Wrong driver (fglrx) version loading at bootup.

I have a Suse 11.4_64 system, KDE 4.7.1

I made a recent change to my graphics hardware, which needed an up to date version of fglrx, so executed the atiupgrade script.
The script executed, installed the updated rpm and the system worked fine.

After a subsequent reboot, i noticed that Desktop effects were not working, looked in /var/log/Xorg.0.log and noticed errors in the fglrx startup - it appeared that the old fglrx was loading.

I ran

find /lib -name fglrx.ko
/lib/modules/2.6.37.6-0.7-desktop/extra/fglrx.ko

And noted that the date code on fglrx.ko was current.
Hmmm

So I ran a controlled experiment

  1. boot system to level 3.
    2 lsmod |grep fglrx
    fglrx size_of_old
    3 rmmod fglrx
  2. modprobe fglrx
    5 lsmod |grep fglrx
    fglrx *size_of_new
    *6 init 5
  3. kdm starts up, kogin*, *desktop effects working (correct fglrx installed)

My question is - where might the old fglrx be hiding* ?

How do I purge it?
*

On 10/13/2011 05:46 AM, cmcgrath5035 wrote:

> which needed an up to date version of fglrx

what were the indications that you needed a different version?


DD
openSUSE®, the “German Automobiles” of operating systems

The “old” version is Catalyst 11.4 (April 2011 Release)
With Radeon 6550 graphics hardware, no OpenGL, errors logged in Xorg.0.log
Desktop is flaky, ultimately locks up.

The “new” version is Catalyst 11.9, runs well.

The “old” version seems stuck in a cache, or perhaps initrd?

On 10/13/2011 01:46 PM, cmcgrath5035 wrote:
>
> Desktop is flaky, ultimately locks up.

i may be wrong, but i believe the best way to a happy face with ATI is
via this page: http://tinyurl.com/37v9y7m


DD
openSUSE®, the “German Automobiles” of operating systems

Thanks, that reference is great in-depth on the topic.

I have been using the atiupgrade script from please_try_again, which has served me very well for the past few months.
I do have the “new”, or should I say “most recent” 11.9 release of fglrx loaded into the system and, as best I can tell, the only fglrx.ko on the file system.

So the real question is, why is the 11.4 (April, 2011) version loading at boot, and where is it coming from.

Is fglrx included in the initrd, loaded by GRUB?

NO.
By the way, as its name indicates, atiupgrade replaces the older driver - and its job is even easier if it was already installed as a package. It just not “updates” the driver but completely deinstalls the older package and installs the new one. The reason for doing so is to allow building the package on openSUSE versions which are not supported by the ati install script. It is usefull at any version change.

The kernel module “fglrx.ko” is not part of any package but gets built when you install the package - not when you create it. That’s why the kernel sources are required. At boot, the service /etc/init.d/boot.fglrxbuild checks it the module is available for the loaded kernel and compiles it if it’s missing - such as after a kernel update.

To look for the fglrx modules available on your system, you can use:

find /lib/modules -name "fglrx.ko" -ls

The glx module included in Catalyst is /usr/lib|lib64/xorg/modules/updates/extensions/fglrx/fglrx-libglx.so.

I was about to say that you can use atiupgrade to downgrade the driver as well …but atipugrade doesn’t work right now: It’s not able to download the ati installer from the ATI web site. Maybe they changed the location … or this script has got too popular and they want people go to their website and get the driver from there. I have to take a look at the URLs.

# **atiupgrade 11.8**
 (1) - analyzing system... 
 (2) - get latest Catalyst version number...
Latest Catalyst version (11.9) already installed. Do you want to continue? [yn]y
 (3) - checking required packages... 
     * kernel-source is aready installed
     * kernel-desktop-devel is aready installed
     * kernel-syms is aready installed
     * gcc is aready installed
     * make is aready installed
 (4) - downloading the ATI installer... 
Connecting... Please wait.
asking libproxy about url 'http://www2.ati.com/drivers/linux/ati-driver-installer-11.8-x86.x86_64.run'
libproxy suggest to use 'direct://'
--2011-10-13 07:40:09--  http://www2.ati.com/drivers/linux/ati-driver-installer-11.8-x86.x86_64.run
Resolving www2.ati.com... 24.244.18.18, 24.244.18.33
Connecting to www2.ati.com|24.244.18.18|:80... connected.
HTTP request sent, awaiting response... 404 Not Found
2011-10-13 07:40:12 ERROR 404: Not Found.

atiupgrade without argument is equivalent to atiupgrade 11.9 and gives the same result. :frowning:

It’s working again.

