Re: [linux-audio-dev] LAAGA - avoiding fragility

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

Subject: Re: [linux-audio-dev] LAAGA - avoiding fragility
From: Paul Davis (pbd_AT_Op.Net)
Date: Mon May 07 2001 - 17:01:21 EEST


Iain writes:

>This is not a new concern (to the general topic) - and some possible ways
>round it have been mooted in previous discussion. I bring it up here in
>case anyone thinks that the API needs to accommodate provisions for it.
>
>==
>
>In the case that a process has SCHED_FIFO - if an element in that process
>crashes it can lock up the entire machine. Taking us back to the kind of
>problem that people criticise MacOS/'doze for.
>
>One solution is to have a separate process monitoring the server and killing
>it if it locks up.
>
>Is there a better way?
>should provision for this be part of the API?
>(or integral to the server).

its definitely a source of concern, though the concern is not the
thread "crashing" but getting into a boundless loop. i don't think it
should be part of the API, but a sensible server implementation might
well fork a super-high-priority RT_FIFO thread that wakes up every 30
seconds to check on the "real" RT_FIFO threads that the server has
started. this has been discussed before, and seems to me to be mostly
an implementation detail, though a very important one for people not
using dual CPU boxes (which, of course, everybody doing serious audio
should be :)

--p


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

This archive was generated by hypermail 2b28 : Mon May 07 2001 - 17:21:36 EEST