Wednesday, 2020-04-15

T42<Rodrigosolari> after build and flasing00:47
T42<Rodrigosolari> this is dmesg log00:47
T42<Rodrigosolari> any ghelp ?01:29
dkterwhere are these messages being mirrored from?01:34
T42<Rodrigosolari> what?01:35
dkteri'm on freenode irc but it looks like there's some bridge to another chat service01:37
T42<Rodrigosolari> yes ! telegram01:37
dkteroh ok01:38
T42David %lastname% was added by: David %lastname%01:39
T42<David %lastname%> cool, found it01:40
T42<kkk_sw> How to create right ril_subscription.conf?05:34
T42<kkk_sw> I have taken this from other device, but ril work unstable05:35
r0kk3rzdepends on what you mean by unstable05:55
T42<Verevka86> @kkk_sw [How to create right ril_subscription.conf?], +++05:55
T42<kkk_sw> @r0kk3rz [depends on what you mean by unstable], Sometimes disconnected, incoming calls dnt work(when anyone try call, line busy and on phone missed call)06:00
r0kk3rzdont use the reply function please06:00
T42<Rodrigosolari> any help06:06
T42<Rodrigosolari> in the logcat sh: /usr/libexec/droid-hybris/system/bin/logcat: /system/bin/linker64: bad ELF interpreter: No such file or directory06:06
T42<kkk_sw> Have you installed lineage before installing sailfish?06:17
T42<kkk_sw> And for me after adding it was fixed06:18
T42<Verevka86> Please fix me help-07:25
T42<Verevka86> Already up-to-date.07:25
T42<Verevka86> Problem: bluez5-libs-devel-5.50+git3-1.10.2.jolla.armv7hl requires bluez5-libs = 5.50+git3, but this requirement cannot be provided07:25
T42<Verevka86>   not installable providers: bluez5-libs-5.50+git3-1.10.2.jolla.armv7hl[jolla]07:25
T42<Verevka86>  Solution 1: Following actions will be done:07:25
T42<Verevka86>   deinstallation of obexd-server-0.48+git17-1.2.2.jolla.armv7hl07:25
T42<Verevka86>   deinstallation of obexd-0.48+git17-1.2.2.jolla.armv7hl07:25
T42<Verevka86>   deinstallation of droid-config-sagit-bluez4-1-1.armv7hl07:25
T42<Verevka86>   deinstallation of bluez-libs-4.101+git77-1.3.1.jolla.armv7hl07:25
T42<Verevka86>   deinstallation of kf5bluezqt-bluez4-5.24.0+git17-1.4.1.jolla.armv7hl07:25
T42<Verevka86>   deinstallation of bluez-4.101+git77-1.3.1.jolla.armv7hl07:25
T42<Verevka86>  Solution 2: do not ask to install a solvable providing pkgconfig(bluez5)07:25
T42<Verevka86>  Solution 3: break bluez5-libs-devel-5.50+git3-1.10.2.jolla.armv7hl by ignoring some of its dependencies07:25
T42<Verevka86> Choose from above solutions by number or cancel [1/2/3/c] (c): c07:25
T42<Rodrigosolari> @kkk_sw [Have you installed lineage before installing s …], Yes07:30
T42<elros34> @Rodrigosolari logcat will not work until droid-hal-init starts working correctly. For now systemd doesn't even start thus no journal. You didn't disable selinux07:47
r0kk3rz@Vereka86 what are you doing?07:50
T42<adampigg> Rokk3rz: u miss typed....i guess they need to choose option 1 if they want bluebinder/bluez5 on that port07:54
T42<Verevka86> @r0kk3rz [@Vereka86 what are you doing?], I collect packages, bluebinder is not going to07:54
T42<Verevka86> @adampigg [Rokk3rz: u miss typed....i guess they need to …], No choice 😔07:56
*** birdzhang1 is now known as birdzhang09:16
T42<adampigg> @Verevka86 why?10:14
T42<adampigg> do you need bluebinder or not?10:14
T42<Verevka86> @adampigg [do you need bluebinder or not?], Yes10:15
T42<adampigg> so, you need to remove bluez4 and use bluez510:15
deathmist breaks every single one of those checks, assuming this is because HABUILD_DEVICE is also $DEVICE so it's not searching 2 files anymore @ as well as line 363 (before "sudo mic create")10:21
T42<Verevka86> @adampigg [so, you need to remove bluez4 and use bluez5], I spent a whole week collecting on these sources, and this problem never occurred 😄10:51
deathmistmal: could you perhaps confirm the issue? this is on dhd submodule master branch and I'm testing 3.3 building after getting past my previous issue thanks to