Thanks, p_t_a, it worked fine two days ago as well…

I understand your description of the process, at a moderate level.

I tried an experiment, I renamed the file /lib/modules/2.6.37.6-0.7-desktop/extra/fglrx.ko and rebooted the system.
As boot was zinging by, I saw /etc/init.d/boot.fglrxbuild detect that there was no fglrx.ko and rebuilt it.
Here is the log

cat /var/log/fglrx-build.log
********************************************************************************
*   Build date: Thu Oct 13 20:31:48 EDT 2011                                   *
********************************************************************************

Used CPUs/Cores for compilation  =>   4 ]


********************************************************************************
*                                                                              *
*   Kernel:    2.6.37.6-0.7-desktop                                            *
*   Source:    /lib/modules/2.6.37.6-0.7-desktop/build                         *
*   Include:   /lib/modules/2.6.37.6-0.7-desktop/source/include                *
*                                                                              *
********************************************************************************



file /lib/modules/2.6.37.6-0.7-desktop/source/include/../arch/x86/include/asm/compat.h says:COMPAT_ALLOC_USER_SPACE=arch_compat_alloc_user_space
/usr/src/kernel-modules/fglrx/2.6.x /
rm -rf *.c *.h *.o *.ko *.a .??* *.symvers
make: Entering directory `/usr/src/linux-2.6.37.6-0.7-obj/x86_64/desktop'
/usr/bin/make -C ../../../linux-2.6.37.6-0.7 O=/usr/src/linux-2.6.37.6-0.7-obj/x86_64/desktop/. 
  LD      /usr/src/kernel-modules/fglrx/2.6.x/built-in.o
  CC [M]  /usr/src/kernel-modules/fglrx/2.6.x/firegl_public.o
  CC [M]  /usr/src/kernel-modules/fglrx/2.6.x/kcl_agp.o
  CC [M]  /usr/src/kernel-modules/fglrx/2.6.x/kcl_acpi.o
  CC [M]  /usr/src/kernel-modules/fglrx/2.6.x/kcl_debug.o
  CC [M]  /usr/src/kernel-modules/fglrx/2.6.x/kcl_ioctl.o
  CC [M]  /usr/src/kernel-modules/fglrx/2.6.x/kcl_io.o
  CC [M]  /usr/src/kernel-modules/fglrx/2.6.x/kcl_pci.o
/usr/src/kernel-modules/fglrx/2.6.x/firegl_public.c: In function 'KCL_GetInitKerPte':
/usr/src/kernel-modules/fglrx/2.6.x/firegl_public.c:2381:5: warning: return makes integer from pointer without a cast
/usr/src/kernel-modules/fglrx/2.6.x/firegl_public.c:2382:5: warning: return makes integer from pointer without a cast
/usr/src/kernel-modules/fglrx/2.6.x/firegl_public.c:2383:5: warning: return makes integer from pointer without a cast
/usr/src/kernel-modules/fglrx/2.6.x/firegl_public.c: In function 'KCL_GetPageTableByVirtAddr':
/usr/src/kernel-modules/fglrx/2.6.x/firegl_public.c:2428:5: warning: return makes integer from pointer without a cast
/usr/src/kernel-modules/fglrx/2.6.x/firegl_public.c:2431:5: warning: return makes integer from pointer without a cast
/usr/src/kernel-modules/fglrx/2.6.x/firegl_public.c:2432:5: warning: return makes integer from pointer without a cast
/usr/src/kernel-modules/fglrx/2.6.x/firegl_public.c: In function 'KCL_TestAndClearPageDirtyFlag':
/usr/src/kernel-modules/fglrx/2.6.x/firegl_public.c:2606:5: warning: return makes integer from pointer without a cast
  CC [M]  /usr/src/kernel-modules/fglrx/2.6.x/kcl_str.o
  CC [M]  /usr/src/kernel-modules/fglrx/2.6.x/kcl_iommu.o
  CC [M]  /usr/src/kernel-modules/fglrx/2.6.x/kcl.o
  CC [M]  /usr/src/kernel-modules/fglrx/2.6.x/kcl_wait.o
  LD [M]  /usr/src/kernel-modules/fglrx/2.6.x/fglrx.o
  Building modules, stage 2.
  MODPOST 1 modules
  CC      /usr/src/kernel-modules/fglrx/2.6.x/fglrx.mod.o
  LD [M]  /usr/src/kernel-modules/fglrx/2.6.x/fglrx.ko
make: Leaving directory `/usr/src/linux-2.6.37.6-0.7-obj/x86_64/desktop'
make: Entering directory `/usr/src/linux-2.6.37.6-0.7-obj/x86_64/desktop'
/usr/bin/make -C ../../../linux-2.6.37.6-0.7 O=/usr/src/linux-2.6.37.6-0.7-obj/x86_64/desktop/. modules_install
  INSTALL /usr/src/kernel-modules/fglrx/2.6.x/fglrx.ko
  DEPMOD  2.6.37.6-0.7-desktop
make: Leaving directory `/usr/src/linux-2.6.37.6-0.7-obj/x86_64/desktop'
rm -rf *.c *.h *.o *.ko *.a .??* *.symvers
/
Calling 'depmod -a 2.6.37.6-0.7-desktop' this may take a while...
WARNING: Can't read module /lib/modules/2.6.37.6-0.7-desktop/weak-updates/updates/hdj_mod.ko: No such file or directory


