The AMD/ATI Bar & Grill

axed · 2011-08-08 10:21:01

Does already someone know, why X crashes all the time when loading fglrx?
Configuration: Hp Envy 14 Switchable Intel/Ati 5650, catalyst-total package from aur, xorg-server 1.10

[    43.197] 
X.Org X Server 1.10.3
Release Date: 2011-07-08
[    43.231] X Protocol Version 11, Revision 0
[    43.231] Build Operating System: Linux 2.6.39-ARCH x86_64 
[    43.231] Current Operating System: Linux bla 3.0-ARCH #1 SMP PREEMPT Sat Aug 6 16:18:35 CEST 2011 x86_64
[    43.231] Kernel command line: cryptdevice=/dev/sda4:root root=/dev/mapper/root ro nomodeset
[    43.231] Build Date: 09 July 2011  08:04:37AM
[    43.231]  
[    43.231] Current version of pixman: 0.22.2
[    43.231]     Before reporting problems, check http://wiki.x.org
    to make sure that you have the latest version.
[    43.231] Markers: (--) probed, (**) from config file, (==) default setting,
    (++) from command line, (!!) notice, (II) informational,
    (WW) warning, (EE) error, (NI) not implemented, (??) unknown.
[    43.232] (==) Log file: "/var/log/Xorg.0.log", Time: Mon Aug  8 12:07:19 2011
[    43.312] (==) Using config directory: "/etc/X11/xorg.conf.d"
[    43.345] (==) ServerLayout "aticonfig Layout"
[    43.345] (**) |-->Screen "aticonfig-Screen[0]-0" (0)
[    43.345] (**) |   |-->Monitor "aticonfig-Monitor[0]-0"
[    43.345] (**) |   |-->Device "aticonfig-Device[0]-0"
[    43.345] (==) Automatically adding devices
[    43.345] (==) Automatically enabling devices
[    43.376] (WW) The directory "/usr/share/fonts/OTF/" does not exist.
[    43.376]     Entry deleted from font path.
[    43.385] (WW) `fonts.dir' not found (or not valid) in "/usr/share/fonts/100dpi/".
[    43.385]     Entry deleted from font path.
[    43.385]     (Run 'mkfontdir' on "/usr/share/fonts/100dpi/").
[    43.385] (WW) `fonts.dir' not found (or not valid) in "/usr/share/fonts/75dpi/".
[    43.385]     Entry deleted from font path.
[    43.385]     (Run 'mkfontdir' on "/usr/share/fonts/75dpi/").
[    43.385] (==) FontPath set to:
    /usr/share/fonts/misc/,
    /usr/share/fonts/TTF/,
    /usr/share/fonts/Type1/
[    43.385] (==) ModulePath set to "/usr/lib/xorg/modules"
[    43.385] (II) The server relies on udev to provide the list of input devices.
    If no devices become available, reconfigure udev or disable AutoAddDevices.
[    43.385] (II) Loader magic: 0x7d3440
[    43.385] (II) Module ABI versions:
[    43.385]     X.Org ANSI C Emulation: 0.4
[    43.385]     X.Org Video Driver: 10.0
[    43.385]     X.Org XInput driver : 12.2
[    43.385]     X.Org Server Extension : 5.0
[    43.386] (--) PCI:*(0:0:2:0) 8086:0046:103c:1436 rev 2, Mem @ 0xc0000000/4194304, 0xb0000000/268435456, I/O @ 0x00004050/8
[    43.386] (--) PCI: (0:1:0:0) 1002:68c1:103c:1436 rev 0, Mem @ 0xa0000000/268435456, 0xc4400000/131072, I/O @ 0x00003000/256, BIOS @ 0x????????/131072
[    43.386] (II) Open ACPI successful (/var/run/acpid.socket)
[    43.386] (II) "extmod" will be loaded by default.
[    43.386] (II) "dbe" will be loaded by default.
[    43.386] (II) "glx" will be loaded by default.
[    43.386] (II) "record" will be loaded by default.
[    43.386] (II) "dri" will be loaded by default.
[    43.386] (II) "dri2" will be loaded by default.
[    43.386] (II) LoadModule: "extmod"
[    43.436] (II) Loading /usr/lib/xorg/modules/extensions/libextmod.so
[    43.449] (II) Module extmod: vendor="X.Org Foundation"
[    43.449]     compiled for 1.10.3, module version = 1.0.0
[    43.449]     Module class: X.Org Server Extension
[    43.449]     ABI class: X.Org Server Extension, version 5.0
[    43.449] (II) Loading extension MIT-SCREEN-SAVER
[    43.449] (II) Loading extension XFree86-VidModeExtension
[    43.449] (II) Loading extension XFree86-DGA
[    43.449] (II) Loading extension DPMS
[    43.449] (II) Loading extension XVideo
[    43.449] (II) Loading extension XVideo-MotionCompensation
[    43.449] (II) Loading extension X-Resource
[    43.449] (II) LoadModule: "dbe"
[    43.449] (II) Loading /usr/lib/xorg/modules/extensions/libdbe.so
[    43.450] (II) Module dbe: vendor="X.Org Foundation"
[    43.450]     compiled for 1.10.3, module version = 1.0.0
[    43.450]     Module class: X.Org Server Extension
[    43.450]     ABI class: X.Org Server Extension, version 5.0
[    43.450] (II) Loading extension DOUBLE-BUFFER
[    43.450] (II) LoadModule: "glx"
[    43.450] (II) Loading /usr/lib/xorg/modules/extensions/libglx.so
[    43.471] (II) Module glx: vendor="X.Org Foundation"
[    43.471]     compiled for 1.10.3, module version = 1.0.0
[    43.471]     ABI class: X.Org Server Extension, version 5.0
[    43.471] (==) AIGLX enabled
[    43.472] (II) Loading extension GLX
[    43.472] (II) LoadModule: "record"
[    43.472] (II) Loading /usr/lib/xorg/modules/extensions/librecord.so
[    43.474] (II) Module record: vendor="X.Org Foundation"
[    43.474]     compiled for 1.10.3, module version = 1.13.0
[    43.474]     Module class: X.Org Server Extension
[    43.474]     ABI class: X.Org Server Extension, version 5.0
[    43.474] (II) Loading extension RECORD
[    43.474] (II) LoadModule: "dri"
[    43.474] (II) Loading /usr/lib/xorg/modules/extensions/libdri.so
[    43.495] (II) Module dri: vendor="X.Org Foundation"
[    43.495]     compiled for 1.10.3, module version = 1.0.0
[    43.495]     ABI class: X.Org Server Extension, version 5.0
[    43.495] (II) Loading extension XFree86-DRI
[    43.495] (II) LoadModule: "dri2"
[    43.495] (II) Loading /usr/lib/xorg/modules/extensions/libdri2.so
[    43.504] (II) Module dri2: vendor="X.Org Foundation"
[    43.504]     compiled for 1.10.3, module version = 1.2.0
[    43.504]     ABI class: X.Org Server Extension, version 5.0
[    43.504] (II) Loading extension DRI2
[    43.504] (II) LoadModule: "fglrx"
[    43.514] (II) Loading /usr/lib/xorg/modules/drivers/fglrx_drv.so
[    43.716] (II) Module fglrx: vendor="FireGL - ATI Technologies Inc."
[    43.729]     compiled for 1.4.99.906, module version = 8.87.5
[    43.729]     Module class: X.Org Video Driver
[    43.731] (II) Loading sub module "fglrxdrm"
[    43.731] (II) LoadModule: "fglrxdrm"
[    43.731] (II) Loading /usr/lib/xorg/modules/linux/libfglrxdrm.so
[    43.761] (II) Module fglrxdrm: vendor="FireGL - ATI Technologies Inc."
[    43.761]     compiled for 1.4.99.906, module version = 8.87.5
[    43.761] (II) ATI Proprietary Linux Driver Version Identifier:8.87.5
[    43.761] (II) ATI Proprietary Linux Driver Release Identifier: 8.872                                
[    43.761] (II) ATI Proprietary Linux Driver Build Date: Jul  7 2011 22:42:39
[    43.761] (++) using VT number 7

