*** cvp <cvp!~cvp@pD9EE2636.dip0.t-ipconnect.de> has joined #sfdroid | 00:10 | |
*** cvp <cvp!~cvp@pD9EE2636.dip0.t-ipconnect.de> has quit IRC (Remote host closed the connection) | 00:47 | |
*** horuxan <horuxan!~horuxanxx@201.46.189.63> has quit IRC (Ping timeout: 240 seconds) | 02:16 | |
*** horuxan <horuxan!~horuxanxx@201.46.189.63> has joined #sfdroid | 02:18 | |
*** neerad <neerad!~neerad@112.133.229.182> has quit IRC (Ping timeout: 255 seconds) | 03:23 | |
*** IRCFrEAK <IRCFrEAK!~gk.1wm.su@2a03:4a80:2:2d3:2d3:9c14:a427:77be> has joined #sfdroid | 05:36 | |
*** IRCFrEAK <IRCFrEAK!~gk.1wm.su@2a03:4a80:2:2d3:2d3:9c14:a427:77be> has left #sfdroid | 05:36 | |
*** Mister_Magister <Mister_Magister!~foidbgen@user-94-254-147-215.play-internet.pl> has joined #sfdroid | 08:43 | |
*** neerad <neerad!~neerad@112.133.229.182> has joined #sfdroid | 09:36 | |
neerad | So apparently SFDROID is running only upto the booting screen of CyanogenMod .... Is this progress? | 09:51 |
---|---|---|
dr_gogeta86 | morning all | 09:57 |
neerad | morning :) | 09:57 |
*** pseudodev <pseudodev!uid205973@gateway/web/irccloud.com/x-elzpcwsousbybvsr> has joined #sfdroid | 10:53 | |
*** ranter <ranter!~ranter@37-33-116-222.bb.dnainternet.fi> has joined #sfdroid | 11:03 | |
*** Mister_Magister_ <Mister_Magister_!~foidbgen@user-94-254-145-70.play-internet.pl> has joined #sfdroid | 11:06 | |
*** Mister_Magister <Mister_Magister!~foidbgen@user-94-254-147-215.play-internet.pl> has quit IRC (Ping timeout: 240 seconds) | 11:07 | |
*** baddea <baddea!2569f918@gateway/web/freenode/ip.37.105.249.24> has joined #sfdroid | 11:18 | |
neerad | http://paste.ubuntu.com/24026293/ Any ideas ? | 11:26 |
neerad | SFDROID just boots up to show this https://drive.google.com/open?id=0B8I9RnxIQmQecDhnck1wR3FkUE0 | 11:29 |
baddea | yep , that releated command and your cm image prefer to using offical with salfishos + offical cm + commands = sfdroid booting success | 11:31 |
neerad | baddea but that's it nothing else happens | 11:31 |
neerad | baddea see the logcat for details and see if you have any ideas ? | 11:32 |
baddea | https://www.youtube.com/watch?v=Ov4o1f61ErI | 11:32 |
baddea | what is your device? | 11:32 |
neerad | Xiaomi Redmi 2 | 11:32 |
baddea | did you replace init in the root for your device as me have oneplus one "bacon.init" | 11:35 |
baddea | ? | 11:35 |
neerad | Mine is not bacon mine is redmi 2(wt88047) | 11:35 |
neerad | Your bacon.init will be containing import init.sfdroid.rc I guess, if there is anything else please pastebin your init file | 11:36 |
baddea | what you need to do make fresh jolla with offical cm and install sfdroid with replace of your init than enter command as unroot and root status | 11:37 |
baddea | than press icon for sfdroid and it most booting | 11:38 |
neerad | I replace what init.rc ? | 11:38 |
baddea | nope | 11:38 |
neerad | then what replaces what ? | 11:38 |
baddea | give your email i try to sending clear instruction | 11:38 |
neerad | nrdnandan@gmail.com | 11:39 |
baddea | i try to talk krnlyng to help you soon but as i know him he is busy these day | 11:39 |
baddea | my email baddea14@gmail.com | 11:40 |
neerad | krnlyng any help bro ? | 11:40 |
neerad | krnlyng I did what you said and here is the result my sfdroid is booting but only booting and nothing else happens for 15 minutes | 11:40 |
baddea | i try to sending instruction it easy not hard but if you following it and not booting there possible that you have hard bugs so krnlyng can help you with that . | 11:40 |
baddea | Good luck | 11:40 |
neerad | okay | 11:41 |
neerad | cool | 11:41 |
baddea | as oneplus one now can run sfdroid but security settings not opening face error message releated to hard bugs but i fix that with other method but problem not fixed . | 11:42 |
baddea | that make allow me to install un market apk app after all . | 11:43 |