Summary report:
================================================================================

   Kernel     => 2.6.37.6-0.7-desktop
   Detected   => RPM package
   Build      =>  OK ]
   Install    =>  OK ]


Here is the result:

ll /lib/modules/2.6.37.6-0.7-desktop/extra/
total 11336
-rw-r--r-- 1 root root 5791673 Oct 13  2011 fglrx.ko
-rw-r--r-- 1 root root 5791673 Oct 12 08:36 fglrx.ko.cjmold

So I conclude I had the correct fglrx.ko, but now have a freshly compiled one.

I had booted to level 3, logged in as root.

lsmod |grep fglrx
fglrx                2740890  316 (note size - this is the fglrx that does not work)
button                  6829  1 fglrx
rmmod fglrx
modprobe fglrx
lsmod |grep fglrx
fglrx                2968771  316 
button                  6829  1 fglrx
init 5
(GUI starts up, log in, everything working well)

Here is what I believe tobe the interesting part of /var/log/boot.msg

<4>    2.015857] fglrx: module license 'Proprietary. (C) 2002 - ATI Technologies, Starnberg, GERMANY' taints kernel.
<4>    2.015860] Disabling lock debugging due to kernel taint
<6>    2.056843] [fglrx] Maximum main memory to use for locked dma buffers: 15203 MBytes.
<6>    2.056855] [fglrx]   vendor: 1002 device: 9640 count: 1
<6>    2.057261] [fglrx] ioport: bar 1, base 0xf000, size: 0x100
<6>    2.057271] pci 0000:00:01.0: PCI INT A -> GSI 18 (level, low) -> IRQ 18
<7>    2.057275] pci 0000:00:01.0: setting latency timer to 64
<6>    2.057690] [fglrx] Kernel PAT support is enabled
<6>    2.057705] [fglrx] module loaded - **fglrx 8.84.5 [Apr  5 2011]** with 1 minors


Here is the interesting part of /var/log/Xorg.0.log, after the startup with init 5

   206.686] (II) LoadModule: "fglrx"
   206.686] (II) Loading /usr/lib64/xorg/modules/drivers/fglrx_drv.so
   206.729] (II) Module fglrx: vendor="FireGL - ATI Technologies Inc."
   206.729]     compiled for 1.4.99.906, module version = 8.89.4
   206.729]     Module class: X.Org Video Driver
   206.729] (II) Loading sub module "fglrxdrm"
   206.729] (II) LoadModule: "fglrxdrm"
   206.729] (II) Loading /usr/lib64/xorg/modules/linux/libfglrxdrm.so
   206.731] (II) Module fglrxdrm: vendor="FireGL - ATI Technologies Inc."
   206.731]     compiled for 1.4.99.906, module version = 8.89.4
   206.731] (II) ATI Proprietary Linux Driver Version Identifier:8.89.4
   206.731] (II) ATI Proprietary Linux Driver Release Identifier: 8.892                                
   206.731] (II) ATI Proprietary Linux Driver Build Date: Sep 14 2011 10:36:30

Summary

At boot
<6>    2.057705] [fglrx] module loaded - fglrx 8.84.5 [Apr  5 2011] with 1 minors
loads and I assume results in 
**lsmod |grep fglrx
fglrx                2740890  316 (note size - this is the fglrx that does not work)
**
After remoiving fglrx and modprobe and init 5, Xorg repoirts
   206.731] (II) Module fglrxdrm: vendor="FireGL - ATI Technologies Inc."
   206.731]     compiled for 1.4.99.906, module version = 8.89.4
which I assume is
**lsmod |grep fglrx
fglrx                2968771  316 **

