cpufreq Archive on lore.kernel.org
 help / color / mirror / Atom feed
From: Thomas Renninger <mail@renninger.de>
To: cpufreq@lists.linux.org.uk
Cc: Dominik Brodowski <linux@dominikbrodowski.de>
Subject: DELL 600/800 - PPC frequency change issues
Date: Thu, 19 Jan 2006 18:59:22 +0100	[thread overview]
Message-ID: <43CFD37A.1040402@renninger.de> (raw)

When unplugging the AC adapter on these machines
following happens in BIOS:

- ACPI processor event fired, frequencies get limited
  to the lowest one(600 MHz) and BIOS already sets lowest freq.
- when waiting about 10 seconds another ACPI processor 
  event is fired, and all frequencies are available again,
  highest freq(1700 MHz) is set.
- when plugging in ac adapter before waiting 10 seconds
  ACPI processor event is happening immediately and all
  frequencies are available again, highest freq(1700 MHz) is set.


I found these bugs in kernel:

- speedstep centrino driver does not recognise that BIOS
  changed the frequency behind it's back.
  It tells you that 600 MHz are already set, and does not invoke
  PRE/POST transition validation. 
  On next frequency settings, the frequency is "out_of_sync" and
  the validater assumes 600 MHz and the frequency stays there.
  Not sure, maybe it stayed on 600 because of next bug or both, still
  the "out_of_sync" is valid and should be handled gracefully:

  Patch [1/2]
  
- userspace governor is using its own cpufreq_policy struct and
  forgets to set max_frequency there in 
  cpufreq_governor_userspace(CPUFREQ_GOV_LIMITS)
  That results in scaling_setspeed staying high in sysfs even the
  frequency has been lowered on _PPC change.
  When _PPC allows all frequencies again, the userspace governor
  uses it's own policy_struct with the old max value again (now the
  low one) and the frequency still stays at lowest frequency.

  Patch [2/2]

I tested with ondemand and userspace governor on a speedstep-centrino
system, both governors had problems with but should work fine with _PPC
changes whether the BIOS prechanges the freq or not.

Could someone please review.

Thanks,

      Thomas

                 reply	other threads:[~2006-01-19 17:59 UTC|newest]

Thread overview: [no followups] expand[flat|nested]  mbox.gz  Atom feed

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=43CFD37A.1040402@renninger.de \
    --to=mail@renninger.de \
    --cc=cpufreq@lists.linux.org.uk \
    --cc=linux@dominikbrodowski.de \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox