powernowd upgrade fails: "cannot create /sys/devices/system/cpu/cpu1//cpufreq/scaling_governor: Directory nonexistent"

[Steve]

Binary package hint: update-manager

Used the upgrade to update to 7.04

The following dialog popped up during install:

Could not install 'powernowd'

The upgrade aborts now. Please report this bug against the 'update-manager' package and include the files in /var/log/dist-upgrade/ in the bugreport.

subprocess post-installation script returned error exit status 2

Last few lines of main.log:

2007-04-19 21:32:04,108 DEBUG removing bad script '/var/lib/dpkg/info/apache2-common.prerm'
2007-04-19 21:52:12,190 DEBUG got a conffile-prompt from dpkg for file: '/etc/bash.bashrc'
2007-04-19 22:03:12,533 DEBUG got a conffile-prompt from dpkg for file: '/etc/modprobe.d/blacklist'
2007-04-19 22:07:29,770 ERROR got an error from dpkg for pkg: 'powernowd': 'subprocess post-installation script returned error exit status 2
'
2007-04-19 22:24:39,682 WARNING no activity on terminal for 240 seconds (Installed ubuntu-desktop)

Relevant lines of term.log:

Setting up powernowd (0.97-1ubuntu7) ...
Installing new version of config file /etc/init.d/powernowd ...
 * Starting powernowd...
 * Starting powernowd... /etc/init.d/powernowd: 156: cannot create /sys/devices/system/cpu/cpu1//cpufreq/scaling_governor: Directory nonexistent
                                                                         [fail]
invoke-rc.d: initscript powernowd, action "start" failed.
dpkg: error processing powernowd (--configure):
 subprocess post-installation script returned error exit status 2

ProblemType: Package
Date: Thu Apr 19 22:07:30 2007
ErrorMessage:
 ErrorMessage: subprocess post-installation script returned error exit status 2
Package: powernowd
SourcePackage: powernowd

[Michael Vogt (mvo)]

Thanks for your bugreport.

Can you please attach the files in /var/log/dist-upgrade ? This will help us diagnose the problem.

Thanks,
 Michael

[Ed C (echehovin)]

I have the same problem while upgrading Ubuntu to version 7.04 on my AMD Athlon 64 X 2 Dual Core. I have attached the requested log files. Please contact me if I can provide additional information.

[Ed C (echehovin)]

It seems I can only attach one log file at a time so here is the "apt.log" file.

[Ed C (echehovin)]

Attaching the term.log file.

[Steve (stb20)]

I'm using a Dell Inspiron 1501 Laptop with a AMD Turion 64 X2 processor.

[Steve (stb20)]

Other file

[Steve (stb20)]

Final file

[Michael Vogt (mvo)]

Thanks for this additional information. Confirming the bug now:

Setting up powernowd (0.97-1ubuntu7) ...
Installing new version of config file /etc/init.d/powernowd ...
 * Starting powernowd...
 * Starting powernowd... /etc/init.d/powernowd: 156: cannot create /sys/devices/system/cpu/cpu1//cpufreq/scaling_governor: Directory nonexistent
                                                                         [fail]
invoke-rc.d: initscript powernowd, action "start" failed.
dpkg: error processing powernowd (--configure):
 subprocess post-installation script returned error exit status 2

[Daniel Hahler (blueyed)]

Can you provide the output of "ls -l /sys/devices/system/cpu/cpu[0-9]*"?
This looks like it works for cpu0, but not cpu1.

[Daniel Hahler (blueyed)]

This sounds like bug 111375. Is it a duplicate?

[James Athey (james-athey)]

I am seeing similar behavior using a fresh install of Gutsy. My CPU is an Athlon 64 X2 3800+.

# /etc/init.d/powernowd start
 * Starting powernowd... /etc/init.d/powernowd: 156: cannot create /sys/devices/system/cpu/cpu0//cpufreq/scaling_governor: Directory nonexistent
 * CPU frequency scaling not supported

# ls -l /sys/devices/system/cpu/cpu*
/sys/devices/system/cpu/cpu0:
total 0
drwxr-xr-x 5 root root 0 2007-12-04 23:11 cache
-r-------- 1 root root 4096 2007-12-04 23:47 crash_notes
drwxr-xr-x 2 root root 0 2007-12-04 23:12 topology

/sys/devices/system/cpu/cpu1:
total 0
drwxr-xr-x 5 root root 0 2007-12-04 23:11 cache
-r-------- 1 root root 4096 2007-12-04 23:47 crash_notes
-rw-r--r-- 1 root root 4096 2007-12-04 23:47 online
drwxr-xr-x 2 root root 0 2007-12-04 23:12 topology

[Ed C (echehovin)]

dAniel,

There was not enough information for bug 111375 to tell if it is the same.
The author just said "I can't get powernowd to work" and then went on to
explain some diagnostics that he tried.

