Sunday, 2019-11-03

T42<Lukapanio> https://bpaste.net/show/DTTSE07:23
rinigusSituation with udisk (or something else) identifying local storage is quite bad. It considered all these local partitions as sdcard ones and shows then in settings storage and backup07:41
rinigusNot something you want anyone to mess with07:42
rinigusHas anyone managed to disable it? With local storage on sda,b,c?07:42
rinigusPro1 devs? adampigg piggz Mister_Magister ^07:45
r0kk3rzweird, you should be able to modify its udev hooks07:46
rinigusr0kk3rz : that's what I tried last night (pasted conf), but without any visible effect07:48
rinigusBut I'll try again. Last time i focused on failed mounts07:50
r0kk3rzwould also be worth seeing what each of those are showing in udevadm07:51
rinigusr0kk3rz : will check it out07:54
kimmolihttps://wiki.archlinux.org/index.php/Udev#Some_devices,_that_should_be_treated_as_removable,_are_not07:56
kimmoli`or the opposite of this?07:56
riniguskimmoli: UDISKS_SYSTEM_INTERNAL I haven't tried. not sure if its for udisks or udisks207:58
r0kk3rzudev stuff should be fairly standard07:59
T42<DylanVanAssche> Are webhooks not available anymore on OBS?08:41
T42<DylanVanAssche> https://webhook.merproject.org/webhook/08:41
T42<DylanVanAssche> Just saw that the live build log works again, awesome :D08:41
*** OhYash1 is now known as ohyash08:41
spiiroin@DylanVanAssche mce does not send dbus messages when volume keys are pressed/released (has no reason to). the volume key related policy messages happen around display blank/unblank i.e. tell ui to ignore volkeys on blank, and conversely start processing them again shortly after display unblank + modifications to that from musinc playback state etc08:57
T42<elros34> rinigus you could also try to add: ATTRS{type}=="SD" to existing udisks rules. Type can be read from  cat /sys/block/*/device/type09:17
T42<Lukapanio> кинь завдання з ікт09:59
T42<Lukapanio> якогось не прийшли09:59
T42<Lukapanio> іщ утвштп гз цшер [   46.552494] droid-hal-init: cannot find '/system/bin/vold_HYBRIS_DISABLED' (No such file or directory), disabling 'vold'10:02
T42<Lukapanio> [   46.553106] droid-hal-init: Not bootcharting.10:02
T42<Lukapanio> so ending with *10:02
T42<Lukapanio> full log https://bpaste.net/show/DTTSE10:02
T42<hacker12455> hmm10:04
T42<hacker12455> [   46.551639] fs_mgr: Cannot open file /fstab.qcom10:04
riniguselros34 r0kk3rz kimmoli : I can avoid udisks2 trying to mount some partitions via systemd by overriding 80-udisks2 (https://github.com/sailfishos-sony-tama/droid-config-sony-tama-pie/blob/master/sparse/etc/udev/rules.d/80-udisks2.rules#L45)10:05
rinigusbut all partitions still show up in Settings/Storage and Backup. With handy commands like "let's format them"10:05
rinigushow can I blacklist partitions shown in settings?10:06
r0kk3rzuse udev rules10:06
rinigusr0kk3rz : any working example? All I tried so far has failed10:10
r0kk3rzive never had this issue so im not 100% exactly how it works10:11
r0kk3rzid expect its picking up on some attribute though10:11
rinigusJoys of closed source part... I'd expect the same, question is what10:12
T42<elros34> @Lukapanio what did you change? Your partitions are not mounted10:14
r0kk3rzmaybe watch dbus when the partitoner thingie is open10:14
T42<elros34> @Lukapanio at least not by mount units10:15
T42<Lukapanio> @elros34 [@Lukapanio what did you change? Your partition …], idk system is mounted10:15
r0kk3rzyou know we do things with overlayfs right?10:16
T42<Lukapanio> @elros34 [@Lukapanio what did you change? Your partition …], moved base to arm32 cm1410:16
r0kk3rzso is the data partition mounted10:16
r0kk3rzdid you get your first boot stage working?10:17
rinigusr0kk3rz: maybe Jolla devs can step up and tell how to blacklist those10:17