T42<elros34> deathmist: I can confirm this grep "file1 file1" issue. I reported it few days ago but there is no hot-hadk anymore so it was probably lost10:55
T42<Nebrassy> after creating files for camera, I can switch between front and back in the app, focus and flash works etc, but can't record video, when trying to it just takes a photo13:17
T42<adampigg> @Nebrassy what files did you create for camera?13:45
T42<adampigg> @Nebrassy you shouldnt need the first 213:51
T42<adampigg> and was the third generated by droid-camres?13:51
T42<Nebrassy> I just followed the pdf13:51
T42<Nebrassy> Yes13:51
T42<Nebrassy> But if I delete the files nothing changes13:52
T42<adampigg> does advanced camera work?13:59
T42<Nebrassy> Haven't tried, where's that?14:00
deathmistDanct12: dummy_netd fixes my 16.0 cellular data issue btw :) I just did `systemctl enable dummy_netd`, rebooted and stuff works! I'm testing SFOS 3.3 if version matters (it shouldn't)14:01
T42<adampigg> @Nebrassy openrepos or store14:02
T42<Nebrassy> I'll try14:04
T42<Verevka86> @adampigg [so, you need to remove bluez4 and use bluez5], How to do it?14:15
T42<Verevka86> @deathmist [mal: could you perhaps confirm the issue? this …], Now I can’t build 3.2 :-( bluez4 came from somewhere and I don’t know how to remove it.14:28
T42<Nebrassy> @adampigg crashes when trying to record14:28
T42<Verevka86> @mal [ …], Big thank 🤝15:02
T42<Rodrigosolari> @Nebrassy [Use journalctl --no-page for now], thys my log17:20
dkterwhy is it only checking CM repos? i set it up with LOS 15.117:33
T42<adampigg> @Nebrassy [@adampigg crashes when trying to record], Logs?  logcat?17:49
maldeathmist: can you explain when that issue happens?18:52
deathmistmal: on build_package -i, -gg etc, look at all the lines affected and basically all the things after them are affected on dhd submodule master branch18:54
deathmistI have a solution on mind, but not sure if it's the best: remove the quotes and do pattern_lookup="/dev/null" IF it's unset after the ls commands, that way the build won't get stuck and still works without patterns etc18:56
malso the issue happens only when there are both DEVICE and HABUILD_DEVICE?18:59
maldeathmist: I can't reproduce the issue, I have both DEVICE and HABUILD_DEVICE and latest dhd submodule and -gg works fine19:00
T42<Rodrigosolari> mal: you can help me?19:01
mal@Rodrigosolari depends on what you are trying to fix19:04
deathmistmal: I only have DEVICE defined, not sure if this affects things; here is my droid-hal repo:
T42<Rodrigosolari> mal: the phone is in the logo, I don't have a logcat, why is it in the beginning, but I already sent the log of "journalctl --no-page"19:06
T42<Rodrigosolari> mal: thys my log19:08
T42<elros34> deathmist how can you have only DEVICE set if set HABUILD_DEVICE=$DEVICE if it's empty19:11
T42<elros34> @Rodrigosolari did you finally disable selinux?19:12
deathmistelros34: I thought mal meant having both DEVICE and HABUILD_DEVICE defined in spec files and such, but yes they are both set when the script is running19:12
deathmistand are equal19:12
malthat is a good question, also is droid-hal-init working, have the android base specific things done from faq19:12
maldeathmist: I have different DEVICE and HABUILD_DEVICE for my device and I have no issues19:13
T42<Rodrigosolari> @elros yes, I was wrong on a line in, it is already disabled19:13
mal@Rodrigosolari did you do then android base specific things from faq?19:13
T42<elros34> @Rodrigosolari so reboot and provide full dmesg and journalctl19:14
T42<Rodrigosolari> mal: Excuse me, I don't understand what you're trying to tell me.19:16
T42<elros34> mal: how this grep can work for you, it's like: grep -q "^- gstreamer1.0-droid" "file1 file1" so grep think it's one file19:17