*** baddea <baddea!2569f918@gateway/web/freenode/ip.37.105.249.24> has quit IRC (Quit: Page closed) | 11:45 | |
saidinesh5 | NeKit: Ping | 13:06 |
NeKit | pong | 13:06 |
neerad | damnnnn | 13:06 |
saidinesh5 | wow that was fast.. did you have to make any patches to init so that you coulld run your androiid x86 in chroot? | 13:07 |
NeKit | I think the trick is that there was only one init running | 13:08 |
NeKit | no droid-hal-init, only Android x86 init in chroot | 13:08 |
saidinesh5 | Ahhh interesting | 13:08 |
neerad | init.rc ? | 13:09 |
*** pseudodev <pseudodev!uid205973@gateway/web/irccloud.com/x-elzpcwsousbybvsr> has quit IRC (Quit: Connection closed for inactivity) | 13:09 | |
saidinesh5 | the thing with Andriod's init is it is trying to remount /dev/ and do all the udev stuff with ueventd.. | 13:09 |
saidinesh5 | and if i tried to use droid-hal-init in chroot by bind mountingg my /dev , sys, proc it is segfaulting | 13:10 |
saidinesh5 | so seeing what magic bits there are inside the droid-hal-init | 13:10 |
NeKit | in theory udev and ueventd should be able to work side by side | 13:12 |
NeKit | at least that's what Ubuntu has, except Android running in LXC | 13:13 |
saidinesh5 | Ahh... but if i bind mount /dev into the chroot, android's init is wiping it out to run it's own ueventd and then kaboom | 13:21 |
saidinesh5 | sailfish becomes unusable | 13:21 |
NeKit | don't mount it | 13:22 |
NeKit | I think it's better to create tmpfs for Android root and extract files there | 13:22 |
*** mal <mal!~mal@dsl-hkibrasgw1-58c392-192.dhcp.inet.fi> has quit IRC (Ping timeout: 245 seconds) | 13:23 | |
saidinesh5 | you mean extract things from android's initrd? | 13:26 |
neerad | saidinesh5 Is sfdroid booting up ? | 13:28 |
saidinesh5 | nope. after healthd, ueventd all others are crashing | 13:28 |
NeKit | yes | 13:29 |
neerad | saidinesh5 seems like mine is booting up but is stuck only in booting up process | 13:29 |
neerad | https://drive.google.com/open?id=0B8I9RnxIQmQecDhnck1wR3FkUE0 | 13:30 |
*** pseudodev <pseudodev!uid205973@gateway/web/irccloud.com/x-wgrjtkjrynwrcyah> has joined #sfdroid | 13:30 | |
saidinesh5 | hmm.. usually that means some process is crashing in Android's init and you're stuck with the bootanim | 13:30 |
neerad | http://paste.ubuntu.com/24026293/ | 13:31 |
saidinesh5 | NeKit: i am a little unsure about that approach mainly because last i tried to run the whole of the android in chroot, it was unable to get some locks on some devices and services were crashing | 13:31 |
saidinesh5 | that's why right now trying to get a very minimal subset of android, which hopefully doesnt conflict with the droid-hal-init inside chroot | 13:32 |
neerad | Nekit saidinesh5 chroot is a bad idea, although it allows it to cherry pick services and modules to load | 13:32 |
saidinesh5 | NeKit: | 13:32 |
saidinesh5 | F/SystemServer(29022): BOOT FAILURE making display ready | 13:32 |
saidinesh5 | F/SystemServer(29022): java.lang.NullPointerException: Attempt to invoke virtual method 'void com.android.server.wm.WindowManagerService.displayReady()' on a null object reference | 13:32 |
saidinesh5 | *neerad^ | 13:32 |
saidinesh5 | neerad: any reason why chroot is a bad idea? | 13:32 |
neerad | saidinesh5 I guess it hasn't worked for any other dev before | 13:33 |
saidinesh5 | windowmanagerservice sounds like a very basic service needed .. but yeah iirc i had similar messages before device starrted heating up | 13:34 |
NeKit | I fear that in case of two inits running there are conflicts with system server | 13:34 |
NeKit | Alien Dalvik has it modified probably | 13:34 |
saidinesh5 | Ahhh | 13:34 |
neerad | Nekit yes the two init causes loading the default and patched modules at the same time and leads to conflict | 13:34 |
