I also noticed when starting jack: Apr 10 11:46:50 cubase pulseaudio[23745]: [pulseaudio] module-jack-sink.c: JACK error >Cannot use real-time scheduling (RR/5)(1: Operation not permitted)< Apr 10 11:46:50 cubase pulseaudio[23745]: [pulseaudio] module-jack-sink.c: JACK error >JackClient::AcquireSelfRealTime error< Although qjackctl reports that jackd *does* run with realtime scheduling. Confusing. Anyway, reproduced the problem. Every so often, I get more of those "cannot get freq" messages and the sound locks up for a few seconds, but then comes back. Then, I got an much longer stream of them (maybe a minute's worth), but the sound and responsiveness of Ardour still came back. Right now, it happened when I was watching YouTube on Chrome via the Pulseaudio Jack sink. At the time of the messages, qjackctl was suddenly using 25% of my decent i5 CPU.When I checked the log, it was full of messages like: Tue Apr 11 16:28:09 2017: Jack: ALSA XRun wait_status = 0 Tue Apr 11 16:28:09 2017: Jack: JackSocketServerChannel::Execute : fPollTable i = 1 fd = 12 Tue Apr 11 16:28:09 2017: Jack: JackRequest::Notification Tue Apr 11 16:28:09 2017: Jack: JackEngine::ClientNotify: no callback for notification = 3 Tue Apr 11 16:28:09 2017: Jack: JackEngine::ClientNotify: no callback for notification = 3 Tue Apr 11 16:28:09 2017: Jack: JackClient::ClientNotify ref = 2 name = dbusapi notify = 3 Tue Apr 11 16:28:09 2017: Jack: JackClient::kXRunCallback Tue Apr 11 16:28:09 2017: Jack: JackEngine::ClientNotify: no callback for notification = 3 Tue Apr 11 16:28:09 2017: Jack: JackEngine::ClientNotify: no callback for notification = 3 Tue Apr 11 16:28:09 2017: Jack: JackExternalClient::ClientNotify ref = 5 client = qjackctl name = qjackctl notify = 3 Tue Apr 11 16:28:09 2017: Jack: JackEngine::ClientNotify: no callback for notification = 3 Tue Apr 11 16:28:09 2017: Jack: JackSocketServerChannel::Execute : fPollTable i = 2 fd = 14 Tue Apr 11 16:28:09 2017: Jack: JackSocketServerChannel::Execute : fPollTable i = 3 fd = 17 Tue Apr 11 16:28:09 2017: Jack: JackSocketServerChannel::Execute : fPollTable i = 4 fd = 20 Tue Apr 11 16:28:09 2017: Jack: JackSocketServerChannel::Execute : fPollTable i = 5 fd = 23 Tue Apr 11 16:28:09 2017: Jack: ALSA XRun wait_status = 0 Tue Apr 11 16:28:09 2017: Jack: JackSocketServerChannel::Execute : fPollTable i = 1 fd = 12 Tue Apr 11 16:28:09 2017: Jack: JackRequest::Notification Tue Apr 11 16:28:09 2017: Jack: JackEngine::ClientNotify: no callback for notification = 3 Tue Apr 11 16:28:09 2017: Jack: JackEngine::ClientNotify: no callback for notification = 3 Tue Apr 11 16:28:09 2017: Jack: JackClient::ClientNotify ref = 2 name = dbusapi notify = 3 Tue Apr 11 16:28:09 2017: Jack: JackClient::kXRunCallback This seems to happen at the same time as the syslog lines from my OP. Apr 11 16:28:09 cubase kernel: [108661.803326] usb 3-3: 1:2: cannot get freq at ep 0x1 Apr 11 16:28:09 cubase kernel: [108661.805662] usb 3-3: 2:2: cannot get freq at ep 0x82 By the time I'd gathered enough log information to paste above, the hang was over, so I haven't yet had a chance to try Clemens' suggestion (If you kill Jack and PulseAudio, do aplay/arecord work?). Anecdotally, from grepping logs, I do believe that "cannot get freq" lines in the syslog happen even when I'm not using jackd, but perhaps with jackd there's more of a problem. I guess I'm none the wiser. I'll try to log times when jackd is and isn't running and match them to the "cannot get freq" lines in the log, but meanwhile I definitely have a problem here when I'm using jackd, and I'm still not sure if it's software-or-hardware-related. Antony On 10 April 2017 at 11:45, Clemens Ladisch &lt;clemens(a)ladisch.de&gt; wrote: -- http://www.linkedin.com/in/antgel http://about.me/antonygelberg

On Tue, April 11, 2017 9:11 am, Antony Gelberg wrote: I just checked my system log when loading the jack-sink module and noticed the same thing: pulseaudio[30057]: [pulseaudio] module-jack-sink.c: JACK error >Cannot use real-time scheduling (RR/15)(1: Operation not permitted)< pulseaudio[30057]: [pulseaudio] module-jack-sink.c: JACK error It doesn't seem to cause any problems, I just played a couple of youtube videos with no underruns. -- Chris Caudle

When troubleshooting something, even a broken radio or coffee machine, you should check on thing after the other. Why do you do random tests running two sound servers at once? You mentioned you want to use Ardour. At least for testing purpose consider to use Ardour just with ALSA, explicitly without launching jack and/or pulseaudio at all. Btw. I'm very surprised that so many people need to do real-time audio recording with pulseaudio. I know, thousands of people want to record skype interview or watch youtube tutorials. I don't know if http://github.com/i-rinat/apulse helps with skype interviews, but at least http://rg3.github.io/youtube-dl/ helps with tutorials.

On Fri, 14 Apr 2017, Chris Caudle wrote: Yes I have the same thing. It does cause problems at low latency (jack at 16/2 for example). Actually, I should clarify, low latency depends on the application talking to PA, Skype for example needs jack to run at 128/2 (at 48000) to function properly. module-jack-sink/source are "example" code. Means not well tested or complete. They try to pull pulse into the same latency as jack is running (watch cpu used by PA as latency in jack changes). Those two modules need some love which I don't think is going to come from the original author. I would suggest that these modules need to include some buffering. My fondest wish would be for PA to see jackd as a device with the same kind of configuration as ALSA devices (and protections... maybe more) I am not sure if merely adding a ring buffer would fix that or not. (might be a start though) I have not personally looked at the code. -- Len Ovens www.ovenwerks.net