[    43.763] (WW) Falling back to old probe method for fglrx
[    43.875] (II) Loading PCS database from /etc/ati/amdpcsdb
[    43.886] (--) Chipset Supported AMD Graphics Processor (0x68C1) found
[    43.889] (WW) fglrx: No matching Device section for instance (BusID PCI:0@1:0:1) found
[    43.889] (II) fglrx: intel VGA device detected, load intel driver.
[    43.889] (II) LoadModule: "intel"
[    43.889] (II) Loading /usr/lib/xorg/modules/drivers/intel_drv.so
[    43.927] (II) Module intel: vendor="X.Org Foundation"
[    43.927]     compiled for 1.10.2, module version = 2.15.0
[    43.927]     Module class: X.Org Video Driver
[    43.927]     ABI class: X.Org Video Driver, version 10.0
[    43.929] ukiDynamicMajor: found major device number 251
[    43.929] ukiDynamicMajor: found major device number 251
[    43.929] ukiOpenByBusid: Searching for BusID PCI:1:0:0
[    43.929] ukiOpenDevice: node name is /dev/ati/card0
[    43.929] ukiOpenDevice: open result is 8, (OK)
[    43.929] ukiOpenByBusid: ukiOpenMinor returns 8
[    43.929] ukiOpenByBusid: ukiGetBusid reports PCI:1:0:0
[    43.932] (II) AMD Video driver is running on a device belonging to a group targeted for this release
[    43.951] (II) AMD Video driver is signed
[    43.952] (II) Loading /usr/lib/xorg/modules/drivers/fglrx_drv.so
[    43.952] (II) Loading /usr/lib/xorg/modules/linux/libfglrxdrm.so
[    43.952] (II) fglrx(0): pEnt->device->identifier=0x18502d0
[    43.954] (II) pEnt->device->identifier=(nil)
[    43.954] (II) fglrx(0): === [xdl_xs110_atiddxPreInit] === begin
[    43.954] (II) fglrx(0): PowerXpress: Integrated GPU is selected.
[    44.835] (II) fglrx(0): PowerXpress: Diagnostic output from /usr/lib64/fglrx/switchlibGL:
Switch to Intel