T42<elros34> unless it find only one file for you because your HABUILD_DEVICE is different then DEVICE19:23
T42<Rodrigosolari> @elros34 [@Rodrigosolari so reboot and provide full dmes …], my Dmesg
T42<Rodrigosolari> my journalctl
T42<elros34> this is useless, I told you to reboot19:30
T42<Rodrigosolari> @elros Excuse me, but restart the phone and I get the same result19:35
T42<elros34> I don't believe you: reboot -f and then grab dmesg19:38
malwhere is the rest of those logs, those are just very short piece from the end19:42
mal@Rodrigosolari you didn't tell which android base you are using19:43
T42<Nebrassy> It's hybris 1619:44
T42<Nebrassy> And I suspect selinux is still enabled19:44
mal@Rodrigosolari use "journalctl -alb --no-pager" for getting journal log19:44
T42<Rodrigosolari> @elros
T42<Rodrigosolari> is dmesg19:47
mal@Rodrigosolari show output of "cat /proc/cmdline" and "zcat /proc/config.gz | grep SELINUX"19:49
T42<Rodrigosolari> @Nebrassy [It's hybris 16],
T42<Rodrigosolari> disable, selinux, and add config in the defconfig for te dissably19:50
T42<Rodrigosolari> mal: lineageos 1619:51
T42<Rodrigosolari> its my android19:51
T42<elros34> is that really whole dmesg because looks like it's cuted and the begining19:52
T42<Rodrigosolari> (Photo, 1280x960)
T42<Rodrigosolari> mal: it seems that yes, selinux is activated, even though I followed the comments, and the guide, and disabled selinux = 0, and added CONFIG_SECURITY_SELINUX_BOOTPARAM = y20:04
T42<Rodrigosolari> in the defconfig20:04
T42<elros34> so what changed? Yesterday you were able to provide usable dmesg20:06
T42<Rodrigosolari> what change, remove #BOARD_KERNEL_CMDLINE + = androidboot.selinux = 0 and added #BOARD_KERNEL_CMDLINE audit = 0, because it was already added, selinux = 0, and in defconfig CONFIG_SECURITY_SELINUX_BOOTPARAM = and, very apart look for the log porters, to see directions to other users, that's where I removed the audit = 0, but whe20:12
T42n I do what I indicate mal: I have androidboot.secureboot = 1, which I imagine selinux is enabled20:12
T42<Rodrigosolari> @elros34 this a log
T42<elros34> CONFIG_SECURITY_SELINUX_BOOTPARAM_VALUE=0 but I think selinux=0 have higher priority so it should be diasabled. The issue is with your logs20:16
mal@Rodrigosolari did you copy the files as said here
mal@elros34 I would also assume the parameter from commandline would override it20:19
mal@Rodrigosolari androidboot.secureboot should not matter, selinux=0 should disable selinux and you seem to have it, check the previous message I wrote20:20
T42<Rodrigosolari> mal: Of course, if you had read the faq, and cloned it, of course also apply patches and the build20:24
T42<Rodrigosolari> I will remove the packages and I will do it again :)20:24
malremove what?20:24
malwhat does "systemctl status droid-hal-init" show on device20:25
T42<Rodrigosolari> mal: remove target and installed
malwhy are you ignoring some of my messages? there is a reason why I ask for some output20:58
T42<Rodrigosolari> Yes and I understand you and I apologize, I do not ignore any of your messages, I read you completely, only that my phone is charging, from trying to charge the battery, give me about 10 minutes more than I am charging it with a fast charge21:01
maldeathmist: I still don't understand the problem you have, I printed the output of pattern_lookup variable in the script and it only has one path21:03
mal@elros34 why does the pattern_lookup contain more than one file for you, can you show your config repo21:05
mal@elros34 oh, so the problem happens only if the variables DEVICE and HABUILD_DEVICE are the same?21:07
deathmistmal: look at that's exactly what it stores in pattern_lookup variable for me21:08
T42<elros34> mal: because I never set HABUILD_DEVICE so set it to same value as $DEVICE: As a result patter_lookup contains path to 2 same pattern files21:08
