The AMD/ATI Bar & Grill

abhinandh · 2011-06-17 19:38:13

[  7764.235] (II) fglrx(0): PowerXpress: Discrete GPU is selected.
[  7764.343] (II) fglrx(0): PowerXpress: Diagnostic output from /usr/lib64/fglrx/switchlibGL:
Switch to AMD

[  7764.350] drmOpenDevice: node name is /dev/dri/card0
[  7764.350] drmOpenDevice: open result is 12, (OK)
[  7764.350] drmOpenByBusid: Searching for BusID pci:0000:00:02.0
[  7764.350] drmOpenDevice: node name is /dev/dri/card0
[  7764.350] drmOpenDevice: open result is 12, (OK)
[  7764.350] drmOpenByBusid: drmOpenMinor returns 12
[  7764.350] drmOpenByBusid: drmGetBusid reports pci:0000:00:02.0
[  7764.350] (**) fglrx(0): Depth 24, (--) framebuffer bpp 32
[  7764.350] (==) fglrx(0): RGB weight 888
[  7764.350] (==) fglrx(0): Default visual is TrueColor
[  7764.350] (**) fglrx(0): Option "Tiling" "off"
[  7764.350] (**) fglrx(0): Option "LinearFramebuffer" "on"
[  7764.350] (**) fglrx(0): Option "Shadow" "off"
[  7764.351] (II) fglrx(0): Integrated Graphics Chipset: Intel(R) Arrandale
[  7764.351] (**) fglrx(0): ChipID override: 0x0046
[  7764.351] (**) fglrx(0): ChipRev override: 2
[  7764.351] (**) fglrx(0): Chipset: "Arrandale"
[  7764.351] 
Backtrace:
[  7764.351] 0: /usr/bin/X (xorg_backtrace+0x26) [0x49f556]
[  7764.351] 1: /usr/bin/X (0x400000+0x60cf9) [0x460cf9]
[  7764.351] 2: /lib/libpthread.so.0 (0x7fea3b029000+0xf750) [0x7fea3b038750]
[  7764.351] 3: /usr/lib/xorg/modules/drivers/intel_drv.so (0x7fea37476000+0xf669) [0x7fea37485669]
[  7764.351] 4: /usr/lib/xorg/modules/drivers/fglrx_drv.so (xdl_xs110_atiddxPxPreInit+0x17d) [0x7fea37c4433d]
[  7764.352] 5: /usr/lib/xorg/modules/drivers/fglrx_drv.so (xdl_xs110_atiddxPreInit+0x213e) [0x7fea37c1aaae]
[  7764.352] 6: /usr/bin/X (InitOutput+0x834) [0x46ffb4]
[  7764.352] 7: /usr/bin/X (0x400000+0x229c3) [0x4229c3]
[  7764.352] 8: /lib/libc.so.6 (__libc_start_main+0xed) [0x7fea39fa9f6d]
[  7764.352] 9: /usr/bin/X (0x400000+0x22e6d) [0x422e6d]
[  7764.352] Segmentation fault at address (nil)
[  7764.352]

Integrated intel works fine, can't start X with the discrete card. Envy 14 with 10.6 catalyst-total.
Is this normal for fglrx?

[  7764.351] (II) fglrx(0): Integrated Graphics Chipset: Intel(R) Arrandale

Cdh · 2011-06-18 07:37:31

Vi0L0 wrote:

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 have seen this some time ago in the nouveau wiki:
http://nouveau.freedesktop.org/wiki/KernelModeSetting

Deactivating KMS and unloading Nouveau
The framebuffer console reserves nouveaufb and therefore you cannot directly unload the Nouveau kernel module. You have to unbind nouveaufb first. When you unbind nouveaufb, the earlier fb driver (not vesafb) or the VGA console driver will take over. Your kernel configuration needs to have CONFIG_VT_HW_CONSOLE_BINDING enabled.
NOTE: For NV50 class hardware and above, the VGA text mode cannot be restored, you will be without a working virtual console until you load a new framebuffer driver! For older hardware the VGA text mode console should be restored fine. If you want to try to restore the VGA text mode yourself, the command vbetool post might be able to do that. Please, read the vbetool manual, before attempting it.
Here is an example script to unload Nouveau KMS:
#!/bin/bash

echo 0 > /sys/class/vtconsole/vtcon1/bind
rmmod nouveau
/etc/init.d/consolefont restart
rmmod ttm
rmmod drm_kms_helper
rmmod drm
The echo line unbinds nouveaufb from the framebuffer console driver (fbcon). Usually it is vtcon1 as in this example, but it may also be another vtcon*. See /sys/class/vtconsole/vtcon*/name which one of them is a "frame buffer device".
The console font trick may be needed with the VGA text mode to restore proper console state (to keep the active line visible), and depends on your distribution.

This shouldn't be that different with intel or radeon...

Vi0L0 · 2011-06-18 12:04:21

powerXpress one more time - dual card systems
yesterday just before midnight we (me and Aarin) were talking with kevlarman on #ati @freenode, and here's what he said:

kevlarman wrote:

afaik you can't get catalyst working without disabling the intel card in hardware

kevlarman wrote:

you may have some luck with 1) removing all traces of the intel driver 2) turning off the intel graphics before xorg starts (see the radeon/intel hack thing), 3) making sure kms is off and radeon is blacklisted

So bassically this whole AMD's powerXpress support seams to be a joke ;P, well at least in the way it works now. All this phoronix news, pxp implementation in every main linux distro, and my work over this - all this means nothing because it shouldn't work (at least in AMD's way) .

