Monday, 2021-02-15

T42<Verevka86> @mal [@Verevka86 to fix the hybris-boot issue on hyb …], it works thanks 🤝04:09
T42<Verevka86> system-as-root required for hybris-17.1?07:46
mal@Verevka86 no, we skip the / mount like this https://github.com/mer-hybris/droid-hal-sony-ganges-pie/blob/master/rpm/droid-hal-common.inc#L32 (do not add /vendor there for you probably)09:19
piggzmorning10:03
T42<edp_17> Hi all, I am building 4.0.1.48 for hammerhead (cm12.1 base) and getting error for libgbinder: https://paste.ubuntu.com/p/R5zPWtVhSR/11:18
T42<edp_17> Is there a workaround? (I have not got this on my Note4-treltexx.)11:18
mal@edp_17 did you update dhd submodule? also did you cleanup the gbinder repo before rebuilding11:24
T42<edp_17> Hi mal, yes I updated all submodules but did not clean up the gbinder repo. I was aware of only removing the hybris/mw/libhybris before start building packages.11:27
T42<edp_17> mal: How can I cleanup the gbinder repo?11:27
malgo to the gbinder source folder and run git clean -x -f -d11:33
T42<edp_17> mal: Thanks, it worked!11:38
T42<Verevka86> Help me please:12:26
T42<Verevka86> https://paste.ubuntu.com/p/2X2yyQtdkz/12:26
T42<Verevka86> @Verevka86 [Help me please:12:36
T42<Verevka86> https://paste.ubuntu.com/p/2X2 …], https://github.com/mer-hybris/droid-hal-device/commit/665f4e284517e8fe2f18f6404f3a6b343556f4e712:36
mal@Verevka86 update dhd submodule and try again13:11
maljust merged a fix for that13:11
kernlingMister_Magister, sorry if you're still on your own time, but did you have time to check the new lh?13:58
Mister_Magisterkernling: will check today… just after work14:17
Mister_Magistergive me roughly 1h14:18
kernlingok no rush14:18
kernlingthanks14:18
Mister_Magisterno, thank you!14:18
T42<edp_17> Hi all. I need a little help. I've installed the build on hammerhead but it stucks on boot logo. (There was no telnet either.) I've managed to get into the device and seen the /system wasn't mounted. I have an error log and the issue (I think) at lines: 691, 692, 693 and 694: https://pastebin.ubuntu.com/p/xNvK9H8CXR/15:36
T42<edp_17> What else I should change in the local repo to make it work?15:37
T42<edp_17> Once I telnet in the device I can manually mount /system. So, I guess, the problem is somewhere in the init routine.15:43
mal@edp_17 when did you last time check your kernel configs with mer kernel check tool?15:43
T42<edp_17> mal: During creation of this build.15:44
malwhich sfos version did you build?15:44
T42<edp_17> I have updated the mer-kernel-check and used the fresh one to see what's missing.15:44
T42<edp_17> 4.0.1.4815:44
malyou probably the first one to test that with such old kernel15:44
T42<edp_17> mal: Oh, yes, the mentioned limitation! Below kernel version 3.10 is not supported.15:46
malmight need some investigation what is needed to get it working15:47
T42<edp_17> That explains why I was able to build 4.0 for Note 4 (treltexx) which has kernel version 3.10.x15:47
T42<edp_17> There is another issue, when I plug the device in to the PC I have about 20 seconds to issue commands before it terminates the connection. The device doesn't get rebot just it loses the connection. In that case I need to unplug and plug back in then assign an IP and bring the USB network up manually. (I wrote a script for it, so that's15:49
T42 quick.) On this, even the 'touch /init_enter_debug2' doesn't help.15:49
malI need to test 4.0.1 on fp2 at some point15:49
T42<edp_17> mal: What kernel version is there on the fp2?15:50
mal3.415:50
T42<edp_17> I see.15:50
T42<edp_17> There is another problem. The new firejail doesn't realy work on the Note4. (I've contacted a firejail developer, hopefully he/she can help out with what's missing from the kernel.)15:51
T42<edp_17> As a workaround, I removed the firejail call from the .desktop files but obviously that is not feasible solution on the long term.15:52
T42<Herrie1982> What specifically did you update on 4.x that would break older kernels? We had issues with newer systemd in LuneOS since we use the one provided by Yocto, but after some kernel patches for namespaces we got it to run with even recent systemd15:55
