Tuesday, 2019-11-05

rinigusmal: have you had a chance to checkout or ask about storage detection by settings/{storage,backup,camera}?13:55
rinigusre: our issue with /dev/sda-sdf disks when on internal storage13:55
malnot yet, completely forgot14:02
malrinigus: do those show at the end of the storage page or how?14:05
rinigusmal: then its good to remind :). as I have only two fs as a "normal" storage (/ and /home) they show up after that.14:06
rinigusmal: its quite a long list of different partitions.14:07
rinigusmal: some of them are of "non-correctly" formatted (shown with red labels) and I think it proposed to format them too14:07
riniguson camera and backup, some of the partitions can be used to store images/backups.14:08
rinigusprobably list of "good" ones - ext4 or similar14:08
rinigusmal ^14:08
malrinigus: so do you see the usual User files and system files and then external ones?14:11
malrinigus: can you show the output of mount?14:16
rinigusmal: can't make screenshot right now, so I try to explain. In Settings/Storage, we have rings describing each partition. For XZ2 (and I take for 1+5 and pro1), we have also rings for those system-used partitions (firmware and such)14:16
malrinigus: but you also have the usual ones?14:17
rinigusmal: mount does not have that many of them. let me see if I can get df from my earlier pastes (at work right now)14:17
rinigusmal: but its surely not as many mounted14:17
rinigusmal: yes, I do14:17
malso those extra ones are listed in the external list at the end14:17
rinigusmal: discussion started on sunday morning14:18
rinigusmal: yes, in this respect they are14:18
malfirst is always user then system and the memory cards etc14:18
rinigusmal: as there is no clear distinction in UI, I cannot tell if its already external. but yes, first was expected ones (home and system, I presume) and then looooong list with funky partitions and uuid's14:19
rinigusas I have on onyx uuid for sd card14:20
malrinigus: everything after system is "external"14:20
malso those are for some reason not seen as "internal" and therefore are shown in the list14:21
rinigusmal: we suspect its related to the fact that internal partitions are on /dev/sd{a,b,c,d}14:21
rinigusmal: I tried to flag them as system ones or asked for them to be ignored by udisk2, but that didn't help much.14:22
rinigusmal: so, question is how are those partitions detected by storage. if there is some flag we can set14:22
malI'll need to find the udev rule changes you made14:23
rinigusmal: I used this: https://github.com/sailfishos-sony-tama/droid-config-sony-tama-pie/blob/master/sparse/etc/udev/rules.d/80-udisks2.rules#L4514:24
riniguslocation as on device14:24
rinigusalso, see http://www.merproject.org/logs/%23sailfishos-porters/%23sailfishos-porters.2019-11-03.log.html#t2019-11-03T11:13:3114:25
malrinigus: are those ones seen in settings sdb* (or other like c and so on) or do sda* also show?14:25
rinigusmal: sda shows up as well. I don't have that many partitions on sdb,c14:26
malrinigus: does it create mount service for those? mount-sd@*.service14:27
rinigusmal: note I am very confident in it, but I cannot check right now. in 1-3 hours can14:27
rinigusmal: with the default rules, it made mount services for couple of them. with the overwritten rules, as in my current config, these are not made14:28
rinigusmal: I noticed the failed mount services, but didn't check the others. with new rules, the failed ones disappeared14:30
malrinigus: ok, does your device support sd cards? if so how do those show in /dev?14:31
rinigusmal: it does, but I haven't inserted any yet14:32
rinigusdeathmist, piggz: have you tried sd cards on your devices? (see discussion above)14:32
T42<adampigg> Later14:35
malrinigus: would be interesting to know how it appear to make sure it is detected correctly in those udev rules14:46
rinigusmal: I will test it then. probably the rules will have to be adjusted, unless it shows up as the one already used there.14:47
rinigusmal: coming back to the question - how is jolla's settings detecting external storage?14:48
deathmistmal: rinigus: my phone doesn't have an SD card slot (thank OnePlus lol), but I've tested USB OTG storage and it shows up normally (along the mess of system partitions) in Settings > Storage and I've backed up to and restored my SFOS user stuff from it too14:50
malrinigus: that information comes from lower level, nemo-qml-plugin-systemsettings and where it gets the data14:51
