Re: a joint letter on low latency and Linux

[Paul Barton-Davis]

>> >I'm adding a feature to our RTLinux consumer. It collects a frame
>> >and dumps it, then it looks at the queue -- if the queue is at a low
>> >water mark, the RT consumer reaches down into Linux, sets need_resched and 
>> 
>> there is no queue. this is real time FX processing or synthesis,
>> remember ? 
>
>I'm confused. I thought that buffering was possible.
>Get data -> send it to processor -> output result
>             couple of seconds of
>             buffer here.

no, the h/w buffer is on the order of 1ms-120ms, depending on how its
configured here. every time the interrupt arrives, you have to (1)
collect the next set of data and (2) deliver a new set. For real time
processing, the buffering between (1) and (2) can't go much above 3ms
before any serious audio user will complain.

the "couple of seconds" of buffer is whats used to store data coming
from disk so that disk i/o jitter is invisible.

--p