Cause
A change in the kernel introducing the intel_pstate driver,
incompatible with how scaling was managed until 6.7, was brought in.
Consequence
The cpuspeed service printed, albeit only cosmetic, errors
during system boot and shutdown.
Fix
Since the intel_pstate driver handles performance scaling on
its own, the service is now effectively disabled on platforms using it.
Result
The service initscript was altered so it bails out on platforms
using the intel_pstate driver. Users may also safely disable it
entirely.
Created attachment 1029918[details]
boot.log
Description of problem:
After the installation of RHEL-6.7-Snapshot-1, during booting into any rc mode list of error messages "/etc/rc5.d/S13cpuspeed: line91: echo: write error: Invalid argument" are seen.
Version-Release number of selected component (if applicable):
cpuspeed-1.5-20.el6_4
How reproducible:
Always while booting into rc modes like rc5, rc4, rc3, rc2 and rc1.
Steps to Reproduce:
1.Install RHEL-6.7-Snapshot-1
2.Boot into the installed OS. (Generally in run level 5)
3.Check the boot.log messages for error messages for current run level.
Actual results:
Error messages by cpuspeed are seen.
"/etc/rc5.d/S13cpuspeed: line91: echo: write error: Invalid argument"
Starting cpuspeed [failed]
Expected results:
No Error messages. cpuspeed should work properly.
Additional info:
#cat /sys/devices/system/cpu/cpu*/cpufreq/scaling_available_governors
performance powersave
#cat /sys/devices/system/cpu/cpu*/cpufreq/scaling_governor
powersave
Seems like /etc/rc5.d/S13cpuspeed is trying to insert values like "userspace" or "ondemand" or "conservative" and is failing because its not supported by the system.
Similar issue was listed in https://access.redhat.com/solutions/328853
Thank you for the report. This is a known issue affecting Intel platforms caused by changes in the kernel.
This is of a purely cosmetic nature and you can safely disable the cpuspeed service as scaling and performance management is now done by the kernel driver directly.
(In reply to Sujith from comment #3)
> In that case, can you please disable this service from init.d, just to
> prevent any confusion from user-end?
Yes, something along those lines is the planned fix for this problem.
Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.
For information on the advisory, and where to find the updated
files, follow the link below.
If the solution does not work for you, open a new bug report.
https://rhn.redhat.com/errata/RHBA-2015-1440.html