So I am left trying to understand where boot is getting fglrx 8.84.5 (old, does not work) when fglrx 8.89.4 is in /lib/modules/2.6.37.6-0.7-desktop/extra/fglrx.ko

I’ll assume that somehow the uninstall of the old driver package did not proceed properly(or completely), but I am not sure what to look for (and where).
This is the RPM that atiupgrade created and installed

rpm -qa |grep fglrx
fglrx64_xpic_SUSE114-8.892-1.x86_64

Recall that this is my A8 APU/Radeon 6550 graphics hardware that did not properly identify itself to atiupgrade.
I allowed the upgrade to proceed by answering “y” when asked.
I don’t see anything else in atiupgrade that should have been affected, but it is fairly complex code so I could be missing something.

You might be booting two different kernels. That would explain why you have two different versions of the module.
Check with:

uname -a

Again, the module is NOT part of the package. It is compiled at installtion time. The script /usr/bin/fglrx-kernel-build.sh (which is called by /etc/init.d/boot.fglrxrebuild) is responsible for compiling the module and removing old ones, not atiupgrade (which just builds and installs the package).

Again, the module is NOT part of the package. It is compiled at installtion time. The script /usr/bin/fglrx-kernel-build.sh (which is called by /etc/init.d/boot.fglrxrebuild) is responsible for compiling the module and removing old ones, not atiupgrade (which just builds and installs the package).

I understand, but it has to compile from the installed rpm, which is

fglrx64_xpic_SUSE114-8.892-1.x86_64

Only one kernel lurking here

uname -a
Linux PVE-LinuxSRV5 2.6.37.6-0.7-desktop #1 SMP PREEMPT 2011-07-21 02:17:24 +0200 x86_64 x86_64 x86_64 GNU/Linux

Am I too focused on fglrx.ko ?

Should I be tracking other components of the “complete” fglrx (all the files in /usr/X11R6/lib and /lib64) ?

Yet the fact remains that the wrong fglrx.ko is being loaded by the boot process, but can’t be found on the file system…

Just to confirm,

When 
lsmod |grep fglrx
fglrx                2968771  371 

This is part of fglrx.ko, correct?

I say "part of" because 
ls -la /lib/modules/2.6.37.6-0.7-desktop/extra/fglrx.ko
-rw-r--r-- 1 root root 5791673 Oct 13  2011 /lib/modules/2.6.37.6-0.7-desktop/extra/fglrx.ko

Is much larger than reported by lsmod

Correct, but atiupgrade is not involved in this process.

I’ll say yes. Look at the result of this command:

readlink /usr/lib64/xorg/modules/updates/extensions/libglx.so

On my system it says:

/usr/lib64/xorg/modules/updates/extensions/fglrx/fglrx-libglx.so

Then look if the file you get (which should not be different from mine) is included in the fglrx package:

rpm -ql fglrx64_xpic_SUSE114-8.892-1.x86_64 | grep $(readlink /usr/lib64/xorg/modules/updates/extensions/libglx.so)

Did you check with:

find /lib/modules -name "fglrx.ko" -ls

Use “find”, not “ls” !

You cannot compare the size of the file on the filesystem with the space the module takes in memory. There is nothing wrong here. These values are correct by the way (same here).

Here is a bundle of all your suggested queries

ls -la /lib/modules/2.6.37.6-0.7-desktop/extra/fglrx.ko
-rw-r--r-- 1 root root 5791673 Oct 13  2011 /lib/modules/2.6.37.6-0.7-desktop/extra/fglrx.ko

carl@PVE-LinuxSRV5:~> readlink /usr/lib64/xorg/modules/updates/extensions/libglx.so
/usr/lib64/xorg/modules/updates/extensions/fglrx/fglrx-libglx.so

carl@PVE-LinuxSRV5:~> rpm -ql fglrx64_xpic_SUSE114-8.892-1.x86_64 | grep $(readlink /usr/lib64/xorg/modules/updates/extensions/libglx.so)
/usr/lib64/xorg/modules/updates/extensions/fglrx/fglrx-libglx.so

carl@PVE-LinuxSRV5:~> find /lib/modules -name "fglrx.ko" -ls
3227776 5668 -rw-r--r--   1 root     root      5791673 Oct 13  2011 /lib/modules/2.6.37.6-0.7-desktop/extra/fglrx.ko

I believe they all look correct, which I expected since the 11.9 version of fglrx, once loaded, seems to work properly.