But oh well maybe it will be better, vgaswitcheroo could help in turning intel card off before switch to amd card, also your last entry Cdh about turning KMS off looks promissing. The battle isn't over, jet...

Ochi · 2011-06-18 12:26:44

Hmmm... I still wonder if the Radeon will ever be able to output video on my particular system (as mentioned before, Dell Vostro 3350, apparently muxer-less) when the Intel is disabled. The reason for this is the following observation:

Before I had tried catalyst, I've been experimenting with the open source drivers for both the Intel and the Radeon. When using them and having KMS enabled, vgaswitcheroo is available on my system which allows me to toggle the Intel and the Radeon cards on and off pretty well.

Now the thing is, as soon as I disable the Intel card using vgaswitcheroo, the screen will just freeze. The system can still be used blindly or via ssh and X can be started as well. When checking the X logs via ssh, you can see that it is apparently running on the Radeon, but the screen will still be totally frozen (to my knowledge, re-enabling the Intel then doesn't really help except that the screen will eventually unfreeze when switching to a VT).

In addition, the open source radeon driver told me that there don't seem to be any outputs attached. Thus my feeling is that the Radeon is just an outputless slave which will have to use some other means to output video. Just disabling the Intel doesn't really help, at least on my system without some kind of hardware switch.

P.S.: I'd like to try kevlarman's hints as well (i.e. trying quite the same using catalyst instead of the open source drivers), but when I disable KMS, intel and radeon, vgaswitcheroo won't be available. I have an acpi_call command available for switching off the _Radeon_ even when vgaswitcheroo is unavailable, but I don't know yet how to disable the _Intel_ without vgaswitcheroo.

P.P.S.: The AMD switching graphics are a joke even under Windows. When I bought the device I thought, alright, maybe I won't be able to use the graphics under Linux from day one but at least they'd work under Windows. But no... I usually use OpenGL applications but OpenGL is currently unsupported with switchable graphics. Fine print on the AMD page: "...OpenGL applications will run on the integrated GPU or the APU as the case may be)...". DirectX works alright somehow, but OpenGL application will run on the Intel and even slower than when I just omit the Catalyst driver. So it doesn't even completely work with the Windows drivers yet! XD Frustrating...

Last edited by Ochi (2011-06-19 00:04:19)

Cdh · 2011-06-21 07:44:29

sumski wrote:

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

I can. log says

kernel: [ 5678.898659] [fglrx:fireglAsyncioIntEnableMsgHandler] *ERROR* IRQMGR returned error 1 when trying to enable interrupt source 20100000

immediately after connecting or disconnecting something to HDMI.

But it somehow works with

chris@chrisl ~ % xrandr --output CRT1 --mode 1600x1200 --right-of LVDS --output LVDS --mode 1600x900
** connect screen to VGA output **

chris@chrisl ~ % xrandr --output CRT1 --mode 1600x1200 --right-of LVDS --output LVDS --mode 1600x900

** remove screen from VGA output  and connect screen to HDMI output **

chris@chrisl ~ % xrandr
Screen 0: minimum 320 x 200, current 3200 x 1200, maximum 4000 x 4000
LVDS connected 1600x900+0+0 (normal left inverted right x axis y axis) 382mm x 214mm
   1600x900       60.3*+
   1440x900       60.3  
   1280x768       60.3  
   1280x720       60.3  
   1024x768       60.3  
   1024x600       60.3  
   800x600        60.3  
   800x480        60.3  
   640x480        60.3  
DFP1 connected 1600x1200+0+0 (normal left inverted right x axis y axis) 408mm x 306mm
[1]    8251 segmentation fault (core dumped)  xrandr
139 chris@chrisl ~ % xrandr
Screen 0: minimum 320 x 200, current 3200 x 1200, maximum 4000 x 4000
LVDS connected 1600x900+0+0 (normal left inverted right x axis y axis) 382mm x 214mm
   1600x900       60.3*+
   1440x900       60.3  
   1280x768       60.3  
   1280x720       60.3  
   1024x768       60.3  
   1024x600       60.3  
   800x600        60.3  
   800x480        60.3  
   640x480        60.3  
DFP1 connected 1600x1200+0+0 (normal left inverted right x axis y axis) 408mm x 306mm
   1600x1200      60.0*+
   1400x1050      60.0  
   1280x1024      75.0     60.0  
   1440x900       60.0  
   1280x960       75.0     60.0  
   1280x768       75.0     60.0  
   1280x720       75.0     60.0  
   1024x768       75.0     70.1     60.0  
   1024x600       75.0     70.1     60.0  
   800x600        75.0     60.3  
   800x480        75.0     60.3  
   640x480        75.0     59.9  
CRT1 disconnected (normal left inverted right x axis y axis)

