) Loading /usr/lib/xorg/modules/extensions/libextmod.so
[    43.820] (II) Module extmod: vendor="X.Org Foundation"
[    43.820]     compiled for 1.10.2, module version = 1.0.0
[    43.820]     Module class: X.Org Server Extension
[    43.820]     ABI class: X.Org Server Extension, version 5.0
[    43.820] (II) Loading extension MIT-SCREEN-SAVER
[    43.820] (II) Loading extension XFree86-VidModeExtension
[    43.820] (II) Loading extension XFree86-DGA
[    43.820] (II) Loading extension DPMS
[    43.820] (II) Loading extension XVideo
[    43.820] (II) Loading extension XVideo-MotionCompensation
[    43.820] (II) Loading extension X-Resource
[    43.820] (II) LoadModule: "dbe"
[    43.820] (II) Loading /usr/lib/xorg/modules/extensions/libdbe.so
[    43.821] (II) Module dbe: vendor="X.Org Foundation"
[    43.821]     compiled for 1.10.2, module version = 1.0.0
[    43.821]     Module class: X.Org Server Extension
[    43.821]     ABI class: X.Org Server Extension, version 5.0
[    43.821] (II) Loading extension DOUBLE-BUFFER
[    43.821] (II) LoadModule: "glx"
[    43.821] (II) Loading /usr/lib/xorg/modules/updates/extensions/libglx.so
[    43.848] (II) Module glx: vendor="FireGL - ATI Technologies Inc."
[    43.849]     compiled for 6.9.0, module version = 1.0.0
[    43.849] (II) Loading extension GLX
[    43.849] (II) LoadModule: "record"
[    43.850] (II) Loading /usr/lib/xorg/modules/extensions/librecord.so
[    43.861] (II) Module record: vendor="X.Org Foundation"
[    43.861]     compiled for 1.10.2, module version = 1.13.0
[    43.861]     Module class: X.Org Server Extension
[    43.861]     ABI class: X.Org Server Extension, version 5.0
[    43.861] (II) Loading extension RECORD
[    43.861] (II) LoadModule: "dri"
[    43.861] (II) Loading /usr/lib/xorg/modules/extensions/libdri.so
[    43.870] (II) Module dri: vendor="X.Org Foundation"
[    43.870]     compiled for 1.10.2, module version = 1.0.0
[    43.870]     ABI class: X.Org Server Extension, version 5.0
[    43.870] (II) Loading extension XFree86-DRI
[    43.870] (II) LoadModule: "dri2"
[    43.871] (II) Loading /usr/lib/xorg/modules/extensions/libdri2.so
[    43.871] (II) Module dri2: vendor="X.Org Foundation"
[    43.871]     compiled for 1.10.2, module version = 1.2.0
[    43.871]     ABI class: X.Org Server Extension, version 5.0
[    43.871] (II) Loading extension DRI2
[    43.871] (II) LoadModule: "fglrx"
[    43.872] (II) Loading /usr/lib/xorg/modules/drivers/fglrx_drv.so
[    44.139] (II) Module fglrx: vendor="FireGL - ATI Technologies Inc."
[    44.152]     compiled for 1.4.99.906, module version = 8.86.5
[    44.152]     Module class: X.Org Video Driver
[    44.153] (II) Loading sub module "fglrxdrm"
[    44.153] (II) LoadModule: "fglrxdrm"
[    44.153] (II) Loading /usr/lib/xorg/modules/linux/libfglrxdrm.so
[    44.180] (II) Module fglrxdrm: vendor="FireGL - ATI Technologies Inc."
[    44.180]     compiled for 1.4.99.906, module version = 8.86.5
[    44.180] (II) ATI Proprietary Linux Driver Version Identifier:8.86.5
[    44.180] (II) ATI Proprietary Linux Driver Release Identifier: 8.861                                
[    44.180] (II) ATI Proprietary Linux Driver Build Date: May 24 2011 22:46:53
[    44.180] (++) using VT number 7

[    44.182] (WW) Falling back to old probe method for fglrx
[    44.226] (II) Loading PCS database from /etc/ati/amdpcsdb
[    44.248] (--) Chipset Supported AMD Graphics Processor (0x68C1) found
[    44.249] (WW) fglrx: No matching Device section for instance (BusID PCI:0@1:0:1) found
[    44.249] (II) fglrx: intel VGA device detected, load intel driver.
[    44.249] (II) LoadModule: "intel"
[    44.249] (II) Loading /usr/lib/xorg/modules/drivers/intel_drv.so
[    44.279] (II) Module intel: vendor="X.Org Foundation"
[    44.279]     compiled for 1.10.2, module version = 2.15.0
[    44.279]     Module class: X.Org Video Driver
[    44.279]     ABI class: X.Org Video Driver, version 10.0
[    44.280] ukiDynamicMajor: found major device number 251
[    44.280] ukiDynamicMajor: found major device number 251
[    44.280] ukiOpenByBusid: Searching for BusID PCI:1:0:0
[    44.280] ukiOpenDevice: node name is /dev/ati/card0
[    44.280] ukiOpenDevice: open result is 8, (OK)
[    44.280] ukiOpenByBusid: ukiOpenMinor returns 8
[    44.280] ukiOpenByBusid: ukiGetBusid reports PCI:1:0:0
[    44.293] (II) AMD Video driver is running on a device belonging to a group targeted for this release
[    44.315] (II) AMD Video driver is signed
[    44.315] (II) Loading /usr/lib/xorg/modules/drivers/fglrx_drv.so
[    44.315] (II) Loading /usr/lib/xorg/modules/linux/libfglrxdrm.so
[    44.315] (II) fglrx(0): pEnt->device->identifier=0x17a2330
[    44.318] (II) pEnt->device->identifier=(nil)
[    44.318] (II) fglrx(0): === [xdl_xs110_atiddxPreInit] === begin
[    44.318] (II) fglrx(0): PowerXpress: Discrete GPU is selected.
[    44.412] (II) fglrx(0): PowerXpress: Diagnostic output from /usr/lib64/fglrx/switchlibGL:
Switch to AMD

