linux-kernel - Re: cpufreq 'choice' Kconfig oddness in 2.6.22-rc6-mm1..

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <17231.1183866569@turing-police.cc.vt.edu>
Date:	Sat, 07 Jul 2007 23:49:29 -0400
From:	Valdis.Kletnieks@...edu
To:	Mattia Dongili <malattia@...ux.it>
Cc:	linux-kernel@...r.kernel.org
Subject: Re: cpufreq 'choice' Kconfig oddness in 2.6.22-rc6-mm1..

On Sat, 07 Jul 2007 15:50:01 +0900, Mattia Dongili said:
> On Sat, Jul 07, 2007 at 12:35:10AM -0400, Valdis.Kletnieks@...edu wrote:
> > So I was looking at drivers/cpufreq/Kconfig, and wondering...
> > 
> > Is there a specific reason why we have a 'choice' clause that allows selecting
> > 'performance' or 'userspace' as the default governor, and no obvious way to
> > select powersave, ondemand, or conservative as a default? Or was this an
> 
> Yes, ondemand and conservative may refuse to load/work on systems with
> a high transition latency and that is sometimes discovered only at
> runtime. This makes them not selectable as default governors.

D'Oh! :)  I was trying to set ondemand as the default, only after verifying
that ondemand works as desired on my laptop.  I forgot that we probably need
to also allow for people who don't test before making things default....

> > oversight I should cook up a patch for?  Looks like the Kconfig and a bit
> > of cut-n-paste in include/linux/cpufreq.h is all that's needed?
> 
> No, some more work is needed (e.g.: falling back to a different governor
> if the selected fails). Thomas Renninger has posted patches on
> cpufreq-devel to accomplish that.

OK - I'll keep an eye open and test them out whenever they hit -mm.
In the meantime, an 'echo ondemand' into the right /sys file is working
for me as a workaround.

Content of type "application/pgp-signature" skipped