[pulseaudio-discuss] should prebuf always be greater than minreq ?

[Arun Raghavan]

On 10 Oct 2014 02:24, "Alban Browaeys" <prahal at yahoo.com> wrote:
>
> Le jeudi 09 octobre 2014 à 18:46 +0530, Arun Raghavan a écrit :
> > Strong nack for fixing without figuring it out. Can we have a bug
> > report please?
> >
> > -- Arun
>
> Agreed about not applying a patch/fix (none were send and I made it
> explicit I had only a workaround).
>
> All I ask is information to know if there might be a bug in pulseaudio.

Sorry if I sounded harsh there, that wasn't the intention.

> I asked if minreq should be <= prebuf , as the only proportion between
> those two I found was to init "prebuf = tlength" and "minreq=tlength/5".
> This 1 to 5 proportion I found in "PULSE_LATENCY_MSEC" code.
>
> And the documentation tell about minreq "will initialize this to a value
> that is deemed sensible by the server".
>
>
> In short I look for what would be "deemed sensible by the server".
>
> Should I send a bug report about that or is the ML more appropriate a
> field ?

There should not be an explicit correlation between minreq and prebuf. The
only real thing to consider is whether it is large enough to guarantee
we'll be able to fill more of the buffer before the prebuf value is
consumed.

If you can construct a small test case with the values that expose this
issue, probably a good idea to file a bug so we can try to dig deeper (and
possibly make a unit test too).

Cheers,
Arun
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.freedesktop.org/archives/pulseaudio-discuss/attachments/20141010/3b6be140/attachment.html>