toltec-base: Disable [email protected] #806

gbyl · 2023-12-27T15:46:23Z

Disable [email protected] in rM1 to avoid the following errors on boot:

systemd[1]: [email protected]: Bound to unit sys-subsystem-net-devices-usb1.device, but unit isn't active.
systemd[1]: Dependency failed for ifplugd on usb1.
systemd[1]: [email protected]: Job [email protected]/start failed with result 'dependency'.

Discussed in #803 (comment)

Eeems

The version number and timestamp for the package need to be updated as well, otherwise it will not be built, and opkg will not detect this as an update.

package/toltec-base/package

Eeems · 2023-12-27T20:20:08Z

Am I correct in assuming that you added systemctl to everything because you were testing with toltecmk? Toltecmk currently doesn't support install-lib or any of the other toltec specific parsing (toltec-dev/build#29). I have an open PR to add them to toltecmk that you can use the build from if you wish: toltec-dev/build#39

gbyl · 2023-12-27T23:14:53Z

Thank you for the review! To be honest, I was not aware of these custom methods. I naively tried them in an interactive shell and they failed. That's why I appended the command name.

To check the postremove() method, I ran systemctl unmask sys-subsystem-net-devices-usb1.device on its own and it outright deleted the file. Is that expected?

Eeems · 2023-12-27T23:22:04Z

To check the postremove() method, I ran systemctl unmask sys-subsystem-net-devices-usb1.device on its own and it outright deleted the file. Is that expected?

I would expect it to remove the symlink to /dev/null that masking would have created.

gbyl · 2023-12-28T15:38:51Z

I thought it was odd too, since Toltec installs are supposed to be 100% reversible. I am not able to re-test this without reflashing. Could you make a backup of that file and test unmasking on your device?

Eeems · 2023-12-28T19:31:51Z

I thought it was odd too, since Toltec installs are supposed to be 100% reversible. I am not able to re-test this without reflashing. Could you make a backup of that file and test unmasking on your device?

Since it's a *.device, it probably is a dynamically created file on boot. Does a systemctl daemon-reload recreate the unit for you? If not, how about a reboot? From a quick skim of an update file, I don't see a specific unit for the device.

gbyl · 2023-12-28T19:57:12Z

I tried both and find / -name *.device confirms that it did not come back. The unmasking command definitely removed it. I think we are running into something similar to this: https://unix.stackexchange.com/a/611152

Eeems · 2023-12-28T20:00:47Z

I tried both and find / -name *.device confirms that it did not come back. The unmasking command definitely removed it. I think we are running into something similar to this: https://unix.stackexchange.com/a/611152

Device files are native to systemd and automatically generated. The files do not exist on the device. Masking one will create a symlink to /dev/null and store that in /etc/systemd/system. Unmasking will remove the symlink, and thus reenable the built in unit.

What does systemctl status sys-subsystem-net-devices-usb1.device return for you?

gbyl · 2023-12-28T20:06:57Z

$ systemctl status sys-subsystem-net-devices-usb1.device
● sys-subsystem-net-devices-usb1.device - /sys/subsystem/net/devices/usb1
     Loaded: loaded
     Active: inactive (dead)

$ systemctl enable sys-subsystem-net-devices-usb1.device
Failed to enable unit: Unit file sys-subsystem-net-devices-usb1.device does not exist.

Eeems · 2023-12-28T20:15:14Z

$ systemctl status sys-subsystem-net-devices-usb1.device
● sys-subsystem-net-devices-usb1.device - /sys/subsystem/net/devices/usb1
     Loaded: loaded
     Active: inactive (dead)

$ systemctl enable sys-subsystem-net-devices-usb1.device
Failed to enable unit: Unit file sys-subsystem-net-devices-usb1.device does not exist.

As this is a device file, you don't enable it, it's just enabled by default, you should be able to start the unit. I would recommend not blocking while starting it, though:

systemctl start --no-block sys-subsystem-net-devices-usb1.device

gbyl · 2023-12-28T20:27:27Z

Unfortunately, I got the same result. The start command did not report an error, but the status remains the same.

Eeems · 2023-12-28T21:32:43Z

Unfortunately, I got the same result. The start command did not report an error, but the status remains the same.