If anyone is interested:

chris@chrisl ~ % gdb xrandr core
GNU gdb (GDB) 7.2
Copyright (C) 2010 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law.  Type "show copying"
and "show warranty" for details.
This GDB was configured as "x86_64-unknown-linux-gnu".
For bug reporting instructions, please see:
<http://www.gnu.org/software/gdb/bugs/>...
Reading symbols from /usr/bin/xrandr...(no debugging symbols found)...done.
[New Thread 8251]

warning: Can't read pathname for load map: Eingabe-/Ausgabefehler.
Reading symbols from /usr/lib/libXrandr.so.2...(no debugging symbols found)...done.
Loaded symbols for /usr/lib/libXrandr.so.2
Reading symbols from /usr/lib/libX11.so.6...(no debugging symbols found)...done.
Loaded symbols for /usr/lib/libX11.so.6
Reading symbols from /lib/libm.so.6...(no debugging symbols found)...done.
Loaded symbols for /lib/libm.so.6
Reading symbols from /lib/libc.so.6...(no debugging symbols found)...done.
Loaded symbols for /lib/libc.so.6
Reading symbols from /usr/lib/libXext.so.6...(no debugging symbols found)...done.
Loaded symbols for /usr/lib/libXext.so.6
Reading symbols from /usr/lib/libXrender.so.1...(no debugging symbols found)...done.
Loaded symbols for /usr/lib/libXrender.so.1
Reading symbols from /usr/lib/libxcb.so.1...(no debugging symbols found)...done.
Loaded symbols for /usr/lib/libxcb.so.1
Reading symbols from /lib/libdl.so.2...(no debugging symbols found)...done.
Loaded symbols for /lib/libdl.so.2
Reading symbols from /lib/ld-linux-x86-64.so.2...(no debugging symbols found)...done.
Loaded symbols for /lib/ld-linux-x86-64.so.2
Reading symbols from /usr/lib/libXau.so.6...(no debugging symbols found)...done.
Loaded symbols for /usr/lib/libXau.so.6
Reading symbols from /usr/lib/libXdmcp.so.6...(no debugging symbols found)...done.
Loaded symbols for /usr/lib/libXdmcp.so.6
Core was generated by `xrandr'.
Program terminated with signal 11, Segmentation fault.
#0  0x000000000040876c in ?? ()
(gdb)

Last edited by Cdh (2011-06-21 07:46:20)

wlog · 2011-06-22 13:04:42

I found that when I use the "pxp_switch_catalyst intel" to switch to open source radeon driver(and enable KMS)
I can't use gnome 3 experience or run 3D OpenGL Program(like Unigine Heaven).

After I remove 20-catalystpxp.conf from /etc/X11/xorg.conf.d, everything works

Before remove 20-catalystpxp.conf, Xorg log had

[     8.240] (II) Loading /usr/lib/xorg/modules/updates/extensions/libglx.so

After remove 20-catalystpxp.conf, Xorg log had

[     9.649] (II) Loading /usr/lib/xorg/modules/extensions/libglx.so

Maybe pxp_switch_catalyst need to remove 20-catalystpxp.conf when switch to intel(or radeon)

Last edited by wlog (2011-06-22 13:19:09)

Cdh · 2011-06-22 16:29:26

Actually it's easier to make HDMI work: Connect screen via VGA, no need to enable it (but maybe it is necessary to run xrandr (even with no arguments)), disconnect the screen vom VGA and connect screen to HDMI. Error gets written to log and nothing happens. Now simply running xrandr will produce

chris@chrisl ~ % xrandr
xrandr: server did not report mode 0xb3 for output DFP1
1 chris@chrisl ~ %

But right after that the HDMI screen will activate and be normally usable.

Last edited by Cdh (2011-06-22 16:31:19)

xscream · 2011-06-23 07:49:44

Hi everybody,

First of all, thx to Vi0L0 for the catalyst-total package, great work, it is simpler than ever.

But I have some question before installing arch on my new computer (HD5850 based) and some issues to show

I've read the previous pages but I did'nt find informations if there are particular instruction to follow on setup catalyst-total

Based on the Ati Catalyst wiki, here is the procedure that I will follow :
- Fresh Install of Arch64 (urxvt,openbox,thunar)
- Then I will install Xorg but I don't know what version : classic or catalyst-maximize-fix ?
- setup catalyst repo above archlinuxfr repo above core/community/... repo and uncomment multilib repo
- install catalyst-total from the repo and run aticonfig ... like the wiki says
- install wm : openbox for my part
- compositing is enabled by default, so I will install cairo-compmgr and conky-lua (to make beautiful effects
- then smplayer configuration like the ATI wiki (vaapi)
- to finish other stuff but not related to this forum

On an other computer, HD 4XXX based, I set up a similar environnement, but I have some issues with catalyst-total installed :
- tear free desktop is good, working flawlessly, but with terminals, there are some displays problems, like when there is an yaourt selection, some part of the terminal are not drawed until I write something.
- with xorg-server(not catalyst-maximize version) + tear free + cairo-compmgr, when I open Thunar (for example) the window gamma/colors (I don't know how to say in english, sorry dudes, I'm french) are very very ligth, but sometimes it's good.
- I can hear texture loading (it's the best way to describe the sound) when I move the mouse on desktop, like in games, it's a little funny at the beginning but non-existent on Win7 on desktop applications ...