So the good driver (11.9) is there and functions well.

Question remains, why on initial boot if the older fglrx module loaded.

What file (or list) tells the boot process what modules to load?

After looking thru /etc/init.d/boot.fglrxrebuild, it appears that it is probably loading fglrx before level 3 is reached.
Maybe I need to be looking there

Yes it does.
Rather than looking in logs, just run:

$ modinfo fglrx | grep -v alias

and see if it reports another module.
(the many aliases are irrelevant. That’s why I grepped them out) .

Modules can be loaded in ramdisk (initrd) but it would be foolish to load fglrx there.
in Yast: Sysconfig Editor -> Kernel -> INITRD_MODULES

To have some control on boot.fglrxrebuild, have a look in:
Kernel -> Graphics card -> ATI .

modinfo fglrx |grep -v alias
filename:       /lib/modules/2.6.37.6-0.7-desktop/extra/fglrx.ko
license:        Proprietary. (C) 2002 - ATI Technologies, Starnberg, GERMANY
description:    ATI Fire GL
author:         Fire GL - ATI Research GmbH, Germany
srcversion:     0369EDAEF049F90A5DDF846
depends:        button
vermagic:       2.6.37.6-0.7-desktop SMP preempt mod_unload modversions 
parm:           firegl:charp

Nothing obvious, although I need to research vermagic a bit

Yast: INITRD_MODULLES and Graphcs-ATI look OK

I tried an experiment, with unexpected results.

I turned off boot.fglrxrebuild by renaming it.

I then rebooted to level 3 - and the “old”, “bad” fglrx is still loading.

So something else is loading it.

Here is the /var/log/boot.message (partial) listing again

  2.008542] fglrx: module license 'Proprietary. (C) 2002 - ATI Technologies, Starnberg, GERMANY' taints kernel.
<4>    2.008547] Disabling lock debugging due to kernel taint
<6>    2.020609] usb 3-2: New USB device found, idVendor=046d, idProduct=c00c
<6>    2.020612] usb 3-2: New USB device strings: Mfr=1, Product=2, SerialNumber=0
<6>    2.020615] usb 3-2: Product: USB Optical Mouse
<6>    2.020616] usb 3-2: Manufacturer: Logitech
<6>    2.028757] input: Logitech USB Optical Mouse as /devices/pci0000:00/0000:00:12.0/usb3/3-2/3-2:1.0/input/input3
<6>    2.028839] generic-usb 0003:046D:C00C.0001: input,hidraw0: USB HID v1.10 Mouse [Logitech USB Optical Mouse] on usb-0000:00:12.0-2/input0
<6>    2.064683] [fglrx] Maximum main memory to use for locked dma buffers: 15203 MBytes.
<6>    2.064693] [fglrx]   vendor: 1002 device: 9640 count: 1
<6>    2.065333] [fglrx] ioport: bar 1, base 0xf000, size: 0x100
<6>    2.065344] pci 0000:00:01.0: PCI INT A -> GSI 18 (level, low) -> IRQ 18
<7>    2.065348] pci 0000:00:01.0: setting latency timer to 64
<6>    2.065721] [fglrx] Kernel PAT support is enabled
<6>    2.065737] [fglrx] module loaded - fglrx 8.84.5 [Apr  5 2011] with 1 minors
<7>   

Hmmm, fglrx also provides keyboad and mouse drivers, correct?
Might that have caused it to sneak into initrd?

How do I rebuild the initrd, just to rule it out?
Or, is there a way to query it’s actual contents?
The Yast: INITRD_MODULLES is what is supposed to be built in, correct?

Google for “see initrd content”!

Yes. Not sure it’s the problem though. Why would you have fglrx in initrd?!
What’s the timestamp of this logfile? Do you mean /var/log/boot.msg? I don’t have /var/log/boot.message.

Found it!

