Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

jackd hangs with JackAudioDriver::ProcessGraphAsyncMaster process error after boot #50

Open
nettings opened this issue May 15, 2020 · 9 comments
Labels
bug Something isn't working

Comments

@nettings
Copy link
Owner

then recovers after > 1min.
May 15 21:42:26 mn-Dantax jackd[643]: JackAudioDriver::ProcessGraphAsyncMaster: Process error
May 15 21:42:26 mn-Dantax jackd[643]: JackAudioDriver::ProcessGraphAsyncMaster: Process error
May 15 21:42:26 mn-Dantax jackd[643]: JackAudioDriver::ProcessGraphAsyncMaster: Process error
[repeats hundreds of times, many per second]

Hardware is Pi4 with DAC+ADC with AMP60 on top, kernel 4.19.97

System software: Raspbian GNU/Linux 10 (buster)
medianet overlay: Tue May 5 23:02:49 2020 +0200 b13c3ad

@nettings nettings added the bug Something isn't working label May 15, 2020
@nettings
Copy link
Owner Author

not 100% reproducible, happened two out of three boots so far.

@nettings
Copy link
Owner Author

nettings commented Jul 11, 2020

This was only ever seen on several DAC+ADC, possibly also without AMP60. No further incidents in a while, all machines had a routine kernel update. Needs to be watched, though.
EDIT: apparently unrelated to hardware

@nettings
Copy link
Owner Author

nettings commented Sep 9, 2020

same issue on AMP4/RPI4
Linux mn-vp200a-l 5.4.51-v7l+

@nettings
Copy link
Owner Author

Usually not self-recovering. Killing mod-host helps most of the time, but not always.

@nettings
Copy link
Owner Author

nettings commented Sep 10, 2020

Funny: is this the race condition I'm seeing? There is a jack_lsp job running in parallel...

Sep 10 22:16:10 mn-vp200a-l zita-n2j[646]: Connect: can't connect named futex name = jack_sem.1001_default_lsp-01 err = No such file or directory

This was the first time such a message occurred.

EDIT: since observed two more times, out of some 20 reboots.

@nettings
Copy link
Owner Author

It looks like jack client creation and/or jack_lsp is racy and can send jackd into this death spiral. After reintroducing waiting for ports, the issue seems to have disappeared or at least become so rare that I haven't observed it since. Needs to be looked at further. See commits c214689 and 4f55138 .

@nettings
Copy link
Owner Author

It still happens when the boot process is stalled because of an fsck or because systemd-timesyncd stalls and then syncs, at which point all Jack clients are long ready and spinning on mn_jack_wait or mn_jack_connect, which seems to trigger the death spiral.

@nettings
Copy link
Owner Author

Not observed since.

@nettings
Copy link
Owner Author

again observed several times on bookworm, only ever with the mod-monitor client.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
bug Something isn't working
Projects
None yet
Development

No branches or pull requests

1 participant