saidinesh5 | yeah right now looking to trimm down everything that's not the init part of init.c | 13:35 |
saidinesh5 | so that all it does is start/stop services | 13:35 |
saidinesh5 | neerad: that's why the init.rc in the chroot is almost a init.sfdroid.rc | 13:36 |
neerad | saidinesh5 I am trying to load one service at a time and see how ti goes | 13:36 |
neerad | it* | 13:36 |
saidinesh5 | but it seems the init itself does more than just starting / stopping services .. mounting devices/special paths etc.. | 13:36 |
neerad | yes indeed it does that saidinesh5 | 13:37 |
*** mal <mal!~mal@dsl-hkibrasgw1-58c392-192.dhcp.inet.fi> has joined #sfdroid | 13:37 | |
neerad | saidinesh5 can it be version mismatch that might be leading to crash after crash? | 13:43 |
saidinesh5 | : java.lang.NullPointerException: Attempt to invoke virtual method 'void com.android.server.wm.WindowManagerService.displayReady()' on a null object reference - why would that be null? | 13:44 |
neerad | No idea saidinesh5.... :( | 13:45 |
saidinesh5 | source code to your rescue... | 13:46 |
saidinesh5 | com.android.server.am.ActivityManagerService.showBootMessage(ActivityManagerService.java:6287) | 13:46 |
saidinesh5 | check that line and see | 13:46 |
neerad | saidinesh5 could it be version mismatch like sledges mal and krnlyng has been telling.... | 13:47 |
*** Mister_Magister_ <Mister_Magister_!~foidbgen@user-94-254-145-70.play-internet.pl> has quit IRC (Read error: Connection reset by peer) | 13:48 | |
*** Mister_Magister <Mister_Magister!~foidbgen@user-94-254-145-70.play-internet.pl> has joined #sfdroid | 13:48 | |
saidinesh5 | could be. could be something else . like a missiing file or something. | 13:48 |
neerad | saidinesh5 I am trying to build a new base that matches SFDROID | 13:48 |
saidinesh5 | you need to check that file's source code to find out | 13:48 |
neerad | hmmm... okay | 13:48 |
saidinesh5 | i have had simmilar issues and my CM base was compiled by myself.. . | 13:48 |
neerad | saidinesh5 still it didn't work | 13:50 |
neerad | saidinesh5 You have cancro I guess | 13:50 |
neerad | *? | 13:50 |
saidinesh5 | nope. i have had similar issues .. with window manager crashing etc. in my logs | 13:51 |
saidinesh5 | just didnt push further cuz the phone was heating up .. and with it trying to use my /data partition the logs were fully useless .. as it was trying to start all my android apps too | 13:51 |
neerad | saidinesh5 No I am asking was the CM_BASE and SFDROID had the same branch and version ? | 13:51 |
saidinesh5 | yep | 13:51 |
saidinesh5 | CM-12.1 YOG7D | 13:51 |
neerad | saidinesh5 Hmmm..... So then you decided to take the "chroot" path ..... !! | 13:52 |
saidinesh5 | yup | 13:52 |
saidinesh5 | well originally LXC.. but that seemed like a lot more work than this | 13:52 |
neerad | saidinesh5 It seems to me somehow system is trying to load from the default modules and not sfdroid modules | 13:53 |
saidinesh5 | yup. iirc i remember writing that down as one more reason for trying out chroot :P | 13:54 |
neerad | Because I had many messages like "Cannot find /usr/libexec/sfdroid/system/Settings/Settings.apk using /usr" | 13:54 |
saidinesh5 | there simply were too many modules to keep track of which service is pulling what from where... | 13:54 |
mal | and there was the bug in paths | 13:55 |
saidinesh5 | ^ | 13:55 |
saidinesh5 | lol | 13:55 |
neerad | saidinesh5 I am currently building my own CM_BASE to test if it really is version mismatch or the issue is somehting else | 13:55 |
mal | not sure if the bug still exists in 12.1 branch | 13:56 |
neerad | mal Which bug are you talking about? | 13:56 |
mal | https://github.com/sfdroid/android_frameworks_base/commit/76b9994afecba9758159916cdff306b63e5f5b15#diff-cb589fb4d249734569e11adf2d1c645aR11974 | 13:57 |
mal | that should have privilegedAppDirSFDROID | 13:57 |
neerad | Great now patch had been faulty | 13:58 |
