| T42 | <Lukapanio> hi again | 06:30 |
|---|---|---|
| T42 | <Lukapanio> i have built | 06:30 |
| T42 | <Lukapanio> i have: https://bpaste.net/show/O62DI | 06:31 |
| T42 | <Lukapanio> in dmesgon mu pc | 06:31 |
| T42 | <Lukapanio> so how to debug? | 06:31 |
| T42 | <Lukapanio> usb net disappear fast | 06:31 |
| T42 | <Lukapanio> I can telnet if I do touch /data/.stowaways/sailfishos/init_enter_debug2 | 07:02 |
| T42 | <Lukapanio> ok enabled journal ctl | 07:06 |
| T42 | <Lukapanio> will left it for some time | 07:06 |
| T42 | <Lukapanio> ok | 07:08 |
| T42 | <Lukapanio> will reboot | 07:08 |
| T42 | <Lukapanio> https://bpaste.net/show/WZ6CG | 07:12 |
| T42 | <Lukapanio> any idea what is wrong? | 07:18 |
| r0kk3rz | you skipped the first boot phase | 07:26 |
| r0kk3rz | thats whats wrong | 07:26 |
| T42 | <Lukapanio> @r0kk3rz where can i get that, this is full output of journalctl | 07:30 |
| T42 | <Lukapanio> why there is no usb after intramfs | 07:39 |
| T42 | <Lukapanio> may be somw int breakes it? | 07:40 |
| r0kk3rz | double check your fixup mountpoints | 07:47 |
| T42 | <elros34> @Lukapanio Are you sure you disabled selinux? Confirm via zcat /proc/config.gz | grep -i selinux and cat /proc/cmdline. Also you need straggler files because droid-hal-init fails | 08:18 |
| T42 | <Lukapanio> i will add all partitions to fixup-mountpoints | 08:20 |
| T42 | <Lukapanio> and yes i have selinux=0 in cmdline | 08:20 |
| *** OhYash1 is now known as ohyash | 08:31 | |
| T42 | <Lukapanio> for fix-mountpoints mod can i rebuild onlyhybris-boot? | 08:33 |
| T42 | <Lukapanio> or all system | 08:33 |
| T42 | <Lukapanio> ? | 08:36 |
| T42 | <elros34> you need to make hybris-hal and package droid-hal but that's not your issue | 08:36 |
| T42 | <elros34> check in dmesg about selinux | 08:37 |
| T42 | <Lukapanio> sh-3.2# dmesg | grep selinux | 08:38 |
| T42 | <Lukapanio> sh-3.2# | 08:38 |
| T42 | <elros34> by default grep is case sensitive, also I told you to check config.gz also to make sure bootparam is enable as usually people forget about it | 08:40 |
| *** OhYash1 is now known as ohyash | 08:41 | |
| T42 | <Lukapanio> sh-3.2# zcat /proc/config.gz | grep -i selinux | 08:58 |
| T42 | <Lukapanio> sh-3.2# | 08:58 |
| T42 | <Lukapanio> @elros34 | 08:58 |
| T42 | <Lukapanio> it is empty | 08:58 |
| T42 | <elros34> check 9.3.1 and make sure your generated out/*/.config is correct | 09:02 |
| T42 | <Lukapanio> hmmm CONFIG_SECURITY_SELINUX=y is comented in my defconig | 09:04 |
| T42 | <Lukapanio> so @elros34 must i uncoment and disable wia bootparam | 09:11 |
| T42 | <Lukapanio> or comented selinux is ok too? | 09:11 |
| T42 | <Lukapanio> droid-hal-init: Failed to initialize property area | 09:11 |
| T42 | <Lukapanio> looks bad | 09:11 |
| T42 | <Lukapanio> and then droid-hal-init.service: Main process exited, code=exited, status=1/FAILURE | 09:12 |
| T42 | <elros34> you need to follow 9.3.1 so selinux is disabled via bootparam, next is droid-hal-init | 09:12 |
| T42 | <Lukapanio> ok | 09:15 |
| T42 | <Lukapanio> rebuilding | 09:15 |
| T42 | <Lukapanio> sh-3.2# zcat /proc/config.gz | grep -i selinux | 09:25 |
| T42 | <Lukapanio> CONFIG_SECURITY_SELINUX=y | 09:25 |
| T42 | <Lukapanio> CONFIG_SECURITY_SELINUX_BOOTPARAM=y | 09:25 |
| T42 | <Lukapanio> CONFIG_SECURITY_SELINUX_BOOTPARAM_VALUE=1 | 09:25 |
| T42 | <Lukapanio> # CONFIG_SECURITY_SELINUX_DISABLE is not set | 09:25 |
| T42 | <Lukapanio> CONFIG_SECURITY_SELINUX_DEVELOP=y | 09:25 |
| T42 | <Lukapanio> CONFIG_SECURITY_SELINUX_AVC_STATS=y | 09:25 |
| T42 | <Lukapanio> CONFIG_SECURITY_SELINUX_CHECKREQPROT_VALUE=1 | 09:25 |
| T42 | <Lukapanio> # CONFIG_SECURITY_SELINUX_POLICYDB_VERSION_MAX is not set | 09:25 |
| T42 | <Lukapanio> CONFIG_DEFAULT_SECURITY_SELINUX=y | 09:25 |
| T42 | <Lukapanio> CONFIG_DEFAULT_SECURITY="selinux" | 09:25 |
| T42 | <Lukapanio> sh-3.2# | 09:25 |
| T42 | <Lukapanio> @elros34 | 09:25 |
| mal | is CONFIG_AUDIT=y? | 09:26 |
| T42 | <elros34> now it must be | 09:26 |
| mal | true, that was from device | 09:26 |
| T42 | <Lukapanio> yep ONFIG_AUDIT=y | 09:28 |
| T42 | <Lukapanio> so @elros34 whats next? | 09:32 |
| T42 | <elros34> not the best instruction but should be enough: https://wiki.merproject.org/wiki/Adaptations/faq-hadk#droid-hal-init:_Failed_to_initialize_property_area | 09:40 |
| T42 | <Lukapanio> @elros34 [not the best instruction but should be enough: …], can it be done from initram fs? | 09:59 |
| T42 | <adampigg> from what i understand, modern android needs twrp delivered as an initramfs, loaded from the regular kernel. has anyone tried booting a twrp initramfs from a sailfish bootimg, making it selectable by holding a key? | 10:03 |
| T42 | <Lukapanio> @elros34 prolem is no usb after running echo "continue" >/init-ctl/stdin | 10:04 |
| T42 | <elros34> it might fix automagically when droid-hal-init will work | 10:08 |