[    44.842] 
Backtrace:
[    44.884] 0: /usr/bin/X (xorg_backtrace+0x26) [0x45d446]
[    44.884] 1: /usr/bin/X (0x400000+0x61869) [0x461869]
[    44.884] 2: /lib/libpthread.so.0 (0x7fd7cc214000+0xf7e0) [0x7fd7cc2237e0]
[    44.884] 3: /usr/lib/xorg/modules/drivers/fglrx_drv.so (xdl_xs110_atiddxPxPreInit+0x17a) [0x7fd7c8b3b5fa]
[    44.884] 4: /usr/lib/xorg/modules/drivers/fglrx_drv.so (xdl_xs110_atiddxPreInit+0x20fe) [0x7fd7c8b11b2e]
[    44.884] 5: /usr/bin/X (InitOutput+0x834) [0x470af4]
[    44.884] 6: /usr/bin/X (0x400000+0x229e3) [0x4229e3]
[    44.884] 7: /lib/libc.so.6 (__libc_start_main+0xed) [0x7fd7cb19217d]
[    44.884] 8: /usr/bin/X (0x400000+0x22e8d) [0x422e8d]
[    44.884] Segmentation fault at address 0x10
[    44.884] 
Fatal server error:
[    44.884] Caught signal 11 (Segmentation fault). Server aborting
[    44.884] 
[    44.884] 
Please consult the The X.Org Foundation support 
     at http://wiki.x.org
 for help. 
[    44.884] Please also check the log file at "/var/log/Xorg.0.log" for additional information.
[    44.884]

First, fglrx complained in the log file that he couldn't find the intel firmware for switchin', so I installed xf86-video-intel and now he stopped complaining about that. ( see [ 43.889] (II) fglrx: intel VGA device detected, load intel driver. )
I will try the remove the intel stuff again and use the powerful card atm only, if it is that what makes fglrx lucky ...

ChemBro · 2011-08-08 10:59:24

Vi0L0 wrote:

linux 3.0 is in [core] repo
more info: http://www.archlinux.org/news/changes-t … filenames/
catalyst package has been updated in both repo and AUR.

Did you forget something? :-)

kernel26-headers --> linux-headers

Please update. I can't reboot at the moment, because I updated the kernel (linux 3.0), but building the modules failed.

EDIT: or I switch to radeon for the time inbetween

Last edited by ChemBro (2011-08-08 11:04:25)

Vi0L0 · 2011-08-08 11:22:59

@axed:
Yes, powerXpress support doesn't work atm, but i'm keeping it in -total because i still got some hope that someone will figure it out ;-).
I also see other things in Arch that will make such switching support not that shiny - like hardcoded paths into /usr/lib/libgl lib in virtualbox and GNOME3.
It's not that bad with virtualbox - it will at least work - you will have to install libgl package so it may use it (i dunno if acceleration will work here) (or make a symlink from catalystpxp's libgl), GNOME3 is a bigger problem - if libgl is installed - it will use it, no matter that catalystpxp's libgl is in LD_LIBRARY_PATH.

Anyway - if you want to work with catalyst-total on machine with intel's gpu you will have to turn intel's card off in bios xD...
I really need to get some free time and work on this pXp... also maybe someone from Łódź City can borrow me a laptop with intel and radeon inside? :-)

Last edited by Vi0L0 (2011-08-08 11:27:40)

Vi0L0 · 2011-08-08 11:26:53

ChemBro wrote:

Vi0L0 wrote:
linux 3.0 is in [core] repo
more info: http://www.archlinux.org/news/changes-t … filenames/
catalyst package has been updated in both repo and AUR.
Did you forget something? :-)
kernel26-headers --> linux-headers
Please update. I can't reboot at the moment, because I updated the kernel (linux 3.0), but building the modules failed.
EDIT: or I switch to radeon for the time inbetween

I did it for catalyst package.
I had no time yesterday to update catalyst-total, i will update and look at it after i get home (~4h).

axed · 2011-08-08 12:16:19

Vi0L0 wrote:

@axed:
Yes, powerXpress support doesn't work atm, ...

I guess the usual it's-from-amd-so-it's-working-but-not-for-you procedure.. meh

Vi0L0 wrote:

Anyway - if you want to work with catalyst-total on machine with intel's gpu you will have to turn intel's card off in bios xD...

Always teh best jokez on teh internetz

Well, I'm looking forward to your success B)
*went meanwhile back to vgaswitcheroo*

ChemBro · 2011-08-08 19:08:18

Thanks for the update. Works now.

yell! · 2011-08-10 12:33:17

one question about catalyst-total and pxp support:
powerexpress is just a script to switch from intel integrated to ati dedicated video chipset?

My laptop has Fusion technology.
I want to switch from integrated Ati 4250 to dedicated video chipset Ati 5470 but through AMDCCLE is not possible (I can select the adapter but it doesn't work, in real time)
Is there support for fusion technology or I made a mistake buying and AMD laptop?

agapito · 2011-08-11 06:18:06

AMD Catalyst™ Preview Driver – OpenGL 4.2 Beta Support for Linux® Based Systems

http://support.amd.com/us/kbarticles/Pa … Linux.aspx

Vi0L0 · 2011-08-12 16:30:56

agapito wrote:

AMD Catalyst™ Preview Driver – OpenGL 4.2 Beta Support for Linux® Based Systems
http://support.amd.com/us/kbarticles/Pa … Linux.aspx

It's versioned as 8.88.8 - so i guess it's catalyst 11.8 with OpenGL 4.2 support.

I'm testing it atm and it doesn't look good.

Still got hd 4850 here and i'm experiencing rly nasty behaviour with KDE:
- when minimizing/maximizing window with button or with AeroSnap (i dunno how's it called in KDE) Xs often are crashing and restarting into kdm's login screen;
- hd movies through mplayer-vaapi are sometimes lagging (audio is fine, video is slow);
- whole kwin rendering is slower;
- wine game (simple World Of Goo) is lagging.