mal | that folder contains Settings.apk | 13:59 |
saidinesh5 | ffbm = android fastboot mode? | 13:59 |
neerad | yeah exactly and SystemUI | 13:59 |
neerad | yep saidinesh5 | 13:59 |
mal | neerad: so try to fix that and see if it help | 13:59 |
saidinesh5 | Ah | 14:00 |
neerad | Okay cool | 14:00 |
saidinesh5 | didnt know init is called even for fastboot mode | 14:00 |
neerad | mal One question how did it work for oneplus-onyx ? | 14:01 |
neerad | mal with the same faulty patch | 14:01 |
neerad | mal Atleast it booted as seen in the youtube video | 14:01 |
mal | neerad: no idea | 14:01 |
saidinesh5 | it was probably looking for teh file in /system , and found it there? | 14:02 |
mal | saidinesh5: but those are patched afaik | 14:02 |
* saidinesh5 sees what patches are there for SystemUI | 14:04 | |
neerad | mal saidinesh5 Somebody has to look at all critical boot up patch again and correspond to the errors in /logcat | 14:05 |
saidinesh5 | https://github.com/sfdroid/android_frameworks_base/commits/mako_12 cant see any systemUI related patches there.. | 14:05 |
neerad | saidinesh5 you are seeing mako_12 see oneplus-onyx | 14:06 |
neerad | I guess that would be appropriate | 14:06 |
mal | neerad: sfdroid-12.1-YOG7D is the to use (copied from onyx) | 14:07 |
neerad | mal Yeah both are same | 14:07 |
neerad | mal saidinesh5 is seeing for mako_12 branch | 14:07 |
mal | that should also be pretty much the same | 14:07 |
neerad | mal except where kimmoli has patched the files | 14:08 |
saidinesh5 | still not many patches there | 14:08 |
mal | I fixed the path error in sfdroid-12.1-YOG7D | 14:20 |
mal | not sure what that will cause but still better to fix it | 14:20 |
saidinesh5 | yeah | 14:20 |
neerad | mal great | 14:23 |
*** mal <mal!~mal@dsl-hkibrasgw1-58c392-192.dhcp.inet.fi> has quit IRC (Ping timeout: 255 seconds) | 14:26 | |
*** neerad <neerad!~neerad@112.133.229.182> has quit IRC (Ping timeout: 255 seconds) | 14:39 | |
*** mal <mal!~mal@dsl-hkibrasgw1-58c392-192.dhcp.inet.fi> has joined #sfdroid | 14:51 | |
*** neerad <neerad!~neerad@112.133.229.182> has joined #sfdroid | 14:51 | |
*** cvp <cvp!~cvp@pD9EE2636.dip0.t-ipconnect.de> has joined #sfdroid | 16:59 | |
*** cvp_ <cvp_!~cvp@pD9EE2C3E.dip0.t-ipconnect.de> has joined #sfdroid | 17:21 | |
*** cvp <cvp!~cvp@pD9EE2636.dip0.t-ipconnect.de> has quit IRC (Ping timeout: 268 seconds) | 17:25 | |
*** cvp_ <cvp_!~cvp@pD9EE2C3E.dip0.t-ipconnect.de> has quit IRC (Remote host closed the connection) | 17:31 | |
neerad | Okay after making my own base error is different http://paste.ubuntu.com/24028499/ | 18:13 |
neerad | Art is failing to start | 18:13 |
neerad | Abort message: 'art/runtime/runtime.cc:932] LoadNativeLibrary failed for "libjavacore.so": dlopen failed: cannot locate symbol "__android_log_close" referenced by "libjavacore.so"...' | 18:15 |
saidinesh5 | that one is because of version mismatch | 19:04 |
saidinesh5 | iirc i had to patch some cpp file for a similar error while working on Sailfish OS | 19:04 |
neerad | weird?. the base and sfdroid had the same version | 19:05 |
saidinesh5 | not sure but i remember specifically patching the logging library because of a missing symbol | 19:06 |
neerad | the old base hadn't the same version but still didn't show up the error | 19:07 |
saidinesh5 | see if your sources have this method: https://github.com/WaRP7/android_system_core/blob/master/liblog/logger_write.c#L138 | 19:09 |
*** neerad <neerad!~neerad@112.133.229.182> has quit IRC (Ping timeout: 260 seconds) | 19:12 | |
saidinesh5 | https://github.com/sfdroid/android_system_core/blob/hammerhead_12/liblog/logd_write.c doesnt have tha tmethod | 19:18 |
*** neerad <neerad!~neerad@112.133.229.182> has joined #sfdroid | 19:23 | |
saidinesh5 | https://github.com/sfdroid/android_system_core/blob/hammerhead_12/liblog/logd_write.c doesnt have tha tmethod | 19:31 |