[    44.420] 
Backtrace:
[    44.448] 0: /usr/bin/Xorg (xorg_backtrace+0x26) [0x49f546]
[    44.448] 1: /usr/bin/Xorg (0x400000+0x60cf9) [0x460cf9]
[    44.448] 2: /lib/libpthread.so.0 (0x7ffb5e442000+0xf7e0) [0x7ffb5e4517e0]
[    44.449] 3: /usr/lib/xorg/modules/drivers/fglrx_drv.so (xdl_xs110_atiddxPxPreInit+0x17a) [0x7ffb5b05a33a]
[    44.449] 4: /usr/lib/xorg/modules/drivers/fglrx_drv.so (xdl_xs110_atiddxPreInit+0x213e) [0x7ffb5b030aae]
[    44.449] 5: /usr/bin/Xorg (InitOutput+0x834) [0x46ffa4]
[    44.449] 6: /usr/bin/Xorg (0x400000+0x229c3) [0x4229c3]
[    44.449] 7: /lib/libc.so.6 (__libc_start_main+0xed) [0x7ffb5d3c017d]
[    44.449] 8: /usr/bin/Xorg (0x400000+0x22e6d) [0x422e6d]
[    44.449] Segmentation fault at address 0x10
[    44.449] 
Fatal server error:
[    44.449] Caught signal 11 (Segmentation fault). Server aborting
[    44.449] 
[    44.449] 
Please consult the The X.Org Foundation support 
     at http://wiki.x.org
 for help. 
[    44.449] Please also check the log file at "/var/log/Xorg.0.log" for additional information.
[    44.449]

P3808 · 2011-06-17 01:47:52

sumski wrote:

sumski wrote:
It seems catalyst 11.6 borked HDMI output, but "only" with X. tty is fine sad
Can anyone reproduce this?

same here... HDMI-cable isn't detected

sumski · 2011-06-17 04:43:54

You can switch to tty2 ,log in , plug in the cable , and go back into X (tty7), and cable get's recognized. Really stupid bug

Last edited by sumski (2011-06-17 04:44:12)

Cdh · 2011-06-17 09:22:59

hal8000 wrote:

(1/1) checking package integrity [#####################################] 100%
(1/1) checking for file conflicts [#####################################] 100%
error: failed to commit transaction (conflicting files)
catalyst-utils: /etc/ati/amdpcsdb.default exists in filesystem

This is your problem: There are some files in your filesystem that the catalyst-utils package wants to overwrite. That is disabled by default.

But /etc/ati certainly belongs to fglrx driver. Maybe you have installed it before without using the package manager?

You can safely remove the whole /etc/ati directory when you are installing catalyst-utils afterwards since it only holds configuration relevant to fglrx.
So

sudo rm -r /etc/ati
sudo pacman -S catalyst-utils

Ochi · 2011-06-17 10:38:54

I'm trying to get catalyst to work on my Vostro 3350 with a Radeon 6470M + Intel (i5 Sandy Bridge) hybrid configuration. First off, aticonfig tells me (even with 11.6) that the card is unsupported. fglrx still seems to detect the card... somehow. The Xorg log reads

"AMD Video driver is running on a device generally supported by this driver".

I've learned that this machine doesn't seem to have a muxer but that the Radeon is attached to the Intel. Is that correct? Maybe this explains why fglrx also tries to load the Intel driver? In the Xorg logs, the driver says quite unambiguously that it detected the Radeon as discrete and the Intel as integrated graphics and that it initiates the loading of the intel driver:

...
[    42.243] (--) Chipset Supported AMD Graphics Processor (0x6760) found
[    42.260] (II) fglrx: intel VGA device detected, load intel driver.
[    42.260] (II) LoadModule: "intel"
[    42.261] (II) Loading /usr/lib/xorg/modules/drivers/intel_drv.so
...

My guess is that it needs the intel driver to output any video with the Radeon at all, namely through the intel (which does not work, see below, but that's my guess).

