Thursday, 2018-08-09

Mukhamedgarifr0kk3rz: dev board with msm8940/8937 family soc08:01
r0kk3rzMukhamedgarif: so its not yet an actual device?08:06
r0kk3rzi assume it has an android adaptation for it08:06
Mukhamedgarifr0kk3rz: yes. It has android adaptation based on android N.08:14
r0kk3rzwhich hybris tree did you use?08:16
Mukhamedgarifr0kk3rz: hybris-1408:19
r0kk3rzdo you have a working lineage os image for your board?08:19
Mukhamedgarifr0kk3rz: no. it is like AOSP.08:21
r0kk3rzyeah so the hybris tree needs to match the base image fairly precisely, so that might not work08:22
r0kk3rzso i would suggest either patching the AOSP tree you have, or build a lineage os image first08:24
Mukhamedgarifr0kk3rz: I'd cherry picked hybris commits into source tree. So after some time, droid-hal-init start well and ssh works. Surfaceflinger doesn't crashes as well as test_hwcomposer. But there is still no any image on the screen.08:26
r0kk3rzwhat do the logs say? like logcat and journalctl08:28
T4<adampigg> wtf08:31
r0kk3rzthanks kimmoli :)08:32
NeoChapay@dampigg what does logs say tu-ru-ru08:34
Yardanicocan the reason of bootloop on my phone be that I removed "androidboot.selinux=permissive" from kernel commandline in ?08:34
Yardanicobut I added selinux=0 into as well08:34
NeoChapay@adampigg what does logs say tu-ru-ru08:34
r0kk3rz@adampigg looks niiice -
Yardanicobtw, for some reason, when building sailfish flashable .zip with "mic" russian keyboard layout gets removed from LXqt so I have to re-add it manually :D08:54
Yardanicohmm I still have a bootloop after adding permissive back09:55
malYardanico: how telling how soon the device reboots so we can guess what the reason is, there some possibilities10:13
mal*how about telling10:13
Yardanicomal, it reboots really fast, like 5-15 seconds after booting, I think it may be a kernel misconfiguration10:14
YardanicoI was able to mask ofono once so that's not it10:14
Yardanicowell, not mask but ln -s it to /dev/null10:15
malthat sounds line selinux issue10:16
Yardanicomal, yeah I think so too10:16
YardanicoI still don't understand should I add selinux=0 into defconfig or BoardConfg because last time I added it to BoardConfig device didn't bootloop but there wasn't "selinux=0" in /proc/cmdline10:17
Yardanicoalso should I keep androidboot.selinux=permissive in BoardConfig?10:17
malyou only do what is told10:19
*** Mikaela_ is now known as Mikaela11:12
Yardanicoso it stillbootloops, probably I need to change CONFIG_SECURITY_SELINUX_BOOTPARAM to `y` BUT it's already set up like that in defconfig but for some reason it11:14
Yardanico*it's not applied11:14
Yardanicoso I change CONFIG_SECURITY_SELINUX_BOOTPARAM to `y` in defconfig, recompile kernel, run kernel check and it still shows a warning about CONFIG_SECURITY_SELINUX_BOOTPARAM11:14
demonaxshHi, everyone11:19
demonaxshAre there any special features of bluetooth configure? I built a kernel with bluez support, but for mtk bluetooth devices it works on a special firmware. Now bluetooth apparently does not even detected, and of course devices are not detected either.11:19
r0kk3rzdemonaxsh: do you know what bt chip you have?11:30
demonaxshnot idea, in the kernel config for bluetooth (wifi etc) driver - CONFIG_MTK_COMBO=y, in the specification for the device I found only Bluetooth 4.0, BLE11:35
r0kk3rzhmm ok11:35
r0kk3rzlooks like mtk does do soc integrated ones11:35
r0kk3rzcheck the boardconfig.mk11:37
r0kk3rzor look for some bluedroid configs11:37
demonaxshin kernel config  also:11:39
demonaxshMTK_BT_BLE_MANAGER_SUPPORT = no11:42
demonaxshMTK_BT_BLUEDROID_A2DP_APTX = no11:42
demonaxshMTK_BT_BLUEDROID_AVRCP_TG_15 = yes11:42
demonaxshMTK_BT_BLUEDROID_HFP_AG_17 = yes11:42
demonaxshMTK_BT_BLUEDROID_PLUS = yes11:42