saidinesh5 | neerad: ^ | 19:32 |
saidinesh5 | [00:39:59] <saidinesh5> see if your sources have this method: https://github.com/WaRP7/android_system_core/blob/master/liblog/logger_write.c#L138 | 19:32 |
saidinesh5 | hmm.. bunch of hacks trying to trick sfdroid into thinking UID 100000 is root .. i dont think i need them either | 20:00 |
saidinesh5 | right now it seems the segfaults are coming from sfdroid looking for /usr/blah blah inside chroot | 20:01 |
saidinesh5 | need to weed them out | 20:01 |
saidinesh5 | also.. I/strace (19553): openat(AT_FDCWD, "/system/lib/hw/gralloc.default.so", O_RDONLY) = 12 | 20:05 |
saidinesh5 | it should open : /system/lib/hw/gralloc.msm8974.so right? | 20:05 |
saidinesh5 | immediately after that : I/strace (19553): writev(12, [{"\4", 1}, {"pL", 2}, {"C\367\251X\237Q\3314", 8}, {"\7", 1}, {"libc\0", 5}, {"Fatal signal 11 (SIGSEGV), code "..., 80}], 6) = 97 | 20:06 |
saidinesh5 | after that: F/dex2oat (19710): art/compiler/driver/compiler_driver.cc:761] Failed to resolve class Landroid/content/ActivityNotFoundException; . I think this is because of missing environment variables | 20:06 |
neerad | hmmm | 20:07 |
saidinesh5 | missing/bad actually | 20:07 |
saidinesh5 | neerad: i am beginning to think chroot is the way to go now. especially because if i wasnt trying this stuff in chroot, id have to put up with a dozen reboots and overheating already | 20:08 |
*** pseudodev <pseudodev!uid205973@gateway/web/irccloud.com/x-wgrjtkjrynwrcyah> has quit IRC (Quit: Connection closed for inactivity) | 20:09 | |
kimmoli | :) | 20:12 |
neerad | saidinesh5 cool man | 20:12 |
saidinesh5 | also neerad ; <saidinesh5> hmm.. bunch of hacks trying to trick sfdroid into thinking UID 100000 is root .. i dont think i need them either might apply to you too as droid-hal-init is running as user root | 20:22 |
neerad | I am not running as chroot saidinesh5 | 20:23 |
saidinesh5 | i know | 20:23 |
neerad | saidinesh5 How to run it as chroot ? | 20:24 |
saidinesh5 | 1 sec. trying to find a line where it was checking if uid == 100000 to return a package | 20:24 |
neerad | yeah it is the latest commit | 20:24 |
saidinesh5 | https://github.com/sfdroid/android_frameworks_base/commit/2e84686fbb42c77a07684a59cf5db2733d073244#diff-82acdfc3b7bf9f8a160cc01c37be2d8eR977 | 20:25 |
saidinesh5 | this one | 20:25 |
neerad | yep | 20:26 |
saidinesh5 | and as for how to run it in chroot, i havent still gotten it to run, so it maybe pointless for you to try. but basically the idea is pack the whole rootfs into sfdroid_core and patch the init to not mount sys, dev, procfs | 20:27 |
saidinesh5 | and use a stripped down init.rc .. similar to init.sfdroid.rc | 20:27 |
saidinesh5 | that's what i am doing | 20:27 |
neerad | okay saidinesh5 .... keep up | 20:29 |
saidinesh5 | https://paste.kde.org/ppdwrb8bs | 20:29 |
saidinesh5 | start-sfdroid.sh | 20:29 |
saidinesh5 | init.rc : https://paste.kde.org/pt9y6wurq | 20:30 |
neerad | Does the error is reduced when running chroot than non-chroot saidinesh5 | 20:30 |
saidinesh5 | honestly no clue. when not running in chroot i couldnt even get proper error logs becuase the services were started up along with the rest of the system, so had to search for them in logcat manually | 20:31 |
saidinesh5 | now i am just staring at the logcat only when i run start-sfdroid.sh | 20:31 |
neerad | hmmm That's a major problem in debugging an issue | 20:31 |
saidinesh5 | i know. it was a bigger problem for me because this is my primary device and the error logs were littered with stuff from my android services | 20:32 |
neerad | saidinesh5 Youcould try it in an emulator | 20:32 |
saidinesh5 | that's what nh1402[m] was trying/planning to | 20:33 |
neerad | Oh okay, and what was the result? | 20:33 |
