snd-aloop Daemon Crashed Due To XRun?

[GraysonPeddie]

Arch Linux
Linux 3.14-rc5 custom-compiled RT kernel (not updated yet but waiting for final release)
Cadence 0.8.1
ALSA 1.0.27
Sound card: Echo Gina3G
Carla (GIT) 753.f2960b6-4
Driver configuration: 96k/1024 samples/2 periods per buffer
DE: GNOME Shell 3.10 -- my favorite desktop environment
CPU/GPU: AMD A10-5700 with Radeon HD 7660D

(Yes, I know, there are some things that I have not updated yet before I start a thread...)

I was watching a series of let's plays in YouTube and when starting to play the next video, wha-- there's no sound. So I checked the logs in Cadence and here it is:

Code: Select all

Mon Mar 24 19:20:41 2014: ERROR: JackEngine::XRun: client = jack2alsa was not finished, state = Running
Mon Mar 24 19:20:41 2014: Disconnecting 'system:capture_1' from 'jack2alsa:playback_1'
Mon Mar 24 19:20:41 2014: ERROR: JackAudioDriver::ProcessGraphAsyncMaster: Process error
Mon Mar 24 19:20:41 2014: ERROR: JackEngine::XRun: client = jack2alsa was not finished, state = Triggered
Mon Mar 24 19:20:41 2014: ERROR: JackAudioDriver::ProcessGraphAsyncMaster: Process error
Mon Mar 24 19:20:41 2014: Disconnecting 'system:capture_2' from 'jack2alsa:playback_2'
Mon Mar 24 19:20:41 2014: ERROR: Failed to find port 'jack2alsa:playback_1' to destroy
Mon Mar 24 19:20:41 2014: ERROR: Failed to find port 'jack2alsa:playback_2' to destroy
Mon Mar 24 19:20:41 2014: ERROR: Failed to find port 'jack2alsa:playback_3' to destroy
Mon Mar 24 19:20:41 2014: ERROR: Failed to find port 'jack2alsa:playback_4' to destroy
Mon Mar 24 19:20:41 2014: ERROR: Failed to find port 'jack2alsa:playback_5' to destroy
Mon Mar 24 19:20:41 2014: ERROR: Failed to find port 'jack2alsa:playback_6' to destroy
Mon Mar 24 19:20:41 2014: Client 'jack2alsa' with PID 1344 is out
Mon Mar 24 19:20:41 2014: Client 'cadence-aloop-daemon' with PID 1339 is out
Mon Mar 24 19:20:42 2014: Disconnecting 'alsa2jack:capture_1' from 'Carla:Calf Equalizer 12 Band:in_l'
Mon Mar 24 19:20:42 2014: Disconnecting 'alsa2jack:capture_1' from 'Carla:Calf Filter:in_l'
Mon Mar 24 19:20:42 2014: Disconnecting 'alsa2jack:capture_2' from 'Carla:Calf Equalizer 12 Band:in_r'
Mon Mar 24 19:20:42 2014: Disconnecting 'alsa2jack:capture_2' from 'Carla:Calf Filter:in_r'
Mon Mar 24 19:20:42 2014: Disconnecting 'alsa2jack:capture_3' from 'Carla:Calf Equalizer 12 Band (2):in_l'
Mon Mar 24 19:20:42 2014: Disconnecting 'alsa2jack:capture_4' from 'Carla:Calf Equalizer 12 Band (2):in_r'
Mon Mar 24 19:20:42 2014: Disconnecting 'alsa2jack:capture_5' from 'Carla:Calf Equalizer 12 Band (3):in_l'
Mon Mar 24 19:20:42 2014: Disconnecting 'alsa2jack:capture_6' from 'system:playback_6'
Mon Mar 24 19:20:42 2014: Client 'alsa2jack' with PID 1343 is out

Does anyone know why does such an xrun cause snd-aloop to crash? It's only got 3 xruns so far.

As for equalizers and filters, I've used them as a way to correct frequencies from flaws that my speakers reveal and Audyssey cannot be used over multi-channel analog inputs. No bass management, either. This is where Carla comes in to fix my problems. JACK through HDMI for audio cannot be used.

[bluebell]

Please see thread "Jack xrun every 1.5-2 seconds. What am I missing?".

Maybe there are processes with a higher realtime priority than jackd.

