linux-kernel - [PATCH 00/10] scheduler-ticker > 1000Hz, nitromethane for linux ;-)

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-next>] [day] [month] [year] [list]

Message-ID: <4F4B7CC6.1030307@tu-ilmenau.de>
Date:	Mon, 27 Feb 2012 13:53:26 +0100
From:	Stephan Bärwolf 
	<stephan.baerwolf@...ilmenau.de>
To:	linux-kernel@...r.kernel.org
Subject: [PATCH 00/10] scheduler-ticker > 1000Hz, nitromethane for linux ;-)

Hi all,

this is a not fully serious (this is why I only post it here without sendig it
directly via mail) patchset for:

        * tune the scheduler at configuration point
        * increase the ticker-frequency way above 1000Hz
        * tune the ticker-freq. from userspace during runtime

Of course this increases your systems overhead, but it enables you to have
more timer-triggered contexswitches per second and so (?)higher(?) interactivity.
(Of course after tuning some scheduler-parameters, min_gran, wakeup_gran ...)

An interesting feature might be the userspace controlling of the tickerfreq, too.
(Without disturbing HZ-dependend kernelcode, of course.)

I don't have very high hopes for an mainline integration ;-) but perhaps
there are people, who want to play with such stuff.
The Code is more or less an excerise and perhaps some reviewing would
be nice - so please contact me.
If (in the very unlikely case) there is enough interest in it, I can cleanup
the code and start maintaining it...

* The patch is diffed against 3.2.7
* The whole patchset is also downloadable as an bzipped archive at:
  http://www.matrixstorm.com/software/linux/nitromethane/linux-3.2.7-patches.tar.bz2

So now, please blame me "we don't need such s*it, it causes only overhead"... ;-)

regards Stephan

pub   4096R/FB9B12DB 2011-10-14 Stephan Bärwolf <stephan.baerwolf@...ilmenau.de>
 Primary key fingerprint: 2252 B41E 12DF 41AD 476F  7F36 084A 0911 FB9B 12DB



Stephan Baerwolf (10):
  menuconfig: implement new submenu for extended scheduler-options
  sched: implement the new menuconfig controlling of sched_tunables
  menuconfig: extend nitro-menu by responsiveness nitro options
  sched: implement the controlling of responsiveness-nitro-options
  menuconfig: extend nitro-menu by HZ-frequency Nitro
  hrtimer: implement the controlling of nitro-hzboost
  menuconfig: extend nitro-menu by priority-nitro
  sched: implement the controlling of priority-nitro
  menuconfig: extend nitro-menu by userspace-multiplier
  hrtimer/sysctl: implement the controlling of userspace multiplier

 arch/x86/Kconfig            |    4 +-
 include/linux/sched.h       |    8 ++
 kernel/Kconfig.hz           |    3 -
 kernel/Kconfig.sched        |  202 +++++++++++++++++++++++++++++++++++++++++++
 kernel/sched.c              |    9 ++
 kernel/sched_fair.c         |   58 +++++++++++--
 kernel/sysctl.c             |   16 ++++
 kernel/time/tick-common.c   |   33 +++++++
 kernel/time/tick-internal.h |   20 ++++
 kernel/time/tick-sched.c    |   51 +++++++++++
 10 files changed, 390 insertions(+), 14 deletions(-)
 create mode 100644 kernel/Kconfig.sched

-- 
1.7.3.4
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/