saidinesh5 | i mean using the emulator image.. well he couldnt build the emmulator for arm achitecture iirc | 20:33 |
neerad | So had the latest commit have any effect on the error logs? | 20:34 |
saidinesh5 | building the image right now. it may take some time | 20:35 |
neerad | See how it goes | 20:35 |
saidinesh5 | did you patch bionic btw? | 20:40 |
neerad | Do i have to? | 20:40 |
neerad | No i didn't | 20:40 |
neerad | sfdroid repos don't contain any | 20:40 |
saidinesh5 | https://github.com/sfdroid/android_frameworks_base/commit/760c7312ef2e549a99a9354322cd84b2141a7962 | 20:41 |
saidinesh5 | https://github.com/sfdroid/android_bionic/commit/ed75d3837bca0df3d7086ca38bb4818f0d37b73d seems to implement that stuff | 20:41 |
saidinesh5 | i am not sure how important this AlarmManagerService is for Android though. | 20:42 |
saidinesh5 | so make sure your bionic provides that stuff | 20:43 |
neerad | saidinesh5 So i have to include each and ever repo in sfdroid | 20:43 |
neerad | android_system_core too ? that too is in sfdroid repo | 20:43 |
saidinesh5 | unlikely .. but basically see if your the commits in the repos you picked have everything available to them | 20:43 |
saidinesh5 | the only commits in android_system_core seem to be for shutting down some logging: https://github.com/sfdroid/android_system_core/commit/8a6286d15509b17c164c997d22d0c2fa74b158d4 | 20:44 |
neerad | hmmm.... okay I will have to see the individual repos of sfdroid then to see if the patches are neccessary or not | 20:44 |
saidinesh5 | (because udev emits messages which android doesnt need) | 20:44 |
neerad | thnx bro saidinesh5 | 20:46 |
neerad | got it | 20:46 |
krnlyng | saidinesh5, the alarmmanager patch is essential | 20:48 |
krnlyng | saidinesh5, but the android_bionic patch is only for mako cm10! | 20:48 |
saidinesh5 | krnlyng: but does bionic need any patche ? | 20:48 |
saidinesh5 | Ahh | 20:48 |
neerad | krnlyng So android_system_core is neccessity or not? | 20:48 |
neerad | patched one I mean | 20:48 |
neerad | the one in sfdroid repos | 20:48 |
krnlyng | neerad, no it's not really necessary | 20:49 |
krnlyng | neerad, it just contains one patch which silences some log message | 20:49 |
saidinesh5 | udev related logmessages | 20:49 |
neerad | What are neccessary repos list them please | 20:49 |
krnlyng | mal, we could also make your yog7 branches the default now i think | 20:50 |
krnlyng | neerad, do you have the build script? | 20:51 |
krnlyng | neerad, this contains the most accurate list | 20:51 |
neerad | yes but it doesn't contain bionic patches as mentioned | 20:51 |
neerad | I used saidinesh5 build script and kimmoli | 20:51 |
saidinesh5 | btw. /system/lib/hw/gralloc.msm8974.so seems to be missing in my final build | 20:51 |
saidinesh5 | that seems important | 20:52 |
krnlyng | saidinesh5, you need extrat-files.sh | 20:52 |
krnlyng | extract* | 20:52 |
saidinesh5 | krnlyng: yeah i used TheMuppets repo | 20:52 |
krnlyng | ok | 20:52 |
saidinesh5 | there are other lights.msm8974 files in there | 20:52 |
saidinesh5 | so it wont work with gralloc.default.so right? | 20:52 |
krnlyng | saidinesh5, you don't copy gralloc anyways | 20:53 |
krnlyng | but i am not sure why it is missing | 20:53 |
krnlyng | saidinesh5, in the non-chroot/non-lxc approach i mean | 20:53 |
saidinesh5 | it is in my device's system partition | 20:53 |
krnlyng | there it is essential ;) | 20:53 |
saidinesh5 | but not sure why it is missing in my build | 20:53 |
saidinesh5 | heh | 20:53 |
krnlyng | in sfdroid we use the one from /system | 20:53 |
krnlyng | as does sfos | 20:53 |
saidinesh5 | yeah .. right now i am bind mounting /vendor into the chroot .. but this library is in /system , which i am providing with the chroot itself | 20:54 |
saidinesh5 | (eventually i'd like to use overlayfs or something so all the unpatched libraries come directly from /system) | 20:54 |