And it's all after just 15th minutes of testing... what a crap

I also run GNOME3 but it's support still sux.

I prepared package catalyst-test (contains catalyst-utils + catalyst-hook + lib32-catalyst-utils) cuz i thought it will be fine and i will move it into AUR's catalyst-test, but hell no, i don't think it's good idea.

Instead you can download PKGBUILD (+ other needed files) from here: http://catalyst.apocalypsus.net/tarball … .2b.tar.gz
Maybe with hd 5xxx/6xxx it will work better...

Cdh · 2011-08-13 07:58:54

I have a hd 6550 = hd 5650 and I can't reproduce the bad behaviour. Kwin seems fine with all effects, but vaapi has rendering problems (flickering that shows earlier frames). But it's the sort I occasionally too had with earlier versions and that was "fixed" with a reboot but I don't want to reboot now

edit: Still have to change back to 11.7: After a while of idling in kde it goes "ASIC hang happened".

Last edited by Cdh (2011-08-13 13:28:23)

bart_vv · 2011-08-13 10:51:49

Vi0L0 wrote:

@axed:
Yes, powerXpress support doesn't work atm, but i'm keeping it in -total because i still got some hope that someone will figure it out ;-).

Hmm, what about that rumour that it's working on ubuntu? Anyone checked to confirm this?

b0uncyfr0 · 2011-08-17 04:35:55

Is that ASIC Hang causing your whole PC to freeze? Im getting it too but i thought my Overclcok on the cpu was bad. It was running fine in Windows but i realised nix might be a bit more sensitive. If its my 6950 im relieved, my OC is still good.

Cdh · 2011-08-17 17:49:05

ASIC hang causes X to completely lock up. Killing the X server might work somehow, but the screen won't refresh and is stuck with the framebuffer from what was displayed while the hang happened.
ssh etc. is working fine though.
Oh, and the failed X is eating 100% cpu after the hang.

edit: http://www2.ati.com/drivers/linux/ati-d … x86_64.run
Officially released. Gnome Shell still not working I heard.

Last edited by Cdh (2011-08-17 17:50:16)

Vi0L0 · 2011-08-17 21:24:28

bart_vv wrote:

Vi0L0 wrote:
@axed:
Yes, powerXpress support doesn't work atm, but i'm keeping it in -total because i still got some hope that someone will figure it out ;-).
Hmm, what about that rumour that it's working on ubuntu? Anyone checked to confirm this?

I wrote mail to ubuntu's developer, asking for help, in response he wrote:

I'm afraid I can't help you, as we use alternatives instead of simple
symlinks to libraries. I can only suggest you to ask the maintainer of
the scripts for SUSE.

So i will ask also at SUSE, thou i didn't hear anything about working pXp on this distro.

catalyst 11.8 has been released
AUR and repo has been updated.
i don't see anything interesting in it, phoronix is saying:

Among the bug-fixes in Catalyst 11.8 is fixing 2D corruption when using Compiz with RandR 1.3 panning enabled, an X Server crash fix for Unity, blank displays when starting X with CrossFire enabled, and missing display properties for DisplayPort.

On phoronix froums there's an info that mouse freeze bug is fixed
+ i don't see any improvements in GNOME3 support.

Also... i'm experiencing this bug:

$ dmesg
(...)
BUG: using smp_processor_id() in preemptible [00000000] code: X/2534
caller is kas_GetExecutionLevel+0xd/0x20 [fglrx]
Pid: 2534, comm: X Tainted: P        W   3.1.0-mainline #1
Call Trace:
 [<ffffffff811b992d>] ? debug_smp_processor_id+0xdd/0x100
 [<ffffffffa0292900>] ? IRQMGR_QueryAsicTimestamp+0x20/0x20 [fglrx]
 [<ffffffffa01bd64d>] ? kas_GetExecutionLevel+0xd/0x20 [fglrx]
 [<ffffffffa01da019>] ? MCIL_GetExecutionLevel+0x39/0x80 [fglrx]
 [<ffffffffa01d7c82>] ? MCIL_QuerySystemInfo+0xb2/0x100 [fglrx]
 [<ffffffffa02a0455>] ? IRQMGR_SynchronizeExecution+0x25/0x80 [fglrx]
 [<ffffffffa02a1988>] ? IRQMGR_GetCurrentThreadId+0x38/0x50 [fglrx]
 [<ffffffffa029f394>] ? CallbackQueueAccess+0x164/0x2a0 [fglrx]
 [<ffffffffa0292900>] ? IRQMGR_QueryAsicTimestamp+0x20/0x20 [fglrx]
 [<ffffffffa01d9b72>] ? MCIL_bMiniportCapEnabled+0x82/0xa0 [fglrx]
 [<ffffffffa0292bbb>] ? RegisterIRQClient+0xdb/0x1c0 [fglrx]
 [<ffffffffa029008e>] ? IRQMGR_Access+0xee/0x190 [fglrx]
 [<ffffffffa01dbdd0>] ? irqmgr_interrupt_handler+0x20/0x20 [fglrx]
 [<ffffffffa01dbb9e>] ? IRQMGR_enable_hifreq_interrupt+0xee/0x150 [fglrx]
 [<ffffffffa01ec7e2>] ? firegl_trace+0x72/0x1e0 [fglrx]
 [<ffffffffa01dbdd0>] ? irqmgr_interrupt_handler+0x20/0x20 [fglrx]
 [<ffffffffa01da5dc>] ? irqmgr_wrap_enable_hifreq_interrupt+0x5c/0x1f0 [fglrx]
 [<ffffffffa01ef32b>] ? firegl_cmmqs_QueryThreadWakeupSource+0xab/0xd0 [fglrx]
 [<ffffffffa01d9366>] ? MCIL_SuspendThread+0x106/0x120 [fglrx]
 [<ffffffffa023dc22>] ? _ZN2OS13suspendThreadEj+0x22/0x30 [fglrx]
 [<ffffffffa023691f>] ? CMMQSWaitOnTsSignal+0xaf/0xd0 [fglrx]
 [<ffffffffa0246032>] ? _Z8uCWDDEQCmjjPvjS_+0xc32/0x10c0 [fglrx]
 [<ffffffffa01eecf2>] ? firegl_cmmqs_CWDDE_32+0x332/0x440 [fglrx]
 [<ffffffffa01ed620>] ? firegl_cmmqs_CWDDE32+0x70/0x100 [fglrx]
 [<ffffffffa01ed5b0>] ? firegl_cmmqs_createdriver+0x170/0x170 [fglrx]
 [<ffffffffa01cc1d8>] ? firegl_ioctl+0x1e8/0x250 [fglrx]
 [<ffffffff810431d9>] ? get_parent_ip+0x9/0x20
 [<ffffffffa01bdc49>] ? ip_firegl_unlocked_ioctl+0x9/0x10 [fglrx]
 [<ffffffff811147ee>] ? do_vfs_ioctl+0x8e/0x4f0
 [<ffffffff8104331f>] ? add_preempt_count+0x7f/0xb0
 [<ffffffff810431d9>] ? get_parent_ip+0x9/0x20
 [<ffffffff81043277>] ? sub_preempt_count+0x87/0xb0
 [<ffffffff8100bbc3>] ? sys_rt_sigreturn+0x1e3/0x200
 [<ffffffff81114ce8>] ? sys_ioctl+0x98/0xa0
 [<ffffffff81358312>] ? system_call_fastpath+0x16/0x1b

it was present also in earlier releases, thou because:

CONFIG_TREE_PREEMPT_RCU=y

is present now in our default kernel configuration of linux package i guess that we will all suffer because of this dmesg's spam. It doesn't harm performance, so it's not that bad.
In bugzilla i can see that there's already a thread for it here, so i guess i will add my comment into it, tommorow.

rowdog · 2011-08-17 22:19:37

11.8 fixed the "mouse goes brain dead over the systray" for my HD 5450!

Warper4 · 2011-08-17 23:55:39

11.8 fixed to mi Segment Fault on X-Plane 9

Gnome 3 continue with rainbow in top bar

Fififox · 2011-08-18 15:13:32

The annoying mouse freeze bug is fixed for my HD5770, AWESOME! I'm a happy user now.

And no trace of any problem in dmesg...

canuckkat · 2011-08-20 04:41:00

Vi0L0 wrote:

- wine game (simple World Of Goo) is lagging.

Just curious, have you tried the linux version of world of goo?

Panzerino · 2011-08-20 09:05:04

After update to the 11.8 version, when some applications try to use OpenGl, the desktop freeze completely i.e. - if I start glxgears after 2 sec the desktop freeze completely, or if the output mode of the mplayer is set to "gl"- the same.
The nasty "mouse bug" is gone.
I'm on xfce4.

Vi0L0, wish to thank you for your efforts! Your enthusiasm is much appreciated!

Vi0L0, I'm on Radeon 6450. ati config --initial does not put any module in "load" section of the xorg.conf. Which modules must be loaded manually?
On the old Catalyst driver, glxgears shows about 3600 fps not bad at all... .

Last edited by Panzerino (2011-08-21 15:37:29)

Vi0L0 · 2011-08-20 09:24:37

Fififox wrote:

And no trace of any problem in dmesg...

Well, then maybe i'm wrong about that preempt_rcu :>

canuckkat wrote:

Just curious, have you tried the linux version of world of goo?

Nope, i'm using exe version to perform a quick test on lib32-catalyst

Panzerino wrote:

After update to the 11.8 version, when some application try to use OpenGl, the desktop freeze completely i.e. - if I start glxgears after 2 sec the desktop freeze completely, or if the output mode of the mplayer is set to "gl"- the same.

Is there any interesting thing in dmesg output or in ~/.xsession-errors file?
Are you using clean catalyst or catalyst-total?
Also show me the output of fglrxinfo and ldd /usr/bin/glxgears and lsmod | grep fglrx.

Panzerino wrote:

Vi0L0, I'm on Radeon 6450. ati config --initial does not put any module in "load" section of the xorg.conf. Which modules must be loaded manually?

Guess you mean "Module" not "Load" Section - then none. It should work fine without any additional modules.

edgydalina · 2011-08-20 11:17:40

Here I am, having the same problem as Panzerino. (Still catalyst-total, I'll try normal catalyst later ...)

Here's the output you requested:

dmesg | grep fglrx:

[    4.526674] fglrx: module license 'Proprietary. (C) 2002 - ATI Technologies, Starnberg, GERMANY' taints kernel.
[    4.551133] [fglrx] Maximum main memory to use for locked dma buffers: 3787 MBytes.
[    4.551194] [fglrx]   vendor: 1002 device: 6779 count: 1
[    4.551443] [fglrx] ioport: bar 4, base 0xe000, size: 0x100
[    4.551665] [fglrx] Kernel PAT support is enabled
[    4.551678] [fglrx] module loaded - fglrx 8.88.7 [Jul 28 2011] with 1 minors
[    4.809138] fglrx_pci 0000:01:00.0: setting latency timer to 64
[   12.619246] fglrx_pci 0000:01:00.0: irq 50 for MSI/MSI-X
[   12.619684] [fglrx] Firegl kernel thread PID: 851
[   12.619716] [fglrx] Firegl kernel thread PID: 852
[   12.619746] [fglrx] Firegl kernel thread PID: 853
[   12.620048] [fglrx] IRQ 50 Enabled
[   12.870648] [fglrx] Gart USWC size:1236 M.
[   12.870652] [fglrx] Gart cacheable size:489 M.
[   12.870656] [fglrx] Reserved FB block: Shared offset:0, size:1000000 
[   12.870658] [fglrx] Reserved FB block: Unshared offset:f921000, size:3df000 
[   12.870659] [fglrx] Reserved FB block: Unshared offset:3fff4000, size:c000

Nothing interesting in .xsession-errors, would be hard to achieve anyway as I can't do anything after the freeze ...

fglrxinfo:

display: :0  screen: 0
OpenGL vendor string: ATI Technologies Inc.
OpenGL renderer string: AMD Radeon HD 6450
OpenGL version string: 4.1.11005 Compatibility Profile Context

ldd /usr/bin/glxgears:

	linux-vdso.so.1 =>  (0x00007fffd39ff000)
	libGL.so.1 => /usr/lib/catalystpxp/libGL.so.1 (0x00007f4f8f4ec000)
	libm.so.6 => /lib/libm.so.6 (0x00007f4f8f26a000)
	libX11.so.6 => /usr/lib/libX11.so.6 (0x00007f4f8ef2e000)
	libc.so.6 => /lib/libc.so.6 (0x00007f4f8ebce000)
	libpthread.so.0 => /lib/libpthread.so.0 (0x00007f4f8e9b1000)
	libXext.so.6 => /usr/lib/libXext.so.6 (0x00007f4f8e79f000)
	libxcb.so.1 => /usr/lib/libxcb.so.1 (0x00007f4f8e584000)
	libdl.so.2 => /lib/libdl.so.2 (0x00007f4f8e380000)
	/lib/ld-linux-x86-64.so.2 (0x00007f4f8f6f2000)
	libXau.so.6 => /usr/lib/libXau.so.6 (0x00007f4f8e17e000)
	libXdmcp.so.6 => /usr/lib/libXdmcp.so.6 (0x00007f4f8df79000)

lsmod | grep fglrx:

fglrx                2869510  56 
button                  4470  1 fglrx

Will now try 'catalyst' ...

UPDATE:

Installed 'catalyst' and 'catalyst-utils'. For the first time in a long time there was a 'libGL' in /usr/lib which I didn't have to copy there manually ...
The [FAIL] when loading modules at boot disappeared.
Rest staid the same (output, I mean) and freeze after some seconds when trying to do anything 3D (not actually sure about 'glxgears', maybe I didn't wait long enough?).

I'm starting to think that for this bug the catalyst developers are to blame. Maybe something card specific?

UPDATE #2:

I'm back on 11.7 but I'll 'tail -f' this thread for this problem.

Last edited by edgydalina (2011-08-20 18:12:29)

Panzerino · 2011-08-20 12:41:50

Vi0L0,

I'm using only catalyst-hook and catalyst-utils. Clean install- no previous drivers installed/uninstalled.
After recent kernel update, catalyst-hook worked properly- there was "ok" after the fglrx module was build.

ldd /usr/bin/glxgears

linux-vdso.so.1 => (0x00007fffae7ff000)
libGL.so.1 => /usr/lib/libGL.so.1 (0x00007f16bd3cf000)
libm.so.6 => /lib/libm.so.6 (0x00007f16bd14d000)
libX11.so.6 => /usr/lib/libX11.so.6 (0x00007f16bce11000)
libc.so.6 => /lib/libc.so.6 (0x00007f16bcab0000)
libpthread.so.0 => /lib/libpthread.so.0 (0x00007f16bc893000)
libXext.so.6 => /usr/lib/libXext.so.6 (0x00007f16bc681000)
libxcb.so.1 => /usr/lib/libxcb.so.1 (0x00007f16bc466000)
libdl.so.2 => /lib/libdl.so.2 (0x00007f16bc262000)
/lib/ld-linux-x86-64.so.2 (0x00007f16bd5d5000)
libXau.so.6 => /usr/lib/libXau.so.6 (0x00007f16bc060000)
libXdmcp.so.6 => /usr/lib/libXdmcp.so.6 (0x00007f16bbe5b000)

*******
lsmod | grep fglrx
fglrx 2869510 108
button 4470 1 fglrx
*******
fglrxinfo
display: :0.0 screen: 0
OpenGL vendor string: ATI Technologies Inc.
OpenGL renderer string: AMD Radeon HD 6450
OpenGL version string: 4.1.11005 Compatibility Profile Context

~/.xsession-errors
>> is empty.

Thank you very much!

Unia · 2011-08-23 13:42:12

Hey Vi0l0,

I recently experienced some system hang ups, mostly when opening Firefox but last time when I opened VLC's preferences window. With my topic opened here I traced back the error message and found the following part at kernel.log. Since it is related to fglrx (at least, I assume from the log) I thought I should ask here.

