Sender | Message | Time |
---|---|---|
19 Nov 2024 | ||
oak | Yeah on LineageOS 19.1 the squid logo ball was only going over the line for like 3-5 times usually, now it takes 20 bounces | 05:35:51 |
atemu12 | You could have a look at logcat and perhaps get a glance of what's taking so long | 05:36:05 |
atemu12 | There isn't anything as nice as systemd-analyze on Android though that I know of | 05:36:26 |
atemu12 | FWIW it's not taking that long on my device | 05:36:40 |
atemu12 | Have you added a bunch of files to userdata in the process of updating by chance? e.g. to create a backup? | 05:37:23 |
atemu12 | I once had super long boot times and it was caused bu Android running some SElinux re-labelling or some shit | 05:37:59 |
atemu12 | And that scales poorly with the amount of files you have | 05:38:19 |
oak | It looks like it is waiting forever for some sensor and then it timeouts and boots finally | 05:38:42 |
atemu12 | Huh | 05:38:53 |
atemu12 | I wonder if that's an upstream LOS bug | 05:38:58 |
oak | 11-19 07:39:20.502 995 995 I sensors-hal: wait_for_mandatory_sensors:507, some mandatory sensors not available yet, will wait for 23 seconds... | 05:40:09 |
oak | It waits until it goes to 0 seconds and then gives up I guess | 05:40:26 |
atemu12 | I hate this kind of error message with a passion | 05:40:58 |
atemu12 | "some" | 05:41:02 |
atemu12 | Gee thanks | 05:41:08 |
atemu12 | And it's like in the best position to tell you because the function logging this error likely has all the devices it's waiting on readily available. | 05:41:41 |
oak | There is sensors list file needs to be updated otherwise it will delay boot for 30 seconds in /vendor/persist/sensors/sensors_list.txt | 05:43:34 |
oak | It actually gives useful error message in the end: ```
| 05:45:29 |
oak | * It actually gives useful error message in the end: ``` 11-19 07:44:52.903 992 992 E sensors-hal: wait_for_mandatory_sensors:531, some mandatory sensors not available even after 30 seconds, giving up. 11-19 07:44:52.903 992 992 E sensors-hal: wait_for_mandatory_sensors:533, 3 missing sensor(s) 11-19 07:44:52.903 992 992 E sensors-hal: wait_for_mandatory_sensors:536, mag 11-19 07:44:52.903 992 992 E sensors-hal: wait_for_mandatory_sensors:536, mag_cal 11-19 07:44:52.903 992 992 E sensors-hal: wait_for_mandatory_sensors:536, rotv | 05:45:42 |
oak | * It actually gives useful error message in the end: 11-19 07:44:52.903 992 992 E sensors-hal: wait_for_mandatory_sensors:531, some mandatory sensors not available even after 30 seconds, giving up. 11-19 07:44:52.903 992 992 E sensors-hal: wait_for_mandatory_sensors:533, 3 missing sensor(s) 11-19 07:44:52.903 992 992 E sensors-hal: wait_for_mandatory_sensors:536, mag 11-19 07:44:52.903 992 992 E sensors-hal: wait_for_mandatory_sensors:536, mag_cal 11-19 07:44:52.903 992 992 E sensors-hal: wait_for_mandatory_sensors:536, rotv | 05:45:48 |
atemu12 | Now the question is whether it actually doesn't have those sensors or whether the HAL/BLOBs are broken | 05:47:23 |
oak | Is there easy way within Robotnix to just get access to the extracted Android source tree? | 06:29:08 |
atemu12 | debug-enter-env | 06:29:30 |
oak | Just found it from the documentation | 06:31:04 |
oak | That sensors-hal error is coming from binary file :( | 06:43:50 |
oak | I tried to ask questions in the #lineageos iRC-channel but luk1337 answered me "if you upgrade to 21/22 this year, ask me again" x) | 16:43:55 |
oak | * I tried to ask questions in the #lineageos IRC-channel but luk1337 answered me "if you upgrade to 21/22 this year, ask me again" x) | 16:44:41 |
oak | I'm wondering should I try upgrading to 21, then I could at the same time test Atemu's open PR | 16:45:26 |
countoren joined the room. | 19:46:58 | |
countoren | Hi all, I have a small staticlly linked cross compiled server that runs on propitiatory AOSP, it works nice with libusb, and rxcpp but now I would like to have write logs to logcat, any of you have something to point me too for liblog for a nix statically linked executable ? | 19:49:43 |