| T42 | <Lukapanio> but how can i execute mask droid-hal-init: ln -s /dev/null /etc/systemd/system/droid-hal-init.service, reboot and then strace -f /sbin/droid-hal-init | 10:09 |
| T42 | <Lukapanio> in intramfs? | 10:09 |
| T42 | <elros34> skip it, for sure you need straggler files. Btw you are not in initframfs but in init-debug script in sfos rootfs | 10:10 |
| T42 | <Lukapanio> ouch yep | 10:10 |
| T42 | <Lukapanio> so @elros34 i must add all files in out/target/product/cedric or files and folders | 10:16 |
| T42 | <Lukapanio> like: | 10:16 |
| T42 | <Lukapanio> %define straggler_files\/init.mmi.boot.sh\/init.mmi.touch.sh\/init.qcom.ssr.sh\/selinux_version\/service_contexts\%{nil} | 10:16 |
| T42 | <Lukapanio> like /file1 file2 file3 or like /file1 /folder1 /file2 | 10:18 |
| T42 | <Lukapanio> or may be /file1 /folder1/file1 | 10:18 |
| T42 | <Lukapanio> ? | 10:18 |
| T42 | <elros34> I don't think folders are needed maybe simlinks like /vendor and for sure init*sh and *context* files | 10:20 |
| T42 | <Lukapanio> problem is there is no context files @elros34 | 10:22 |
| T42 | <hacker12455> You sure? | 10:22 |
| T42 | <hacker12455> I think I had some context files, not sure | 10:22 |
| T42 | <Lukapanio> yep @hacker12455 check on server | 10:22 |
| T42 | <hacker12455> Later | 10:24 |
| T42 | <Lukapanio> @elros34 | 10:37 |
| T42 | <Lukapanio> HABUILD_SDK [cedric] hello@Strangers-In-My-Network:~/hadk/out/target/product/cedric/root$ grep -r context | 10:37 |
| T42 | <Lukapanio> Binary file sbin/charger matches | 10:37 |
| T42 | <Lukapanio> init.rc: # Set the security context of /adb_keys if present. | 10:37 |
| T42 | <Lukapanio> init.rc: # Set the security context of /postinstall if present. | 10:37 |
| T42 | <Lukapanio> init.rc: # Set SELinux security contexts on upgrade or policy update. | 10:37 |
| T42 | <Lukapanio> Binary file init matches | 10:37 |
| T42 | <Lukapanio> add init file? | 10:39 |
| T42 | <elros34> if you mean /init then no | 10:40 |
| T42 | <Lukapanio> yep i mean /init | 10:40 |
| T42 | <Lukapanio> but where are contexts? | 10:41 |
| T42 | <elros34> I thought instruction is clear but looks like no. If there are not such a files then move on. You can always strace droid-hal-init if it will still fail | 10:42 |
| T42 | <Lukapanio> so next is add droid-hal-$DEVICE-detritus package to patterns | 10:44 |
| T42 | <Lukapanio> ? | 10:44 |
| T42 | <elros34> sure but you can manually copy (sdcard or via twrp) and install this package before creating whole new image | 10:45 |
| T42 | <adampigg> mal, lbt, abranson, 3.2 isnt on obs yet!! | 11:51 |
| T42 | <eugenio_g7> :D | 11:51 |
| abranson | this is really turning into a telegram channel :D | 11:52 |
| mal | @adampigg :D | 11:53 |
| mal | abranson: it's a bit sad in my opinion :) | 11:53 |
| abranson | well, it does make it easier for people to follow long term. but we should never drop irc if there's no better open standard. | 11:54 |
| T42 | <eugenio_g7> Sorry about that, I should take the plunge and move everything over matrix with real bridges :D | 11:54 |
| T42 | <adampigg> irc is convenient when im at home on a pc, telegram more convenient when away to keep up with the conversation | 11:56 |
| mal | @adampigg screen+irssi and ssh ftw | 11:58 |
| r0kk3rz | rip irc | 12:00 |
| T42 | <adampigg> mal: and there you have summarised in 5 words why telegram is more convenient ;) | 12:01 |
| mal | @adampigg https://xkcd.com/1782/ | 12:02 |
| T42 | <adampigg> ha, ive seen that...... | 12:04 |
| T42 | <adampigg> perhaps we should/ all move to slack, or teams! | 12:05 |
| T42 | <adampigg> i wonder if the BT improvements allow starfish to be used now? | 12:05 |
| mal | @adampigg was that the asteroid client? we were thinking about that also | 12:09 |
| T42 | <adampigg> yes | 12:09 |
| T42 | <Lukapanio> @elros34 repo problem: nothing provides droid-hal-cedric-detritus needed by pattern:jolla-hw-adaptation-cedric-1-1.noarch, | 12:43 |
| T42 | <elros34> have you built it? | 12:46 |
| T42 | <Lukapanio> i have run build_packages.sh | 12:46 |
| T42 | <elros34> so make sure package exists | 12:47 |
| T42 | <Lukapanio> sorry for stupid question but how? | 12:48 |
| T42 | <elros34> in your local repo: droid-local-repo/$DEVICE/droid-hal-$DEVICE/ | 12:49 |
| T42 | <Lukapanio> hmmm it is not there | 12:50 |
| T42 | <Lukapanio> how must it build @elros34 ? | 12:58 |
| T42 | <elros34> are you sure you have followed 7.2.2 first dot carefully? | 12:59 |
| T42 | <Lukapanio> шрфму - droid-hal-cedric-detritus | 13:02 |
| T42 | <Lukapanio> in Jolla-hw-adaptation-cedric.yaml | 13:03 |
| T42 | <elros34> hmm I think I found why straggler files are not auto detected. It's because busybox's diff use diffrent output format "-" instead "< " | 13:03 |
| T42 | <Lukapanio> but i have no %define straggler_files\ | 13:04 |
| T42 | <Lukapanio> becouse i cant finde that files | 13:04 |
| T42 | <Lukapanio> that files seems to be packed | 13:04 |