maldeathmist: do you also have the same issue that rinigus has?14:51
deathmistyep, OnePlus 514:52
rinigusmal: pro1 has the same issue14:52
deathmistafaik for me all of sd[a-f]* is shown as I can see some labels from https://pastebin.com/iUPQSmJX etc14:53
deathmistthose are all meant to be system partitions ofc14:53
rinigusmal: that's good taking that it is open source!14:54
maldeathmist: can you grab udevadm info things for one of those partitions14:55
malrinigus: settings app didn't have any smart things there, just using PartitionModels with differenty types14:56
rinigusmal: so, we have to figure out how its populated and change settings or code accordingly. I will be able to grab udevadm info when I get back home14:59
T42<nitin03 %lastname%> build package for poco f1 ..I was unable to telnet this time I did "touch init_enter_debug2" in sfos directory...then I was able to telnet into the device..got a shell...enter devel-su..but it asks for a password15:02
T42<hacker12455> Can you get to GUI?15:02
T42<nitin03 %lastname%> no15:03
malrinigus: I think it needs some udev magic15:03
T42<hacker12455> Hmm15:03
T42<hacker12455> nitin03 what base?15:03
mal@nitin03 telnet is already root, no need to use devel-su in there15:04
rinigusmal: reading the code, looks like it. will have to see what this magic could be15:04
T42<nitin03 %lastname%> I did edit Storage=automatic in journald.conf...but reboot command doesn't work15:05
T42<nitin03 %lastname%> it says Failed to talk to init daemon.15:05
T42<nitin03 %lastname%> @hacker12455 [Hmm], lineage16.015:06
T42<elros34> use reboot -f :P15:06
mal@nitin03 if you do "touch init_enter_debug2" that disables systemd so the boot stops very early15:07
mal@nitin03 what things have you added to config repo, some changes are mentioned in faq, not sure if all15:07
T42<nitin03 %lastname%> files you suggested...like ld.config.28.txt and in its corresspoding file and symlink file in lib/systemd/system folder15:12
maland the .rc files?15:13
mal@nitin03 can you tell how you used that ld.config.28.txt, you added some .mount file for it?15:14
T42<nitin03 %lastname%> @mal [and the .rc files?], yes I added mount file plus the symlink in local.fs.target.wants15:15
mal@nitin03 you also added the .rc files mentioned here https://github.com/mer-hybris/hadk-faq#hybris-16-0 ?15:18
T42<nitin03 %lastname%> filename system-etc-ld.config.28.txt.mount...and its content https://pastebin.com/syQEx43F15:18
T42<nitin03 %lastname%> @mal [@nitin03 you also added the .rc files mentione …], yeah i already have these 3 files15:19
maldid you add anything to straggler files in droid-hal spec file?15:20
T42<nitin03 %lastname%> @mal [did you add anything to straggler files in dro …], yes15:20
malhopefully logs tell something, you need to remove init_enter_debug2 after you changed journal logging, you could also try masking droid-hal-init to see if that helps15:23
T42<nitin03 %lastname%> okk...thanks so much15:24
deathmistmal: here's the output of udevadm info for my /vendor partition for example https://pastebin.com/cqp6LGfb15:46
T42Thinhx2 was added by: Thinhx215:47
T42<Thinhx2> @nitin03 %lastname% [okk...thanks so much], Sailfish booted in your poco f1?15:50
T42<nitin03 %lastname%> @Thinhx2 [Sailfish booted in your poco f1?], no15:50
T42<Thinhx2> Same mi8se15:51
T42<nitin03 %lastname%> @Thinhx2 [Same mi8se], at which stage you r stuck?15:52
T42<Thinhx2> Libstick is crash15:52
T42<Thinhx2> Fwd from Thinhx2: Telnet,led,log debug,usb port,battery,network fine,button fine15:53
deathmist@Thinhx2 give journalctl -b output, also what base?15:53
T42<nitin03 %lastname%> @Thinhx2 [Telnet,led,log debug,usb port,battery,network …], so it booted?15:53
T42<Thinhx2> @deathmist [@Thinhx2 give journalctl -b output, also what …], Los16 bro15:54
T42<Thinhx2> @nitin03 %lastname% [so it booted?], Stuck mi logo15:54
T42<Thinhx2> https://pastebin.com/8A3x2FjW15:55
deathmistat least droidmedia is not built but that shouldn't be an issue iirc, can you link your droid-config repo?15:59
deathmistand give output of "mount" command15:59
T42<Thinhx2> @deathmist [at least droidmedia is not built but that shou …], https://github.com/BusterGR18/droid-config-sirius16:01
T42<Thinhx2> Here for you16:01
rinigusdeathmist: vendor partition is probably not the best example, it maybe checked for mount. There was some coffee in repo referenced by mal. What about some other, obscure partition?16:04