T42<Herrie1982> We have systemd 246 running with 3.4 kernels at our end15:55
mal@Herrie1982 systemd16:07
mal@Herrie1982 show the patches please16:08
T42<Herrie1982> @mal [@Herrie1982 show the patches please], Well it even depends on which version of 3.4 it seems ;)16:09
T42<Herrie1982> There's some clutter in these trees because of GCC fixes, but well: https://github.com/shr-distribution/linux/commits/tenderloin/3.4/cm-12.1-halium16:09
T42<Herrie1982> https://github.com/shr-distribution/linux/commit/b840c1217fad4dd0db7e47aa671c5b0c848381a416:09
T42<Herrie1982> https://github.com/shr-distribution/linux/commit/c34cbea158fdce44085781a8318dcc4d3503239c16:10
T42<Herrie1982> https://github.com/shr-distribution/linux/commit/f311986ee22b148876c327a9d9a29a240a62c5a516:10
T42<Herrie1982> https://github.com/shr-distribution/linux/commit/01141a353241d545a2aaf9addaa32afe39a9856216:10
malyeah, I think I have seen those before when I did anbox hacking long time ago16:10
Mister_Magisterkernlyng: what logs do you require16:11
mal@Herrie1982 that last one is new to me16:11
kernlyngMister_Magister, i assume it means it doesn't work?16:11
Mister_Magisternope16:11
T42<Herrie1982> The vfs_fstat one is not in all versions16:11
Mister_Magistersegfault16:11
T42<Herrie1982> Some newer 3.4 had some changes for that16:11
Mister_Magisterboth debug outputs are nil16:11
kernlyngMister_Magister, can you run it through gdb16:12
Mister_Magistersure16:12
T42<Herrie1982> https://github.com/Halium/android_kernel_lge_hammerhead/commits/halium-5.1 is for Hammerhead, specifically https://github.com/Halium/android_kernel_lge_hammerhead/commit/17a8d0630b424d66981900e701ae2a452d4b8726 is slightly different there16:12
mal@Herrie1982 hmm, could that explain issues I have with 3.4.113 on fp216:12
kernlyngMister_Magister, EGL_PLATFORM=hwcomposer gdb test_hwcomposer16:12
kernlyngetc etc16:12
Mister_Magisteri know i know16:12
Mister_Magisterhttps://paste.opensuse.org/bf420e6516:13
T42<Herrie1982> You should be able to get it to boot with 3.4 kernel really with those 3/4 patches for namespaces and the fstat16:13
T42<Herrie1982> We didn't test it on a whole lot of devices, but LG Hammerhead, Onyx (oneplus) and good old HP Touchpad (tenderloin)16:13
mal@Herrie1982 for some reason I also had to revert "net: Loopback ifindex is constant now"16:14
T42<Herrie1982> We use your kernel checking script at our end, since we also use similar setup, only run Android in LXC like UBPorts, for the rest we share a lot more with Mer, so your defconfig works for us normally ;)16:14
malok, I will test those patches16:15
T42<Herrie1982> @mal [@Herrie1982 for some reason I also had to reve …], Well you have a lot more stuff in SFOS vs us in LuneOS, so that could explain, but for systemd these patches should do really16:17
T42<Herrie1982> The VFS one was the one that was giving us headaches at some point especially due to the different implementations in different 3.4 versions16:18
mal@Herrie1982 I also haven't been able to get UI working yet with that 3.4.113 kernel (that is lineage-17.1 kernel with older config to keep it compatible with old android libs)16:20
T42<Herrie1982> Well Hammerhead/Onyx are on 3.4.0 even, so there probably isn't a very hard requirement for a "newer" 3.4.16:26
T42<Herrie1982> At least not for systemd with the right patches16:26
T42<Verevka86> I try to run hybris-17.1, there is no image on the screen, I can log in via ssh - https://paste.ubuntu.com/p/6FjvRtHrDf/16:29
T42<Verevka86> how do I view the startup logs for debugging?16:30
mal@Herrie1982 the reason for the updated kernel is not systemd but the fact that it has some security fixes and also stability fixes for camera etc16:36
T42<Herrie1982> @mal [@Herrie1982 the reason for the updated kernel …], OK but if it doesn't want to display any UI that's not very helpful ;)16:37
mal@Verevka86 journalctl -b --no-pager and if boot it far enough /usr/libexec/droid-hybris/system/bin/logcat16:38