T42<elros34> @Lukapanio 14.1? so why do you have android.hardware.keymaster@3.0-service-qti.rc in logs? you ar emixing something10:18
T42<Lukapanio> hmm10:20
T42<Lukapanio> something built wrong10:21
T42<Lukapanio> los 14 is installed10:21
T42<Lukapanio> ro.cm.build.version=14.110:21
T42<elros34> I am talking about changes in you droid-config10:21
deathmistrinigus r0kk3rz: I've had the exact same issue since I started porting for my OP5, I've just ignored it as it's not that big of a deal, but it could be quite dangerous if the format buttons actually work and you just accidentally wipe a critical system partition (I have 70 partitions so it also makes navigating a bit annoying)10:23
r0kk3rzinteresting10:25
r0kk3rzdo they also get exposed under sda .etc?10:26
deathmistyep https://pastebin.com/iUPQSmJX10:27
r0kk3rzi wonder if that app is doing something hacky...10:29
rinigusdeathmist: we have to fix it. Not it's also exposed in camera. So, probably we have to do something with udev, as pointed out by others10:30
r0kk3rzrinigus: not if they're doing something hacky and picking things in the /sdx range ;)10:38
r0kk3rzmight need mal to check it out10:38
rinigusr0kk3rz: let's hope not.10:45
rinigusr0kk3rz: don't see anything fired on dbus when opening settings nor settings/storage. as reported by `dbus-monitor`10:53
r0kk3rzhmmm10:53
r0kk3rzthen maybe they arent using the udisks2 dbus api10:53
deathmistso I looked at lsblk, when I plugged in a USB drive it showed up as "sdg" so device internal partitions in my case are "sd[a-f]*", will now have to learn a lot more about rules and such as I haven't worked with udev much at all before ^^11:12
rinigusr0kk3rz: maybe. what's also disturbing is that udisks2 just ignores ENV{UDISKS_IGNORE}="1" and still creates entries under its dbus tree11:13
rinigusalthough, when checked with `udevadm info -q all -n /dev/sda5`, this var is set11:14
T42<Umeaboy> Is there a list of unofficially supported Sailfish devices besides the adaptations wiki?11:18
rinigusdeathmist: where are your repos? would like to see what have you included in sparse11:37
vknecht@Umeaboy: https://en.wikipedia.org/wiki/Sailfish_OS#Community_enthusiasts'_ports_to_devices_from_other_vendors11:37
deathmistriningus: here https://git.io/fji3O my etc/ is be a kinda polluted with a lot of unnecessary audio/media/codec symlinks, should probably clean that mess up sometime soon :)11:39
rinigusdeathmist: thanks! looks to me that many put lots of links to /vendor/etc and I wondered what the policy/recommendation regarding that is11:55
deathmistI probably don't need half of them, will have to remove most ones and see what regressions come up and add them back as needed. some symlinks on /etc by default go to /system/etc/ and at least on my device many of the files are on /vendor/etc/ instead11:58
rinigusdeathmist: I seem to have similar configs in vendor. maybe should add them as well and see if it does any better11:59
Thaodanrinigus: Have you fixed your find-mmc-bypartlabel problem?14:26
ThaodanI've done some work on that: https://github.com/Thaodan/initrd-helpers/commit/051c712ef758fbd9f70335ba3ec82420297c56c214:27
deathmistmal: do you know if the ld.config*.txt patch is required universally on 16.0? I'm preparing a PR for the FAQ doc regarding that and fixing audio with the symlink that worked for both me and piggz already. maybe unrelated, but what about adding libhybris to >=15.1 hybris manifests so it doesn't have to be manually cloned by default? :)  it has14:31
deathmistworked in a manifest for me on both bases https://git.io/JegZr14:31
rinigusThaodan: I did fix it in my local copy, but haven't published anything. for me it was just replacement with sda, yours looks more universal14:47
rinigusfeel free to push for it14:47