rinigusNot mounted one16:05
deathmistrinigus: mal: https://pastebin.com/cqp6LGfb there, updated with my "LOGO" partition instead which definitely isn't mounted16:07
malalso check info from one of those disk (like sdb or something)16:08
maldeathmist: is that shown in settings storage list?16:08
T42<Thinhx2> What i need to do now,m16:11
malis that hybris-16.0?16:11
T42<Thinhx2> Right16:11
T42<Thinhx2> Use linage source 16.0 android 916:12
deathmistI honestly don't know, all unmounted ones except a few say "Unsupported SAMSUNG memory card" and "The filesystem is incompatible. Formatting is needed.". from the mounted ones I see /system, firmware, dsp, vendor. updated same paste with sdb info16:13
T42<Thinhx2> @deathmist [I honestly don't know, all unmounted ones exce …], Oh shit,my mi8se using samsung emmc16:15
mal@Thinhx2 you are missing the mount service for ld.config.vndk_lite.txt, the droid-hal-early-init.sh is not working well and should be replaced with a mount service (adapt this https://pastebin.com/syQEx43F, name should be system-etc-ld.config.vndk_lite.txt.mount)16:15
malalthough not sure if that is even needed16:16
T42<Thinhx2> Thank you so much, thank all16:16
T42<Thinhx2> Let me try16:16
mal@Thinhx2 you need to provide logs, journalctl -b and if possible /usr/libexec/droid-hybris/system/bin/logcat16:17
rinigusSounds to me that we have to blacklist devices on udev level as udisks_ignore was pretty much ignored by udisks and it still created entries under dbus interface. (As checked this weekend).16:17
rinigusdeathmist mal /\16:18
rinigusAlthough, checking on onyx, there are descriptions for partitions on dbus that are not down in settings..16:22
vknechtgotta hope this assertive regexp is not the cause of your trouble (just wondering, can't really get all it implies...) https://git.merproject.org/mer-core/nemo-qml-plugin-systemsettings/blob/master/src/udisks2monitor_p.h#L4816:57
malvknecht: rinigus https://git.merproject.org/mer-core/nemo-qml-plugin-systemsettings/blob/master/src/partitionmanager_p.h#L4617:01
rinigusvknecht: good find! corresponding commit: https://git.merproject.org/mer-core/nemo-qml-plugin-systemsettings/commit/5772467e04150535b97f6a5796bce9d204a46f3b17:01
malrinigus: we still need to know how sd card is shown before we do anything to those lines17:02
malprobably needs some configuration option to use a custom regex?17:02
deathmistmal: damn it seems lists my internal device partitions "sd[a-f]*" as external, could this be it? when I plugged in a USB drive via OTG it showed up as sdg17:03
maldeathmist: I think it's because of that17:04
deathmistguess I could compile a custom version right now to test it for my own device17:04
deathmistfor me adjusting it the sdX range to sd[g-z] should be good enough17:04
malyes, but a proper fix is needed17:05
rinigusmal: I'll test. It's just after rebuild yesterday, I have boot loops due to ofono. Ofono is disabled by moving it's service. On to of it, i lost wifi and I'm trying to get some logs to figure out why17:31
deathmistmal: rinigus: yep, after the change all seems to work perfectly :)  https://imgur.com/a/2k63nPk17:32
rinigusAs for fix, we should probably check for udisk2 or udev env variable. That way we can skip all this hard coded regex from code17:32
maldeathmist: now that is missing the internal storage parts, user and system17:53
deathmisttbh I don't know how it's supposed to look, I've only browsed the internal storage using the entry there, what is it meant to look like?17:54
*** rinigus97 is now known as rinigus_17:55
maldeathmist: it should have "User data", "System data" and then variable amount of external devices depending on what is supported by the device17:55
maldeathmist: I assume that "Mass memory" is root or something?17:56
malwhich would suggest it is also detected incorrectly17:56
rinigus_mal: Mass memory is in the case it doesn't detect 2 partitions (/ and /home).18:00
deathmistmal: I guess it's actually "User data", when I go to "file manager" there it goes to "User data", not sure how I would rename those properly since I don't remember seeing those before either in the menus. I assume "User data" = /home/nemo and "System data" = / then?18:01
deathmistI added a few more screenshots there too now18:01
rinigus_mal: same is on onyx18:01