demonaxshMTK_BT_CHIP = MTK_CONSYS_MT673511:42
demonaxshMTK_BT_SUPPORT = yes11:42
demonaxshit's from ProjectConfig.mk11:43
Mukhamedgarifr0kk3rz: it is an attempt to run bootanimation.12:14
r0kk3rzMukhamedgarif: what am i looking at?12:25
Yardanicoehhh, I don't know.. in the first day of porting sailfish os I managed to easily boot to telnet stage but not it bootloops12:29
r0kk3rzYardanico: obviously you've been dutifully commiting changes to git somewhere so you can track what you changed12:30
r0kk3rzright? ;)12:30
Yardanicor0kk3rz, not really :D I will probably just revert kernel defconfig and boardconfig changes12:31
Yardanicoso I have CONFIG_AUDIT=y in defconfig (also CONFIG_SECURITY_SELINUX_BOOTPARAM=y) and selinux=0 in BOARD_KERNEL_CMDLINE in boardconfig, that's right or not?12:32
Mukhamedgarifr0kk3rz: I tried to run bootanimation. It is logcat.12:35
r0kk3rzwhy are you running the boot animation?12:35
r0kk3rzand not like, test_hwcomposer or something12:35
Mukhamedgarifr0kk3rz: I tried both. No video output.12:37
Yardanicoso HADK says to do the same if bootloop happens but it seems it doesn't help12:37
r0kk3rzMukhamedgarif: yeah but logs :P12:39
malYardanico: did you do any other selinux or audit related changes in defconfig?12:40
r0kk3rzdemonaxsh: got your kernel repo somewhere?12:42
Yardanicomal, I think I didn't, I'll remove all defconfig changes, and manually fix ERRORs from kernel check and then push it to github so I can show you12:42
Mukhamedgarifr0kk3rz:  here is logcat for "EGL_PLATFORM=hwcomposer test_hwcomposer"
r0kk3rzmmm that doesnt look very happy12:49
YardanicoI think I would just remove all sailfish os HADK files (because I already have all repos) and re-do some steps from HADK12:57
demonaxshkernel repo I got from vendor device, but in my opinion it is also in public domain, alps-3.18 mediatek13:24
r0kk3rzdemonaxsh: link :P13:28
demonaxshi do not find, what you need? config or all in kernel repo?13:34
r0kk3rzall of it13:37
demonaxshI hope I will not have problems with any licenses))13:39
demonaxshI also found /system/etc/bluetooth and there is also /etc/bluetooth, maybe all of exists in /system/etc/blutooth should be in /etc/bluetooth?13:42
malprobably no13:43
Yardanicoif I run `make hybris-hal`, compiled kernel will contain changes I've made in BoardConfig, right?13:45
malyes, and the you of course need to package it again with -d so it will be included in the next image13:46
Yardanicomal, well I want to experiment just with kernel (by using fastboot flash) without reassembling .zip every time13:46
malwell then it's fine, just note that in case you test some features built as modules it will more complicated than that13:48
Yardanicoso it still bootloops, kernel commit for sfos changes - , and on this line  I added "selinux=0" after "androidboot.selinux=permissive"13:49
malwas bootparam change taken into kernel, check the .config in out/13:53
Yardanicomal, oh, KERNEL_OBJ/.config contains "CONFIG_SECURITY_SELINUX_BOOTPARAM_VALUE=1", that's bad?13:57
malthe commandline should override that13:57
OhYashmal : None of the sensors work in a newer build that worked on previous one. (Proximity, Gyro & accelerometer worked previously not now). What area should I check?14:17
malOhYash: faq, ctrl+f ->sensors14:19
OhYashah shit forgot. sorry to disturb.14:20
rydareyou dare disturb the great mal?15:15
rydareoops i disturbed him with highlight15:16
rydaresorry m*l15:16
Yardanicoso I re-installed all SDK tolling etc, so now my device boots to terminal and dmesg says that it booted in SELinux permissive mode16:44
Yardanicothat's bad, right?16:44
Yardanico(i didn't add selinux=0 to cmdline yet)16:44
malnot sure what you mean, did you do anything to the adaptation packages?16:48
