Sunday, 2019-07-07

MDTech-us_MANHey all02:01
MDTech-us_MANfor some reason every time I try to do build_packages, it upgrades droid-hal-nash-devel02:01
MDTech-us_MANthen it always ends with:02:02
MDTech-us_MANThere are some running programs that might use files deleted by recent upgrade. You may wish to check and restart some of them. Run 'zypper ps -s' to list these programs.02:02
MDTech-us_MAN* Check /home/maxwell/hadk/external/libhybris.log for full log.02:02
MDTech-us_MAN!! can't install the package02:02
merbotMDTech-us_MAN: Error: "!" is not a valid command.02:02
Mister_MagisterMDTech-us_MAN: for most of europe it's bast midnight02:17
Mister_Magisteru need to wait untill everybody wakes up02:17
Mister_Magister"for some reason every time I try to do build_packages, it upgrades droid-hal-nash-devel" as it's supposed to02:17
Mister_MagisterMDTech-us_MAN: that's not error ignore it but look into libhybris log cause building libhybris fails02:18
MDTech-us_MANI decided to try doing a fresh build and when I got to the build_packages part, after following the instructions in the faq regarding ssu-kickstart I get the following failed dependencies:06:04
MDTech-us_MANerror: Failed build dependencies:06:04
MDTech-us_MANpkgconfig(android-headers) is needed by droid-config-nash-1-1.armv7hl06:04
MDTech-us_MANrepomd-pattern-builder is needed by droid-config-nash-1-1.armv7hl06:04
MDTech-us_MANssu-kickstart-configuration-jolla is needed by droid-config-nash-1-1.armv7hl06:04
deathmistmorning! I'd like to make my life a little easier and add a few files to the installer zip file to fix tar on TWRP for example, how would one accomplish that? I already found the script at hybris/hybris-boot/updater-script and tried "tar -xf sfos.tar.bz2 ./bin/tar" but it seems to take forever to extract the single file08:01
vknechtdeathmist, see how hybris/droid-configs/droid-configs-device/kickstart/pack_package-droid-updater and iiuc and am not mistaken, you can tweak it by copying it to hybris/droid-configs/kickstart/pack_package-droid-updater08:07
malMDTech-us_MAN: check faq for related notes08:43
T42<DSstill01> how to revert a deleted file in an obs package?08:48
malwhat do you mean?08:49
T42<DSstill01> i deleted a file in droid-hal in obs and i want to restore it08:50
T42<DSstill01> anyone which has access to common repo can rebuild libgrilio? because i need 1.35 for latest ofono09:03
T42<DSstill01> i didn't want to make a package in my repo/adding that repo. it would be better to update it on common09:10
mal@DSstill01 why do you want to use latest ofono09:16
T42<DSstill01> Because i cherry-picked latest commits, including the one which changes libgrilio version :p09:17
T42<DSstill01> I can add a revision but meh09:17
vknechtI think there's no choice, had to build libgrilio locally too, to satisfy dependencies09:17
monich@DSstill01 yeah, don't pull anything from that repo automatically, it's a testing place, it's often broken. I gave you the link in case if you want to just download the rpm09:18
vknechtif I'm correct, ofono required this new libgrilio was tagged, so it ends up in common, but libgrilio is too old09:20
T42<DSstill01> Hmm, at this point i add a revision to ofono _service and i will remove once libgrilio is updated in common repo09:20
T42<DSstill01> It's tagged but they didn't rebuild it09:20
monichanother approach is to downgrade ofono to mer/1.21+git5109:23
monichit's mer/1.21+git52 which introduced dependency on libgrilio 1.0.3509:25
T42<DSstill01> For now i added a revision so that it gets everything before libgrilio 1.0.35 commit. I think that's the best approach09:26
vknechtis there a constraint on OBS, so that we have to add new packages from web interface rather than locally with osc ?09:35
mal@DSstill01 you didn't answer why you needed those changes09:36
T42<DSstill01> i don't need them09:37
malthen why did you build newer ofono?09:37
malusually ofono should be kept at the version released with the sfos version you use, unless for a good reason09:38
T42<DSstill01> i didn't09:38
T42<DSstill01> i'm building it with old libgrilio09:38
mal@DSstill01 then I don't understand what your problem is09:38
malwhy are you building ofono09:38
T42<DSstill01> because i have 2 professional commits needed for my device09:39
vknecht@DSstill01: iiuc it's the reverse, you have to build specific grilio, not ofono09:39
malso then patch those on top the ofono version had09:39
malthen all dependencies are right09:40
T42<DSstill01> mal: no because they are only for samsung devices while ofono in common repo is unified09:41