malyeah, that was probably tested mainly on device with different values for those21:08
malwhich is that case also on my fp221:09
maldeathmist: one way to fix that would be to add a check if the DEVICE and HABUILD_DEVICE are the same and use a different ls command in that case21:10
malbut that would be a bit ugly21:11
T42<Rodrigosolari> mal: sh-3.2# systemctl status droid-hal-init21:11
T42<Rodrigosolari> ● droid-hal-init.service - droid-hal-init21:11
T42<Rodrigosolari>    Loaded: loaded (/lib/systemd/system/droid-hal-init.service; disabled; vendor preset: enabled)21:11
T42<Rodrigosolari>    Active: inactive (dead)21:11
T42<elros34> maybe find -regex should be used instead ls21:11
mal@elros34 or maybe "ls foo | sort -u"21:14
malor something like that to get unique ones21:14
malthat sort -u also would seem to work21:15
mal@Rodrigosolari so there is one of your problems, droid-hal-init should be running21:15
mal@Rodrigosolari have you verified that partitions are correctly mounted, show output of "mount"21:16
maldeathmist: this seems to work for me pattern_lookup=$(ls "$ANDROID_ROOT"/hybris/droid-configs/patterns/jolla-hw-adaptation-{$DEVICE,$HABUILD_DEVICE}.yaml | sort -u 2>/dev/null)21:17
T42<elros34> I think head -n1 would be better because pattern_lookup can't contains 2 even unique paths after that commit:
maltrue, either one should work21:22
mal@elros34 also | uniq seems to work21:25
malso we have many options21:25
T42<Rodrigosolari> mal: this my log of mount
maldeathmist: just curious, how is that mainlining going?21:26
deathmistmal: I'd probably have to work on display next because there is no way framebuffer is any good for more than a demo :D21:27
deathmistbut about the fixes you suggested, if both jolla-hw-adaptation-{$DEVICE,$HABUILD_DEVICE}.yaml exist it would still fail though even if they were unique filenames21:28
maldeathmist: yeah, display can a bit of a pain, I have been helping with fp2 mainline kernel and that is quite difficult21:28
maldeathmist: if both exist then someone made a bad mistake21:28
maldeathmist: or is there some case where both of those exist and that it would make sense21:29
T42<Rodrigosolari> mal: my droid-haal-init is chage, is running21:31
maldeathmist: fp2 has quite many things working in mainline kernel, wlan, bt and even modem and other stuff is up in wip code21:31
mal@Rodrigosolari just to be sure, check that "ls /usr/libexec/droid-hybris/system/etc/init/" returns correct things21:32
mal@Rodrigosolari does /usr/libexec/droid-hybris/system/bin/logcat work now?21:32
deathmistmal: I can't think of a situation where both those would be unique and both files would exist so maybe it's just fine to go with the " | uniq" solution as that would at least fix the regression21:32
T42<Rodrigosolari> mal: yes i work logcat21:33
T42<Rodrigosolari> :)21:33
T42<Rodrigosolari> thanks very tanks21:33
mal@Rodrigosolari what did you change to make that work?21:33
maldeathmist: yeah, I can't see a reasonable setup where those both would exist, unless some android build would be used for two devices both with same HABUILD_DEVICE but only one with DEVICE with the same one, but not sure if such a setup would exist21:35
maldeathmist: if you want to make a pull request for that go ahead, otherwise I'll try to do it tomorrow21:36
T42<Rodrigosolari> mal: I rebuilt the packages with the new target, selinux, it was deactivated, I checked it, I flashed a new created package, because I created a new zip package, and I flashed it with potato, since I use its sources21:37
T42herrkeemail was added by: herrkeemail21:37
T42<Rodrigosolari> mal: in the logcat obtains 12-04 17:44:25.045  3195  3195 W ServiceManagement: Waited for hwservicemanager.ready for a second, waiting another...21:42