And this is after you've also rebooted your device?

gbyl · 2023-12-28T21:49:32Z

That's right. I'm hoping it's only something local to my device from too much poking around. If you are not able to reproduce this, we can safely assume that, and resume the review of this PR

Eeems · 2023-12-28T23:47:26Z

That's right. I'm hoping it's only something local to my device from too much poking around. If you are not able to reproduce this, we can safely assume that, and resume the review of this PR

I'd like to make sure this is working as expected before merging. I haven't tested this package yet on my device as I've been busy with other things.

What happens if you reload udev? udevadm control --reload-rules && udevadm trigger

gbyl · 2023-12-29T01:58:32Z

No worries. I ran that command and got no output. The status remains the same too. I'm not sure how to troubleshoot further without a reflash to better emulate the experience of the average Toltec user.

Eeems · 2023-12-29T02:04:19Z

No worries. I ran that command and got no output. The status remains the same too. I'm not sure how to troubleshoot further without a reflash to better emulate the experience of the average Toltec user.

You can use codexctl to install specific OS versions easily. From there you can just reenable toltec and be back where you were. Minus any manual changes to the root partition you made.

gbyl · 2023-12-29T03:55:14Z

Excellent. I can confirm it works fine on a fresh install:

$ systemctl status sys-subsystem-net-devices-usb1.device
● sys-subsystem-net-devices-usb1.device - /sys/subsystem/net/devices/usb1
     Loaded: loaded
     Active: inactive (dead)

systemd[1]: sys-subsystem-net-devices-usb1.device: Job sys-subsystem-net-devices-usb1.device/start timed out.
systemd[1]: Timed out waiting for device /sys/subsystem/net/devices/usb1.
systemd[1]: sys-subsystem-net-devices-usb1.device: Job sys-subsystem-net-devices-usb1.device/start failed with result 'timeout'.

Then manually replaying the steps Toltec goes through:

$ systemctl mask sys-subsystem-net-devices-usb1.device
Created symlink /etc/systemd/system/sys-subsystem-net-devices-usb1.device → /dev/null.

$ systemctl unmask sys-subsystem-net-devices-usb1.device
Removed /etc/systemd/system/sys-subsystem-net-devices-usb1.device.

Despite the "removed" message, this time it does come back on the next reboot as expected.

Eeems · 2023-12-29T03:59:04Z

Despite the "removed" message, this time it does come back on the next reboot as expected.

If you look closely at the message produced by the systemctl mask command you'll see that the file being removed by unmask was actually just a symlink to /dev/null. So it's not actually removing the unit file, just the symlink used to instruct systemd that it's masked.

I'm not entirely sure what was happening for you previously, as unmasking it should have been enough for systemd to just automatically start seeing it again. Since this is a device, and not a normal unit file.

gbyl · 2023-12-29T04:26:08Z

Agreed. That makes perfect sense now. Is there anything else that should be changed in this PR?

Eeems · 2023-12-29T04:27:21Z

I'll let you know when I get around to testing it :)

Eeems

Tested and works fine on my rM1, someone will need to retest this on a rM2

Eeems · 2024-01-13T00:32:43Z

Tested on my new rM2 and it works 😎

* toltec-base: disable [email protected] --------- Co-authored-by: gbyl <[email protected]> Co-authored-by: Nathaniel van Diepen <[email protected]>

### New Packages - `move-logs-to-opt` - 0.0.1-1 (#821) - Move log files to `/opt` to avoid filling up the root partition. - `oxide-extra` and `liboxide-dev` - 2.7-3 (#760) ### Updated Packages - `folly` - 0.0.1-4 (#823) - `koreader` - 2024.01-1 (#826) - `linux-stracciatella` - 5.4.70.4-1 (#808) - `oxide`, `oxide-utils`, `inject_evdev`, and `liboxide` - 2.7-3 (#760) - `puzzles` - 0.2.4-4 (#785) - `rm2-suspend-fix` - 0.0.0-3 (#821) - `genie` - 0.1.7-1 (#825) - `harmony` - 0.2.3-1 (#825) - `iago` - 0.1.2-1 (#825) - `remux` - 0.3.0-1 (#825) - `simple` - 0.2.1-1 (#825) - `tilem` - 0.1.1-1 (#764) - `toltec-base` - 1.3-1 (#806) - `yaft` - 0.1.1 (#763) - `zoneinfo-utils` - 2023c-1 (#702) - `rmfakecloud-proxy` - 0.0.3-5 (#757 #831) ### Packages removed - `erode`, `tarnish`, `decay`, and `corrupt` - Replaced by `oxide` - `fret`, and `anxiety` - Replaced by `oxide-extra` - `rot` - Replaced by `oxide-utils`

toltec-base: disable [email protected]

f00f574

gbyl force-pushed the disable/busybox-ifplugd@usb1 branch from 8d78ff2 to f00f574 Compare December 27, 2023 16:18

add missing systemctl calls in if-statements

4a5cb6f

Eeems requested changes Dec 27, 2023

View reviewed changes

package/toltec-base/package Outdated Show resolved Hide resolved

package/toltec-base/package Outdated Show resolved Hide resolved

package/toltec-base/package Outdated Show resolved Hide resolved

package/toltec-base/package Outdated Show resolved Hide resolved

Eeems added the packages Add or improve packages of the repository label Dec 27, 2023

