linux-kernel - Re: scheduling policy was: Re: 2.6.37 considered stable?

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]

Date:	Sat, 5 Feb 2011 13:27:05 +0100
From:	"Dr. Martin Rogge" <marogge@...inehome.de>
To:	Mike Galbraith <efault@....de>
Cc:	Gene Heskett <gene.heskett@...il.com>, linux-kernel@...r.kernel.org
Subject: Re: scheduling policy    was: Re: 2.6.37 considered stable?

Mike Galbraith wrote:

>> Gene, have you tried schedtool to set the scheduling policy of the
>> process to SCHED_FIFO or even SCHED_ISO?
> 
> /*
>  * Scheduling policies
>  */
> #define SCHED_NORMAL            0
> #define SCHED_FIFO              1
> #define SCHED_RR                2
> #define SCHED_BATCH             3
> /* SCHED_ISO: reserved but not implemented yet */
> #define SCHED_IDLE

When I mentioned SCHED_ISO I was making the mental note "provided you are 
running a scheduler that supports SCHED_ISO" as I do. 

> You only want to run apps RT if they really really need it though.  In
> many cases, running an app RT will reduce performance, and may well
> increase latencies.  (deterministic != quick like a bunny;)

I agree with that statement, and on all machines I own the video applications 
run under SCHED_NORMAL nice 0. However, on machines where the audio/video 
performance is borderline, it may help to tweak the scheduling policy (save 
buying new hardware).

Regarding "quick like a bunny": depending on the KPI (forgive the corporate 
lingo ;-) you are looking at, it means different things. For me as a desktop 
user, it is determined by the "snappiness" and "smoothness" of the user 
interface and all other applications with direct user interaction. Throughput 
KPIs are less relevant, although still important for foreground applications.

A solution that gives me that, to the extent that compiling large projects in 
the background is unnoticable to the user experience, is superior to others. 
Unfortunately, in my experience, auto-this and auto-that is not always the 
answer.

Anyway, all I wanted to do is to encourage Gene to try something out and see 
if it helps.

Martin

--
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/