Subject: Re: [linux-audio-dev] mutex in process callback?
From: Lukas Degener (AFBLukas_AT_gmx.de)
Date: Thu Mar 27 2003 - 17:20:40 EET
>
>
>if i understand what you're asking correctly, you could just stash a
>copy of the current pointer at the start of the process callback. then
>(assuming there is only one pointer), the operation is atomic with
>respect to the process callback: either it happens before you get the
>copy of the pointer or not at all (since your copy of the pointer
>won't be updated until the next time the callback is run).
>
Ah... you're right. Actualy this is pretty much straight forward.
Funny i didn't think of it myself :-)
There's one thing however, i am not completly sure about:
What happens if the working thread swaps buffers two times within one
cycle of the
playback thread? I'll try to do a simple diagram:
(Time) ---->
Playback
---------------------------------------------------------------------------------------------->
bang!
front==A ...
reading.....................................................still
reading from A
Worker
----------------------------------------------------->Worker------------------------------>
bang!
back==B, ...working on B...done...swap
back==A!...working on A
May not be a very common situation, but who knows. So i would have to
assure that
the worker thread does not process more than one "job" during one
playback cycle.
Or use more than two buffers.
Or did i miss something here?
>if more
>than one pointer is involved stick the sets into arrays and stash the
>index to the current set instead of the pointers themselves.
>
This will propably not be the case. At least i think so right now.
>hopefully that makes sense and is applicable.
>
It is indeed. I will try it.
Ah... another reply just droped into my mailbox <beep!> ... let's see
what Paul suggests. :-)
Thanks Rob,
Lukas
This archive was generated by hypermail 2b28 : Thu Mar 27 2003 - 17:34:58 EET