I think this will be enough for today, thx again for your work Vi0L0, and all the community to help and improve packaging.

xscream, maybe not alone on arch in little britanny, now I go to work ^^

sonicbhoc · 2011-06-28 05:23:58

I am having the same problem another poster is having in which the intel driver is loaded, and then X crashes and burns. At least, that's what I think is happening... Anyway, here is my Xorg.0.log output:

[   291.399] 
X.Org X Server 1.10.2
Release Date: 2011-05-28
[   291.400] X Protocol Version 11, Revision 0
[   291.400] Build Operating System: Linux 2.6.38-ARCH i686 
[   291.400] Current Operating System: Linux beast 2.6.39-ARCH #1 SMP PREEMPT Tue Jun 7 05:49:02 UTC 2011 i686
[   291.400] Kernel command line: root=/dev/mapper/archfs-root ro nomodeset
[   291.400] Build Date: 30 May 2011  08:16:10AM
[   291.400]  
[   291.400] Current version of pixman: 0.22.0
[   291.401]    Before reporting problems, check http://wiki.x.org
        to make sure that you have the latest version.
[   291.401] Markers: (--) probed, (**) from config file, (==) default setting,
        (++) from command line, (!!) notice, (II) informational,
        (WW) warning, (EE) error, (NI) not implemented, (??) unknown.
[   291.402] (==) Log file: "/var/log/Xorg.0.log", Time: Tue Jun 28 00:46:30 2011
[   291.472] (==) Using config file: "/etc/X11/xorg.conf"
[   291.473] (==) Using config directory: "/etc/X11/xorg.conf.d"
[   291.515] (==) ServerLayout "aticonfig Layout"
[   291.515] (**) |-->Screen "aticonfig-Screen[0]-0" (0)
[   291.515] (**) |   |-->Monitor "aticonfig-Monitor[0]-0"
[   291.528] (**) |   |-->Device "aticonfig-Device[0]-0"
[   291.528] (==) Automatically adding devices
[   291.528] (==) Automatically enabling devices
[   291.596] (WW) The directory "/usr/share/fonts/OTF/" does not exist.
[   291.596]    Entry deleted from font path.
[   291.596] (WW) The directory "/usr/share/fonts/Type1/" does not exist.
[   291.596]    Entry deleted from font path.
[   291.597] (==) FontPath set to:
        /usr/share/fonts/misc/,
        /usr/share/fonts/TTF/,
        /usr/share/fonts/100dpi/,
        /usr/share/fonts/75dpi/
[   291.597] (**) ModulePath set to "/usr/lib/xorg/modules/updates,/usr/lib/xorg/modules"
[   291.597] (II) The server relies on udev to provide the list of input devices.
        If no devices become available, reconfigure udev or disable AutoAddDevices.
[   291.597] (II) Loader magic: 0x822e120
[   291.597] (II) Module ABI versions:
[   291.597]    X.Org ANSI C Emulation: 0.4
[   291.597]    X.Org Video Driver: 10.0
[   291.597]    X.Org XInput driver : 12.2
[   291.597]    X.Org Server Extension : 5.0
[   291.599] (--) PCI:*(0:0:2:0) 8086:0046:17aa:3952 rev 24, Mem @ 0xcfc00000/4194304, 0xd0000000/268435456, I/O @ 0x00001800/8
[   291.599] (--) PCI: (0:1:0:0) 1002:68c0:17aa:3952 rev 0, Mem @ 0xe0000000/268435456, 0xcf1e0000/131072, I/O @ 0x00002000/256, BIOS @ 0x????????/131072
[   291.599] (WW) Open ACPI failed (/var/run/acpid.socket) (No such file or directory)
[   291.599] (II) "extmod" will be loaded by default.
[   291.599] (II) "dbe" will be loaded by default.
[   291.599] (II) "glx" will be loaded by default.
[   291.599] (II) "record" will be loaded by default.
[   291.599] (II) "dri" will be loaded by default.
[   291.599] (II) "dri2" will be loaded by default.
[   291.599] (II) LoadModule: "extmod"
[   291.619] (II) Loading /usr/lib/xorg/modules/extensions/libextmod.so
[   291.629] (II) Module extmod: vendor="X.Org Foundation"
[   291.629]    compiled for 1.10.2, module version = 1.0.0
[   291.629]    Module class: X.Org Server Extension
[   291.630]    ABI class: X.Org Server Extension, version 5.0
[   291.630] (II) Loading extension MIT-SCREEN-SAVER
[   291.630] (II) Loading extension XFree86-VidModeExtension
[   291.630] (II) Loading extension XFree86-DGA
[   291.630] (II) Loading extension DPMS
[   291.630] (II) Loading extension XVideo
[   291.630] (II) Loading extension XVideo-MotionCompensation
[   291.630] (II) Loading extension X-Resource
[   291.630] (II) LoadModule: "dbe"
[   291.630] (II) Loading /usr/lib/xorg/modules/extensions/libdbe.so
[   291.638] (II) Module dbe: vendor="X.Org Foundation"
[   291.638]    compiled for 1.10.2, module version = 1.0.0
[   291.638]    Module class: X.Org Server Extension
[   291.638]    ABI class: X.Org Server Extension, version 5.0
[   291.638] (II) Loading extension DOUBLE-BUFFER
[   291.638] (II) LoadModule: "glx"
[   291.638] (II) Loading /usr/lib/xorg/modules/updates/extensions/libglx.so
[   291.667] (II) Module glx: vendor="FireGL - ATI Technologies Inc."
[   291.667]    compiled for 6.9.0, module version = 1.0.0
[   291.667] (II) Loading extension GLX
[   291.667] (II) LoadModule: "record"
[   291.668] (II) Loading /usr/lib/xorg/modules/extensions/librecord.so
[   291.680] (II) Module record: vendor="X.Org Foundation"
[   291.680]    compiled for 1.10.2, module version = 1.13.0
[   291.680]    Module class: X.Org Server Extension
[   291.680]    ABI class: X.Org Server Extension, version 5.0
[   291.680] (II) Loading extension RECORD
[   291.680] (II) LoadModule: "dri"
[   291.681] (II) Loading /usr/lib/xorg/modules/extensions/libdri.so
[   291.694] (II) Module dri: vendor="X.Org Foundation"
[   291.694]    compiled for 1.10.2, module version = 1.0.0
[   291.694]    ABI class: X.Org Server Extension, version 5.0
[   291.694] (II) Loading extension XFree86-DRI
[   291.694] (II) LoadModule: "dri2"
[   291.695] (II) Loading /usr/lib/xorg/modules/extensions/libdri2.so
[   291.695] (II) Module dri2: vendor="X.Org Foundation"
[   291.696]    compiled for 1.10.2, module version = 1.2.0
[   291.696]    ABI class: X.Org Server Extension, version 5.0
[   291.696] (II) Loading extension DRI2
[   291.696] (II) LoadModule: "fglrx"
[   291.696] (II) Loading /usr/lib/xorg/modules/drivers/fglrx_drv.so
[   291.934] (II) Module fglrx: vendor="FireGL - ATI Technologies Inc."
[   291.941]    compiled for 1.4.99.906, module version = 8.86.5
[   291.941]    Module class: X.Org Video Driver
[   291.957] (II) Loading sub module "fglrxdrm"
[   291.957] (II) LoadModule: "fglrxdrm"
[   291.957] (II) Loading /usr/lib/xorg/modules/linux/libfglrxdrm.so
[   291.991] (II) Module fglrxdrm: vendor="FireGL - ATI Technologies Inc."
[   291.991]    compiled for 1.4.99.906, module version = 8.86.5
[   291.991] (II) ATI Proprietary Linux Driver Version Identifier:8.86.5
[   291.991] (II) ATI Proprietary Linux Driver Release Identifier: 8.861                                
[   291.991] (II) ATI Proprietary Linux Driver Build Date: May 24 2011 22:46:44
[   291.991] (--) using VT number 7