| T42 | <Lukapanio> and no context files | 13:05 |
| T42 | <elros34> show your rpm/droid-hal- | 13:05 |
| T42 | <elros34> $DEVICE.spec | 13:05 |
| T42 | <Lukapanio> @elros34 [show your rpm/droid-hal- | 13:07 |
| T42 | <Lukapanio> $DEVICE.spec], https://bpaste.net/show/RENK6 | 13:07 |
| T42 | <elros34> @Lukapanio could you test something? sb2 -t $VENDOR-$DEVICE-armv7hl -m sdk-build -R. install some editor: zypper install vim; then edit /usr/lib/rpm/check-files according to https://pastebin.com/Jy6Jz3pc | 13:13 |
| T42 | <elros34> then run rpm/dhd/helpers/build_packages.sh —droid-hal | 13:14 |
| T42 | <Lukapanio> @elros34 [then run rpm/dhd/helpers/build_packages.sh —dr …], rpm/dhd/helpers/build_packages.sh —droid-hal | 13:17 |
| T42 | <Lukapanio> unknown option(s): —droid-hal | 13:17 |
| T42 | <Lukapanio> ouch | 13:18 |
| T42 | <Lukapanio> iknow | 13:18 |
| T42 | <Lukapanio> not - but -- | 13:18 |
| T42 | <Lukapanio> yay | 13:20 |
| T42 | <Lukapanio> error: Installed (but unpackaged) file(s) found: | 13:20 |
| T42 | <Lukapanio> /bugreports | 13:20 |
| T42 | <Lukapanio> /d | 13:20 |
| T42 | <Lukapanio> /sdcard | 13:20 |
| T42 | <elros34> yes finally, that is good! | 13:20 |
| T42 | <Lukapanio> thanks | 13:20 |
| T42 | <Lukapanio> will add them now | 13:20 |
| T42 | <elros34> but i expected different output, these files are not critical so grab new journalctl | 13:22 |
| T42 | <Lukapanio> after rebuilding or now? | 13:22 |
| T42 | <Lukapanio> with old rootfs | 13:23 |
| T42 | <elros34> mal: lack of "Installed (but unpackaged) file(s) found" is caused by busybox's grep which use different output format. Not sure how to fix it properly: https://pastebin.com/Jy6Jz3pc | 13:23 |
| T42 | <elros34> @Lukapanio you can make on device ln -s /sys/kernel/debug /d but I doubt it change anything | 13:25 |
| T42 | <Lukapanio> thanks @elros34 finalyi have droid-hal-cedric-detritus-0.0.6-201910291322.armv7hl.rpm | 13:25 |
| T42 | <Lukapanio> must i rebuild rootfs or just install droid-hal-cedric-detritus-0.0.6-201910291322.armv7hl.rpm | 13:25 |
| T42 | <Lukapanio> ? | 13:25 |
| T42 | <elros34> like I said these 3 symlinks doesn't looks important | 13:26 |
| T42 | <Lukapanio> but what is problem then? | 13:27 |
| rinigus | hi! as I have android's KeyMasterHalDevice crashing on boot: is it of any importance for sfos? | 13:27 |
| T42 | <elros34> @Lukapanio imposible to know without logs, debugging | 13:28 |
| T42 | <Lukapanio> @elros34 [@Lukapanio imposible to know without logs, deb …], journalctl? | 13:29 |
| mal | @elros34 ok, needs some testing | 13:29 |
| T42 | <Lukapanio> ok will shut down and send logs | 13:30 |
| T42 | <elros34> There many patches in rpm sources so maybe just adding one more | 13:31 |
| rinigus | I can see some discussion regarding keymaster, but not really sure whats up with it. | 13:32 |
| T42 | <Lukapanio> https://pastebin.com/Kdbw8B8r | 13:32 |
| T42 | <elros34> Missed 291 kernel messages: provide dmesg | 13:34 |
| vknecht | rinigus, seems it had to be disabled in oreo based ports, dunno about pie ( https://github.com/mer-hybris/droid-config-sony-nile/blob/master/sparse/usr/libexec/droid-hybris/system/etc/init/android.hardware.keymaster%403.0-service-qti.rc ) | 13:35 |
| T42 | <Lukapanio> @elros34 [Missed 291 kernel messages: provide dmesg], https://pastebin.com/7GAZCEA9 | 13:37 |
| rinigus | vknecht: according to Mister_Magister (https://piggz.co.uk/sailfishos-porters-archive/index.php?log=2019-08-17.txt#line178), he managed to get gui working with that starting. | 13:37 |
| rinigus | ... on Pie | 13:37 |
| T42 | <Lukapanio> but after debug shell no usb | 13:39 |
| Mister_Magister | rinigus: what i | 13:39 |
| Mister_Magister | rinigus: no that was just coincidence | 13:39 |
| Mister_Magister | gui was starting weirdly after 30 minutes of boot for whatever reason | 13:39 |
| rinigus | Mister_Magister: | 13:40 |
| T42 | <elros34> @Lukapanio still dmesg missing initial logs so it's useless | 13:40 |
| Mister_Magister | rinigus: | 13:40 |
| rinigus | Mister_Magister: ok, I see. do you have that service enabled or disabled | 13:40 |
| rinigus | ? | 13:40 |
| T42 | <Lukapanio> @elros34 [@Lukapanio still dmesg missing initial logs s …], ouch seems so but why? | 13:40 |
| Mister_Magister | rinigus: phone is dead but enabled | 13:40 |
| Mister_Magister | does keymaster not work for you? | 13:40 |
| rinigus | Mister_Magister: it has the same android.hardware.keymaster@3.0-impl: Configure failed in logcat as it was for you | 13:41 |
| Mister_Magister | rinigus: ah yes you need to set android and SPL in kernel | 13:41 |
| Mister_Magister | most likely | 13:41 |
| Mister_Magister | not sure but most likely | 13:41 |
| Mister_Magister | cause you are on 15.1 right? | 13:41 |
| rinigus | nope, its aosp9. what's 'set android and SPL in kernel'? | 13:42 |
| Mister_Magister | rinigus: what was initial android version on that phone | 13:42 |