vknechtoh well, ignore me :-)09:41
malthere is no ofono in common repo09:41
mal@DSstill01 you are missing the point09:41
malalso you should talk to monich whether those patches could be included to ofono09:41
T42<DSstill01> aww, did you remove it? i remember i had to increase my ofono version not to conflict with the sf one09:42
T42<DSstill01> they're this and this:
mal@DSstill01 you can still increase the version number even if you use ofono version from 3.0.3 as base for your patches09:44
mal@DSstill01 those packages should not even have been in devel common09:45
malsome webhook again added those there09:45
T42<DSstill01> i remember there was an ofono somewhere and I had to increase the version number to avoid installing the sf one09:46
T42<DSstill01> probably they removed it09:46
malcome on, you are completely missing my point09:47
T42<DSstill01> you said that those packages (ofono) shouldn't be in common09:48
malyes, but that was not the point, which was that you applied your patches on top of some newer ofono version instead of 1.21+git44 which is in and therefore you have dependency issues09:50
T42<DSstill01> no, i don't have dependency errors, because i put a revision param in _services that excludes newer libgrilio commits09:51
T42<DSstill01> and my 2 commits were before those changes09:52
malwhat the hell are you talking about09:52
malwhat has libgrilio have to with any of this09:52
T42<DSstill01> check and
malI really don't understand what the problem is09:53
T42<DSstill01> none now09:53
malI have no idea what it ever was09:54
T42<DSstill01> i fixed it, i was just asking to update libgrilio in common repo, but since you said no, i excluded the commits that uses grilio 1.35 by putting a revision09:55
T42<DSstill01> that's all lol09:55
mallike I said, you should have just taken 1.21+git44 of ofono and applied your two patches on top of that, bumped the tag version and build09:56
malthat is the proper way09:56
T42<DSstill01> that would require to delete the branch/repo every time that ofono gets updated, so, i cherry-pick newest commits on top of older ones09:57
malthat is why I said to talk to monich09:57
malno need to delete the branch, just create a new branch for new release and then build that on OBS09:58
T42<DSstill01> i would talk to him but he wouldn't accept that because it needs a flag not to include those for normal devices09:59
malhmm, there are other quirks in ofono already, if that is enabled only if some quirk is enabled I wouldn't be a problem10:00
mal*it wouldn't10:00
T42<DSstill01> hmm10:01
malif you look here for quirks
T42<DSstill01> ah yes10:02
vknecht@DSstill01: upstream, upstream, upstream ! Not only is it professional, but you'd be famous too :-)10:02
T42<DSstill01> awwwwwwwww yeeeeah10:02
T42<DSstill01> i think i'm already famous, tho10:02
mal@DSstill01 you were thinking about including that to mer-core the wrong way, it would be specifically enabled when needed not enabled by default10:03
maland the config file is easy for doing that10:04
T42<DSstill01> no, i was thinking about makefiles flags10:04
malyes, that is not the proper way10:04
malin this case10:04
T42<DSstill01> allright10:04
malmaybe have a look at how those other quirks are used and adapt your patches use the config file10:05
T42<DSstill01> yes, i will try to improve my patch (the newest one) a bit first10:06
malI also used to patch ofono for one old device10:06
malI just updated the branch for every release10:07
malthat change was so ugly it couldn't be included in ofono10:07
T42<DSstill01> but would that be a problem for my users? I mean, 1.64 to 1.21...would they miss sf ofono updates until ofono reaches v1.64?10:08
T42<DSstill01> mine is ugly as well, but it works10:08
malwell you shouldn't have bumped the version so much10:08
T42<DSstill01> LOL10:09
malusually packages can be downgraded also during update10:09
T42<DSstill01> i wanted to be sure..10:09
malyou can test that10:09
T42<DSstill01> not all users will manually downgrade it10:09
T42<DSstill01> it needs to be an automated process that is done when running zypper dup10:10
malI meant that you change the version and then do normal update and see if the new version is installed10:10
malyou just have to test it10:11
T42<DSstill01> you mean the tag of the fork ?10:11
T42<DSstill01> ok will try10:12
T42<DSstill01> i have a problem with obs that adds 1.64 even if i deleted all tags except 1.2010:24
T42<DSstill01> ok, i fixed it, it's building 1.2010:28
malyou should not go that low, then it will install the one from jolla repos10:37
