Subject: [linux-audio-dev] Re: rtcd vs hz > 100!
From: Paul Barton-Davis (pbd_AT_Op.Net)
Date: ke loka 27 1999 - 21:03:12 EDT
>Hmm! I'm worried that milliseconds may end up too constraining.
>Also, one-shots with absolute times seems more reliable to me since it
>factors out the time to get the request to rtcd.
but for periodic clients, that is almost completely irrelevant: they
will set themselves up for rtc notifications just before they "hunker
down for the long haul" - any delay in sending the request happens
only once. if the interval is large, there will be no effect at
all. if the interval is small, they might take a hit for the the first
notification.
as for milliseconds, i'm not stuck on them. reasonably long durations
happen to fit into a 32 bit integer. we could use timespec for the
interval too - thats fine. gets rid of the union too - oh, how i miss
anonymous unions in C, just another reason to love C++ :)
>There's one idea I'm afraid to mention since I think *you'll* probably
>want it! :) rtcd could (optionally) signal pthread condition variables
>in shared memory.
not me! the overhead for handling a condition variable is quite a bit
higher, since it involves sending a signal, and a (pthread-internal)
signal handler on the other end deciding what to do.
This archive was generated by hypermail 2b28 : pe maalis 10 2000 - 07:28:00 EST