| rinigus | Mister_Magister: according to tech specs, should be android 8 | 13:43 |
| Mister_Magister | it's sony? | 13:43 |
| Mister_Magister | another lazy ass devs? | 13:43 |
| rinigus | ... its sony xperia xz2 | 13:44 |
| rinigus | ... with lovely bugs in bootloader of all people | 13:44 |
| Mister_Magister | rinigus: but be careful i got my bootloader blocked and phone bricked due to that lol. gimme some time i will find you link k? | 13:44 |
| rinigus | Mister_Magister: thanks | 13:44 |
| Mister_Magister | first grab hex editor | 13:45 |
| Mister_Magister | https://github.com/phhusson/device_phh_treble/blob/android-9.0/fixSPL/getSPL.c | 13:46 |
| Mister_Magister | here you go | 13:46 |
| Mister_Magister | you need to reverse that and set android and SPL version in kernel | 13:46 |
| Mister_Magister | SPL aka Security Patch Level of vendor and boot has to match | 13:47 |
| Mister_Magister | that requirements was in 15.1 but was removed an 16.0 and lazy ass devs didn't remove it. also sfos bootimage doesn't add those | 13:47 |
| rinigus | Mister_Magister: and I don't follow - how would that help me? right now its booting, just without gui :) | 13:48 |
| T42 | <Lukapanio> why dmesg is not full? | 13:49 |
| rinigus | vknecht: but that service doesn't seem to be disabled in https://github.com/mer-hybris/droid-config-sony-ganges-pie/blob/master/sparse/usr/libexec/droid-hybris/system/etc/init/disabled_services.rc | 13:49 |
| Mister_Magister | rinigus: it will make keymaster working | 13:49 |
| Mister_Magister | gui requires keymaster working | 13:49 |
| Mister_Magister | i almost got gui but when i added spl manually to kernel it broke bootloader and it now freezes before fastboot and i ran out of 5 boots and now i have brick | 13:50 |
| rinigus | Mister_Magister: ok , rather weird though. let me read up a bit on that. | 13:50 |
| Mister_Magister | rinigus: you can either try my advice or not lol nobody is forcing you | 13:51 |
| Mister_Magister | tho that's what happened | 13:51 |
| Mister_Magister | sony and asus devs are lazy | 13:51 |
| rinigus | Mister_Magister: on sony, after you run out of these 5 boots, you do get corrupt device warning, but can still get to fastboot | 13:51 |
| Mister_Magister | rinigus: not if your bootloader is frozen due to your bootimage | 13:52 |
| rinigus | no forcing perceived either :) | 13:52 |
| Mister_Magister | ofcourse on asus fastboot is there too | 13:52 |
| T42 | <Lukapanio> if device is brick may be flash bootloader with edl | 13:52 |
| Mister_Magister | only if it wasn't blocked due to my bootimage | 13:52 |
| Mister_Magister | yes i know @Lukapanio | 13:52 |
| Mister_Magister | tho i need to remove back to get to it | 13:52 |
| Mister_Magister | which is glass | 13:52 |
| T42 | <Lukapanio> ouch | 13:53 |
| T42 | <Lukapanio> buy new back | 13:53 |
| Mister_Magister | i read that some pepopl made xiaomi edl cable and it worked | 13:53 |
| Mister_Magister | but | 13:53 |
| Mister_Magister | i need to have device charged and turned off | 13:53 |
| Mister_Magister | and if it's charged i can't turn it off without unplugging battery | 13:53 |
| T42 | <Lukapanio> idk i can reboot to edl from twrp with adb reboot edl | 13:53 |
| Mister_Magister | imagine i can't get into twrp because i can't flash it because kernel is freezed before fastboot? | 13:54 |
| T42 | <Lukapanio> before fastboot it is not even linux i think it must be lk | 13:54 |
| Mister_Magister | i've went through this topic completely | 13:55 |
| Mister_Magister | rinigus: so if you try watchout | 13:55 |
| T42 | <Lukapanio> but why is it frozen? | 13:56 |
| rinigus | Mister_Magister: I will, surely. maybe first read up on keymaster and see if I can make it work differently. somehow its working on aosp9, isn't it? | 13:56 |
| Mister_Magister | rinigus: it is working because versions match | 13:56 |
| Mister_Magister | like i told you | 13:56 |
| Mister_Magister | also | 13:56 |
| T42 | <Lukapanio> but why dmesg is not full @elros34 ? | 13:57 |
| Mister_Magister | rinigus: phh-treble has script to fix SPL version. you can either modify kernel image for quickfix or make script fixing it | 13:58 |
| vknecht | @Lukapanio: gotta be quickier to telnet and type command after boot :-) | 14:01 |
| vknecht | or get dmesg buffer bigger iirc | 14:02 |
| rinigus | Mister_Magister: thanks. (will go and read up on it to avoid locks) | 14:05 |
| Mister_Magister | u can go ask on phh-treble if you don't trust me (and if you don't get banned) | 14:08 |
| rinigus | Mister_Magister: interesting reading - https://github.com/sonyxperiadev/bug_tracker/issues/165 | 14:15 |
| Mister_Magister | rinigus: almost exactly what i told u :P | 14:17 |
| mal | @elros34 btw, you can add optional things to that regex in that sed so it could support both ways | 14:18 |