T42<elros34> looks like always 5s loop delay :P14:51
maldeathmist: I think most devices need the ld.config mount14:58
maldeathmist: the audio lib symlink looks strange14:58
deathmistyeah it really is, but it fixed our issue of audio.primary HW module failing to load (as seen in logcat), may need to wait a bit more to see if it affects more devices too before adding that14:59
malis it trying to load wrong lib?15:01
deathmistboth devices that the fix worked on are msm8998 platform tho so it may just be a quirk for the QCOM SD 835 or something15:01
malbefore the symlink, or from wrong location?15:01
Thaodanrinigus: I already did, please test if you can15:02
Thaodanhttps://github.com/sailfishos/initrd-helpers/pull/1615:02
rinigusThaodan: will do with the next image15:05
Thaodanrinigus: Comment on that PR please when you got results.15:05
Thaodangood15:05
rinigussure15:05
T42<Lukapanio> https://bpaste.net/show/DF4UO15:27
T42<Lukapanio> here is full log15:27
T42<Lukapanio> @elros34  no more keymaster15:27
T42<hacker12455> Fwd from hacker12455: [   40.565593] droid-hal-init: Starting service 'droid_init_done'...15:37
T42<hacker12455> [   40.566557] fs_mgr: Cannot open file fstab.qcom15:37
T42<hacker12455> [   40.566609] droid-hal-init: fs_mgr_mount_all returned an error15:37
malthat is not an error, we don't use fstab in sfos15:39
T42<hacker12455> alright15:41
T42<Lukapanio> so where is error16:02
T42<Lukapanio> ?16:02
T42<Lukapanio> https://bpaste.net/show/NRFUC16:09
T42<Lukapanio> journal16:09
rinigusdeathmist: what was the audio.primary problem and what was the fix? I can check my logcat16:12
deathmistrinigus: audio did not work for me on 16.0 regardless of having all proper configs, I had to do a `ln -s /system/lib/vndk-28/libaudioroute.so /usr/libexec/droid-hybris/system/lib/libaudioroute.so` and after a reboot audio magically started working :p16:14
rinigusdeathmist: for me, audio works without that, as tested via videos. But haven't tried yet for phone calls16:18
deathmistshouldn't be an issue then, call audio is handled via afglue/hidl PA modules, for me pulseaudio didn't work at all before that symlink16:18
mal@Lukapanio which android base version?16:19
T42<Lukapanio> @mal cm-1416:20
T42<Lukapanio> android 716:20
maldid you add the .rc files using the script in faq https://github.com/mer-hybris/hadk-faq/blob/master/README.rst#hybris-14-116:21
T42<Lukapanio> ouch no thanks16:21
T42<hacker12455> the script is supposed to output nothing?16:43
T42<hacker12455> as in nothing in the terminal16:43
vknechtwell, no, see its souce ;-)16:44
vknechtI guess it would only print anything if mkdir or ln commands went wrong (ie. not run from $ANDROID_ROOT dir)16:47
rinigusjust for the record in the logs -  /sys/fs/selinux is mounted in android. same as was required to get test_hwc2 working on tama/akari when running with selinux16:59
malrinigus: I'll check what should mount it in sfos17:04
rinigusmal: thanks. I wrote a small script for it as systemctl refused to mount it via .mount17:05
rinigusmal: although, it could be that its not mounted on xperia 10 in sfos17:05
malrinigus: which systemd version do you have on the device?17:05
malrinigus: it's mounted on x1017:05
rinigusmal: good to know its mounted. I have the one without selinux support (-SELINUX flag), as in 3.1.0.1217:06
malrinigus: so probably that is the reason, it should be at least 225+git19 for selinux support17:08
rinigusmal: let's hope so - then it will resolve automatically with the update17:09
rinigusI just have to remember to remove the hack/script17:09
rinigusmal: would you mind to check, when in office, how Settings/Storage (or Camera or Backup) discover external sd cards. as discussed earlier today, we have on several devices with internal storage on /dev/sda-sdf considered as sd cards17:11