malrinigus_: yes, seems like it, I only checked official device for that18:02
maldeathmist: the "Mass memory" is correct18:02
malin ported devices18:02
vknechtah crap, was hoping to use partitionmanager's externalStoragesPopulated signal, but it's too new :-(18:08
vknechtgrr, seems to lead to harbourd invalidation anyway... too bad, seemed to work...18:42
T42<nitin03 %lastname%> do I need to build udevadm also?18:51
T42<nitin03 %lastname%> it isn't mentioned anywhere in the guide18:51
mal@nitin03 that is not relevant for you18:54
malwe were talking about a different issue here18:54
T42<nitin03 %lastname%> @mal [@nitin03 that is not relevant for you], oh18:59
T42<nitin03 %lastname%> I was reading hadk faq page18:59
T42<nitin03 %lastname%> It was mentioned there18:59
malwhich part of it?19:00
malwhat issue are you trying to fix now?19:00
T42<nitin03 %lastname%> @mal [which part of it?], It was a general question19:02
T42<nitin03 %lastname%> it was mentioned in debugging tips19:02
T42<nitin03 %lastname%> So I asked if I will ever need it19:02
piggzrinigus: so, im inserting an sd card to see what happens on pro1?19:03
piggzUDEV  [15865.971553] add      /devices/soc/c0a4900.sdhci/mmc_host/mmc0/mmc0:e624/block/mmcblk0 (block)19:11
piggzUDEV  [15866.247304] add      /devices/soc/c0a4900.sdhci/mmc_host/mmc0/mmc0:e624/block/mmcblk0/mmcblk0p1 (block)19:11
piggzso, system is /dev/sd[a=f]19:11
piggzand, sd is mmcblk019:11
riniguspiggz: interesting, does it show up in settings/storage then?19:12
piggzrinigus: i dont think it does19:12
piggzbut, the list is so long, its hard to be sure!19:12
rinigusmal: see above, on pro1. I'm still fighting with my wifi, will be able to test later19:16
riniguspiggz: do you know if it's mounted on insertion?19:17
piggzdoesnt seem to be19:17
piggzi did get a notification on insertion tho19:17
piggz"memory card inserted SDCARD"19:18
piggzclicking notification takes to storage settings page19:18
malrinigus: ok, that will need some configuration because normally mmcblk0 is assumed to be the internal storage19:21
piggzmal: anything i can do?19:22
malpiggz: you mean for the storage handling? check the repo mentioned earlier and see if can see some easy way to make the internal/external detection configurable19:23
T42<adampigg> (Photo, 2560x1440) https://irc.thaodan.de/.imgstore/mxqpbJ04CD.png19:23
T42<adampigg> Kimmoli^^19:23
T42<adampigg> Mal, ok19:24
T42<adampigg> The nemo one?19:24
mal@nitin03 can you show output of dmesg and also /usr/libexec/droid-hybris/system/bin/logcat (assuming that works)19:24
malpiggz: https://git.merproject.org/mer-core/nemo-qml-plugin-systemsettings/commit/5772467e04150535b97f6a5796bce9d204a46f3b19:24
T42<DylanVanAssche> Out of curiosity, the Aliendalvik 8.1 depends it on Android base adaptation or just the Linux kernel (like lxc support)?19:26
rinigusmal piggz: yes, looks like we need ability to configure that. ideally, it should be via udisks2, so we could drop all regex handling over there.19:33
malrinigus: that is one option, I'll discuss this internally tomorrow19:34
rinigusmal: thanks!19:35
T42<nitin03 %lastname%> @mal [@nitin03 can you show output of dmesg and also …], https://pastebin.com/FwBfvRbw dmesg19:38
malhmm, no messages from droid-hal-init19:50
piggzmal: rinigus: perhaps use the HintSystem property?19:50
malpiggz: what is that?19:52
piggzmal: a property for block devices in udesks19:53
piggzit seems to be set correctly for me for /dev/sd? and mmcblk0p019:53
riniguspiggz: I did try on sunday, but it somehow wasn't taking it into account19:53
piggzie, true for sd?, false for sdcard19:53
kimmoli@adampigg sähkölinnun kakkaa :) on my old KE20 corolla, i had a stone under drivers seat to keep it not falling down. Just changed to larger when it did drop through the floor19:57
T42<nitin03 %lastname%> @mal [hmm, no messages from droid-hal-init], any way to solve it?20:32
malnot sure, that just makes debugging a bit more difficult, did /usr/libexec/droid-hybris/system/bin/logcat work?20:35
T42<nitin03 %lastname%> @mal [not sure, that just makes debugging a bit more …], no20:35
T42<nitin03 %lastname%> will masking droid-hal-init help?20:36
maldo you still have the file /init_enter_debug2?20:37
T42<nitin03 %lastname%> @mal [do you still have the file /init_enter_debug2?], yes20:37
malwell then obviously logcat cannot work as that stops boot very early20:37
T42<nitin03 %lastname%> @mal [well then obviously logcat cannot work as that …], problem is if I remove it20:38
