Subject: Re: [linux-audio-dev] Re: high level laaga
From: Steve Harris (S.W.Harris_AT_ecs.soton.ac.uk)
Date: Fri May 11 2001 - 17:50:59 EEST
On Fri, May 11, 2001 at 03:17:48PM +0200, Abramo Bagnara wrote:
> > Kai's site describes the problems with using IPC for the
> > interconnection. So we need something else.
>
> the theoretical extra cost is 2*(N-1) context switch (where N is process
> count). I'd like to know if indeed this break everything.
>
> That apart I'm strongly convinced that the right architecture need to
> handle both cases (one-thread and multiple-process) appropriately.
My feeling is that its better to provide only one method of achieving the
same end, I think the complications of allowing both single and multiple
process methods would be large.
However, if I understand the model correctly, there is nothing preventing
a LAAGA application from scribbing on data malloc'd by another LAAGA app
in the single thread model, isn't this likly to be a problem? It reasonable
to shout at people who write buggy, simple LADSPA plugins, but full on HD
recorders (for example) are a damn sight more complicated and could easily
tread on each others toes. Not to mention the extra complications for
debugging.
- Steve
This archive was generated by hypermail 2b28 : Fri May 11 2001 - 17:11:23 EEST