[   291.994] (WW) Falling back to old probe method for fglrx
[   292.043] (II) Loading PCS database from /etc/ati/amdpcsdb
[   292.067] (--) Chipset Supported AMD Graphics Processor (0x68C0) found
[   292.067] (WW) fglrx: No matching Device section for instance (BusID PCI:0@1:0:1) found
[   292.067] (II) fglrx: intel VGA device detected, load intel driver.
[   292.067] (II) LoadModule: "intel"
[   292.067] (II) Loading /usr/lib/xorg/modules/drivers/intel_drv.so
[   292.097] (II) Module intel: vendor="X.Org Foundation"
[   292.097]    compiled for 1.10.2, module version = 2.15.0
[   292.097]    Module class: X.Org Video Driver
[   292.097]    ABI class: X.Org Video Driver, version 10.0
[   292.099] ukiDynamicMajor: found major device number 252
[   292.099] ukiDynamicMajor: found major device number 252
[   292.099] ukiOpenByBusid: Searching for BusID PCI:1:0:0
[   292.099] ukiOpenDevice: node name is /dev/ati/card0
[   292.099] ukiOpenDevice: open result is 9, (OK)
[   292.099] ukiOpenByBusid: ukiOpenMinor returns 9
[   292.099] ukiOpenByBusid: ukiGetBusid reports PCI:1:0:0
[   292.101] (II) AMD Video driver is running on a device belonging to a group targeted for this release
[   292.114] (II) AMD Video driver is signed
[   292.115] (II) Loading /usr/lib/xorg/modules/drivers/fglrx_drv.so
[   292.115] (II) Loading /usr/lib/xorg/modules/linux/libfglrxdrm.so
[   292.115] (II) fglrx(0): pEnt->device->identifier=0x93029a8
[   292.118] (II) pEnt->device->identifier=(nil)
[   292.118] (II) fglrx(0): === [xdl_xs110_atiddxPreInit] === begin
[   292.118] (II) fglrx(0): PowerXpress: Discrete GPU is selected.
[   293.323] (II) fglrx(0): PowerXpress: Diagnostic output from /usr/lib/fglrx/switchlibGL:
Switch to AMD

[   293.335] 
Backtrace:
[   293.335] 0: Xorg (xorg_backtrace+0x37) [0x80e8bb7]
[   293.335] 1: Xorg (0x8048000+0x5ef9a) [0x80a6f9a]
[   293.335] 2: (vdso) (__kernel_rt_sigreturn+0x0) [0xb77a140c]
[   293.336] 3: /usr/lib/xorg/modules/drivers/fglrx_drv.so (xdl_xs110_atiddxPxPreInit+0x195) [0xb69d08a5]
[   293.336] 4: /usr/lib/xorg/modules/drivers/fglrx_drv.so (xdl_xs110_atiddxPreInit+0x2554) [0xb69a3204]
[   293.336] 5: Xorg (InitOutput+0x853) [0x80b7753]
[   293.336] Segmentation fault at address 0x8
[   293.336] 
Fatal server error:
[   293.336] Caught signal 11 (Segmentation fault). Server aborting
[   293.336] 
[   293.336] 
Please consult the The X.Org Foundation support 
         at http://wiki.x.org
 for help. 
[   293.336] Please also check the log file at "/var/log/Xorg.0.log" for additional information.
[   293.336]

Is there any way to prevent the i915 driver from loading at boot? I've tried using rmmod but all I get is something about the module being in use (but it doesn't say what's using it). I've tried adding it to the /etc/rc.conf modules with an ! in front of it, which apparently is supposed to blacklist it, but in this case it didn't work.

xscream · 2011-06-28 07:32:54

Recently, all modules blacklisting from rc.conf has been disabled.
So adding i915 to rc.conf is useless.