I did not experience any problems with "powernowd" when I upgraded to
Gutsy.

uname -a
Linux Seafarer 2.6.22-14-generic #1 SMP Sun Oct 14 23:05:12 GMT 2007 i686
GNU/Linux

sh -x /usr/share/powernowd/cpufreq-detect.sh
+ /usr/sbin/laptop-detect
+ CPUINFO=/proc/cpuinfo
+ IOPORTS=/proc/ioports
+ [ ! -f /proc/cpuinfo ]
+ grep ^model name /proc/cpuinfo
+ head -1
+ sed -e s/^.*: //;
+ MODEL_NAME=AMD Athlon(tm) 64 X2 Dual Core Processor 3800+
+ grep -E ^cpud[^:]+: /proc/cpuinfo
+ head -1
+ sed -e s/^.*: //;
+ CPU=
+ grep -E ^vendor_id[^:]+: /proc/cpuinfo
+ head -1
+ sed -e s/^.*: //;
+ VENDOR_ID=AuthenticAMD
+ sed -e /^cpu family/ {s/.*: //;p;Q};d /proc/cpuinfo
+ CPU_FAMILY=15
+ MODULE=none
+ MODULE_FALLBACK=acpi-cpufreq
+ [ -f /proc/ioports ]
+ grep -q Intel .*ICH /proc/ioports
+ PIII_MODULE=speedstep-smi
+ MODULE=powernow-k8

Respectfully,
Ed

--- dAniel hAhler <email address hidden> wrote:

> This sounds like bug 111375. Is it a duplicate?
>
> --
> powernowd: "cannot create
> /sys/devices/system/cpu/cpu1//cpufreq/scaling_governor: Directory
> nonexistent"
> https://bugs.launchpad.net/bugs/107820
> You received this bug notification because you are a direct subscriber
> of the bug.
>

      ____________________________________________________________________________________
Be a better friend, newshound, and
know-it-all with Yahoo! Mobile. Try it now. http://mobile.yahoo.com/;_ylt=Ahu06i62sR8HDtDypao8Wcj9tAcJ

[Daniel Hahler (blueyed)]

Ed, I think bug 111375 is related to the x86_64 (amd64) kernel, but you are using the i686/generic one, just like me (I have a AMD Athlon(tm) 64 Processor 3000+ and am not affected by this bug). It appears to be a similar issue.

I think here are already similar issues being mixed up.
The really strange thing in the initial report from Steve (and also reported from Michael), is that it fails for "cpu1" and not for "cpu0" already. But there has been no "ls"-feedback from them yet.

The general problem appears to be that the "cpu directories" are not made available by the kernel (which is the same as bug 111375, but may be slightly different). This would be a kernel bug then.

Please post the output of:
$ sudo modprobe -r powernow-k8
$ sudo modprobe powernow-k8
$ dmesg | grep powernow

[Daniel Hahler (blueyed)]

The powernowd bug (upgrade/install fails) should be fixed by the upload to my PPA: https://launchpad.net/~blueyed/+archive/+index?field.name_filter=powernow
Please test it and provide feedback if it works (better). Thanks.
(The actual patch is attached to bug 67341)

[Ed C (echehovin)]

--- dAniel hAhler <email address hidden> wrote:
> Please post the output of:
> $ sudo modprobe -r powernow-k8
> $ sudo modprobe powernow-k8
> $ dmesg | grep powernow

$ sudo modprobe -r powernow-k8
FATAL: Module powernow_k8 is in use.

$ sudo modprobe powernow-k8

$ dmesg | grep powernow
[ 19.636000] powernow-k8: Found 2 AMD Athlon(tm) 64 X2 Dual Core
Processor 3800+ processors (version 2.00.00)
[ 19.636000] powernow-k8: 0 : fid 0xc (2000 MHz), vid 0xa
[ 19.636000] powernow-k8: 1 : fid 0xa (1800 MHz), vid 0xc
[ 19.636000] powernow-k8: 2 : fid 0x2 (1000 MHz), vid 0x12

      ____________________________________________________________________________________
Looking for last minute shopping deals?
Find them fast with Yahoo! Search. http://tools.search.yahoo.com/newsearch/category.php?category=shopping

[Daniel Hahler (blueyed)]

Ed, I'm not getting the "FATAL: Module powernow_k8 is in use." error when removing the module.
Because this fails for you the dmesg output only contains the lines from booting up, which appear to look OK.
What does "ls -l /sys/devices/system/cpu/cpu*" say for you?
What's the error when doing "sudo invoke-rc.d powernowd restart"?

[Ed C (echehovin)]

Daniel,

Below is the output you requested. I am happy to do whatever I can to
help out.

Respectfully,
Ed

--- dAniel hAhler <email address hidden> wrote:

> What does "ls -l /sys/devices/system/cpu/cpu*" say for you?
> What's the error when doing "sudo invoke-rc.d powernowd restart"?