In reading man mkinitrd, I noticed that their is a build log in /var/log/YAST2/mkinitrd.log

It seems to be continuously appended to, here is some recent activity containing the smoking gun:

File is /var/log/YaST2/mkinitrd.log 
2011-09-09 20:45:04 --------------------------------------------------------------------------
2011-09-09 20:45:04 mkinitrd called with arguments -s 1600x1200
2011-09-09 20:45:06 [BLOCK] /dev/sdd -> ahci
2011-09-09 20:45:06 [BLOCK] /dev/sdd -> sd_mod    
2011-09-09 20:45:07 [MODULES]    02-start.sh: pata_atiixp thermal ata_generic processor fan
2011-09-09 20:45:07 [MODULES]    02-start.sh:   
2011-09-09 20:45:07 [MODULES]    03-rtc.sh: rtc_cmos  
2011-09-09 20:45:07 [MODULES]    03-storage.sh: ext4      
2011-09-09 20:45:07 [MODULES]    05-kms.sh: fglrx    
2011-09-09 20:45:07 [MODULES]    11-block.sh: ahci sd_mod  
2011-09-09 20:45:07 [MODULES]    11-usb.sh: usbcore
2011-09-09 20:45:07 [MODULES]    11-usb.sh: ohci_hcd
2011-09-09 20:45:07 [MODULES]    11-usb.sh: uhci-hcd
2011-09-09 20:45:07 [MODULES]    11-usb.sh: ehci_hcd
2011-09-09 20:45:07 [MODULES]    11-usb.sh: usbhid                      
2011-09-09 20:45:07 [MODULES]    'modinfo -k "2.6.37.6-0.7-desktop" -F supported'  returned with an error.
2011-09-09 20:45:07 [MOUNT] Root:    /dev/disk/by-id/ata-Hitachi_HDT725032VLA360_VFM201R8081X8N-part2  Shared libs:    
2011-09-09 20:45:08 /lib/udev/bluetooth_serial is a script
2011-09-09 20:45:08 /lib/udev/bluetooth.sh is a script
2011-09-09 20:45:08 /lib/udev/collect_lvm is a script
2011-09-09 20:45:08 /lib/udev/findkeyboards is a script
2011-09-09 20:45:08 /lib/udev/idedma.sh is a script
2011-09-09 20:45:09 /lib/udev/iwlwifi-led.sh is a script
2011-09-09 20:45:09 /lib/udev/keyboard-force-release.sh is a script
2011-09-09 20:45:09 /lib/udev/kpartx_id is a script
2011-09-09 20:45:09 /lib/udev/udev-add-printer is a script
2011-09-09 20:45:09 /lib/udev/usb_modeswitch is a script
2011-09-09 20:45:09 /lib/udev/write_cd_rules is a script
2011-09-09 20:45:09 /lib/udev/write_net_rules is a script
2011-09-09 20:45:09 /lib/mkinitrd/bin/ipconfig.sh is a script
2011-09-09 20:45:09 /sbin/ifup is a script
2011-09-09 20:45:09 /lib/mkinitrd/bin/ipconfig.sh is a script
2011-09-09 20:45:09 /lib/mkinitrd/bin/linuxrc is a script
2011-09-09 20:45:09 /usr/bin/on_ac_power is a script  lib64/ld-2.11.3.so   lib64/libacl.so.1.1.0   lib64/libattr.so.1.1.0   lib64/libblkid.so.1.1.0   lib64/libcap.so.2.19   lib64/libcom_err.so.2.1   lib64/libcrypto.so.1.0.0   lib64/libc-2.11.3.so   lib64/libdevmapper.so.1.02   lib64/libdl-2.11.3.so   lib64/libe2p.so.2.3   lib64/libext2fs.so.2.4   lib64/libgcrypt.so.11.6.0   lib64/libgio-2.0.so.0.2800.0   lib64/libglib-2.0.so.0.2800.0   lib64/libgmodule-2.0.so.0.2800.0   lib64/libgobject-2.0.so.0.2800.0   lib64/libgpg-error.so.0.7.0   lib64/libgthread-2.0.so.0.2800.0   lib64/libkeyutils-1.3.so   lib64/liblzo2.so.2.0.0   lib64/libmount.so.1.1.0   lib64/libm-2.11.3.so   lib64/libncurses.so.5.7   lib64/libpcre.so.0.0.1   lib64/libpthread-2.11.3.so   lib64/libreadline.so.6.1   lib64/libresolv-2.11.3.so   lib64/librt-2.11.3.so   lib64/libselinux.so.1   lib64/libsepol.so.1   lib64/libssl.so.1.0.0   lib64/libudev.so.0.10.0   lib64/libutil-2.11.3.so   lib64/libuuid.so.1.3.0   lib64/libz.so.1.2.5   usr/lib64/libatasmart.so.4.0.3   usr/lib64/libcups.so.2   usr/lib64/libdal-0.3.so.0.0.0   usr/lib64/libdirect-1.4.so.5.0.0   usr/lib64/libdirectfb-1.4.so.5.0.0   usr/lib64/libfusion-1.4.so.5.0.0   usr/lib64/libgdk_pixbuf-2.0.so.0.2200.1   usr/lib64/libgnutls.so.26.14.12   usr/lib64/libgpod.so.4.3.1   usr/lib64/libgssapi_krb5.so.2.2   usr/lib64/libimobiledevice.so.1.0.4   usr/lib64/libk5crypto.so.3.1   usr/lib64/libkrb5.so.3.3   usr/lib64/libkrb5support.so.0.1   usr/lib64/libmtp.so.8.3.6   usr/lib64/libparted.so.0.0.1   usr/lib64/libplist.so.1.1.3   usr/lib64/libreiserfs-0.3.so.0.0.0   usr/lib64/libsgutils2.so.2.0.0   usr/lib64/libsplashycnf.so.1.0.0   usr/lib64/libsplashy.so.1.0.0   usr/lib64/libsqlite3.so.0.8.6   usr/lib64/libtasn1.so.3.1.5   usr/lib64/libusb-0.1.so.4.4.4   usr/lib64/libusb-1.0.so.0.0.0   usr/lib64/libusbmuxd.so.1.0.7   usr/lib64/libxml2.so.2.7.8   lib64/libnss_dns-2.11.3.so   lib64/libnss_dns.so.2   lib64/libnss_files-2.11.3.so   lib64/libnss_files.so.2   lib64/libgcc_s.so.1 