You have two solutions:

- Blacklist, but the module can be loaded later, create a file name i915.conf in /etc/modprobe.d/ and add this

blacklist mod1

- Or permanently (can never be loaded by you, root or another module add this instead of the previous line :

install mod1 /bin/false

More info : http://www.archlinux.org/news/changes-t … cklisting/ and see man modprobe.d

Last edited by xscream (2011-06-28 07:33:22)

sonicbhoc · 2011-06-28 12:58:04

I blacklisted that module, and it doesn't load at boot.

But X insists on loading it when started. I don't want to make it unloadable, because I want to use the switchable graphics. What I do want is for X to not load the Intel driver when the AMD card is on.

Radioactiveman · 2011-06-28 16:43:25

Does Catalyst already work with kernel26-3.0? Does the 2.6.39 patch apply correctly?

Vi0L0 · 2011-06-28 16:58:07

First of all - sorry for being absent, i was quiet bussy lately.

xscream wrote:

- Then I will install Xorg but I don't know what version : classic or catalyst-maximize-fix ?

I preffer classic. On my hd4850 it's working well.
Your steps looks fine.

xscream wrote:

On an other computer, HD 4XXX based, I set up a similar environnement, but I have some issues with catalyst-total installed :
- tear free desktop is good, working flawlessly, but with terminals, there are some displays problems, like when there is an yaourt selection, some part of the terminal are not drawed until I write something.
- with xorg-server(not catalyst-maximize version) + tear free + cairo-compmgr, when I open Thunar (for example) the window gamma/colors (I don't know how to say in english, sorry dudes, I'm french) are very very ligth, but sometimes it's good.
- I can hear texture loading (it's the best way to describe the sound) when I move the mouse on desktop, like in games, it's a little funny at the beginning but non-existent on Win7 on desktop applications ...

I never used openbox, thou i think that those issues shouldn't occur, feel free to write bugreports at http://ati.cchtml.com

@sonicbhoc:
like i already said powerXpress support isn't working like it should. I though that it's impossible to resovle but yesterday i red @ bbs.archlinux.pl that it's working fine on ubuntu. LOL, i'm confused... so my work on this isn't over, i'm not kind of person who gives up quickly ;P

Radioactiveman wrote:

Does Catalyst already work with kernel26-3.0? Does the 2.6.39 patch apply correctly?

Yes, i'm using 3.0 rc5 now and it's fine.

xscream · 2011-06-28 20:14:32

sonicbhoc wrote:

But X insists on loading it when started. I don't want to make it unloadable, because I want to use the switchable graphics. What I do want is for X to not load the Intel driver when the AMD card is on.

I don't have my Aspire 4820TG for now (Ati Switchable Intel/HD5650M) but I haven't try to use catalyst, I use only Intel and disable ATI (with radeon not Catalyst) to minimize the battery usage.
I will try when it get back from maintenance.

But I managed with rc.local and acpi_call module to disable during work use the ati card, and enable before shutdown, to prevent a hardware panic because radeon module can not communicate with ATI hardware ! It confused.
I will post when I get my laptop back.

Vi0L0 wrote:

I never used openbox, thou i think that those issues shouldn't occur, feel free to write bugreports at http://ati.cchtml.com

I'll work a few days more and then report a bug.

Vi0L0 wrote:

Yes, i'm using 3.0 rc5 now and it's fine.

Maybe I will try this kernel later.

I will also try to remove xorg-catalyst and test classic version.
But every update, it offers me to replace libgl with catalyst and catalyst-utils ... I have enabled catalyst repo above other repositories, like [archlinuxfr]

I will post a screenshot when I get back to my desktop computer in a few time. Stay tuned ^^

Vi0L0 · 2011-06-29 11:17:40

xscream wrote:

Vi0L0 wrote:
Yes, i'm using 3.0 rc5 now and it's fine.
Maybe I will try this kernel later.

It basically shouldn't change anything, so this step is unnecessary.
I'm using mainline kernel mainly to check is fglrx module compiling on it well, if not i'm trying to provide a patch
+ i'm building mainline kernel on custom config which is different in many ways than arch one - with it and new features offered by fresh kernel whole system's performance is better.

xscream wrote:

I will also try to remove xorg-catalyst and test classic version.
But every update, it offers me to replace libgl with catalyst and catalyst-utils ... I have enabled catalyst repo above other repositories, like [archlinuxfr]

There is no catalyst-total package (and probably won't be, ever) in [catalyst] repo so pacman is trying to get catalyst-utils (it has no pXp support) from the repo which is conflicting with libgl. So my suggestion is simple - if you got catalyst-total then feel free to comment [catalyst] in pacman.conf - especially when you did already install xvba-video package from it.

xscream · 2011-06-29 12:36:56

Vi0L0 wrote:

There is no catalyst-total package (and probably won't be, ever) in [catalyst] repo so pacman is trying to get catalyst-utils (it has no pXp support) from the repo which is conflicting with libgl. So my suggestion is simple - if you got catalyst-total then feel free to comment [catalyst] in pacman.conf - especially when you did already install xvba-video package from it.

