Re: [Jack-Devel] BUG REPORT : Overflow of XRUNs after a certain time.

> Looks like the problem is associated with raw MIDI.
> 
> I have jackd running on my test system without the -Xraw option. It has
> now run well past the point where the lockup occurred with -Xraw, without
> any problem.

I confirm this. Which seems indeed logical as the problem occurs with no 
client recorded apart from the MIDI ports.

Meanwhile, I am tracing the problem and would be grateful if someone familiar 
with jack's code could confirm :

a/ The XRUN messages as reported in the OP are printed by :
- [engine.c] jack_run_one_cycle
- because engine -> spare_usecs <= delayed_usecs

b/ The only occurrence of a call to jack_run_one_cycle is within :
- [engine.c] jack_run_cycle

c/ The only occurrence of a call to jack_run_cycle is the normal return of :
- [alsa_driver.c] alsa_driver_run_cycle

d/ The only place in which the delayed_usecs variable used in a/ is set is :
- [alsa_driver.c] alsa_driver_wait when called by the alsa_driver_run_cycle 
procedure.

Regards,

aCOSwt

1338362921.4108_0.ltw:2, <201205300928.31410.acoswt at orange dot fr>