mal@Herrie1982 needs some more debugging, starting the UI causes device to reboot which is not very nice16:39
T42<Verevka86> @mal [@Verevka86 journalctl -b --no-pager and if boo …], logo screen and LED lit, no reboots16:40
T42<Verevka86> this is all that I have managed to get so far :(17:37
T42<Verevka86> pstore-17:37
T42<Verevka86> (Document) https://irc.thaodan.de/.imgstore/xbe0uDXwnw.bin17:38
kernlingMister_Magister, that's odd17:56
Mister_Magisteryou think?17:58
kernlingMister_Magister, you're installing lh manually right?17:58
Mister_Magisteryup17:59
Mister_Magisterwithout reboot17:59
kernlingMister_Magister, which packages are you installing when you do that?17:59
Mister_Magisterwait17:59
Mister_Magisternow that i look17:59
Mister_Magisteri only sent like 2 packages17:59
Mister_Magisterthats weird17:59
Mister_Magisterwait i did smth wrong, my bad18:00
kernlingok18:01
Mister_Magistertest_hwc DOES NOT SEGFAULT18:02
Mister_Magisterkernling: WOOOOOO18:02
kernlingyay18:02
Mister_Magisteri mean it doesn't display anything either but WOOOOO18:02
kernlingwhat about lipstick?18:03
Mister_Magisterlet me try just rebooting18:03
kernlingok18:03
Mister_MagisterWE GOT GUI BOIS18:03
Mister_Magisterkernling: WOOOOOO18:04
Mister_Magistersuccess18:04
Mister_Magisterat last18:04
Mister_Magisterwe waited entire weekend for this moment18:04
Mister_Magisterkernling: whats next step? u gonna remove debug code, make clean commit to libhybris, its gonna be merged and then i can build libhybris?18:22
kernlingyeah something like that18:25
T42<thisisthethirdshoukosaccount> when running this, rpm/dhd/helpers/build_packages.sh --version20:23
T42<thisisthethirdshoukosaccount>  I get permission denied20:23
T42<thisisthethirdshoukosaccount> PlatformSDK shouko@shouko-virtualbox:~/hadk$ rpm/dhd/helpers/build_packages.sh --version20:23
T42<thisisthethirdshoukosaccount> dirname: missing operand20:23
T42<thisisthethirdshoukosaccount> Try 'dirname --help' for more information.20:23
T42<thisisthethirdshoukosaccount> * Building rpm/droid-hal-version-jasmine_sprout.spec20:23
T42<thisisthethirdshoukosaccount> ./rpm/dhd/helpers/util.sh: line 330: //droid-hal-version-jasmine_sprout.log: Permission denied20:24
T42<thisisthethirdshoukosaccount> !! building o20:24
maldo you have ANDROID_ROOT environmental variable set correctly?20:30
kernlingMister_Magister, https://github.com/libhybris/libhybris/pull/47620:47
Mister_Magisterkernling: <320:47
Mister_Magistershould i mention kernling that HYBRIS_TRACE_HOOKED=1 HYBRIS_TRACE_UNHOOKED=1 HYBRIS_TRACE_DYNHOOKED=1 give bus error still?20:48
kernlingMister_Magister, yeah idk why, it works if we add HYBRIS_LOGGING_LEVEL=debug20:50
kernlingcan you verify that still?20:50
kernlingHYBRIS_LOGGING_LEVEL=debug EGL_PLATFORM=hwcomposer HYBRIS_TRACE_HOOKED=1 HYBRIS_TRACE_UNHOOKED=1 HYBRIS_TRACE_DYNHOOKED=1 test_hwcomposer20:50
kernlingsomething like that^20:50
Mister_Magisteryee seems like working20:53
Mister_Magisteri got wall of text since 15 seconds20:53
kernlinghehe20:53
kernlingok20:53
T42<thisisthethirdshoukosaccount> mal: yes, since I have built other rpms successfully21:02
malMister_Magister: there is now a build a clean version of the changes, please test when you have time21:40
mal@elros34 you can also test the new version, same branch from my repo21:41
Mister_Magistermal: i can now just zypper dup! thats game changer21:48
T42<thisisthethirdshoukosaccount> @thisisthethirdshoukosaccount [when running this, rpm/dhd/helpers/build_packa …], I was trying to build sfos 4.0.121:49
malis the droid-hal-version submodule properly initialized?21:51
Mister_Magistermal: works21:51
T42<elros34> mal: thx, I need to buid it later on obs because when I used local build I had strange results21:51
Mister_Magisteris krnlyng and kernling same person?21:52
krnlyngMister_Magister, kernling is my clone21:52
Mister_Magisterunderstandable, have a nice day21:53
krnlyngit's just a alternate irc nick21:53
Mister_Magisterkrnlyng: i know i just didn't recoginze u22:04

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