Thx for the advice. I will post after a fresh install and full configuration.

Are they hidden options in the kernel linked to ATI catalyst support ? I never tried to make a custom kernel because some people tell me it's not worth it (have I correctly used the word ?).

I will focus on video playing, music and apps, and keep windows for games !

Last edited by xscream (2011-06-29 14:42:34)

Vi0L0 · 2011-06-29 13:53:17

xscream wrote:

Are they hidden options in the kernel linked to ATI catalyst support ?

No.

xscream wrote:

I never tried to make a custom kernel because some people tell me it's not worth it (have I correctly use the word ?).

In most cases those people are right (i understood it, so it's fine for me:) ).

Last edited by Vi0L0 (2011-06-29 13:53:45)

atanas · 2011-06-30 12:31:03

Hello guys,
I just want to inform you that there will be server maintenance at 03.07.2011, 00:00h. Our repository will be down approximately 2 or 3 hours. Sorry for the inconvenience.

volalto · 2011-07-05 11:40:30

Hello everybody,

I recently purchased an Acer TimelineX 4820TG (Intel + Radeon HD5470), and I've been investigating the hybrid graphics issue since then.

I'm currently using open-source radeon and intel drivers + in-kernel vga-switcheroo to manage my two cards. I'd like to move to an intel + Catalyst setup, which seems to work, according to many posts in this thread.

I'd really appreciate some hints on where to start, since I didn't find any howto or guide around, and reading through the 100+ pages in the thread got me a little confused. We could try to start a wiki-page for hybrid-newbies like me.

Thanks for your help!

xscream · 2011-07-06 07:52:49

Hi,

There is a Wiki page about Intel/Radeon
https://wiki.archlinux.org/index.php/Acer_TimelineX

maybe it's time to add a Catalyst part in this wiki and discuss it on the forum !

I'm also a 4840TG user, but I didn't use it since last month, because of the top cover and a RMA to Acer !

Now I cannot boot because of MODULES depreciation in rc.conf, I think I have to reset and do a fresh start.

Last edited by xscream (2011-07-06 08:00:32)

ilusi0n · 2011-07-06 20:07:09

Hello guys;

Is anyone having problems with fglrx these days? I search on google and there is people reporting but not sure if there is a worarroud. But fglrx hard lockup on shutdown or reboot. I wanted to know if anyone had a similiar problem or solved it. The problem was reported in ubuntu as far as I can tell. I saw it (one of the examples) here: https://bugs.launchpad.net/ubuntu/+bug/750437
I search on arch forum but seems no one is having this problems...
Thanks, if anyonr can help : )

Last edited by ilusi0n (2011-07-06 20:12:44)

volalto · 2011-07-07 13:08:46

I tried to install catalyst drivers using the information I got so far...but it didn't go as expected (Acer TimelineX 4820TG, Intel + Radeon HD 5470).

What I did:
- removed open-source driver (xf86-video-ati, ati-dri, ecc)
- installed catalyst-total (Catalyst 11.6)
- disabled the script previously powering-off radeon
- reboot adding nomodeset
- pxp_switch_catalyst amd
- started X (transition to runlevel 5)

Fglrx crashed and Xorg shows black screen:

[   111.261] (WW) Falling back to old probe method for fglrx
[   111.308] (II) Loading PCS database from /etc/ati/amdpcsdb
[   111.329] (--) Chipset Supported AMD Graphics Processor (0x68E0) found
[   111.330] (WW) fglrx: No matching Device section for instance (BusID PCI:0@1:0:1) found
[   111.330] (II) fglrx: intel VGA device detected, load intel driver.
[   111.330] (II) LoadModule: "intel"
[   111.330] (II) Loading /usr/lib/xorg/modules/drivers/intel_drv.so
[   111.367] (II) Module intel: vendor="X.Org Foundation"
[   111.367]     compiled for 1.10.2, module version = 2.15.0
[   111.367]     Module class: X.Org Video Driver
[   111.367]     ABI class: X.Org Video Driver, version 10.0
[   111.369] ukiDynamicMajor: found major device number 252
[   111.369] ukiDynamicMajor: found major device number 252
[   111.369] ukiOpenByBusid: Searching for BusID PCI:1:0:0
[   111.369] ukiOpenDevice: node name is /dev/ati/card0
[   111.369] ukiOpenDevice: open result is 7, (OK)
[   111.369] ukiOpenByBusid: ukiOpenMinor returns 7
[   111.369] ukiOpenByBusid: ukiGetBusid reports PCI:1:0:0
[   111.383] (II) AMD Video driver is running on a device belonging to a group targeted for this release
[   111.398] (II) AMD Video driver is signed
[   111.398] (II) Loading /usr/lib/xorg/modules/drivers/fglrx_drv.so
[   111.398] (II) Loading /usr/lib/xorg/modules/linux/libfglrxdrm.so
[   111.398] (II) fglrx(0): pEnt->device->identifier=0x11738b0
[   111.401] (II) pEnt->device->identifier=(nil)
[   111.401] (II) fglrx(0): === [xdl_xs110_atiddxPreInit] === begin
[   111.401] (II) fglrx(0): PowerXpress: Discrete GPU is selected.
[   111.546] (II) fglrx(0): PowerXpress: Diagnostic output from /usr/lib64/fglrx/switchlibGL:
Switch to AMD

