[pulseaudio-discuss] [PATCH 3/4] core: no rewinding on volume change if the sink does not support it.

Jyri Sarha oku at iki.fi
Mon Apr 11 02:14:04 PDT 2011

On Fri, 8 Apr 2011, Colin Guthrie wrote:
> 'Twas brillig, and oku at iki.fi at 08/04/11 15:18 did gyre and gimble:
>> From: Jyri Sarha <jyri.sarha at nokia.com>
> Would it not be possible to move this check into the function itself and
> just make it a noop?

The reason for this patch in the first place was the load spikes when a
user took his time to adjust volume while listening music. This fixed
the problem and did not give it any further consideration, but I can
now see that it is not the most complete fix.

It appears to me that if "flush" parameter of pa_sink_input_request_rewind() 
call is not true (IOW, reuse the rewound samples) then we could ignore
the the whole request if sink does not support rewinding. However, the
logic of different boolean parameters for pa_sink_input_request_rewind() 
is not completely clear to me.

What do you think?

> I appreciate there would be FCO but it would reduce the likelihood of
> this check being missed on some other calls... unless that is desirable
> sometimes?

As far as I can see the only cases we want to rewind, if sink does not
support it, is when we want to throw away the already queued samples
and not just process them again with new SW volume etc.


More information about the pulseaudio-discuss mailing list