malYardanico: why are you only doing part of the instructions? it clearly says to add selinux=016:48
Yardanicomal, ok, I'll add selinux=016:49
Yardanicomal, ok, so now I added "selinux=0", and device still reboots (but I can connect through telnet for a couple of seconds), so dmesg now says "[    0.005842] SELinux:  Disabled at boot."17:02
Yardanicocomplete dmesg -
malYardanico: is that 14.1 based port?17:20
Yardanicomal, yes17:21
malYardanico: you didn't seem to run the symlink script mentioned in 14.1 porting section of faq (linked in channel topic)17:23
Yardanicohmm, I think I did but I'll try again17:23
maloh, you did, sorry17:23
malmissed that part of the log17:23
Yardanicohow can I dump all systemd logs to a file? journalctl > out.txt ?17:26
malYardanico: maybe journalctl -n1000 or something and the output it to file, some people use also --no-pager as option17:29
Yardanicomal, and what if I just copy /var/log/journal folder?   can I access contents on a PC?17:30
Yardanicoah nvm
malI think the issue could be that some firmware files are not found due to missing symlinks or something, or some needed partitions are not mounted17:33
Yardanicomal, hmm, I'll try to add some partitions to fixup-mountpoints17:34
malYardanico: what partitions did you add before?17:35
Yardanicomal, just same as mido (my device seem to have same partition table as mido because devices are pretty close hardware-wise)17:35
malmake sure the partition table really is the same17:36
Yardanicomal, yes I looked at all partition names17:36
malother option is that the udev rules are not working as they should, try to get output of ls -lR /dev/block/ and also journal log might tell something17:38
Yardanicooh no journalctl is a bit broken because when the device loads it has incorrect clock but then it syncs (maybe with device's built-in RTC) time and journalctl is broken into two parts17:38
maldoesn't matter17:38
Yardanicoit even tried to load libraries for "hwcomposer device" but all libs were not found17:39
Yardanicomal, this part of the log
Yardanicoso it seems like the last part of the log is always a calltrace about unbalanced IRQ 55 or 7317:45
Yardanicologs from 3 boots (or 4) without the first part (because it's separated)
Yardanicooh yeah I found some fixes for these IRQs in another kernel repo for my device17:47
YardanicoI didn't try this repo yet, I just think it might not work (because it's currently used for 8.1 treblized ROM)17:47
Yardanicobut I'll try it now17:48
Yardanicoyay another kernel seems to have solved  this issue! it doesn't reboot now. so splash logo was removed from the screen, it turned off, turned on, and turned off again. gonna see the logs18:16
malhmm, ok18:17
Yardanicohmm, test_hwcomposer can't find (and some other libs), but they are located in /system/vendor/lib6418:40
malwondering if those are needed or not18:42
Yardanicobtw, when I run test_hwcomposer backlight turns on for a second18:45
YardanicoI'll capture strace18:45
malI assume you have set the correct env var before that?18:45
Yardanicomal, you mean LD_LIBRARY_PATH or what?18:45
Yardanicoah, yeah, of course18:45
malwtth correct value18:45
Yardanicocorrect value?18:45
Yardanicofaq-hadk says "EGL_PLATFORM=hwcomposer test_hwcomposer"18:46
malsome device have had broken 32-bit hwcomposer and there is another way to get graphics working if that is the case, but logs would be useful18:46
YardanicoI run strace like that? "EGL_PLATFORM=hwcomposer strace test_hwcomposer"18:47
Yardanicohere's the strace (it's big)
malok, but also get dmesg, journalctl and /usr/libexec/droid-hybris/system/bin/logcat18:48
Yardanicoalso logcat is spammed by adsp18:48
Yardanicooh well, no, I forgot to say that dmesg is spammed by something too :)18:49
