Re: [Jack-Devel] driver_null_cycle
I think I understand what the driver_wait function is doing ...
It appears that the driver_wait function is trying to work out how much
time has passed since the last transport_cycle_start and if it has taken
too long, then it reports an error.
But the null_cycle is still confusing.
Matt
On 15/01/14 14:53, Matt Flax wrote:
> Is Jack requesting samples to be dropped in the null_cycle ? For
> example if I am using a ring_buffer with a DMA loading it ... is it
> necessary to fast forward the beginning of the ring buffer by nframes ?
>
> W.r.t. the driver_wait function, on the understanding that the system
> I am looking at uses a ring buffer and DMA, in your opinion, is it
> possible to re-use the dummy_driver_wait function ?
>
> thanks
> Matt
>
>
>
> On 15/01/14 14:11, Paul Davis wrote:
>>
>>
>>
>> On Tue, Jan 14, 2014 at 10:01 PM, Matt Flax <[hidden]
>> <mailto:[hidden]>> wrote:
>>
>> Hi there,
>>
>> Can someone please clarify the driver_null_cycle function for me ?
>>
>> The function is defined as so :
>> static int you_driver_null_cycle (you_driver_t *driver,
>> jack_nframes_t nframes);
>>
>> My question is w.r.t. the nframes. Is this function meant to
>> flush nframes from the audio stream ?
>>
>> Here is what is written in include/driver.h
>> The JACK engine will call this after the wait function (above) has
>> been called, but for some reason the engine is unable to execute
>> a full "cycle". the driver should do whatever is necessary to
>> keep itself running correctly, but cannot reference ports
>> or other JACK data structures in any way.
>>
>>
>> "the driver should do whatever is necessary to keep itself running
>> correctly"
>>
>> that's about the core of it. what that is depends on the driver and
>> in turn on the device.
>>
>
1389768318.17078_0.ltw:2,a <52D62DC8.6030501 at flatmax dot org>