$ ls -l /sys/devices/system/cpu/cpu*
/sys/devices/system/cpu/cpu0:
total 0
drwxr-xr-x 5 root root 0 2007-12-08 15:38 cache
drwxr-xr-x 4 root root 0 2007-12-08 20:38 cpufreq
-r-------- 1 root root 4096 2007-12-08 20:38 crash_notes
drwxr-xr-x 2 root root 0 2007-12-08 20:38 topology

/sys/devices/system/cpu/cpu1:
total 0
drwxr-xr-x 5 root root 0 2007-12-08 15:38 cache
lrwxrwxrwx 1 root root 0 2007-12-08 20:38 cpufreq ->
../../../../devices/system/cpu/cpu0/cpufreq
-r-------- 1 root root 4096 2007-12-08 20:38 crash_notes
-rw-r--r-- 1 root root 4096 2007-12-08 20:38 online
drwxr-xr-x 2 root root 0 2007-12-08 20:38 topology

$ sudo invoke-rc.d powernowd restart
 * Stopping powernowd: [
OK ]
 * Starting powernowd... [
OK ]

      ____________________________________________________________________________________
Looking for last minute shopping deals?
Find them fast with Yahoo! Search. http://tools.search.yahoo.com/newsearch/category.php?category=shopping

[Daniel Hahler (blueyed)]

Oh... I've missed the point that it fails apparently only during upgrade for you. Is this correct?

I don't know why this may happen during upgrade, but it should be covered then by the set of fixes, which are attached to bug 67341 - there's a check added, which prevents the "cannot create" error and therefor the failure of the script.

[Ed C (echehovin)]

dAniel,

The only time I recall experiencing the problem was when I upgraded to
Feisty. I did not have the problem when I upgraded to Gutsy.

I noticed that there is a double forward slash "//" before "cpufreq" in
the directory path that is giving the "Directory nonexistent" error. Any
idea of the purpose for the "//"? It seems to work if I cd to that path.

Also I noticed that on my machine the directory
/sys/devices/system/cpu/cpu1/cpufreq is a symbolic link to the same
directory for cpu0. Is this directory a symbolic link on your machine?

--- dAniel hAhler <email address hidden> wrote:

> Oh... I've missed the point that it fails apparently only during upgrade
> for you. Is this correct?
>
> I don't know why this may happen during upgrade, but it should be
> covered then by the set of fixes, which are attached to bug 67341 -
> there's a check added, which prevents the "cannot create" error and
> therefor the failure of the script.
>
> ** Summary changed:
>
> - powernowd: "cannot create
> /sys/devices/system/cpu/cpu1//cpufreq/scaling_governor: Directory
> nonexistent"
> + powernowd upgrade fails: "cannot create
> /sys/devices/system/cpu/cpu1//cpufreq/scaling_governor: Directory
> nonexistent"
>
> --
> powernowd upgrade fails: "cannot create
> /sys/devices/system/cpu/cpu1//cpufreq/scaling_governor: Directory
> nonexistent"
> https://bugs.launchpad.net/bugs/107820
> You received this bug notification because you are a direct subscriber
> of the bug.
>

      ____________________________________________________________________________________
Looking for last minute shopping deals?
Find them fast with Yahoo! Search. http://tools.search.yahoo.com/newsearch/category.php?category=shopping

[dlstyley (deaston)]

I'm seeing the same thing on fresh install of Gutsy. I get the exact same error (with the double slashes in the path). It's worth mentioning, however that this is on an old laptop that has a Celeron processor. Yet I do have the same symptom.

[Steve Kowalik (stevenk)]

powernowd (0.97-2ubuntu2) hardy; urgency=low

  [ dAniel hAhler ]
  * debian/init.d:
    - skip use_ondemand if $OPTIONS != "-q" are given (LP: #67341)
    - activate [...]/cpufreq/ondemand/ignore_nice_load in
      use_ondemand, when using ondemand governor instead of powernowd
      (LP: #60042)
    - Use /sys/devices/system/cpu/cpu[0-9]*/ to match CPUs and
      test if it is a directory (LP: #173711, LP: #107820)
    - Do not use "double slashes" in paths to avoid confusion
  * debian/cpufreq-detect.sh:
    - load e_powersaver if CPU has est flag (LP: #117446)
    - fix comparison for CPU_FAMILY=6 (LP: #162524)
    - use /bin/sh instead of /bin/bash for clarification
      (LP: #162524)
    (Patch from Jacob Nielsen, Ralph Janke)

  [ Steve Kowalik ]
  * Correct spelling of Original in Original-Maintainer.

 -- Steve Kowalik <email address hidden> Tue, 18 Dec 2007 19:49:36 +1100

[Kenneth Belitzky (httpdss)]

isnt this a duplicate of bug 85946 ?

[Daniel Hahler (blueyed)]

Kenny, yes, it's similar. I've also closed it now and left instructions about reporting any remaining issues (which are likely kernel problems then).