| rinigus | Mister_Magister: pretty much. I'll ask sony devs if there is a gentle way to enable it :) | 14:18 |
| Mister_Magister | enable what | 14:19 |
| rinigus | keymaster config | 14:21 |
| T42 | <birdzhang> can you upgrade to 3.2.0.12? | 14:23 |
| T42 | <birdzhang> Error: Location 'https://store-repository.jolla.com/releases/3.2.0.12/jolla-hw/adaptation-qualcomm-sbj/armv7hl/repodata/repomd.xml?credentials=store' is temporarily unaccessible | 14:23 |
| rinigus | Mister_Magister: : ... according to error message, its INVALID_ARGUMENT . let's see what's given as arguments in its init | 14:23 |
| T42 | <elros34> mal: I use sed -n 's|.*'"${RPM_BUILD_ROOT}"'\(.*\)$| \1|gp' so both busybox should works | 14:23 |
| Mister_Magister | erm rinigus u need that version in kernel otherwise it won't work | 14:23 |
| Mister_Magister | or patch kernel on live like phh did it | 14:24 |
| T42 | <elros34> I mean both diff* | 14:24 |
| Mister_Magister | there isn't any enabling config | 14:24 |
| Mister_Magister | they have to match. if they match it will work | 14:24 |
| Mister_Magister | rinigus: keymaster sends both bootimage and vendorimage to whatever security it was called (above system level in bootloader level) and it checks if they are correct | 14:25 |
| Mister_Magister | you can't overcome that in any way | 14:25 |
| mal | @elros34 that is also possible, assuming there can be no other error which could match that | 14:25 |
| Mister_Magister | patch manually or do it live are the only options. otherwise phh-treble would do it differently | 14:25 |
| rinigus | Mister_Magister: ok. but now how can patching of hybris-boot (assuming we are talking about it) lock bootloader? if I understood your issue correctly | 14:27 |
| rinigus | ... with 5z | 14:27 |
| Mister_Magister | rinigus: for some reason bootimg patched by me freezed bootloader | 14:27 |
| Mister_Magister | when phhussion did it it didn't | 14:28 |
| rinigus | Mister_Magister: and that's what I don't understand. | 14:28 |
| Mister_Magister | i don't understand either | 14:29 |
| Mister_Magister | noone understands | 14:29 |
| Mister_Magister | just watch out | 14:29 |
| Mister_Magister | you can still flash boot from booted sfos | 14:29 |
| Mister_Magister | or switch slot and then flash boot | 14:29 |
| rinigus | Mister_Magister: thanks! and how do find these magic spl/android numbers? | 14:30 |
| rinigus | Mister_Magister: I cannot switch slots. only slot_a works (thanks to bootloader) | 14:30 |
| T42 | <elros34> mal: good point, I can't use .* | 14:30 |
| Mister_Magister | rinigus: android is 9.0.0 | 14:31 |
| Mister_Magister | and vendor patch level is in vendor/build.prop or smth | 14:32 |
| Mister_Magister | if i find some patched bootloader i can help you later | 14:32 |
| Mister_Magister | i'm at work currently | 14:32 |
| rinigus | Mister_Magister: I will look for those. thanks! no worries, I'll do it later | 14:32 |
| T42 | <Lukapanio> @elros34 here is full dmesg https://pastebin.com/pGhW0MrT | 15:24 |
| T42 | <Lukapanio> so any idea what is wrong? | 15:36 |
| T42 | <elros34> dmesg looks good but looks like your system partition is not mounted. DO you have even /lib/systemd/system/system.mount unit? | 15:37 |
| T42 | <Lukapanio> @elros34 [dmesg looks good but looks like your system pa …], hmmmm no | 15:42 |
| T42 | <Lukapanio> иге црн, | 15:45 |
| T42 | <Lukapanio> but why?* | 15:45 |
| T42 | <Lukapanio> fixup-mountpoints problem? | 15:45 |
| T42 | <elros34> search for *fstab* and uevent*rc in out directory | 15:45 |
| T42 | <Lukapanio> @elros34 [search for *fstab* and uevent*rc in out direct …], ./vendor/etc/fstab.qcom | 15:50 |
| T42 | <Lukapanio> ./system/vendor/ueventd.rc | 15:50 |
| T42 | <elros34> is this somwhere in out/ directory? | 15:51 |
| T42 | <Lukapanio> idk why it create 12mb vendor.img | 15:51 |
| T42 | <Lukapanio> @elros34 whats next with this files? | 15:51 |
| T42 | <Lukapanio> where must i place them? | 15:53 |
| T42 | <Lukapanio> @elros34 [is this somwhere in out/ directory?], yep | 15:54 |
| T42 | <Lukapanio> and ueventd.rc is in /data/.stowaways/sailfishos too | 15:55 |
| T42 | <elros34> hmm then maybe check $ANDROID_ROOT/droid-hal-$DEVICE.log to figure out why system.mount is not created | 15:58 |
| rinigus | Mister_Magister: solution that made keymaster to start was adding `--os_version 9 --os_patch_level 2019-08-01` to https://github.com/sailfishos-sony-tama/droid-hal-img-boot-sony-tama-pie/blob/master/rpm/droid-hal-akari-img-boot.spec#L3 . Thanks for help and help of Marijn pointing to mkbootimg from sony devs channel! | 16:00 |
| rinigus | no gui still | 16:00 |
| T42 | <elros34> verbose obs failed: any idea https://build.sailfishos.org/package/show/home:elros34:sailfish-devel/rpm? | 16:15 |
| Mister_Magister | rinigus: told you it needs to be added to kernel | 16:23 |
| Mister_Magister | one way or another | 16:23 |
| Mister_Magister | still have bricked device xd | 16:23 |
