From: Pedro Lopez-C. <ped...@gm...> - 2009-03-04 08:47:06
|
Hi, On Wed, Mar 4, 2009 at 12:45 AM, Tim <ter...@ro...> wrote: > Hi, Pedro. > I tried your patch, but I discovered that when muse is run > with the dummy audio driver (-a switch), muse shortly > quits with "WatchDog: fatal error, realtime task timeout". > > I can see that the patch does in fact choose RTC, instead of > system timer, but beyond that I'm not sure how to fix this. > I tried with and without kernel RTC module loaded. > Can you think of anything? Seems like a really useful patch... > Here is what I get WITH the patch, followed by WITHOUT: > (Don't mind the Lash or VST stuff)... > > WITH PATCH: > -------------------------- [...] > AlsaTimer::AlsaTimer(this=0xb518834c) called > AlsaTimer::initTimer(this=0xb518834c) > AlsaTimer::initTimer(): best available ALSA timer: RTC timer > AlsaTimer::setTimerFreq(this=0xb518834c) > AlsaTimer::setTimerFreq(): Setting ticks (period) to 40 ticks > AlsaTimer::startTimer(this=0xb518834c): handle=0x86d9160 [...] > WITHOUT PATCH: > -------------------------- [...] > AlsaTimer::AlsaTimer(this=0xb510c34c) called > AlsaTimer::initTimer(this=0xb510c34c) > AlsaTimer::initTimer(): best available ALSA timer: system timer I've added the latter debug message in my patch to know the results of the exploration, so I don't understand how it went into the debug output if you compiled muse without the patch? A second thought is that an ALSA timer is not needed when the RTC device has successfully opened. My patch didn't fix this behavior, though. I hope to find some time to take a look to this. Regards, Pedro |