PINE64 dev

12 Dec 2024 03:27

[M] <Y​oda> 6.11 is EOL now

PINE64 dev

12 Dec 2024 01:29

[M] <D​anct12> got a issue with kernel 6.11.5, there are errors about audio:

```
Dec 12 00:26:49 danctnix kernel: hdmi-audio-codec hdmi-audio-codec.0.auto: ASoC: error at snd_soc_dai_digital_mute on i2s-hifi: -19
Dec 12 00:26:49 danctnix kernel: ff8a0000.i2s-i2s-hifi: ASoC: error at __soc_pcm_hw_params on ff8a0000.i2s-i2s-hifi: -19
Dec 12 00:26:49 danctnix kernel: hdmi-audio-codec hdmi-audio-codec.0.auto: ASoC: error at snd_soc_dai_digital_mute on i2s-hifi: -19
Dec 12 00:26:49 danctnix kernel: hdmi-audio-codec hdmi-audio-codec.0.auto: ASoC: error at snd_soc_dai_hw_params on i2s-hifi: -19
Dec 12 00:26:49 danctnix kernel: hdmi-audio-codec hdmi-audio-codec.0.auto: ASoC: error at snd_soc_dai_digital_mute on i2s-hifi: -19
Dec 12 00:26:49 danctnix kernel: ff8a0000.i2s-i2s-hifi: ASoC: error at __soc_pcm_hw_params on ff8a0000.i2s-i2s-hifi: -19
```

PINE64 dev

04 Dec 2024 01:08

[I] <d​iederik> indeed. It's already deprecated which normally means it's on its way out (which could take years, but still)

PINE64 dev

03 Dec 2024 23:30

[I] <d​iederik> "Legacy cgroup v1 memory controller" ... sounds more like a (poor) workaround

PINE64 dev

03 Dec 2024 17:37

[M] <p​iggz> lets see if config_memcg_v1 fixes it

PINE64 dev

03 Dec 2024 08:18

[M] <p​iggz> so, im closer, but next want to see what changed in 6.12

PINE64 dev

03 Dec 2024 08:17

[M] <p​iggz> megi: user reported that the cgroup/memory/noswapping exists on 6.11 but not 6.12...

PINE64 dev

03 Dec 2024 00:58

[I] <m​egi> I thought /sys/fs/cgroup is mounted by systemd daemon itself, before anything else, and not by any service though

PINE64 dev

03 Dec 2024 00:55

[I] <m​egi> I guess it's some special cgroup with swappiness set very low on your system?

PINE64 dev

02 Dec 2024 22:19

[M] <p​iggz> the error is a timing one / systemd issue .. 1, it shouldnt be trying to work with that folder, and 2), on the ppp, it boots by-chance, becuase the failing service on the pp starts later in the boot on the ppp

PINE64 dev

02 Dec 2024 21:37

[M] <p​iggz> Dec 02 20:31:33 PinePhone systemd-tmpfiles[2428]: Failed to create directory or subvolume "/sys/fs/cgroup/memory/noswapping": No such file or directory
Dec 02 20:31:33 PinePhone systemd[1]: systemd-tmpfiles-setup.service: Main process exited, code=exited, status=1/FAILURE
Dec 02 20:31:33 PinePhone systemd[1]: systemd-tmpfiles-setup.service: Failed with result 'exit-code'.
Dec 02 20:31:33 PinePhone systemd[1]: Failed to start Create Volatile Files and Directories.

PINE64 dev

02 Dec 2024 08:48

[M] <p​iggz> Maybe the audit system can help me

PINE64 dev

02 Dec 2024 08:44

[M] <p​iggz> And that showed the file content as expected!

PINE64 dev

02 Dec 2024 08:43

[M] <p​iggz> The program is a script, so i added some debug, to print the mount list, folder content etc

PINE64 dev

01 Dec 2024 23:00

[I] <m​egi> fails to write with what error?

PINE64 dev

12 Dec 2024 01:29

[M] <D​anct12> it spams non-stop

PINE64 dev

04 Dec 2024 08:04

[M] <p​iggz> Yes, ive mentioned this already :)

PINE64 dev

04 Dec 2024 01:04

[I] <m​egi> maybe sailfish needs to move to cgroups v2

PINE64 dev

03 Dec 2024 19:23

[M] <p​iggz> yes, that fixed it

PINE64 dev

03 Dec 2024 14:45

[M] <p​iggz> megi: you know of any changes to cgroups for 6.12, or a new config option needed?

PINE64 dev

03 Dec 2024 08:17

[M] <p​iggz> sailfish sets up that file to prevent critical processess being swapped

PINE64 dev

03 Dec 2024 01:05

[I] <m​egi> yeah https://github.com/systemd/systemd/blob/80ea2271913fdbe48b413759a46b4479b07579e1/src/shared/mount-setup.c#L106

PINE64 dev

03 Dec 2024 00:56

[I] <m​egi> and /sys/fs/cgroup is not yet moutned when tmpfiles runs

PINE64 dev

03 Dec 2024 00:52

[I] <m​egi> dunno, I don't seem to have noswapping in any of my /usr/lib/tmpfiles.d/ on arch linux

PINE64 dev

02 Dec 2024 21:38

[M] <p​iggz> becuase this service fails, it re-runs after the initial-bootstate serice, and clears the /run folder

PINE64 dev

02 Dec 2024 21:36

[M] <p​iggz> megi: so, i think this is the cause of the boot failure:

PINE64 dev

02 Dec 2024 08:45

[M] <p​iggz> Yet when i ssh in, it has gone!

PINE64 dev

02 Dec 2024 08:43

[M] <p​iggz> And then print the file after its written

PINE64 dev

02 Dec 2024 08:42

[M] <p​iggz> Megi: well, it gets weirder....

PINE64 dev

01 Dec 2024 21:58

[M] <p​iggz> this is the diff of the configs, nothing particularly jumps out to me https://www.diffchecker.com/6MSmBuV7/