| T42 | <Lukapanio> @elros34 [hmm then maybe check $ANDROID_ROOT/droid-hal-$ …], cant finde it there | 16:38 |
| Thaodan | Mister_Magister: i got that too because my device is still in early stages of aosp | 17:29 |
| Mister_Magister | rinigus: lmao i actually helped lol | 17:30 |
| Thaodan | Mister_Magister: isn't that what you describe about booting part of avb 2.0 (which would also allow relocking with user key's if it weren't for Sony) | 17:31 |
| rinigus | Mister_Magister: yes, thanks for help! | 17:31 |
| Mister_Magister | Thaodan: dunno what u r talking about sorry | 17:32 |
| Thaodan | rinigus: is there an irc channel for Sony devs? | 17:32 |
| Thaodan | rinigus: maybe we should gather such kind of information in a wiki. | 17:33 |
| rinigus | Thaodan: not to my knowledge. they use slack for it and I've asked for access through xda PM | 17:34 |
| Thaodan | Mister_Magister: the verification of boot images | 17:34 |
| Mister_Magister | ah stil dunno | 17:34 |
| Mister_Magister | Thaodan: u just need rinigus fix right? | 17:34 |
| rinigus | Thaodan: we have wiki already under github/mer-hybris. that would be decent place to add | 17:35 |
| Thaodan | rinigus: please do that. does the gui works now? Also a good place is to lock at the device repo from Sony for boot parameters. | 17:36 |
| Thaodan | Mister_Magister: I don't need the fix currently as my device is to wip but better have that knowledge for Android 10 | 17:36 |
| Mister_Magister | well i don't understand you at all hd | 17:37 |
| Mister_Magister | hf* | 17:37 |
| rinigus | Thaodan: you mean vbmeta/dtbo and such? no gui yet, have to do something else for bit of time | 17:37 |
| Mister_Magister | rinigus: i still need to unbrick my device xd | 17:37 |
| Thaodan | rinigus: yeah especially for the kernel building process because this stuff is no longer intree | 17:39 |
| Thaodan | Mister_Magister: Id just liked to now if this(keymaster) stuff is required for gui that I can prepare and look for changes between Android 9 and 10 | 17:41 |
| Mister_Magister | know* | 17:41 |
| Mister_Magister | Thaodan: it most likely is required for gui | 17:41 |
| Mister_Magister | hwc wasn't staritng for me without it iirc | 17:41 |
| Mister_Magister | well i never got gui after all so can't confirm | 17:41 |
| Mister_Magister | but it's surely crucial step | 17:42 |
| Mister_Magister | as it blocks services | 17:42 |
| Thaodan | this stuff is so badly documented like the init split for example | 17:42 |
| Mister_Magister | Thaodan: if your device didn't came out with 15.1 on board then you probably don't need it | 17:42 |
| Mister_Magister | that requirement was removed in 16.0 and devices that came out with 16.0 don't have it | 17:42 |
| Mister_Magister | Thaodan: documented… well i told rinigus a story | 17:43 |
| Mister_Magister | it was required in 15.1 but google didn't specify how to do it so vendors did whatever they wanted and were too lazy to remove it in 16 as it was removed in 16 | 17:43 |
| Mister_Magister | annying companies | 17:43 |
| Thaodan | Mister_Magister: mine came with pie but the support for the soc wasn't in aosp at that time | 17:44 |
| Mister_Magister | pie is 16? | 17:45 |
| deathmist | yep, LineageOS 16 / Android 9 | 17:46 |
| T42 | <Lukapanio> so @elros34 i cant finde system.mount in log | 18:02 |
| Mister_Magister | Thaodan: then you don't need it at all | 18:04 |
| Mister_Magister | it was removed in 16.0 | 18:04 |
| T42 | <elros34> @Lukapanio sure otherwise it would be created but the question is why. Here it should find your fstab: https://github.com/mer-hybris/droid-hal-device/blob/master/droid-hal-device.inc#L443 | 18:20 |
| piggz | mal: vknecht: which of your PRs should come first? I suspect either way will be a lot of conflicts | 18:23 |
| vknecht | let the cleanup go first, I'll adapt :-) | 18:25 |
| mal | piggz: did we agree on the cleanup already? | 18:27 |
| T42 | <elros34> mal: here is patch for rpm https://build.sailfishos.org/package/show/home:elros34:sailfish-devel/rpm but I can't build it | 18:33 |
| piggz | mal: well, we may aswell go with it, im willing to follow the style guide if it says no ; | 18:37 |
| piggz | still needs testd though | 18:37 |
| T42 | <Lukapanio> @elros34 i have https://bpaste.net/show/U2S26 in log | 18:55 |
| T42 | <elros34> then I don't know. I take your fstab from here https://github.com/luka177/android_device_motorola_cedric/blob/lineage-15.1/rootdir/etc/fstab.qcom and mount unit is generated | 19:07 |
| piggz | mal: ive ran through every function with the cleanup PR, no errors logged on the console, so it seems fine | 19:08 |
| mal | piggz: merge if you feel it's good | 19:09 |
| T42 | <hacker12455> @elros34 maybe it's some config issue? | 19:09 |
| T42 | <elros34> I don't think so, he can always create system.mount manually based on other created mount units, it's easy | 19:11 |