However, I'm experiencing the same segmentation fault as Aarin right after the "Switch to AMD" message:

[    42.326] (II) fglrx(0): PowerXpress: Discrete GPU is selected.
[    43.066] (II) fglrx(0): PowerXpress: Diagnostic output from /usr/lib64/fglr$
Switch to AMD

[    43.075]
Backtrace:
[    43.121] 0: /usr/bin/X (xorg_backtrace+0x26) [0x49f546]
[    43.121] 1: /usr/bin/X (0x400000+0x60cf9) [0x460cf9]
[    43.121] 2: /lib/libpthread.so.0 (0x7f344d1c2000+0xf7e0) [0x7f344d1d17e0]
[    43.121] 3: /usr/lib/xorg/modules/drivers/fglrx_drv.so (xdl_xs110_atiddxPxP$
[    43.122] 4: /usr/lib/xorg/modules/drivers/fglrx_drv.so (xdl_xs110_atiddxPre$
[    43.122] 5: /usr/bin/X (InitOutput+0x834) [0x46ffa4]
[    43.122] 6: /usr/bin/X (0x400000+0x229c3) [0x4229c3]
[    43.122] 7: /lib/libc.so.6 (__libc_start_main+0xed) [0x7f344c14017d]
[    43.122] 8: /usr/bin/X (0x400000+0x22e6d) [0x422e6d]
[    43.122] Segmentation fault at address 0x10
[    43.122]
Fatal server error:
[    43.122] Caught signal 11 (Segmentation fault). Server aborting

I'm using "nomodeset". If I don't do this, I'm getting a "Screen 1 deleted because of no matching config section" kind of error and a different segmentation fault.

Last edited by Ochi (2011-06-17 10:47:34)

P3808 · 2011-06-17 15:28:41

sumski wrote:

You can switch to tty2 ,log in , plug in the cable , and go back into X (tty7), and cable get's recognized. Really stupid bug

thx, I already read that above... I just wanted you to know that others have this bug too

sumski · 2011-06-17 15:46:50

Yeah , i should really report that. Is anyone from AMD/ATI reading unofficial bugzilla?

Vi0L0 · 2011-06-17 16:11:45

sumski wrote:

Yeah , i should really report that. Is anyone from AMD/ATI reading unofficial bugzilla?

Yes, in fact they are sometimes answering (or requesting more info) in comments.

@Aarin & @Ochi:

Ochi wrote:

My guess is that it needs the intel driver to output any video with the Radeon at all, namely through the intel (which does not work, see below, but that's my guess).

I guess you could be right.
I'm still concerned how is it possible to switch between intel/catalyst without rebooting (i heard powerXpress switch should work this way) because of 2 reasons:
- catalyst cannot work with KMS;
- intel cannot work without KMS.
I don't see any mistake in my pXp implementation (especialy when it's basing on SUSE, and it's doing same things like original ATI sripts do), everything should work like ATI want it to work, for me it's looking like upstream bug so i will try to talk with ATI (i've started at #ati @ freenode).

Arch Linux

#3051 2011-06-15 17:00:03

Re: The AMD/ATI Bar & Grill

#3052 2011-06-15 18:06:26

Re: The AMD/ATI Bar & Grill

#3053 2011-06-15 18:39:21

Re: The AMD/ATI Bar & Grill

#3054 2011-06-15 18:54:26

Re: The AMD/ATI Bar & Grill

#3055 2011-06-15 19:11:57

Re: The AMD/ATI Bar & Grill

#3056 2011-06-15 20:22:44

Re: The AMD/ATI Bar & Grill

#3057 2011-06-15 20:39:02

Re: The AMD/ATI Bar & Grill

#3058 2011-06-16 02:07:22

Re: The AMD/ATI Bar & Grill

#3059 2011-06-16 05:49:14

Re: The AMD/ATI Bar & Grill

#3060 2011-06-16 08:09:14

Re: The AMD/ATI Bar & Grill

#3061 2011-06-16 09:41:43

Re: The AMD/ATI Bar & Grill

#3062 2011-06-16 10:40:37

Re: The AMD/ATI Bar & Grill

#3063 2011-06-16 10:50:46

Re: The AMD/ATI Bar & Grill

#3064 2011-06-16 11:57:35

Re: The AMD/ATI Bar & Grill

#3065 2011-06-16 16:15:48

Re: The AMD/ATI Bar & Grill

#3066 2011-06-16 22:32:02

Re: The AMD/ATI Bar & Grill

#3067 2011-06-16 23:32:20

Re: The AMD/ATI Bar & Grill

#3068 2011-06-17 00:51:00

Re: The AMD/ATI Bar & Grill

#3069 2011-06-17 01:47:52

Re: The AMD/ATI Bar & Grill

#3070 2011-06-17 04:43:54

Re: The AMD/ATI Bar & Grill

#3071 2011-06-17 09:22:59

Re: The AMD/ATI Bar & Grill

#3072 2011-06-17 10:38:54

Re: The AMD/ATI Bar & Grill

#3073 2011-06-17 15:28:41

Re: The AMD/ATI Bar & Grill

#3074 2011-06-17 15:46:50

Re: The AMD/ATI Bar & Grill

#3075 2011-06-17 16:11:45

Re: The AMD/ATI Bar & Grill

Board footer