T42<DSstill01> which is what i want hehe10:37
liuxit still in bootloops for packages built on obs. i'm comparing rpms beteen obs's and local built's. until now, I found obs droid-config-aries.rpm contains extra adaptation-community.ini compared to local one. have no idea why obs built rpm has this file but local does not have10:39
T42<elros34> how about debugging why it bootloop? If it's cm11 then the first thing to check is libhybris10:53
malliux: yes, OBS builds have one repo more (which is the OBS project for your device)10:54
malliux: it comes from
malliux: are droid-hal rpms the same version?10:55
liuxelros34, i'm trying to debug it now. but it reboot after seconds the diconnect of telnet, and before usb net work. i cannot get error of reboot point10:55
liuxmal, yes, all same10:56
T42<elros34> liux: which telnet 23 or 2323?10:56
liuxelros34: telnet 2323. it can be connected then disconneced after seconds.10:58
malliux: do you have time to run dmesg -w in there10:59
liuxelros34: only two time i'm lucky to see the language selection appeared. maybe next time i can get journal log11:00
liuxmal, let me try11:01
maljournalctl -f is also a good option11:01
T42<elros34> touch /data/.stowaways/sailfishos/init_enter_debug2 so you can easily grab logs11:02
liuxmal, elros34, i've got some log from dmesg and journal:
T42<DSstill01> monich: if i added a config for this: would you merge it to upstream?11:10
T42<elros34> liux: faq line 57 iirc is solution to QPA-HWC: err in present returned -22 also for udev spam. What libhybris version do you have in local repo?11:19
liuxelros34, it's tag:
liuxwith commit 5a0019d067f010e341ef22f80b472e4c6fe5d97211:21
maltry journalctl -b -f to get earlier journal log parts11:22
malliux: verify the libhybris revision by going to libhybris source folder and run git status and git show HEAD11:23
T42<elros34> reboot when lipstick starts looks like well know problem with libhybris on cm11 but if you claim 0.5.31 worked for you earlier then maybe it's fixed now.11:26
liuxmal, it's commit 5a0019d067f010e341ef22f80b472e4c6fe5d972 (HEAD -> master, tag:, origin/master, origin/HEAD)11:26
liuxmal, the libhybris contains submodule with same name... perhaps my obs configuration does not make obs fetch same code as local repo11:31
malliux: the reason why I asked was that sometimes local source can have wrong submodule version11:32
liuxelros34, could you help show me the patch for the problem? i'll check if obs fetch different source code which does not contain the patch11:34
liuxmal, the HEAD of submodule is "commit fae0a52632893cbf3bb7ce0762147b7067db176b (HEAD)"11:35
malok, so that is not the problem then11:35
mal@elros34 I doubt that is the issue since his local builds work11:36
T42<elros34> yeah, I thought the same11:37
malliux: recheck your local config repo with git status for any uncommitted changes11:39
liuxmal, ok11:40
T42<elros34> liux: does /proc/last_kmsg after it reboots? if no try to get whole dmesg. dmesg -w might not work so use: while true; do dmesg -c; done11:40
vknechtplease patternize ?11:41
liuxis the patternize necessary for building? I 've never applied for patternize. does it matter?11:43
malvknecht: why home build and not take into use the actual f5121 project11:44
vknechtwell, still experimental, and not using the same base, so probably not ready to replace current project I guess11:46
malvknecht: which base are you using11:46
vknechttook time to make a big description ;-)11:46
malis everything working in that?11:47
vknechtwell, probably not yet ; put it in OBS to check if it fixed my local build problems, aka. droid-config-f5121 rpm not getting used by mic11:48
vknechtso almost all sparse files missing11:48
vknechtthus no gui11:48
liuxelros34, it's hard to connect with telnet or usb net because it always disconnect in a short time. let me try11:49
vknechtdidn't try to build image from it yet, have to shower and eat ;-)11:49
malliux: did you get full journal log yet with journalctl -b -f11:50
T42<elros34> liux: did you touch that file I said?11:51
liuxmal, elros34, not yet. i'm still trying to connect it with telnet or usbnet during bootloops11:54
T42<elros34> you can create it in twrp11:56
malI would also try to disable droid-hal-init for test purposes12:00
liuxelros34, just touch it but still hard to connect device12:02
T42<elros34> it should stop booting, does it reboot now?12:05
T42<dashinfantry> who test  sailfish  os on xperia  10?12:06
liuxelros34, does not reboot now but no way to connect in it. sorry didn't notice your msg on irc just now12:26
