Sunday, 2022-02-27

T42<HengYeDev> @elros34 reboot didn't help13:15
T42<HengYeDev> Is it bad that droid-bootctl and start-user-session failed?13:15
T42<HengYeDev> seems like droid-bootctl is not needed without slots? should i mask it?13:16
T42<elros34> If you do not need to mark boot as successfully to avoid switching to slob b then I guess you do not need it. start-user-session failed should be normal if you mask user@100000 I think13:19
T42<HengYeDev> It is also curious that there are very little differences between logcat of normal boot and actdead mode when minimer is run13:38
T42<HengYeDev> actdead:
T42<HengYeDev> regular:
T42<HengYeDev> except for some ordering, that is13:38
Thaodanbootctl is needed to tell the bootloader the boot was sucessfull13:53
ThaodanIf not marked the boot slot will be market as unbootable13:53
ThaodanI guess that even happens on single slot boot13:53
T42<edp_17> Thaodan, @elros34 : Is there a list of processes that are not starting when I mask user@100000.service?15:08
T42<HengYeDev> the only difference in logcat that seem to be of significance is G2D0-RGB_PRI threadLoop is started happens before hwcomposer is started in normal boot but after it is started in actdead15:11
T42<HengYeDev> i'm not sure if this is any use or if i need to look at other logs?15:11
T42<HengYeDev> aslso, in dmesg we do not have these errors in actdead:
T42<HengYeDev> is it possible that hardware has put the screen in different mode for actdead/regular?15:17
T42<elros34> @edp_17 probably most user services: /usr/lib/systemd/user/. systemctl  status might be useful. I like systemd-analyze plot > graph.svg15:33
T42<edp_17> Thanks @elros34 . I've ran it on my host. Interesting how many processes there are.15:43
T42<edp_17> My idea is to mask each of them one by one and see which one stops the reboot.15:43
T42<edp_17> The 'systemd-analyze plot > n7000.svg' command gives me 'Bootup is not yet finished. Please try again later.'15:58
T42<edp_17> The device is not rebooting, only lipstick crashes and starts over and over again.15:59
T42<edp_17> The 'systemd-analyze --system' command gives me the same. It looks like there is a problem with that. I'll look into it later and report.16:00
T42<HengYeDev> Must CONFIG_MEMCG_KMEM be enabled in my case? i put it in defconfig, but after build its not in .config17:00
T42<HengYeDev> also, decon-0 already ON state in normal boot but not in charging mode17:08
jules[m]Hi all17:54
jules[m]I try to debug camera support on the oneplus5 port17:54
jules[m]Camera app launch and no error. But app only display that.17:54
* jules[m] uploaded an image: (1589KiB) < >17:54
jules[m]Anyone have a tips for debug ?17:54
Thaodan@edp_17: Only when  you have the unit runnig, but broadly speaking its the hole user session18:09
Thaodan@jules[m]: Logcat and see what camera server tells you18:10
T42<elros34> @HengYeDev this config is not must have for systemd. Do you have CONFIG_MEMCG enabled? Check other dependencies which CONFIG_MEMCG_KMEM have19:00
T42<elros34> It will be probably hard but what  about making this decon driver as module or find in sysfs whether it has some control/reset interfecae19:02
T42<elros34> @HengYeDev  If you are lucky then maybe using driver or whole source from lineage-18 branch might have some fixes19:18
T42<elros34> you should have something like decon_log_level in sysfs19:28
T42<elros34> I think this _decon_enable() is triggered by /dev/fb operation, check if masking mce or anything else which might open it makes any difference19:36
T42<HengYeDev> I will try the driver replacement19:58
T42<HengYeDev> mce is also running in actdead mode according to systemctl list-units19:58
jules[m]<Thaodan> "@jules[m]: Logcat and see what..." <- Thanks i will try ..19:59

Generated by 2.17.1 by Marius Gedminas - find it at!