Monday, 2022-05-16

T42<edp_17> mal : Good morning. Yes, dsme is still crashing. logcat:
T42<edp_17> journal:
T42<edp_17> I've tried to start dsme without and with debug as @elros34 suggested. Here is the log from console:
T42<edp_17> @elros34 : Good morning.05:41
T42<my90ds> Please tell me how to fix this (
T42<edp_17> @my90ds : The same command is working for me so the issue will be your connection.05:53
T42<edp_17> @my90ds : Try again.05:53
T42<my90ds> Tried repeatedly. There are no problems with the Internet. The problem is only with the execution of this command. (re @edp_17: @my90ds : Try again.)05:54
T42<edp_17> @my90ds : Use this: repo init -u -b hybris-16.005:56
T42<my90ds> Thank you. It worked. (re @edp_17: @my90ds : Use this: ...)05:58
T42<elros34> @edp_17 hi, "-- -p /usr/lib/dsme/" is not part of gdb but dsme command so include it if you start dsme manually. I assume nothing really crash right?06:30
T42<edp_17> @elros34 : No, if I include those to the dsme call, there is no crash:
T42<edp_17> But the device keeps rebooting. :(07:22
T42<edp_17> When I mask dsme.service, it reboots much quicker (I cannot even telnet in)07:22
T42<elros34> so mask something else like droid-hal-init/user@100000 to ease debugging. Get rid of StartLimitAction=reboot, PrivateNetwork=true and other sandboxing options too see if that will help07:32
T42<elros34> from dsme.service07:32
*** Daaanct12 is now known as Daanct1207:35
T42<Asgard_f0rk> Can someone guide me how can I situp my PC for porting sailfishos07:36
T42<edp_17> @Asgard_f0rk : Follow the HADK:
T42<edp_17> @elros34 : I've masked user@100000 and droid-hal-init. Didn't help, the device keeps rebooting. Now, I amend dsme.service as you requested.07:47
T42<edp_17> dsme service before:
T42<edp_17> and after:
T42<edp_17> Result: blue led is blinking, I have telnet and seemingly the reboot stopped.07:50
T42<edp_17> I guess because I removed this: StartLimitAction=reboot07:51
T42<edp_17> Now, should I start droid-hal-init and user@100000 services?07:51
T42<elros34> check whether it still crash07:58
T42<edp_17> Here is the journal since the last reboot:
T42<edp_17> It looks like not crashing, doesn't it?08:16
T42<edp_17> I've unmasked droid-hal-init and user@100000 services and got this in journal:
T42<edp_17> Wait, both services are still masked.08:48
T42<edp_17> Yeah, when I unmasked them the device keeps rebooting before I can get in.08:53
T42<edp_17> When user@100000 was masked and droid-hal-init was unmasked, the device kept rebooting.09:43
T42<edp_17> When user@100000 was unmasked and droid-hal-init was masked, the reboot stopped.09:43
T42<edp_17> @elros34 ^09:43
T42<edp_17> I thought trying to start droid-hal-init manually.09:48
T42<edp_17> Seen that it first starts that bind mounts ld.config.29.txt. (I don't need this as don't have ld.config.29.txt)09:49
T42<edp_17> Then it starts When tried to execute the commands manually, I got this :
T42<edp_17> Okay, if I manually start all steps from /usr/bin/droid/, lipstick starts and the device doesn't reboot.10:00
T42<edp_17> What about if I increase these in droid-hal-init.service? : TimeoutSec=1510:07
T42<edp_17> StartLimitInterval=6010:07
T42<edp_17> StartLimitBurst=210:07
T42<edp_17> Might it be the device need more time for something but dorid-hal-init doesn't wait long enough?10:07
T42<edp_17> full journal:
*** attah_ is now known as attah10:13
*** attah_ is now known as attah10:41
T42<edp_17> How can I know whether I need bluez4 or bluez5?13:26
T42<edp_17> Should have searched logs first: "kernel version. 3.4 can't use bluez5", so it will be bluez4. :)13:28
malyes, it is dependent on kernel version, 3.4 can also use bluez5 if you use backported kernel bluetooth drivers13:28
T42<edp_17> mal : I have not backported anything but the kernel might have that driver. How can I check?13:30
T42<edp_17> I've checked this in my kernel and all are there.
T42<edp_17> So, it means I already have backported bluez5 driver that I can use, doesn't it?13:36
T42<edp_17> mal ^13:36
malcan you give a link to the kernel sources13:44
T42<edp_17> mal : Sure.
T42<edp_17> lineage-17.1 branch:
T42<edp_17> For building broadcom-bluetooth (bluez5), I need to install bluez5-libs-devel. When trying to install, I get:
T42<edp_17> Can I simply chose solution 1?13:51
malso you have broadcom bluetooth?13:53
malyes, you can select 113:54
T42<edp_17> Yep. Thanks for confirming.13:56
T42<edp_17> mal : Do you have idea why droid-hal-init crashes when it starts as service but it runs happily when started manually (I mean command by command from the script)?13:57
malhow does it crash?14:04
malone thing I could think of is systemd sandboxing, if it's because of that it would suggest some issue in kernel14:07
malyou can try if it work without sandboxing by commenting out line 15-17 in /usr/lib/systemd/system/droid-hal-init.service i.e.
malI will be away for an hour or so14:09
T42<edp_17> mal : No worries. I'll catch you up later. I'll try what you recommended.14:10
T42<edp_17> mal : I do not have those "Protect*" lines in my /usr/lib/systemd/system/droid-hal-init.service.14:12
T42<edp_17> my current droid-hal-init.service file is:
T42<edp_17> But the device kept rebooting with this one too.14:13
T42<edp_17> Oh, and dsme.service looks like this :
T42<edp_17> @pketo, please add Galaxy S4 access to the store. Thanks. Device model: Galaxy S4 (jfltexx / jfltexx)14:18
mal@edp_17 does dsme fail if you have those sandboxing lines enabled?16:03
T42<edp_17> mal : Yes it does. Moreover all apps that use sandboxing do crash and not start. (e.g. :sailfish-browser)16:39
T42<edp_17> Well, almost all because gallery and calendar do start.16:49
T42<elros34> @edp_17 try to set this to -1: then get dmesg from dsme failure, droid-hal-init or anything else which fails.18:50
T42<elros34> Would be good if you can find any other device with 3.4 kernel which works fine on sfos 4.4.0. That way you could compare missing/breaking commits18:52
T42<elros34> I can see lipstick complains about missing: libandroidicu.so18:58
malah, that needs the usual symlink probably19:22
malfp2 has 3.4 kernel and works fine, I had to patch the kernel a bit though19:23
T42<edp_17> It was complainig when dsme was masked but then this error gone when I enabled dsme and masked droid-hal-init and manually started its commands one by one. (re @elros34: I can see lipstick c...)19:26
T42<edp_17> I'll try to set that in kernel and get some logs.19:27
T42<edp_17> mal : what did you patch in your kernel?19:27
T42<edp_17> @elros34 : what does that audit_initialized do and how could it help?19:31
T42<elros34> it should make our dmesg readable, currently it's flooded with useless messeages so there is no way to see even boot messages, not to mention errors when dsme/droid-hal-init fails19:32
T42<edp_17> I see, thanks.19:35
T42<edp_17> @elros34 : I've changed that line like this. Is it okay?
T42<elros34> I hope so22:56

Generated by 2.17.1 by Marius Gedminas - find it at!