liux elros34, does not reboot now but no way to connect in it. sorry didn't notice your msg on irc just now12:26
T42<elros34> liux: try this
piggzmal: any more thoughts in my audio library loading problem?12:33
malnot really, did you give full logs12:37
piggzmal: its fixed!12:39
piggzTheKit had the fix all along12:39
piggzln -s /system/lib/vndk-28/ /usr/libexec/droid-hybris/system/lib12:39
malpiggz: are there different files?12:40
piggzhow that was worked out i dont know ... didnt see any complaints in logs12:40
piggzmal: searching fs12:40
piggzmal: no, just these12:42
piggznothing else in /system or /vendor12:42
malhmm, wondering why it doesn't find those by default12:44
piggzbut, next i think ill do the keyboard ... but after ive been out on MTB!12:46
piggzsee if i can find those keycode in a DTS file somewhere12:46
KALUbEmal: If you can spare a minute. I've made some progress with the sprdwl.ko module. There's a binary on android called macloader which calls insmod on the sprdwl module (That's why it wasn't loaded by default on sailfish). However it then tries to access a wlan0 socket which it can't find. strace:
KALUbEAlso relevant logcat output:
malKALUbE: that tries to insmod but the module is not present12:56
malwhere is the module on the device12:56
KALUbEOh, sorry that's an older log. I fixed that by symlinking it to the right place12:56
malboth logs have the same error12:57
KALUbESymlinking /lib/modules/3.10.blah/sprdwl.ko back to /system/lib/modules/sprdwl.ko creates this:
KALUbEChecking output of lsmod before and after shows that the module DOES get loaded13:06
KALUbEBut logcat has
KALUbEAnd the file does exist
malKALUbE: so what happens now if you modprobe the module again?13:10
malor is the module loaded already after that13:11
KALUbEmodprobing returns nothing and exits with status 1. I assume that means module already loaded?13:12
malso no wlan inferface or device?13:14
KALUbEHuh, I had two levels of symbolic links. That seems like the cause13:15
KALUbEmali.ko wasn't getting loaded either, dmesg showed droid-hal-init complaining about too many levels of symbolic links13:17
KALUbEOk, fixed that weird issue... This is output I saw yesterday
malwhy are you showing old logs13:21
KALUbEI mean, it's the same output I got yesterday. I forgot to log it. Sorry13:22
malthat log still complains about missing module13:28
KALUbEThat's what I don't get, it opens the module, calls insmod succesfully (lsmod shows that it gets loaded) and then it complains that it failed13:30
malKALUbE: show latest output13:36
malwith correct module13:36
KALUbEstrace: logcat: lsmod:
KALUbEmal: correct logcat
malwhy are you not giving dmesg13:42
malthe strace still says "return of insmod : ret = 0, No s" which is hiding some part but I think it might be "No such file"13:43
KALUbEYeah, you can see the output in logcat13:43
malI thought you fixed that13:43
malyou said you added the file13:44
malshow ls -l /system/lib/modules/sprdwl.ko13:44
malmaybe it doesn't like symlinks13:44
KALUbEI copied the file just in case, strace shows it opens /system/lib/modules/sprdwl.ko succesfully, and calls read and init_module which both execute fine13:45
KALUbESo I don't understand why it would say no such file or directory, it opens the file...13:46
malshow the strace of that situation13:47
KALUbEthat's the one I linked above13:47
malbut you said it opens it it successfully13:47
KALUbEline 460 and 472?13:48
KALUbEit opens as fd 4 I think, and then calls read on it. Although I might be misunderstanding system calls13:48
malnot sure if you really need that binary13:48
KALUbEMaybe not, it does seem to expect the socket to exist already. So it's the module that's having issues13:49
malit would be nice to know what android does13:52
malwhich module it loads etc13:52
vknechtmal, do you have patternizing powers ? :-)13:54
KALUbEmal: I'll grab some android logs13:58
malvknecht: nope13:58
malvknecht: you can add patterns manually for testing13:59
deathmist"No update candidate for 'pulseaudio-modules-droid-glue-12.2.4-1.armv7hl'. The highest available version is already installed." any clues?14:00
malis that a problem?14:00
deathmistyeah, it stop mic build14:01
malhow about you show mic output14:01
vknechtgunzip droid-local-repo/$DEVICE/repodata/whatever-patterns.xml.gz, split each patten in its own file and put those in a disabled _pattern package ?14:05
malsome people have done it14:11
MDTech-us_MANmal: I get that error when I run the mentioned commans14:13
vknechtok thanks, reading on it14:13
