Bug 5860 - ondemand and speedstep-ich fail on inspiron 8500
Summary: ondemand and speedstep-ich fail on inspiron 8500
Status: CLOSED PATCH_ALREADY_AVAILABLE
Alias: None
Product: Power Management
Classification: Unclassified
Component: cpufreq (show other bugs)
Hardware: i386 Linux
: P2 normal
Assignee: Éric Piel
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2006-01-10 05:37 UTC by Simon
Modified: 2006-01-17 08:25 UTC (History)
0 users

See Also:
Kernel Version: 2.6.14/15
Subsystem:
Regression: ---
Bisected commit-id:


Attachments

Description Simon 2006-01-10 05:37:01 UTC
Most recent kernel where this bug did not occur:
Distribution: Fedora Core 4
Hardware Environment: Dell Inspiron 8500
Software Environment:
Problem Description:

The ondemand (and conservative) scaling governors can be loaded from the modules
and are listed as available governors, but they will not work. echo "ondemand" >
scaling_governor has no effect.
Fedora kernels ship with the speedstep-ich module built into the kernel. Other
drivers are provided as modules, but cannot be loaded as the speedstep-ich
claims it works with my chipset and so blocks the module loading.

I did file a bug report with redhat  which contains some more details.

https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=165494

Steps to reproduce:
1) Install FC4 on a I8500. 
2) Load the ondemand module. 
3) Attempt to use the ondemand governor.
Comment 1 Simon 2006-01-17 06:15:47 UTC
I have also tested this on kernel 2.6.15 and the speedstep-ich module is still
exhibits the same behaviour.
Comment 2 Éric Piel 2006-01-17 06:50:24 UTC
Yes, that's completely "normal". speedstep-ich was not setting information about
the transition latency in the 2.6.15 neither. Mattia Dongili has done the work
to fix this... and you are (somehow) lucky, it's available in the 2.6.16-rc1
that Linus has released an hour ago ;-)
Comment 3 Simon 2006-01-17 08:15:34 UTC
Nice, speedstep-ich in 2.6.16-rc1 worked as expected
Comment 4 Éric Piel 2006-01-17 08:25:03 UTC
Good! Well, now next one is speedstep-smi, but then it would first require to
see  if it's possible to reduce the maximum transition latency (which is 1s)...

Note You need to log in before you can comment on or make changes to this bug.