[GraysonPeddie]

I don't have xruns every few seconds, but I'll see what I can do. Thanks.

EDIT: The realtime priority to the right of the realtime checkbox and above the "client timeout (ms)" is set to 10. Is that the same as setting a -P switch to 10? If it is, I might try setting it to 95 as detailed in the other thread.

Code: Select all

$ ps -eo rtprio,pid,tid,class,ni,pri,psr,pcpu,stat,comm | sort -n | grep -ve "^ *-"
RTPRIO   PID   TID CLS  NI PRI PSR %CPU STAT COMMAND
    50    36    36 FF    -  90   3  0.0 S    irq/9-acpi
    50    38    38 FF    -  90   3  0.0 S    irq/40-AMD-Vi
    50    48    48 FF    -  90   3  0.0 S    irq/41-PCIe PME
    50    49    49 FF    -  90   3  0.0 S    irq/42-PCIe PME
    50    50    50 FF    -  90   3  0.0 S    irq/8-rtc0
    50    76    76 FF    -  90   1  0.0 S    irq/1-i8042
    50    86    86 FF    -  90   3  0.0 S    irq/43-xhci_hcd
    50    87    87 FF    -  90   3  0.0 S    irq/44-xhci_hcd
    50    88    88 FF    -  90   3  0.0 S    irq/45-xhci_hcd
    50    89    89 FF    -  90   3  0.0 S    irq/46-xhci_hcd
    50    90    90 FF    -  90   3  0.0 S    irq/47-xhci_hcd
    50    91    91 FF    -  90   2  0.0 S    irq/48-xhci_hcd
    50    92    92 FF    -  90   3  0.0 S    irq/49-xhci_hcd
    50    93    93 FF    -  90   3  0.0 S    irq/50-xhci_hcd
    50    94    94 FF    -  90   3  0.0 S    irq/51-xhci_hcd
    50    95    95 FF    -  90   2  0.0 S    irq/52-xhci_hcd
    50    96    96 FF    -  90   0  0.0 S    irq/61-xhci_hcd
    50    97    97 FF    -  90   0  0.0 S    irq/53-ahci
    50    98    98 FF    -  90   0  0.0 S    irq/54-ahci1
    50    99    99 FF    -  90   0  0.0 S    irq/55-ahci2
    50   107   107 FF    -  90   0  0.0 S    irq/18-ohci_hcd
    50   109   109 FF    -  90   3  0.0 S    irq/18-ohci_hcd
    50   111   111 FF    -  90   0  0.0 S    irq/17-ehci_hcd
    50   112   112 FF    -  90   3  0.0 S    irq/17-ehci_hcd
    50   113   113 FF    -  90   1  0.0 S    irq/18-ohci_hcd
    50   204   204 FF    -  90   3  0.0 S    irq/62-snd_hda_
    50   205   205 FF    -  90   3  0.0 S    irq/16-snd_hda_
    50   237   237 FF    -  90   2  0.3 S    irq/20-snd_echo
    50   275   275 FF    -  90   1  0.3 S    irq/64-radeon
    50   413   413 FF    -  90   0  0.0 S    irq/63-enp1s0
    99    10    10 FF    - 139   0  0.0 S    migration/0
    99    11    11 FF    - 139   0  0.0 S    watchdog/0
    99    12    12 FF    - 139   1  0.0 S    watchdog/1
    99    13    13 FF    - 139   1  0.0 S    migration/1
    99    17    17 FF    - 139   2  0.0 S    watchdog/2
    99    18    18 FF    - 139   2  0.0 S    migration/2
    99    22    22 FF    - 139   3  0.0 S    watchdog/3
    99    23    23 FF    - 139   3  0.0 S    migration/3

I also have PREEMPT for uname -a but not PREEMPT-RT.

[bluebell]

10 ist the standard rtprio of jackd. Entering 95 in QjackCtl should be the same as using -P95.

You can see if that's true by issuing the ps-command while jackd is running:

ps -eLo rtprio,tid,pid,cmd | sort -n

[GraysonPeddie]

Okay, I got a 95 for jackdbus which I'm using with Cadence in Arch Linux (Cadence is from KXStudio, so I'm not using QJackCtl).

The problem that I had above did not happen that often, but maybe going from 10 to 95 will fix the problem that I had. Maybe.

Thanks.