MDTech-us_MANthe ones under "Building droid-config package for Sailfish OS"14:14
MDTech-us_MANerror: Failed build dependencies:14:14
MDTech-us_MANpkgconfig(android-headers) is needed by droid-config-nash-1-1.armv7hl14:14
MDTech-us_MANrepomd-pattern-builder is needed by droid-config-nash-1-1.armv7hl14:14
MDTech-us_MANssu-kickstart-configuration-jolla is needed by droid-config-nash-1-1.armv7hl14:14
MDTech-us_MANBuilding target platforms: armv7hl-meego-linux14:14
MDTech-us_MANBuilding for target armv7hl-meego-linux14:14
MDTech-us_MAN* Check /home/sailfish-build/hadk/hybris/droid-configs.log for full log.14:14
MDTech-us_MANI get those errors when I run "rpm/dhd/helpers/ -c"14:14
MDTech-us_MANafter doing that git reset command14:15
maland you have properly built droid-hal rpms?14:15
MDTech-us_MANyou mean the "make hybris-hal"?14:16
MDTech-us_MANor do you mean running "rpm/dhd/helpers/" without the -c14:16
mallatter one14:17
MDTech-us_MANthe problem there is that I get the following error:14:18
MDTech-us_MAN IMPORTANT: some devices in your Android tree are 64bit targets. If your device is aarch64,14:18
MDTech-us_MAN            please define droid_target_aarch64 in your .spec, otherwise define droid_target_armv7hl14:18
MDTech-us_MAN NOTE: Currently there is no Sailfish OS ARM 64bit target, so leave PORT_ARCH as armv7hl14:18
MDTech-us_MAN       Mixed builds of 64bit Android+Linux Kernel and 32bit Sailfish OS work just fine.14:19
MDTech-us_MANI added "%define droid_target_aarch64 1" to rpm/droid-hal-nash.spec but it doesn't seem to be taking affect14:19
MDTech-us_MANahh, nevermind. it seems I added it after the include instead of before it14:21
MDTech-us_MANthats why it didn't find it14:21
MDTech-us_MANthanks for the help14:23
malalso please don't paste many lines here, usually if you have more than 3 lines pastebin or similar service is better14:24
KALUbEmal: I have dmesg and logcat from android: and Interestingly that macloader binary has the same output but wifi works so you were right about it not being important14:26
KALUbEThe driver is calls SC2331 in dmesg I think14:26
vknechthmm, how to get OBS to select the binder spec variant for geoclue, name the package geoclue-providers-hybris-binder instead of just geoclue-providers-hybris ?14:30
vknechtor something in meta or project config tabs ?14:31
deathmistmal: does mic create a full log somewhere I'm not aware of? that's all I get on the terminal when I try to build the rootfs14:31
maldeathmist: rebuild pulseaudio-modules-droid14:37
maldeathmist: I removed pulseaudio wrong devel common and you happened to build before that14:37
maland now it won't find it14:37
vknechtok, selection seems to be from package name14:38
MDTech-us_MANmal: so, here's the next error... its when running build_packages:
vknechtMDTech-us_MAN, try rpm/dhd/helpers/ --mw=
MDTech-us_MANtrying it...15:06
vknechtthen --mw (and you can skip those already build)15:07
MDTech-us_MAN"error: RPC failed; result=22, HTTP code = 404"15:07
MDTech-us_MANwhen its cloning15:07
MDTech-us_MANtryign with .git on the end15:08
MDTech-us_MANall good now15:09
MDTech-us_MANofono-ril-binder-plugin built successfully now15:12
vknechtcrap, with OBS I still get the mic error about droid-config-f5121-bluez5 not installable (and probably with the -sailfish too after disabling the -bluez5 one):15:29
vknechtrepo problem: pattern:jolla-configuration-f5121-1-1.noarch requires droid-config-f5121-bluez5, but this requirement cannot be provided, not installable providers: droid-config-f5121-bluez5-1+master.20190707092513.786300a-1.1.3.jolla.armv7hl[adaptation-community-f5121-]15:29
vknechtwith those excluded from pattern, the image builds fine15:30
* vknecht goes for a walk15:30
malvknecht: do you have some conflict in spec15:35
vknechtmal, in a spec file from droid-configs you mean ?15:37
malvknecht: did you try the usual debug method mentioned in hadk pdf15:41
MDTech-us_MANSo I've been letting it build for a while and it seem stuck on "rpm/ngfd-plugin-native-vibrator.spec". If I look in hybris/mw/ngfd-plugin-droid-vibrator.log, the last line I see is "-- Checking for module 'glib-2.0'" and its been like that for a while now. is this normal?15:41
malit tells what to do if it gets stuck15:41
malread the warning it prints15:41
vknechtwell, it's different from the classical "uninstallable provider" ; the droid-config-f5121-bluez5/sailfish rpm are there15:42
vknechtiirc I did the debug nevertheless, which endend up with mic installed just the ...-bluez5 rpm15:43
vknechtso it's something different15:43