Yardanicomal, dmesg/journalctl/logcat -
malcan you see any adsp firmware files anywhere on the device, some vendor partition or /system/etc/firmware or something19:14
Yardanicomal, well there are some (some in other dirs and some files in /firmware/image)
malcan you get the ls -lR /dev/block I asked earlier19:18
malYardanico: the kernel messages look terrible, are you sure that kernel is ok?19:19
Yardanicomal, well it doesn't error like that on real system, I might try one another kernel too19:20
YardanicoI mean on android19:20
malwhere do you get the kernels?19:21
malwhy are there many kernels available?19:21
Yardanicomal, because there are several developers working on custom ROMs for my device, I can also try official kernel by xiaomi... but I think it would be worse19:21
wiktorek140Yardanico: have u correctly mounted fsg?19:29
r0kk3rzhmm xiaomi or xda special, thats a tough choice19:29
r0kk3rzid probably take xiaomi19:29
Yardanicowiktorek140, what is fsg? :P19:29
Yardanicoto fixup-mountpoints you mean? i'll try. can I do that on the device without rebuilding whole .zip?19:30
wiktorek140Yardanico: partition, on some devices (for example surnia) it required to bootup gui19:30
Yardanicoand can I edit it on device without remaking whole zip?19:31
wiktorek140Yardanico: u can try but it require to rebuild kernel, and edit /lib/systemd/system/fsg.mount on device19:31
Yardanicorebuild kernel? with what changes?19:32
wiktorek140add fsg partition to fixup-mountpoint and then rebuild kernel19:33
YardanicoI thought that it wouldn't be very hard to port sfos to my device because santoni and mido are already ported :)19:33
Yardanicoah, ok19:33
r0kk3rzwe have had reasonable success with xiaomi devices19:34
wiktorek140Yardanico: and consider adding "droid-hal-markw-detritus" to patterns19:36
Yardanicowiktorek140, it's already here19:36
wiktorek140oh, sorry, i dont supose to find it on upper part of file ;)19:37
Yardanicowiktorek140, so I just edit fixup-mountpoint and then `make hybris-boot`?19:39
wiktorek140Yardanico: yes, and then flash hybris-boot.img to device19:39
Yardanicothanks, I'll try19:39
malwiktorek140: Yardanico after editing fixup flashing kernel is not enough19:42
YardanicoI just think that it wouldn't help me because it's not required for santoni/mido19:42
Yardanico(and btw, all custom kernel for my device were developed from a custom mido kernel)19:43
Yardanico"This appears in dmesg, no GUI is displayed kgsl kgsl-3d0: |_load_firmware| request_firmware(a530_pm4.fw) failed: -2" i have same error!19:45
Yardanicoso I need to enable some flags19:46
YardanicoGUI WORKS19:50
YardanicoHOLY SH*T19:50
Yardanicoand touchscreen works!19:50
Yardanicoand audio works!19:52
Yardanicoand sms work (i received some right now)!19:53
r0kk3rzYardanico: see that wasnt hard :P20:11
Yardanicor0kk3rz, yes it wasn't :)20:12
Yardanicothank you all, I think I can probably fix almost all other issue via mido/santoni repos20:13
Yardanicoomg i'm so happy20:13
malYardanico: so my guess of some firmware problem was correct, but the reason was in kernel not in userspace20:19
Yardanicomal, yeah, thank you very much, I'm amazed how calm you are when people (like me) ask you a lot of stupid questions :)20:19
malYardanico: I usually only get annoyed if people don't even try to find the solution themselves but instead assume we just tell everything that is needed20:21
r0kk3rzand then dont follow instructions20:22
malYardanico: some things to look into are chapters 13.2 and 13.3 in hadk pdf20:33
Yardanicomal, ok, thanks20:33
YardanicoI probably wouldn't do anything more today :)20:33
malprobably a good idea20:35

Generated by 2.17.1 by Marius Gedminas - find it at!