krnlyng | saidinesh5, you could copy it in your install script... | 20:55 |
saidinesh5 | not sure where it is coming from actually... | 20:55 |
saidinesh5 | it isnt in themuppets repo i am using | 20:55 |
krnlyng | which device was it again? | 20:56 |
saidinesh5 | Xiaomi Mi3 (cancro) | 20:56 |
neerad | Can I use android_bionic from sfdroid repos, it has nexus4 as it's default branch or patch files individually | 20:57 |
saidinesh5 | no gralloc.msm8974.so in https://github.com/CyanogenMod/android_device_xiaomi_cancro/blob/stable/cm-12.1-YOG7D/device-proprietary-files.txt | 20:58 |
saidinesh5 | neerad: you dont have to use bionic in your build | 20:59 |
saidinesh5 | it is for SFDroid based on CM 10 for Nexus 4 apparently | 20:59 |
neerad | What about the patches? | 20:59 |
krnlyng | neerad, you don't need them, they are already in upstream | 20:59 |
saidinesh5 | there is only 1 patch. which adds timerfd related functions to it | 20:59 |
saidinesh5 | it is available in CM12.1's bionic | 20:59 |
neerad | Still sfdroid fails | 21:00 |
neerad | it boots up but that is all it is | 21:00 |
saidinesh5 | for some other reason.. | 21:00 |
neerad | No welcome screen nothing | 21:00 |
krnlyng | saidinesh5, i think gralloc should actually be built by the android build system. did you build the whole android? | 21:01 |
saidinesh5 | krnlyng: yep i built the whole android. all i get is gralloc.default.so | 21:01 |
saidinesh5 | after loading which surfaceflinger seems to segfault | 21:02 |
saidinesh5 | although let me check if my makefile missed out some cyanogenmod bits | 21:02 |
saidinesh5 | ah it is part of hardware/qcom/display.. | 21:04 |
saidinesh5 | yup it was missing | 21:05 |
mal | krnlyng: good idea | 21:08 |
*** taaem <taaem!~taaem@unaffiliated/taaem> has quit IRC (Ping timeout: 276 seconds) | 21:12 | |
saidinesh5 | Although now i think i should be looking at/for any changes that bionic needs to actually play well with the changes needed by sailfishos | 21:29 |
*** taaem <taaem!~taaem@unaffiliated/taaem> has joined #sfdroid | 21:30 | |
saidinesh5 | interesting: I/strace ( 2303): E/sharebuffer( 2328): error connecting to renderer: No such file or directory | 21:40 |
saidinesh5 | woah | 21:40 |
saidinesh5 | system 2328 0.9 0.4 34780 8448 pts/2 Sl 03:09 0:00 \_ /system/bin/surfaceflinger | 21:40 |
saidinesh5 | didnt see that coming | 21:40 |
neerad | reboot it I had the same error and it was gone after restart | 21:41 |
saidinesh5 | netd, vold, lmkd, healthd, surfaceflinger are running now | 21:41 |
saidinesh5 | neerad: that error is because sfdroid renderer wasnt running | 21:41 |
saidinesh5 | i wasnt actually expecting surfaceflinger to start. hence the woah :P | 21:41 |
neerad | saidinesh5 yes but in my case i hadn't rebooted it after installing it | 21:41 |
saidinesh5 | although the init is segfaulting now. so zygote etc.. didnt start | 21:42 |
neerad | ps aux | grep zygote returns nothing ? | 21:42 |
saidinesh5 | unlikely : Abort message: 'art/runtime/thread.cc:1143] No pending exception expected: java.lang.ClassNotFoundException: Didn't find class "com.android.commands.am.Am" on path: DexPathList[[],nativeLibraryDirectories=[/system/lib, /vendor/lib]]' | 21:42 |
saidinesh5 | i probably missed packing some libraries | 21:44 |
saidinesh5 | will have to check | 21:44 |
saidinesh5 | init itself was segfaulting .. so need to stop that | 21:46 |
saidinesh5 | that could be causing an issue | 21:46 |
saidinesh5 | now need to figure out why init, mediaserver,zygote, keystore are crashing | 21:49 |
*** krnlyng_ <krnlyng_!~liar@77.116.6.239.wireless.dyn.drei.com> has joined #sfdroid | 22:28 | |
*** krnlyng <krnlyng!~liar@77.117.103.246.wireless.dyn.drei.com> has quit IRC (Ping timeout: 260 seconds) | 22:31 | |
saidinesh5 | I also should figure out how to let sfdroid_renderer send commands via. am | 22:33 |