T42<nitin03 %lastname%> I can't telnet into the device20:38
T42<nitin03 %lastname%> i did ran journalctl -fa ...should i share it?20:38
malfrom which kind of boot?20:39
malI think the issue could be usb-moded configuration which might break usb20:39
malmasking droid-hal-init could be one thing to do and then remove /init_enter_debug2 and see if you can still telnet20:40
T42<nitin03 %lastname%> ok let me try20:41
T42<nitin03 %lastname%> https://pastebin.com/UsWLaHWF20:42
T42<nitin03 %lastname%> I was able to telnet...try running logcat...it returns logcat read failure20:47
mallogcat requires droid-hal-init to be running20:47
malthis is just one step in debugging process20:47
T42<nitin03 %lastname%> systemctl status droid-hal-init...returns this  droid-hal-init.service20:51
T42<nitin03 %lastname%>    Loaded: masked (/dev/null)20:51
T42<nitin03 %lastname%>    Active: inactive (dead)20:51
T42<nitin03 %lastname%> Warning: droid-hal-init.service changed on disk. Run 'systemctl daemon-reload' to reload units.20:51
riniguspiggz mal: when checking internal storage, HintSystem was set to true for its partitions. good thing is that it can be easily set in udev rules via udisks_system making it easy to adjust20:52
T42<nitin03 %lastname%> in dmesg droid-hal-init: Failed to initialize property area..this shows up20:52
piggzrinigus: yes, thats why i think thats a way to go20:56
piggzrinigus: mal: maybe this will help https://github.com/canonical-system-enablement/udisks2-examples/blob/master/query-block-devices.c21:01
T42<hacker12455> Update: fixed the ofono issue after adding droidmedia on cedric21:04
T42<hacker12455> I needed to remove and reinsert the SIM card21:04
ThaodanDoes avbtool get run during the creation of boot.img for devices? Mine won't boot with the image I build during droid-img-boot compared to the android image21:15
ThaodanI noticed aosp does run out/host/linux-x86/bin/avbtool add_hash_footer --image  out/target/product/griffin/boot.img --partition_size 0x04000000 --partition_name boot  --prop com.android.build.boot.os_version:10 after creating the image21:17
riniguslooks I just had issues with prev image. clean build, wifi / bt back and I have correct imei as well (as well as mac & bt addresses as on aosp).21:36
rinigusnow would have to test how cellular works...21:37
rinigusThaodan: not sure it does.21:40
ThaodanMine doesn't boot used the same kernel as in sfos, i'll test later if that helps.21:41
ThaodanNeed someone else that wants to step into the tarkness of android 10 :D21:41
rinigusThaodan: have you removed all possible extra filesystems in your config? I had too large kernel with nfs and other goodies included (opened an issue regarding it on sony's github)21:43
Thaodanrinigus: only nfs as you did but the exact same kernel builds aosp..21:44
ThaodanAlso I didn't have --os_version 10 --os_patch_level 2019-09-05  in my boot pkg21:46
rinigusThaodan: I think there was more, like cifs and something else. I found it by mainly commenting out as much extra options as I could. But mine was showing 'corrupted device'. do you make it to sony logo?21:46
Thaodanrinigus: yes the screen stays black for a while and then reboots21:47
ThaodanI wish I had uart aviable21:47
rinigusThaodan: then its maybe something else. recovery?21:48
rinigus... trying to boot recovery?21:48
ThaodanI should try? I already boot to hybris-recovery to debug stuff21:49
piggzrinigus: mal: my reading of the plugin code indicates that internally, it ignores the properties from udisks, and just sets system/external based on its own criteria21:49
riniguspiggz: does it at least load from udisks?21:50
piggzrinigus: there is a monitor that listens to udisk add events21:51
riniguspiggz: thanks for looking into it!21:55
T42kisekinopureya was added by: kisekinopureya22:10
r0kk3rzpiggz: what criteria?22:19
piggzr0kk3rz: i mean that it has its own abstraction of the storageTypes (external/system etc) and doesnt seem to query udisks for them22:26
r0kk3rzwhere is the codes?22:27
piggzr0kk3rz: https://git.merproject.org/mer-core/nemo-qml-plugin-systemsettings/commit/5772467e04150535b97f6a5796bce9d204a46f3b22:48
r0kk3rzah, so blame rainemak ;)22:52
r0kk3rzis my regex fu failing, or does that look like it matches pretty much everything other than mmcblk0* disk paths22:56

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