Eeems modified the milestones: 2024-W01 Merge Window, 2023-W52 Merge Window Dec 27, 2023

gbyl added 4 commits December 27, 2023 16:56

Revert "add missing systemctl calls in if-statements"

b382bfc

update version and timestamp

4be6f7f

review: fix if-statements

4ed07a3

remove trailing space

7c1d203

Eeems self-requested a review December 27, 2023 23:22

fix typo

05d2d56

Eeems modified the milestones: 2023-W52 Merge Window, 2024-W01 Merge Window Dec 31, 2023

Eeems added 2 commits January 5, 2024 16:20

Merge branch 'testing' into disable/busybox-ifplugd@usb1

e8de7ca

Merge branch 'testing' into disable/busybox-ifplugd@usb1

f160c1f

Eeems approved these changes Jan 5, 2024

View reviewed changes

Eeems modified the milestones: 2024-W01 Merge Window, 2024-W02 Merge Window Jan 7, 2024

Merge branch 'testing' into disable/busybox-ifplugd@usb1

37a5fc6

Eeems merged commit 21f43f0 into toltec-dev:testing Jan 13, 2024
3 checks passed

gbyl deleted the disable/busybox-ifplugd@usb1 branch January 13, 2024 04:48

Eeems added a commit that referenced this pull request Jan 28, 2024

toltec-base: Disable [email protected] (#806)

d5f53a1

* toltec-base: disable [email protected] --------- Co-authored-by: gbyl <[email protected]> Co-authored-by: Nathaniel van Diepen <[email protected]>

Eeems mentioned this pull request Jan 28, 2024

Stable merge for week 4 of 2024 #830

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

toltec-base: Disable [email protected] #806

toltec-base: Disable [email protected] #806

gbyl commented Dec 27, 2023

Eeems left a comment

Eeems commented Dec 27, 2023

gbyl commented Dec 27, 2023

Eeems commented Dec 27, 2023

gbyl commented Dec 28, 2023

Eeems commented Dec 28, 2023

gbyl commented Dec 28, 2023

Eeems commented Dec 28, 2023

gbyl commented Dec 28, 2023 •

edited

Loading

Eeems commented Dec 28, 2023

gbyl commented Dec 28, 2023

Eeems commented Dec 28, 2023

gbyl commented Dec 28, 2023

Eeems commented Dec 28, 2023

gbyl commented Dec 29, 2023

Eeems commented Dec 29, 2023

gbyl commented Dec 29, 2023

Eeems commented Dec 29, 2023

gbyl commented Dec 29, 2023

Eeems commented Dec 29, 2023

Eeems left a comment

Eeems commented Jan 13, 2024

toltec-base: Disable [email protected] #806

toltec-base: Disable [email protected] #806

Conversation

gbyl commented Dec 27, 2023

Eeems left a comment

Choose a reason for hiding this comment

Eeems commented Dec 27, 2023

gbyl commented Dec 27, 2023

Eeems commented Dec 27, 2023

gbyl commented Dec 28, 2023

Eeems commented Dec 28, 2023

gbyl commented Dec 28, 2023

Eeems commented Dec 28, 2023

gbyl commented Dec 28, 2023 • edited Loading

Eeems commented Dec 28, 2023

gbyl commented Dec 28, 2023

Eeems commented Dec 28, 2023

gbyl commented Dec 28, 2023

Eeems commented Dec 28, 2023

gbyl commented Dec 29, 2023

Eeems commented Dec 29, 2023

gbyl commented Dec 29, 2023

Eeems commented Dec 29, 2023

gbyl commented Dec 29, 2023

Eeems commented Dec 29, 2023

Eeems left a comment

Choose a reason for hiding this comment

Eeems commented Jan 13, 2024

gbyl commented Dec 28, 2023 •

edited

Loading