Aug 17 16:59:06 localhost kernel: [ 5775.598155] [fglrx] IRQ 45 Disabled
Aug 17 16:59:06 localhost kernel: [ 5775.598172] BUG: scheduling while atomic: X/865/0x00000002
Aug 17 16:59:06 localhost kernel: [ 5775.598173] Modules linked in: dm_mod fuse ipv6 cpufreq_ondemand uvcvideo videodev usbhid media v4l2_compat_ioctl32 hid joydev uas usb_storage keucr(C) arc4 snd_hda_codec_hdmi snd_hda_codec_realtek broadcom ath9k mac80211 ath9k_common ath9k_hw snd_hda_intel snd_hda_codec ath snd_hwdep snd_pcm cfg80211 snd_timer i2c_i801 tg3 intel_agp snd iTCO_wdt psmouse ehci_hcd vboxdrv acer_wmi sparse_keymap evdev sg pcspkr serio_raw i2c_core intel_ips libphy rfkill intel_gtt iTCO_vendor_support usbcore wmi soundcore mei(C) snd_page_alloc thermal battery ac video acpi_cpufreq freq_table processor mperf fglrx(P) button ext4 mbcache jbd2 crc16 sr_mod sd_mod cdrom ahci libahci libata scsi_mod
Aug 17 16:59:06 localhost kernel: [ 5775.598224] Pid: 865, comm: X Tainted: P         C  3.0-ARCH #1
Aug 17 16:59:06 localhost kernel: [ 5775.598226] Call Trace:
Aug 17 16:59:06 localhost kernel: [ 5775.598236]  [<ffffffff813e9ed3>] __schedule_bug+0x5f/0x64
Aug 17 16:59:06 localhost kernel: [ 5775.598241]  [<ffffffff813f0381>] schedule+0x831/0x9f0
Aug 17 16:59:06 localhost kernel: [ 5775.598280]  [<ffffffffa014e2e2>] ? firegl_trace+0x72/0x1e0 [fglrx]
Aug 17 16:59:06 localhost kernel: [ 5775.598284]  [<ffffffff81045128>] ? __wake_up_common+0x58/0x90
Aug 17 16:59:06 localhost kernel: [ 5775.598308]  [<ffffffffa014e2e2>] ? firegl_trace+0x72/0x1e0 [fglrx]
Aug 17 16:59:06 localhost kernel: [ 5775.598312]  [<ffffffff813f2eb7>] ? _raw_spin_unlock_irqrestore+0x37/0x40
Aug 17 16:59:06 localhost kernel: [ 5775.598315]  [<ffffffff81046e13>] ? __wake_up+0x53/0x70
Aug 17 16:59:06 localhost kernel: [ 5775.598317]  [<ffffffff813f2eb7>] ? _raw_spin_unlock_irqrestore+0x37/0x40
Aug 17 16:59:06 localhost kernel: [ 5775.598321]  [<ffffffff8107f470>] ? prepare_to_wait+0x60/0x90
Aug 17 16:59:06 localhost kernel: [ 5775.598338]  [<ffffffffa012355a>] KAS_Thread_WaitForFinish+0x8a/0xc0 [fglrx]
Aug 17 16:59:06 localhost kernel: [ 5775.598341]  [<ffffffff8107f680>] ? abort_exclusive_wait+0xb0/0xb0
Aug 17 16:59:06 localhost kernel: [ 5775.598362]  [<ffffffffa013a567>] MCIL_WorkerThread+0x177/0x230 [fglrx]
Aug 17 16:59:06 localhost kernel: [ 5775.598384]  [<ffffffffa013bc06>] ? MCIL_GetExecutionLevel+0x66/0x80 [fglrx]
Aug 17 16:59:06 localhost kernel: [ 5775.598416]  [<ffffffffa01ee7a1>] ? PassiveRing_Cleanup+0x61/0xc0 [fglrx]
Aug 17 16:59:06 localhost kernel: [ 5775.598444]  [<ffffffffa01fd3e8>] DispatchRing_Cleanup+0x48/0x70 [fglrx]
Aug 17 16:59:06 localhost kernel: [ 5775.598447]  [<ffffffff81141933>] ? kfree+0x163/0x180
Aug 17 16:59:06 localhost kernel: [ 5775.598450]  [<ffffffff810bfee6>] ? free_irq+0x56/0xb0
Aug 17 16:59:06 localhost kernel: [ 5775.598481]  [<ffffffffa01edde0>] ? IRQMGR_FreeResources+0x20/0x60 [fglrx]
Aug 17 16:59:06 localhost kernel: [ 5775.598504]  [<ffffffffa013cff5>] ? irqmgr_wrap_shutdown+0x125/0x180 [fglrx]
Aug 17 16:59:06 localhost kernel: [ 5775.598522]  [<ffffffffa012ea06>] ? firegl_takedown+0x7d6/0x8a0 [fglrx]
Aug 17 16:59:06 localhost kernel: [ 5775.598548]  [<ffffffffa0204202>] ? amdPcsEngine_InitDatabase+0x12/0x20 [fglrx]
Aug 17 16:59:06 localhost kernel: [ 5775.598567]  [<ffffffffa012db15>] ? firegl_release+0x125/0x1c0 [fglrx]
Aug 17 16:59:06 localhost kernel: [ 5775.598580]  [<ffffffffa011eee1>] ? ip_firegl_release+0x11/0x20 [fglrx]
Aug 17 16:59:06 localhost kernel: [ 5775.598584]  [<ffffffff811590fa>] ? fput+0xda/0x210
Aug 17 16:59:06 localhost kernel: [ 5775.598586]  [<ffffffff81155996>] ? filp_close+0x66/0x90
Aug 17 16:59:06 localhost kernel: [ 5775.598589]  [<ffffffff81060000>] ? put_files_struct+0xa0/0x110
Aug 17 16:59:06 localhost kernel: [ 5775.598591]  [<ffffffff81060132>] ? exit_files+0x52/0x60
Aug 17 16:59:06 localhost kernel: [ 5775.598593]  [<ffffffff8106063a>] ? do_exit+0x18a/0x8f0
Aug 17 16:59:06 localhost kernel: [ 5775.598595]  [<ffffffff810610f4>] ? do_group_exit+0x44/0xa0
Aug 17 16:59:06 localhost kernel: [ 5775.598598]  [<ffffffff81071fee>] ? get_signal_to_deliver+0x33e/0x510
Aug 17 16:59:06 localhost kernel: [ 5775.598601]  [<ffffffff8100b1af>] ? do_signal+0x6f/0x780
Aug 17 16:59:06 localhost kernel: [ 5775.598603]  [<ffffffff8106ffd7>] ? __set_task_blocked+0x37/0x80
Aug 17 16:59:06 localhost kernel: [ 5775.598605]  [<ffffffff81072362>] ? set_current_blocked+0x52/0x60
Aug 17 16:59:06 localhost kernel: [ 5775.598608]  [<ffffffff810723ba>] ? sigprocmask+0x4a/0x90
Aug 17 16:59:06 localhost kernel: [ 5775.598609]  [<ffffffff81071488>] ? do_tkill+0x88/0xa0
Aug 17 16:59:06 localhost kernel: [ 5775.598611]  [<ffffffff8106ffd7>] ? __set_task_blocked+0x37/0x80
Aug 17 16:59:06 localhost kernel: [ 5775.598613]  [<ffffffff8100b945>] ? do_notify_resume+0x65/0x80
Aug 17 16:59:06 localhost kernel: [ 5775.598615]  [<ffffffff810723ba>] ? sigprocmask+0x4a/0x90
Aug 17 16:59:06 localhost kernel: [ 5775.598618]  [<ffffffff813f4050>] ? int_signal+0x12/0x17