saidinesh5 | especially since sfdroid disables adb | 22:37 |
krnlyng_ | saidinesh5, you don't need adb for am | 22:37 |
saidinesh5 | although i think it would be easiest if we enabled adb inside the chroot | 22:37 |
saidinesh5 | krnlyng_: basically how do i make sure am something is run within the same chroot as the chroot which sfdroid is running in? | 22:38 |
saidinesh5 | https://github.com/sfdroid/sfdroid_renderer/blob/21b96b2c45dcf2c837056f33b7407e7d65a4369e/sparse/usr/bin/am | 22:38 |
saidinesh5 | wrt this | 22:38 |
saidinesh5 | okay ouch. sailfish os display was eaten away .. i think by surfaceflinger | 22:40 |
krnlyng_ | do you not have the sharebuffer patches? | 22:40 |
saidinesh5 | i do | 22:41 |
krnlyng_ | in surfaceflinger too? | 22:41 |
saidinesh5 | https://paste.kde.org/pjjur6fzx are my running processes btw. | 22:41 |
saidinesh5 | yep. | 22:41 |
* saidinesh5 checks | 22:41 | |
saidinesh5 | $ag sharebuffer | 22:43 |
saidinesh5 | services/surfaceflinger/DisplayHardware/HWComposer.h | 22:43 |
saidinesh5 | and services/surfaceflinger/DisplayHardware/HWComposer.cpp | 22:43 |
saidinesh5 | when i kill the init, my display unblanks | 22:44 |
saidinesh5 | well when i kill the gdbserver running the init | 22:44 |
saidinesh5 | krnlyng_: any idea how i should compile init with debug flags? | 22:45 |
saidinesh5 | as of now my gdb backtrace is totally worthless in debugging init | 22:45 |
saidinesh5 | Also: W/sharebuffer( 6928): connecting to renderer | 22:46 |
saidinesh5 | E/sharebuffer( 6928): error connecting to renderer: No such file or directory | 22:46 |
saidinesh5 | V/SurfaceFlinger( 6928): FPS: 1 | 22:46 |
saidinesh5 | And | 22:46 |
saidinesh5 | the whole logcat: https://paste.kde.org/pdb25gqsk | 22:47 |
neerad | Your phone seems to be under heavy load might me heating up | 22:47 |
saidinesh5 | W/Environment( 6926): EXTERNAL_STORAGE undefined; falling back to default | 22:47 |
saidinesh5 | D/AndroidRuntime( 6926): Shutting down VM | 22:47 |
saidinesh5 | where did you get that from neerad? | 22:47 |
neerad | FPS : 1 | 22:47 |
saidinesh5 | Ah lol | 22:47 |
saidinesh5 | no | 22:47 |
saidinesh5 | that seems unrelated | 22:47 |
saidinesh5 | probably because sfdroid_renderer didnt connect to it | 22:48 |
saidinesh5 | once i edited the am to point to proper classpath: D/AndroidRuntime( 8107): Calling main entry com.android.commands.am.Am | 22:51 |
saidinesh5 | I/art ( 8107): System.exit called, status: 1 | 22:51 |
saidinesh5 | I/AndroidRuntime( 8107): VM exiting with result code 1. | 22:51 |
saidinesh5 | although that is understandable as zygote isnt running | 22:56 |
neerad | Zygote is crashing or not there at all? | 22:56 |
saidinesh5 | not sure actually | 22:57 |
saidinesh5 | init crashed | 22:57 |
saidinesh5 | so .. | 22:57 |
neerad | Yeah zygote was not there at all | 22:57 |
saidinesh5 | could be | 22:57 |
saidinesh5 | init crashed after starting https://paste.kde.org/pjjur6fzx these processes for sure | 22:57 |
*** Mister_Magister <Mister_Magister!~foidbgen@user-94-254-145-70.play-internet.pl> has quit IRC (Quit: Konversation terminated!) | 23:03 | |
saidinesh5 | hmm.. | 23:39 |
saidinesh5 | read(8, "nodev\tsysfs\nnodev\trootfs\nnodev\tb"..., 1024) = 315 | 23:39 |
saidinesh5 | read(8, "", 1024) = 0 | 23:39 |
saidinesh5 | close(8) = 0 | 23:39 |
saidinesh5 | --- SIGSEGV {si_signo=SIGSEGV, si_code=SEGV_ACCERR, si_addr=0xb6f19000} --- | 23:39 |
saidinesh5 | +++ killed by SIGSEGV +++ | 23:39 |
saidinesh5 | ./start-sfdroid.sh: line 51: 15238 Segmentation fault chroot $SFDROID_ROOTFS_PATH /system/xbin/strace /init | 23:39 |
saidinesh5 | strace doesnt seem that helpful: https://paste.kde.org/p5cgxj3pj | 23:43 |
saidinesh5 | interesting: https://www.youtube.com/watch?v=vyBJW3VGZ4A https://pivotce.com/2014/02/17/android-chroot/ | 23:51 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!