2011-09-10 15:37:58 --------------------------------------------------------------------------
2011-09-10 15:37:58 mkinitrd called with arguments -s 1600x1200
2011-09-10 18:23:06 --------------------------------------------------------------------------
2011-09-10 18:23:06 mkinitrd called with arguments -s 1600x1200
2011-09-10 18:27:58 --------------------------------------------------------------------------
2011-09-10 18:27:58 mkinitrd called with arguments -s 1600x1200
2011-09-10 18:31:10 --------------------------------------------------------------------------
2011-09-10 18:31:10 mkinitrd called with arguments -s 1600x1200
2011-09-27 08:37:21 --------------------------------------------------------------------------
2011-09-27 08:37:21 mkinitrd called with arguments 
2011-09-27 08:37:22 [BLOCK] /dev/sdb -> ahci
2011-09-27 08:37:22 [BLOCK] /dev/sdb -> sd_mod    
2011-09-27 08:37:22 [MODULES]    02-start.sh: pata_atiixp thermal ata_generic processor fan
2011-09-27 08:37:22 [MODULES]    02-start.sh:   
2011-09-27 08:37:22 [MODULES]    03-rtc.sh: rtc_cmos  
2011-09-27 08:37:22 [MODULES]    03-storage.sh:       
2011-09-27 08:37:22 [MODULES]    05-kms.sh: fglrx    
2011-09-27 08:37:22 [MODULES]    11-block.sh: ahci sd_mod  
2011-09-27 08:37:22 [MODULES]    11-usb.sh: usbcore
2011-09-27 08:37:22 [MODULES]    11-usb.sh: ohci_hcd
2011-09-27 08:37:22 [MODULES]    11-usb.sh: uhci-hcd
2011-09-27 08:37:22 [MODULES]    11-usb.sh: ehci_hcd
2011-09-27 08:37:22 [MODULES]    11-usb.sh: usbhid                      
2011-09-27 08:37:23 [MODULES]    'modinfo -k "2.6.37.6-0.7-desktop" -F supported'  returned with an error.
2011-09-27 08:37:23 [MOUNT] Root:    /dev/disk/by-id/ata-M4-CT064M4SSD2_0000000011210310CCC6-part3  Shared libs:    
2011-09-27 08:37:26 /lib/udev/bluetooth.sh is a script
2011-09-27 08:37:26 /lib/udev/bluetooth_serial is a script
2011-09-27 08:37:26 /lib/udev/collect_lvm is a script
2011-09-27 08:37:26 /lib/udev/findkeyboards is a script
2011-09-27 08:37:26 /lib/udev/idedma.sh is a script
2011-09-27 08:37:26 /lib/udev/iwlwifi-led.sh is a script
2011-09-27 08:37:26 /lib/udev/keyboard-force-release.sh is a script
2011-09-27 08:37:26 /lib/udev/kpartx_id is a script
2011-09-27 08:37:26 /lib/udev/udev-add-printer is a script
2011-09-27 08:37:26 /lib/udev/usb_modeswitch is a script
2011-09-27 08:37:26 /lib/udev/write_cd_rules is a script
2011-09-27 08:37:26 /lib/udev/write_net_rules is a script
2011-09-27 08:37:26 /lib/mkinitrd/bin/ipconfig.sh is a script
2011-09-27 08:37:26 /sbin/ifup is a script
2011-09-27 08:37:26 /lib/mkinitrd/bin/ipconfig.sh is a script
2011-09-27 08:37:26 /lib/mkinitrd/bin/linuxrc is a script