Vi0L0 · 2011-08-24 07:31:26

@Unia:
the bug is caused by fglrx and should be fixed by AMD itself. You will find more details about this bug here: http://arstechnica.com/civis/viewtopic. … 6&t=157629
I'm not sure what's causing this, it's related with smp and the bug i was wroting about in my previous post no. 3140 here, thou.
You haven't experiencing this bug before 3.0 kernel, have you?
Please write a bug report at http://ati.cchtml.com

edgydalina · 2011-08-24 15:42:47

What about Panzerino's and my bug? Anything new/old?
My suggestion would be ... to wait ... for a new version ...

Arch Linux

#3126 2011-08-08 10:21:01

Re: The AMD/ATI Bar & Grill

#3127 2011-08-08 10:59:24

Re: The AMD/ATI Bar & Grill

#3128 2011-08-08 11:22:59

Re: The AMD/ATI Bar & Grill

#3129 2011-08-08 11:26:53

Re: The AMD/ATI Bar & Grill

#3130 2011-08-08 12:16:19

Re: The AMD/ATI Bar & Grill

#3131 2011-08-08 19:08:18

Re: The AMD/ATI Bar & Grill

#3132 2011-08-10 12:33:17

Re: The AMD/ATI Bar & Grill

#3133 2011-08-11 06:18:06

Re: The AMD/ATI Bar & Grill

#3134 2011-08-12 16:30:56

Re: The AMD/ATI Bar & Grill

#3135 2011-08-13 07:58:54

Re: The AMD/ATI Bar & Grill

#3136 2011-08-13 10:51:49

Re: The AMD/ATI Bar & Grill

#3137 2011-08-17 04:35:55

Re: The AMD/ATI Bar & Grill

#3138 2011-08-17 17:49:05

Re: The AMD/ATI Bar & Grill

#3139 2011-08-17 21:24:28

Re: The AMD/ATI Bar & Grill

#3140 2011-08-17 22:19:37

Re: The AMD/ATI Bar & Grill

#3141 2011-08-17 23:55:39

Re: The AMD/ATI Bar & Grill

#3142 2011-08-18 15:13:32

Re: The AMD/ATI Bar & Grill

#3143 2011-08-20 04:41:00

Re: The AMD/ATI Bar & Grill

#3144 2011-08-20 09:05:04

Re: The AMD/ATI Bar & Grill

#3145 2011-08-20 09:24:37

Re: The AMD/ATI Bar & Grill

#3146 2011-08-20 11:17:40

Re: The AMD/ATI Bar & Grill

#3147 2011-08-20 12:41:50

Re: The AMD/ATI Bar & Grill

#3148 2011-08-23 13:42:12

Re: The AMD/ATI Bar & Grill

#3149 2011-08-24 07:31:26

Re: The AMD/ATI Bar & Grill

#3150 2011-08-24 15:42:47

Re: The AMD/ATI Bar & Grill

Board footer