[   111.552] 
Backtrace:
[   111.588] 0: /usr/bin/X (xorg_backtrace+0x26) [0x49f556]
[   111.588] 1: /usr/bin/X (0x400000+0x60cf9) [0x460cf9]
[   111.588] 2: /lib/libpthread.so.0 (0x7fcb2ae24000+0xf7e0) [0x7fcb2ae337e0]
[   111.588] 3: /usr/lib/xorg/modules/drivers/fglrx_drv.so (xdl_xs110_atiddxPxPreInit+0x17a) [0x7fcb27a3c33a]
[   111.588] 4: /usr/lib/xorg/modules/drivers/fglrx_drv.so (xdl_xs110_atiddxPreInit+0x213e) [0x7fcb27a12aae]
[   111.589] 5: /usr/bin/X (InitOutput+0x834) [0x46ffb4]
[   111.589] 6: /usr/bin/X (0x400000+0x229c3) [0x4229c3]
[   111.589] 7: /lib/libc.so.6 (__libc_start_main+0xed) [0x7fcb29da217d]
[   111.589] 8: /usr/bin/X (0x400000+0x22e6d) [0x422e6d]
[   111.589] Segmentation fault at address 0x10
[   111.589] 
Fatal server error:
[   111.589] Caught signal 11 (Segmentation fault). Server aborting
[   111.589] 
[   111.589] 
Please consult the The X.Org Foundation support 
     at http://wiki.x.org
 for help. 
[   111.589] Please also check the log file at "/var/log/Xorg.0.log" for additional information.
[   111.589]

pxp_switch_catalyst intel --> xorg works, even though with a strange resolution (I guess it's nomodeset's fault)

If I switch BIOS setting to "Discrete Graphic Card only".. catalyst works!

Any idea? Am I missing anything?

awpe · 2011-07-08 21:58:06

Can anyone please say where I can get normal configs for X server for mobility hd5870 (asus g73jh),

the problem:
I've installed arch with xf86-video-ati there was very good 2d acceleration and also fading effects worked well but there wasn't anything like 3d support or powersaving ( fans were turned on 100% )
then I've Installed catalyst-total from aur run aticonfig --initial and rebooted, afterthen I received power saving (fans were not so noisy) and 3d support too BUT 2d accelerate and vsync were turned off...
On ubuntu proprietary driver works great in all ways(2d/3d/vsync/fans), so I've tried to find its configs and wasn't succeed in it - there was nothing I could use in /etc/X11/xorg.conf, and I've heard that ubuntu has many patches so it is very difficult to use its cfgs on other distributives.

(or post link to that info)

volalto · 2011-07-10 09:25:51

@awpe

Did you try catalyst control centre (gui application you can find in system menu). TearFree desktop feature seems to include what you need.

awpe · 2011-07-10 21:02:55

Yes, I've tried and wasn't succeed making away with 2d lags . But this could be a solution if catalyst has cfgs for amdcccle I can try to copy them from ubuntu. What are thinking about this?

Arch Linux

#3076 2011-06-17 19:38:13

Re: The AMD/ATI Bar & Grill

#3077 2011-06-18 07:37:31

Re: The AMD/ATI Bar & Grill

#3078 2011-06-18 12:04:21

Re: The AMD/ATI Bar & Grill

#3079 2011-06-18 12:26:44

Re: The AMD/ATI Bar & Grill

#3080 2011-06-21 07:44:29

Re: The AMD/ATI Bar & Grill

#3081 2011-06-22 13:04:42

Re: The AMD/ATI Bar & Grill

#3082 2011-06-22 16:29:26

Re: The AMD/ATI Bar & Grill

#3083 2011-06-23 07:49:44

Re: The AMD/ATI Bar & Grill

#3084 2011-06-28 05:23:58

Re: The AMD/ATI Bar & Grill

#3085 2011-06-28 07:32:54

Re: The AMD/ATI Bar & Grill

#3086 2011-06-28 12:58:04

Re: The AMD/ATI Bar & Grill

#3087 2011-06-28 16:43:25

Re: The AMD/ATI Bar & Grill

#3088 2011-06-28 16:58:07

Re: The AMD/ATI Bar & Grill

#3089 2011-06-28 20:14:32

Re: The AMD/ATI Bar & Grill

#3090 2011-06-29 11:17:40

Re: The AMD/ATI Bar & Grill

#3091 2011-06-29 12:36:56

Re: The AMD/ATI Bar & Grill

#3092 2011-06-29 13:53:17

Re: The AMD/ATI Bar & Grill

#3093 2011-06-30 12:31:03

Re: The AMD/ATI Bar & Grill

#3094 2011-07-05 11:40:30

Re: The AMD/ATI Bar & Grill

#3095 2011-07-06 07:52:49

Re: The AMD/ATI Bar & Grill

#3096 2011-07-06 20:07:09

Re: The AMD/ATI Bar & Grill

#3097 2011-07-07 13:08:46

Re: The AMD/ATI Bar & Grill

#3098 2011-07-08 21:58:06

Re: The AMD/ATI Bar & Grill

#3099 2011-07-10 09:25:51

Re: The AMD/ATI Bar & Grill

#3100 2011-07-10 21:02:55

Re: The AMD/ATI Bar & Grill

Board footer