2011-09-27 08:37:27 /usr/bin/on_ac_power is a script  lib64/ld-2.11.3.so   lib64/libacl.so.1.1.0   lib64/libattr.so.1.1.0   lib64/libblkid.so.1.1.0   lib64/libc-2.11.3.so   lib64/libcap.so.2.19   lib64/libcom_err.so.2.1   lib64/libcrypto.so.1.0.0   lib64/libdevmapper.so.1.02   lib64/libdl-2.11.3.so   lib64/libe2p.so.2.3   lib64/libext2fs.so.2.4   lib64/libgcrypt.so.11.6.0   lib64/libgio-2.0.so.0.2800.0   lib64/libglib-2.0.so.0.2800.0   lib64/libgmodule-2.0.so.0.2800.0   lib64/libgobject-2.0.so.0.2800.0   lib64/libgpg-error.so.0.7.0   lib64/libgthread-2.0.so.0.2800.0   lib64/libkeyutils-1.3.so   lib64/liblzo2.so.2.0.0   lib64/libm-2.11.3.so   lib64/libmount.so.1.1.0   lib64/libncurses.so.5.7   lib64/libpcre.so.0.0.1   lib64/libpthread-2.11.3.so   lib64/libreadline.so.6.1   lib64/libresolv-2.11.3.so   lib64/librt-2.11.3.so   lib64/libselinux.so.1   lib64/libsepol.so.1   lib64/libssl.so.1.0.0   lib64/libudev.so.0.10.0   lib64/libutil-2.11.3.so   lib64/libuuid.so.1.3.0   lib64/libz.so.1.2.5   usr/lib64/libatasmart.so.4.0.3   usr/lib64/libcups.so.2   usr/lib64/libdal-0.3.so.0.0.0   usr/lib64/libdirect-1.4.so.5.0.0   usr/lib64/libdirectfb-1.4.so.5.0.0   usr/lib64/libfusion-1.4.so.5.0.0   usr/lib64/libgdk_pixbuf-2.0.so.0.2200.1   usr/lib64/libgnutls.so.26.14.12   usr/lib64/libgpod.so.4.3.1   usr/lib64/libgssapi_krb5.so.2.2   usr/lib64/libimobiledevice.so.1.0.4   usr/lib64/libk5crypto.so.3.1   usr/lib64/libkrb5.so.3.3   usr/lib64/libkrb5support.so.0.1   usr/lib64/libmtp.so.8.3.6   usr/lib64/libparted.so.0.0.1   usr/lib64/libplist.so.1.1.3   usr/lib64/libreiserfs-0.3.so.0.0.0   usr/lib64/libsgutils2.so.2.0.0   usr/lib64/libsplashy.so.1.0.0   usr/lib64/libsplashycnf.so.1.0.0   usr/lib64/libsqlite3.so.0.8.6   usr/lib64/libtasn1.so.3.1.5   usr/lib64/libusb-0.1.so.4.4.4   usr/lib64/libusb-1.0.so.0.0.0   usr/lib64/libusbmuxd.so.1.0.7   usr/lib64/libxml2.so.2.7.8   lib64/libnss_dns-2.11.3.so   lib64/libnss_dns.so.2   lib64/libnss_files-2.11.3.so   lib64/libnss_files.so.2   lib64/libgcc_s.so.1 


There seems to be a lot of action in this log, not intentionally by me but perhaps as a side effect of other configuration activity.(e.g. by SuseConfig?, etc)

  1. So I do need to rebuild the initrd.
    Is this appropriate code?
mkinitrd  -k /boot/vmlinuz-2.6.37.6-0.7-desktop -i /boot/initrd-2.6.37.6-0.7-desktop
  1. How does boot.fglrxrebuild deal with this - when it recompiles for good reason, such as when it detects a new rpm has been loaded by atiupgrade, seems mkinitrd would need to be triggered, but apparently is not…?

Google for “see initrd content”!

Thanks for the hint
initrd is a compressed gzip.

I made a copy in a play area, used Ark first to decompress it then Ark again to view the contents.
Yes, the old fglrx.ko is there.