Re: [linux-audio-dev] user lowlatency kernel experience

New Message Reply About this list Date view Thread view Subject view Author view Other groups

Subject: Re: [linux-audio-dev] user lowlatency kernel experience
From: Joakim Verona (joakim_AT_verona.se)
Date: Thu Jun 14 2001 - 15:27:58 EEST


Yes,

Well perhaps it is great, it's just that when the machine dies because
of a runaway process, i'm not happy. I'm not used to this under linux,
and it was mainly this that made me switch from windows/amiga environments.

the amiga os(based on tripos), for instance, achieved a lot in very
little ram. it had the fatal flaw that all processes had to cooperate,
or the machine would die. it wasnt as bad as win311, but anoying anyway.

i just want a higher level supervisor to run, that is external to all
applications and applications *never* can fool, that can fix problems if
they occur.

Perhaps it is possible to apporximate this with my existing apps using
this apporach:

1) start a daemon with the highest priority, that in turn starts
timidity and csound.

2) implement watchdog in the daemon, and if the children never yields,
kill them.

i understand that this is supposed to be handled by the various api
proposals discussed here, im just thinking aloud from my limited
experience of this problem domain.

/joakim

Benno Senoner wrote:

> On Wednesday 13 June 2001 15:30, Joakim Verona wrote:
>
>
>>so, since people here seem to think that the low-latency kernel is
>>really great, i suspect i have missed something. is there a way to get
>>processes to get sceduled promptly, but at the same time not hog the cpu
>>if they try?
>>
>
> not seems to think, IT IS GREAT
> :-)
>
> Regarding your problems: it's not the fault of the lowlatency kernel that
> your application hangs but it is the application itself that must handle CPU
> overload situations properly.
> For example the app could run a higher priority watchdog thread to see if a
> normal thread (without realtime privileges) still gives signs of life.
> If the CPU gets overloaded the app could either stop the audio, shut down
> itself , mute a few voices/ audio stream or whatever.
>
> Developing "hard" realtime applications is not an easy task, especially in
> the audio world where the CPU load varies in an unpredictable fashion.
>
> Benno.
>

-- 
Joakim Verona

www.verona.se joakim_AT_verona.se


New Message Reply About this list Date view Thread view Subject view Author view Other groups

This archive was generated by hypermail 2b28 : Thu Jun 14 2001 - 17:15:39 EEST