malvknecht: show your .ks15:43
malvknecht: just use it15:43
maland see what happens15:43
malvknecht: but you have the normal not installable provider15:44
MDTech-us_MANmal: where does it print this warning?15:45
MDTech-us_MANdon't see anythign like that in the log or build_packages output15:45
vknechtextracted from and corrected to use home OBS:15:45
vknechtuninstallable != not installable, from what I've understood15:46
vknechtbut ok I'll do it after walking and taking the sun a bit :-)15:47
vknechtbtw I added you as maintainer for this OBS home project, in case you need/want to tweak it15:49
malMDTech-us_MAN: either in the build log or to where you build it15:51
liuxmal, elros34, I've uploaded my local libhybris*.rpm and disable libhybris build on obs. then I make image with home repo. The image works fine just like my local built15:52
malliux: so verify again the local libhybris packages15:52
malI mean the local repo15:52
malvknecht: can't you just try it anyway15:53
T42<DSstill01> how to update patterns in obs?15:54
MDTech-us_MANdidn't see that anywhere:
malliux: pastebin output of git remote -v, git status, git rev-parse HEAD in libhybris folder and same in libhybris/libhybris submodule15:55
malMDTech-us_MAN: I gave you the warning message already15:55
MDTech-us_MANjust letting you know it didn't show up anywhere15:55
liuxmal, i'm sure i have not changed source in libhybris manually.  some file was created by maybe. i'll paste git diff on pastebin15:57
malyes, git diff of both is also wanted15:57
malliux: well there are not many possibilities what could be wrong15:57
MDTech-us_MANand stopping it and trying again did work, the build is continuing on. thanks.15:58
liuxi have no idea who made these changes and how they affect build result16:04
malliux: which android base was that, cm-11.0?16:17
malliux: try in OBS to use revision 20bcd5be11fcd82aa9e2e995c4bb37af9c94cbad16:19
liuxmal, yes, cm-11.016:20
liuxok, I'll try it on obs16:21
liuxthe building is on goning. may i know the reason of this issue? i've searched on local scripts but still nothing got until now16:26
MDTech-us_MANand again, I built the image, wiped everything, installed lineageos 15.1, then the built ZIP file (modded to use a working busybox) then I reboot to system and I see the lineage OS logo. is this to be expected? or am I supposed to be seeing a silfish logo?16:33
MDTech-us_MANdmesg on the host only shows "USB disconnect" from when I rebooted to system16:34
MDTech-us_MANnothing after16:35
MDTech-us_MANwiped everything again did the same and now it doesn't show any loading screen and goes straight to the bootloader16:43
MDTech-us_MANthird time I got the lineage os loading screen16:52
malliux: there is bug in libhybris, I thought it was only for hybris-12.1 but seems it's also for 11.016:57
malMDTech-us_MAN: there is no sailfish boot logo usually16:58
malMDTech-us_MAN: what exactly do you mean by lineage loading screen16:58
MDTech-us_MANthe boot screed16:58
MDTech-us_MANam i supposed to be seeing somethign in the host's dmesg?16:59
malyes, check for usb related messages16:59
MDTech-us_MANor only at a later point in the boot process16:59
malvery early in boot16:59
MDTech-us_MANso far nothign16:59
malnothing usb related?16:59
MDTech-us_MANlast I see is USB disconnect,16:59
MDTech-us_MANfrom when I rebooted the deviece17:00
malbut before that, show what you get17:00
MDTech-us_MANthose are the last few17:01
malso nothing related to Mer17:01
malwondering if the correct kernel gets flashed or maybe bootloader somehow reverts the kernel17:02
malI think there have been devices that don't like unsigned kernel images17:02
MDTech-us_MANI've installed other unofficial roms here before and had no issues17:03
MDTech-us_MANwould I be able to verify the kernel from recovery?17:04
MDTech-us_MANare there any boot logs that might be visible somewhere?17:05
T42Rajpratik71 was added by: Rajpratik7117:09
*** justgerd9 is now known as Gerd17:22
*** Gerd is now known as JustGerd17:22
MDTech-us_MANthis is actually pretty interesting17:25
MDTech-us_MANmal: now that I look at the other ROMs that I flashed, in the META-INF folder I see CERT.RSA, CERT.SF, MANIFEST.MF17:26
MDTech-us_MANI suspect that means those were actually signed17:27
MDTech-us_MANis there a way to sign the sf zip?17:32
MDTech-us_MANor work around this?17:32
mccrearyelros34: I replaced my idiosyncratic recovery with twrp-3.0.2-0-asanti_c.img from, and did a full installation from scratch. I had no trouble with the touchscreen on a HW A Photon Q18:07
