Hm... I could not notice anything particular strange in your code. I suggest then to debug driver a bit. Add debug printouts (cmn_err calls) to the places where underrun counters can be incremented (do_outputintr and audio_space_in_queue), and to the places where they can be nullified (you can find them in oss_audio_core.c). At least you will know if underruns are detected by OSS and if so where underrun counters are cleared. My guess is that they are detected but cleared somewhere. Maybe I'm wrong.
Regards, Zakhar Levchenko On 10/9/08, Stéphane Letz <[EMAIL PROTECTED]> wrote: > > > Le 9 oct. 08 à 11:30, Zakhar Levchenko a écrit : > > > > Hi Stephane, > > > > I think it may help if you provide your code if it's not a big > > secret (at least main parts like device initialization, read/write > > loop). I guess the problem is rather in your application than in > > driver. Theorectically OSS may not always detect underruns (for > > example when interrupt handlers where underruns/overruns are > > detected are executed with considerable delay caused by system load > > or similar reasons, the application cannot fill audio buffer in time > > but fills it before interrupt handler is invoked). But I hardly can > > imagine any situation (especially with high system load) when > > nothing is detected at all. > > > > Regards, > > Zakhar Levchenko > > > > > > The code is the OSS backend for jackdmp: > > > http://trac.jackaudio.org/browser/jack2/branches/solaris/solaris/oss/JackOSSDriver.h > > http://trac.jackaudio.org/browser/jack2/branches/solaris/solaris/oss/JackOSSDriver.cpp > > > Stephane Letz > _______________________________________________ > oss-devel mailing list > oss-devel@mailman.opensound.com > http://mailman.opensound.com/mailman/listinfo/oss-devel >
_______________________________________________ oss-devel mailing list oss-devel@mailman.opensound.com http://mailman.opensound.com/mailman/listinfo/oss-devel