mal@Rodrigosolari you are again ignoring my requests 00:32 < mal> @Rodrigosolari just to be sure, check that "ls /usr/libexec/droid-hybris/system/etc/init/" returns correct things21:43
T42<Rodrigosolari> obtains sh-3.2# ls /usr/libexec/droid-hybris/system/etc/init21:45
T42<Rodrigosolari> disabled_services.rc  hwservicemanager.rc  hybris_extras.rc  servicemanager.rc  vndservicemanager.rc21:45
T42<Rodrigosolari> mal:21:46
malis hwservicemanager process running?21:49
T42<Rodrigosolari> mal: I see that with systemctl?21:53
deathmistmal: PR for uniq pipe or fix unique files both existing? :p21:53
T42<Rodrigosolari> ok mal: in my logcat obtains --------- beginning of crash21:56
T42<Rodrigosolari> 12-04 17:15:14.470  3111  3111 F linker  : CANNOT LINK EXECUTABLE "/system/bin/hwservicemanager": library "/usr/libexec/droid-hybris/system/lib64/" needed or dlopened by "/system/bin/hwservicemanager" is not accessible for the namespace "(default)"21:56
T42<Nebrassy> @Rodrigosolari post full logcat21:57
T42<Nebrassy> And if journalctl works now, post that too21:57
maldeathmist: I think we don't have to support case if two files with different names exist because that should not happen21:59
deathmist@Rodrigosolari you need ld.config.28.txt bind-mount unit in your droid-config repo sparse files (you can do changes on the device for testing temporarily) -> + + (note: this last one is a symbolic link!)21:59
maldeathmist: that case if two files with different names exist would need more changes in other places and if such a situation happens in the future then it can be fixed but no point in doing it now22:00
maldeathmist: also if you like maybe make a PR to faq hybris-16.0 section about that ld.config.28.txt22:01
T42<Nebrassy> Since this is a common problem, maybe do something like22:05
T42<Nebrassy> ?22:05
deathmist@Nebrassy SFOS != Android, and I've seen many devices here that haven't needed these ld.config patches either22:12
malyeah, not all devices need it22:12
T42<Nebrassy> mal: any devices with enforced vndk that don't need it?22:14
malno idea22:15
T42<Nebrassy> ie devices that don't have ld.config.vndk_lite.txt, that doesn't enforce paths22:15
malI'm not so familiar with different setups android devices have22:26
T42<Nebrassy> Fairly certain all devices not using BOARD_VNDK_22:33
T42<Nebrassy> RUNTIME_DISABLE will need that patch22:33
dkterok i'm a little lost i think i have the wrong version of... something, somehow22:40
dktertrying to run breakfast22:41
dkterit's looking for in device/*/mako, i don't have that but i do have lineage.mk22:41
dkterwhat am i doing wrong?22:42
T42<Nebrassy> dukter: Which android version, and which hybris branch?22:43
dkterlineage 15.122:43
dkterthis is my manifest
dktershould i not include the bottom 3 items?22:44
T42<Nebrassy> Which hybris branch did you use when you did repo init22:44
dkteruhh hang on22:45
T42<elros34> these droid repos are very old, you shoudn't use them. Instead create new according to HADK22:46
T42<Nebrassy> Do22:46
T42<Nebrassy> repo init -u -b hybris-15.122:46
T42<Nebrassy> And repo sync again22:46
dkteroh ok i'm pretty sure i did the wrong one22:47
deathmistdkter: you can just `history | grep init` to check if you chose 15.122:47
dkteri think i might have accidentally done 11, because i started with 11 and then decided to go with 15.122:48
dkterthanks @Nebrassy22:48
deathmistdkter: you may also need in your manifest for the HAL build stuff to pass22:50
dkterdeathmist: oh really? why?22:51
dkteri copied my manifest from one on a mer-hybris repo22:52
T42<Rodrigosolari> i try now, erase all, and install lineage os, and flashing sailfish os, and hwservicemanager is runing23:46
T42<Rodrigosolari> @Nebrassy this my logcat output
T42<Nebrassy> Do the ld.config changes @Rodrigosolari23:49
T42<Rodrigosolari> yes @Nebrassy23:56

Generated by 2.17.1 by Marius Gedminas - find it at!