mccrearySo let's chalk this all up to using a 'bad' recovery build. My only remaining feedback would be to emphasize installing the latest recovery from in your already very good instructions18:08
mccrearyThanks again for working on SailfishOS for moto_msm8960_jbbl :-)18:09
malMDTech-us_MAN: you only need to sign the kernel, there is a way to do that, probably a web search will tell how18:14
malliux: any luck with the older libhybris18:15
T42<elros34> mccreary: great, I will add recommended twrp to instruction. Just curious, have you find out which property was wrong?18:27
T42<elros34> MDTech-us_MAN: flash hybris-boot to clear any boot failure flags and then boot device with recovery button pressed.18:30
vknechtso here's two mic runs with 2 KS : 1st with @Jolla Configuration f5121 with the error we spoke about previously, 2nd with KS referencing only offending droid-config-f5121-bluez5 :
mccrearyI did not find the underlying bug, but I believe it only affected the cm-11 installation.18:38
mccrearyIn the instructions for updating to 3.0.3, there is a reference to I assume this is a typo, and was intended?18:39
malmccreary: what instructions are you talking about18:45
T42<elros34> mccreary: use, it's more up to date18:46
mccrearyOK :-)18:47
mccrearymal, I was referring to elros34's instructions from
T42<elros34> don't forget about patched glibc repo otherwise device will not boot. I couldn't add it to device common because of dependecies issues.18:48
mccrearyelros34, thanks for the reminder. BTW, nadlabak worked on a 3.4 kernel for the Photon Q:
mccrearyIt has some stability bugs, so was never used for production builds19:01
T42<elros34> I know, shame it was never finished like 2GB ram mod19:04
piggzi think im managing to work out how this keyboard slide code works!19:21
vknechtpreparing a topic for next meeting (thursday July 11th/22 Messidor) about this aosp8 rebase for X/X Compact I've been working on... anyone willing to be a « substitute » ?19:24
vknechtmore info:
vknechtwill add a bit more in this overview shortly, like the fact LVM is disabled atm (tho it works, but makes it difficult to debug since no telnet past switchroot), and not using droid-system rpm tho I made repo for it19:29
MDTech-us_MANelros34: same thing. I see the lineage os boot screen and nothing in dmesg19:46
MDTech-us_MANmal: I can't find any info on signing kernels specifically. only zip files.19:47
malMDTech-us_MAN: really, first search I made gave a lot of links, "android sign boot.img"19:52
T42<elros34> not sure what you mean by os boot screen. You should see only bootloader logo not lineage animation. Does your device use skip_initramfs?19:52
malthat is one option for the issue19:53
malbut still it should remain at bootloader logo19:53
MDTech-us_MANfor me it goes past the bootloader logo19:53
MDTech-us_MANand is stuck on the lineage OS loading screen19:53
malwhich would suggest that it didn't like you kernel19:54
MDTech-us_MANyeah, I was searchign for signing kernel, not signing boot img, thats why I didn't find anything. attempting to do it now.19:56
MDTech-us_MANthings just got weirder20:15
MDTech-us_MANit actually booted into lineage os20:16
MDTech-us_MANto the lineage os setup wizard20:16
MDTech-us_MANbut with a bunch of error messages popping up20:16
T42<elros34> so if you are in android now: cat /proc/cmdline | grep skip_initramfs20:17
MDTech-us_MANthen it rebooted into recovery on its own20:17
MDTech-us_MANI can't bet out of the setup wizard20:18
MDTech-us_MANthe error messages are covering it20:18
MDTech-us_MANand within 60 seconds it just goes to recovery20:19
MDTech-us_MANand then recovery restarts and it tries to boot system again20:19
MDTech-us_MANand the loop continues20:20
vknechtanyone with enough TJC karma to add this at ?20:21
vknechtvknecht (karma: 1, badges: ● 1 ● 1) lol20:22
T42<elros34> MDTech-us_MAN: you can flash lineage kernel, boot and check that command or just revert commit which add skip_initramfs in your kernel: init/initramfs.c20:22
MDTech-us_MANI'll just revert the commit20:24
MDTech-us_MANI just need to rebuild hybris-boot and then do -d, right?20:26
T42<elros34> just flash hybris-boot you dont need to build whole new image20:27
MDTech-us_MANokay... I flashed the new out/target/product/nash/hybris-boot.signed.img to both boot_a and boot_b and now it gets stuck at the bootloader screen and doesn't get to the loading screen20:35
MDTech-us_MANin around 60 seconds it the device resets and the same happens again20:36
MDTech-us_MANnothing in the host dmesg related to Mer20:37
T42<elros34> progress :)20:39