| T42 | <hacker12455> oh neat | 19:13 |
| piggz | mal: not an issue, as i know its the "more correct" way ... but i was never sure why QObject *obj; was better than QObject* obj; ... as the type is supposed to be QObject-Pointer, so always seemed normal for me for the * to be beside the type | 19:13 |
| mal | piggz: I have always used that QObject *obj way, that was also from autoformatter | 19:18 |
| piggz | mal: stroustrup implies im correct :) http://www.stroustrup.com/bs_faq2.html#whitespace | 19:21 |
| mal | piggz: there are many different styles :) | 19:23 |
| T42 | <Lukapanio> @elros34 i must mount systempartition to /system? | 19:26 |
| T42 | <elros34> take a look at your fstab | 19:26 |
| piggz | mal: done. vknecht: pls rebase :) | 19:27 |
| mal | and use correct style for any new code | 19:28 |
| piggz | yes, so, from now on, save to use auto formatter ? | 19:30 |
| piggz | mal: should an on-device update to 3.2 be save for ports, without hal/mw rebuild ? | 19:34 |
| vknecht | will do | 19:35 |
| mal | piggz: not sure, quite a lot of changes to packages so might be scary update | 19:38 |
| mal | piggz: maybe I should test on my secondary fp2 | 19:39 |
| T42 | <Lukapanio> @elros34 like https://bpaste.net/show/MSWY2 | 19:39 |
| piggz | mal: oh, pls do, before i wreck my mido :) | 19:39 |
| T42 | <Lukapanio> ? | 19:39 |
| piggz | i was about to take the plunge, but if you have a donor device! | 19:39 |
| T42 | <elros34> @Lukapanio yeah, something like that | 19:42 |
| mal | piggz: version --dup running, we'll see what happens :D | 19:43 |
| mal | piggz: my secondary fp2 is completely empty device, just clean sfos installation, I used that to test flashing new images (among other things like mainline testing) | 19:43 |
| *** electro is now known as Guest47448 | 19:49 | |
| T42 | <Lukapanio> @elros34 [@Lukapanio yeah, something like that], new journalctl https://bpaste.net/show/LFJ4U | 19:51 |
| T42 | <hacker12455> Jan 01 02:05:14 Sailfish droid-hal-init: Failed to initialize property area | 19:55 |
| T42 | <hacker12455> Jan 01 02:05:14 Sailfish systemd[1]: droid-hal-init.service: Main process exited, code=exited, status=1/FAILURE | 19:55 |
| deathmist | @Lukapanio I remember you not having any *propety_context files in out/, so check if "/system/etc/selinux/plat_property_contexts" for example exists as that's what I have symlinked on / | 19:57 |
| T42 | <elros34> system partition is still not mounted | 20:05 |
| T42 | <hacker12455> hmm it can't seem to find it | 20:08 |
| Thaodan | if someone needs to generate vendor patches that are applied on top of AOSP like Jolla does for Sony you at this PR: | 20:20 |
| Thaodan | https://github.com/sonyxperiadev/device-sony-kumano/commit/defcac48571e8a9fe845c52e99cc8cf00daeef54 | 20:20 |
| Thaodan | eh here: https://github.com/mer-hybris/droid-hal-source/pull/7 | 20:21 |
| mal | piggz: it booted but audio and camera don't work, some other issues also | 20:22 |
| piggz | mal: interesting, thoughts on what caused that? | 20:22 |
| mal | not yet | 20:23 |
| T42 | <Lukapanio> @elros34 [system partition is still not mounted], hmmm but why? | 20:23 |
| vknecht | just wondering, should there really be an empty line here ? https://github.com/piggz/harbour-advanced-camera/blob/1e418434ef22b5cd379f8a8f2e227ac0c1ebe41d/qml/pages/GalleryUI.qml#L12 | 20:24 |
| T42 | <hacker12455> does sailfish automatically use the mount units in a directory? | 20:24 |
| piggz | vknecht: well, according to mal :) | 20:25 |
| mal | vknecht: blame the qt creator auto formatter | 20:25 |
| vknecht | ok no problem ; btw, is there some setting to check in QtCreator for better compliance ? | 20:26 |
| vknecht | or is it all on by default ? | 20:26 |
| mal | not sure | 20:27 |
| mal | piggz: also volume buttons are broken, interesting that system sounds like when plugging usb work but settings app or csd doesn't | 20:28 |
| piggz | ill hold off updating then for now! | 20:28 |
| mal | piggz: I assume some of our middleware depends on some older version of something and something wasn't updated because of that | 20:29 |
| piggz | yup | 20:29 |
| mal | piggz: I'll try one more thing, I might have been using testing repos for adaptation which have older packages so trying using devel | 20:36 |
| mal | piggz: ok, that fixed both cam and audio and buttons, so you need to have new enough pulse droid plugin and gst-droid etc | 20:41 |
| piggz | mal: still no build logs available on obs :( | 20:41 |
| mal | piggz: so looks like that everything might be working | 20:41 |
| piggz | mal: cool, will update | 20:41 |
| piggz | mal: scrap that, ive got a log now | 20:42 |
| mal | piggz: live build log is not working afaik | 20:43 |
| vknecht | snapshot pr rebased | 20:49 |
| mal | vknecht: some style nitpicking, in c++ files we have { in same line as if or else | 20:51 |
| mal | vknecht: check some other file for example | 20:52 |