rinigus... as a result - user maybe able to format or mount  them (haven't tested that theory)17:11
malok, need to check17:12
rinigusmal: thanks!17:13
malrinigus: 3.2.0 does bring that git19 or systemd based on the tjc changelog17:15
mal*of systemd17:15
vknechtis the new selinux support pie-base specific, or should it work on earlier bases too ?17:16
rinigusThaodan: your script doesn't work, just tested. probably due to for loop17:17
rinigus`for mmc_sysfs in /sys/class/block/mmcblk0p*/*/ /sys/class/block/sda/*/; do`17:17
rinigusThaodan: shouldn't it be /sys/class/block/sd* (we have devices with not just sda)?17:19
malvknecht: the android init patch would need to be ported to older bases17:19
Thaodanrinigus: yes it should be17:19
Thaodanhowever it should be /sys/class/block/sd*/*/ to enforce that it lists only dirs17:20
rinigusThaodan: will try that next time17:21
T42<Lukapanio> @mal so done like in faq and i have https://bpaste.net/show/6QPUC17:32
Thaodanrinigus: the error was elsewhere please look again17:50
ThaodanAdded fixes for that17:50
mal@Lukapanio what did you build after that?17:59
T42<Lukapanio> @mal rootfs18:01
mal@Lukapanio I mean did you run "build_packages.sh -c"18:02
mal@Lukapanio did you run the script in ANDROID_ROOT?18:02
T42<hacker12455> @mal [@Lukapanio did you run the script in ANDROID_R …], yes18:03
malcheck that the content was really created18:10
rinigusfor some reason, my lock screen is not showing up, just a rotating circle (waiting). if I don't setup pin, I can swipe left/right to unlock. with pin, no luck - only reflash.18:44
rinigusdisclaimer: I run with selinux18:44
riniguswhat's the name of that process which is expected to show lockscreen? can I run it from terminal for strace?18:45
vknechtprobably encsfa from /lib/systemd/system/jolla-devicelock-encsfa.service18:48
rinigusvknecht: what is it?18:49
vknechtwell, service provided by jolla-devicelock-daemon-encsfa package18:49
vknechtiirc mce might be involved to (the tklock stuff in logs)18:51
T42<adampigg> Rinigus, did you figure out block device issue?19:04
rinigusadampigg: as i had several issues with block devices, let me guess. you mean getting internal storage as external in settings backup, storage, and camera?19:07
rinigusThat one no. But mal will check out internals when he can19:07
rinigusadampigg: or do you have a solution for it?19:11
T42<adampigg> No, was hoping you had19:11
rinigusadampigg: let's hope we get more info tomorrow ...19:18
T42<hacker12455> how do I set a password for devel-su?19:59
Thaodandeveloper options20:00
T42<hacker12455> ok found it20:01
ThaodanI'lll reboot my server in about 5min so the bot will be unaviable for a few minutes after that.20:09
spiiroinrinigus: lockscreen ui is drawn by lipstick. but if you get that spinner on bootup -> AFAIK that is lipstick waiting for modem(s) to get ready -> rild etc22:15
rinigusspiiroin: I was just going to write that it was solved by fixing `fixup_ril` service to not fire :)22:16
rinigusas I use xperia 10 as a base, it has this magic to support  1 and 2 sim card devices.22:16
riniguswhich wasn't working on my half-removed config.22:17
riniguswill have to look into it a bit later to add support to 2-sim variants. so, that part is working fine now22:17
rinigusthanks!22:17
rinigusis it expected that imei is not shown in settings/about if there is no sim card inserted?22:20
malrinigus: imei should show22:21
rinigusmal: then I will debug rild next. have logcat from android as well, so I can compare those and see what's missing22:22
rinigusbut its late over here already for it. gn!22:22
r0kk3rzthis is curious https://github.com/ubports/installer-configs/blob/master/v1/lenok.json22:38
r0kk3rza gui installer for ported devices22:39

Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!