MDTech-us_MANdefine progress :)20:39
MDTech-us_MANany debug info I can extract in this state?20:39
T42<elros34> here you go:
T42<elros34> show your device repo20:40
MDTech-us_MANI didn't commit the fixed "fixup-mountpoints yet though20:41
T42<elros34> what about android device repo20:42
MDTech-us_MANwhich folder would it be cloned to?20:42
T42<elros34> device/"vendor"/"device"20:43
MDTech-us_MANthat's the repo currently used20:46
MDTech-us_MANwhoa. after the device reset one of these times, I got the following error: AP Fastboot Flash Mode (Secure)20:49
MDTech-us_MANNo bootable A/B slot20:49
MDTech-us_MANFailed to boot Linux, falling back to fastboot20:49
MDTech-us_MANFastboot Reason: Fall-through from normal boot mode20:50
T42<elros34> yeah it might fail to boot too many times, you need to flash kernel again20:51
MDTech-us_MANthat's somethign like what I suspected.20:51
MDTech-us_MANjust need to fix the actual problem first20:52
T42<elros34> try to replace every "rndis_bam.rndis" with "gsi.rndis" in hybris/hybris-boot/init-script. Then make hybris-boot and flash it20:52
MDTech-us_MANalmost the same21:01
MDTech-us_MANnow it vibrates after aroudn 5 seconds21:01
MDTech-us_MANthen around 60 seconds it resets21:01
T42<elros34> nothing in dmesg now?21:03
MDTech-us_MANdoes it save logs anywhere?21:07
MDTech-us_MANhere is the part that I changed in the init-script:
MDTech-us_MANand the diff:
MDTech-us_MANmaybe I messed somethign up there?21:10
T42<elros34> It's still in initramfs so no logs unless you manually mount for example sd card and put some info there21:10
MDTech-us_MANI do have an sd card in the phoen21:11
T42<elros34> you could also try to enable CONFIGFS_RNDIS/QCRNDIS or RMNET_BAM in defconfig. I am pretty sure there was somebody with simillar issue not so long time ago21:12
piggzmal: im now getting events like this21:12
piggzsh-3.2# evdev_trace -t /dev/input/event221:12
piggz/dev/input/event2: 1562533888.433 - 0x05/EV_SW - 0x00a/SW_KEYPAD_SLIDE - 121:12
piggz/dev/input/event2: 1562533888.433 - 0x00/EV_SYN - 0x000/SYN_REPORT - 021:12
piggz/dev/input/event2: 1562533891.393 - 0x05/EV_SW - 0x00a/SW_KEYPAD_SLIDE - 021:12
piggzbut, on-screen kaybord not showing up21:12
Mister_Magisterpiggz: gj21:14
MDTech-us_MANelros34: should I keep those changes in the init-script?21:14
T42<elros34> you can append gsi instead replacing rndis_bam21:16
MDTech-us_MANelros32: is it bad to enable all 3?21:18
MDTech-us_MANrndis isn't building21:23
MDTech-us_MAN../../../../../../kernel/motorola/msm8998/drivers/usb/gadget/function/rndis.c:46:1: error: use of undeclared identifier 'KBUILD_MODNAME'21:23
MDTech-us_MANand a few more21:23
T42<elros34> iirc somebody with motorola had similar error but don't know whether he fixed it or not21:26
MDTech-us_MANI had an error when I was buildign the kernel before somewhere in bluetooth related stuff and I solved it by copyign a portion of the code from the main linux code repo21:28
MDTech-us_MANshould I try disablign CONFIGFS_RNDIS/QCRNDIS and enabling RMNET_BAM for now?21:28
T42<elros34> dunno you need to try, you could also boot android and figure aout correct usb setup:
T42<elros34> there is also chance that you have /sys/class/android_usb/android0 directory and hybris-boot use android_usb instead configfs at all:
piggzmal: so, i added some debug lof to mce, and it looks ok ... even the screen powers on/off when sliding keyboard,
piggzbut, no OSK is shown21:45
piggzcould it be becuase the device the generates the slide, is different to the keyboard, so, it thinks the keyboard is seperate and always on21:45
piggzmal: yes, it must event_input.c21:52
piggz    /** Check if another device node is supposed to provide slide status */21:53
T42<elros34> that shoudn't be a problem I have SW_LID in event0 and hwkbd in event 1. There is /run/stat/namespaces/maemo_InternalKeyboard/Open maybe this is used to show/hide vkb21:53
piggz@elros34 what device is that?21:55
T42<elros34> Photon Q21:56
piggzother thing that works without config is the keyboard backlight22:00
piggzjust need to figure out the damn on-screen kdb!22:01
malpiggz: ah, check xperia 2011 repos, that tell how to config it if slide is from different device22:02
malpiggz: so just check the names from evdev_trace -i to see which ones to use22:07
MDTech-us_MANelros34: I managed to get it to build, but now it just gets stuck at the bootloader screen. and nothing in the host dmesg23:18
MDTech-us_MANflashed lineage and trying to look at how it works23:20

Generated by 2.17.1 by Marius Gedminas - find it at!