| vknecht | ok, will see later, already on sdcard one | 20:52 |
| piggz | mal: update started.... thoughts on this error? %setup -q -n %{name}-%{version}/upstream | 21:00 |
| piggz | sry... | 21:00 |
| piggz | https://build.merproject.org/package/live_build_log/home:piggz/libnoson/sailfish_latest_armv7hl/armv8el | 21:01 |
| vknecht | is there a way to apply source formatting to whole project at once, or only select all of each file one by one and Edit->Advanced->Auto-indent selection ? | 21:06 |
| mal | vknecht: I think I used astyle plugin | 21:11 |
| mal | for c++ side | 21:11 |
| mal | and "Qml/JS->Format file" for qml | 21:12 |
| vknecht | hmm, can't find this astyle, guess it's an extra ? reinstalled app sdk earlier today, doesn't seem present (have C++ Beatifier there, disabled, tho) | 21:19 |
| vknecht | *Beautifier | 21:20 |
| mal | yes, you might need to install the astyle package to your OS first | 21:20 |
| mal | before it gets active | 21:21 |
| vknecht | alright, thanks | 21:21 |
| mal | the artistic style under beautifier | 21:21 |
| rinigus | as test_hwcomposer is still crashing on akari, I am posting strace https://dpaste.de/sJbx/raw and corresponding logcat section https://dpaste.de/bXmm/raw | 21:24 |
| vknecht | seems beautifier/astyle needs a config file | 21:29 |
| mal | rinigus: have you made sure all ueventd*.rc files were copied to out/ | 21:37 |
| mal | rinigus: that suggest a permission issue which could be due to missing udev rules which sometimes are caused by missing ueventd*.rc | 21:37 |
| mal | during build | 21:37 |
| rinigus | mal: you mean during system and vendor image builds? | 21:38 |
| mal | no during hybris-hal | 21:39 |
| rinigus | but, as such, no, I haven't made sure all is copied. I will have to check it then | 21:39 |
| rinigus | mal: and how would you check for it? compare out/ of aosp and sfos hybris-hal? | 21:40 |
| mal | rinigus: check output of make modules for any ueventd*.rc and if there is one try make for that module, of course first check find out/ -name 'uevent*.rc' for comparison | 21:40 |
| mal | rinigus: if you have both aosp and sfos out/ then use that find for both | 21:41 |
| rinigus | mal: thanks, will do that then. I do have the both trees (we actually need plane aosp9 for system & vendor image building, as I follow xperia 10 instructions) | 21:42 |
| rinigus | plain aosp9 :) | 21:42 |
| mal | ok, that is quicker to compare | 21:42 |
| rinigus | will start with that tomorrow evening... thanks for checking out the logs! | 21:43 |
| vknecht | hmm, looks like qml/js formatter simplifies | 21:45 |
| vknecht | for (var i = 0; i < modelStorage.rowCount(); i++) | 21:45 |
| vknecht | to | 21:45 |
| vknecht | for (; i < modelStorage.rowCount(); i++) | 21:45 |
| vknecht | but np I caught it | 21:45 |
| T42 | <elros34> rinigus: didn't your previous strace of test_hwc2 fail right /selinux? That could be important | 21:45 |
| mal | vknecht: is i defined somewhere earlier | 21:47 |
| mal | @elros34 I noticed that in earlier strace also | 21:47 |
| mal | new strace didn't fail in same place | 21:48 |
| vknecht | not sure, but at runtime sure got ReferenceError: i is not defined | 21:48 |
| mal | quite odd | 21:48 |
| mal | vknecht: which part of the code? | 21:48 |
| rinigus | elros34: thats strace for hwc2 https://dpaste.de/65A4 | 21:49 |
| mal | rinigus: that looks bad | 21:49 |
| vknecht | this one, https://github.com/piggz/harbour-advanced-camera/blob/370681159dfe8c1c5af00b50b37feec1aa6b0eb5/qml/pages/SettingsOverlay.qml#L537 | 21:51 |
| vknecht | and it also mods the findIndex for grid setting in same file | 21:52 |
| rinigus | mal: it was run using EGL_PLATFORM=hwcomposer test_hwc2 . as for bad, what exactly ? | 21:53 |
| rinigus | lib path? | 21:53 |
| mal | vknecht: wondering why it didn't do it for me | 21:53 |
| mal | rinigus: the final messages are about selinux | 21:54 |
| rinigus | mal: yes, I still have selinux enabled. as I haven't got yet recommendation in disabling it | 21:54 |
| mal | rinigus: I'm quite sure I suggested that to you earlier | 21:55 |
| mal | although in theory it should work assume you have all needed parts | 21:55 |
| rinigus | mal: so, that's an assumption. and its hard to be sure that I do have all the parts. I did base it all on xperia 10 and all repos are there, but its still questionable. | 21:57 |
| rinigus | you did suggest with similar 'in theory' remark. in this respect we can conclude that I should proceed with dropping selinux and rebuilding without it | 21:58 |
| rinigus | mal: /sys/fs/selinux is empty on device, but exists. there is no /selinux | 22:03 |
| rinigus | mal: as soon as I mounted selnux (mount -t selinuxfs none /sys/fs/selinux), test_hwc2 worked! | 22:08 |
| mal | rinigus: hmm, wondering what should mount that | 22:19 |
| vknecht | PRs updated, 'night | 22:23 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!