Monday, 2018-08-27

guhlmal would you be willing to explain the steps for building for hybris-15.1 (again)08:05
guhlwell as far as I see ghosalmartin rebased the android repos from mer-hybris yesterday so I will switch to his repos08:21
guhlmal, for example when I try to make hybris-hal I run into08:37
guhlninja: error: '/home/guhl/mer/chiron.los15.1/droid/out/host/common/obj/JAVA_LIBRARIES/apksig_intermediates/classes.jar', needed by '/home/guhl/mer/chiron.los15.1/droid/out/host/common/obj/JAVA_LIBRARIES/signapk_intermediates/classes-full-debug.jar', missing and no known rule to make it08:38
guhlI could add platform/tools/apksig in the local manifest but did you need that?08:39
guhlwell i have to add all these repos to my local_manifest to be able to start a make hybris-hal09:43
guhland then it fails with09:44
guhlsystem/core/libpixelflinger/codeflinger/CodeCache.cpp:64:10: fatal error: '../../../../external/dlmalloc/malloc.c' file not found09:44
guhlso i will also add <project path="external/dlmalloc" name="platform/external/dlmalloc" groups="pdk" remote="aosp" />09:45
guhlrunning into10:05
guhlout/target/product/chiron/obj/KERNEL_OBJ/usr/include/sound/asound.h:441:18: error: field has incomplete type 'struct timespec'10:05
guhl--> will #include <time.h> in asound.h to fix that10:06
guhlhmm that should be fixed already in asound.h10:14
ghosalmartinguhl, am running into same issue tbh. feels like were missing a dep like rsync or unzip11:58
ghosalmartinfunnily enough for the Pixel XL hybris-hal built fine11:59
ghosalmartinabranson, also any idea why this is failing? it was building fine a few months ago here
guhlghosalmartin, atm i am back at the problem with assemble_vintf12:03
ghosalmartinguhl, what was the issue there?12:03
guhlthis somehow comes from /home/guhl/mer/chiron.los15.1/droid/build/target/board/Android.mk12:04
ghosalmartinhmmm are you using my repos or fresh from LOS?12:06
guhlnow i am using your repos12:06
ghosalmartinmine might need rebasing tbh. i can see them being somewhat out of date with LOS12:07
guhlbut i had the same using the original ones from the hybris manifest12:07
ghosalmartinahh okays12:07
guhli thought you rebased them yesterday12:07
ghosalmartinnot the device and kernel. that was just the mer-hybris ones12:08
guhlthe KERNEL_CONFIG_DATA variable seems to be empty as far is I can see12:08
guhldevice i am using from LOS, kernel from you12:09
ghosalmartinlemme sync the repos your suggested nad see12:09
ghosalmartinam on works internet to should be a breeze :P12:10
ghosalmartina thing I noticed is that the Chiron boots way faster than the Marlin. Probably due to the kernel updates12:12
guhli can boot the hybris-boot.img (via fastboot) and telnet into the device12:13
guhlusing the current init-script from mer-hybris/hybris-boot (master)12:14
ghosalmartinguhl, yeah they added configfs support. can you test something and enable the CONFIG_VT flag and see if the kernel just corrupts please12:16
ghosalmartinif i try and continue init it reboots and fails12:16
ghosalmartinwhich is normal am guessing since theres no rootfs12:16
ghosalmartini just want to eliminate my machine from the equation12:17
guhl(that's what I had in my LOS 14 config12:23
ghosalmartini tried all the combinations12:26
ghosalmartinthe minute CONFIG_VT was enabled for me, the entire kernel was corrupt and would instantly reboot12:26
ghosalmartini have the same issue for the Pixel XL atm12:27
ghosalmartinguhl, had this before? ninja: error: '/home/mgh01/hybris-15.1/out/target/product/chiron/system/usr/share/vim/autoload/spacehi.vim', needed by '/home/mgh01/hybris-15.1/out/target/product/chiron/obj/NOTICE.xml', missing and no known rule to make it12:29
guhlyes, and i fixed the very crude12:30
ghosalmartinwas it the include?12:30
guhlby copying the /system/usr/share/vim directory from my LOS build12:30
ghosalmartinahh :P12:31
guhli did not figure out how to make the vim runtime files install12:31
guhlmake vim will build vim but will not install the runtime files12:31
guhlsometimes is really would like to understand the current android build system a bit better12:32
ghosalmartinis it even documented anywhere?12:32
guhlprobably :P12:34
ghosalmartincan I just build it locally and copy the files?12:35
ghosalmartinhmm if it built fine for the pixel. its probably some dep declared under /device that we dont need12:35
guhlfor sure but the device files are already so complicated (in my opinion)12:37
ghosalmartinthey always are, a quick grep for vim yields nothing12:37
ghosalmartinthe nexus series used to have tidy device files12:37
guhlif you don't want to mess with the system/usr/share/vim directory for now I can upload it somewhere for you12:44
ghosalmartinif you dont mind12:45
guhlthat is the vim directory12:48
ghosalmartinta :)12:48
ghosalmartinturns out you cant run 2 builds at the same time with the same codebase :P12:49
guhlwtf it took forever to rebuild the kernel/bootimage and now when I boot it it still does not have _VT set ?!12:53
ghosalmartinguhl, yeah it takes forever with CONFIG_VT and interesting12:54
guhlno the thing is the make hybris-boot took forever and now I booted the result with fastboot12:55
guhland can telnet into it12:55
ghosalmartinbut it CONFIG_VT seT?12:55
guhland cat /proc/config.gz | gunzip | grep CONFIG_VT says12:55
r0kk3rzmark of the beast12:55
guhlCONFIG_VT is not set12:56
guhlwhat did it do all the time12:56
ghosalmartindoes anyone know what we need CONFIG_VT for? it says virtual consoles...but which :P12:56
ghosalmartinI thought itd be for telnet but clearly not12:57
guhlwell i cleaned the KERNEL_OBJ directory and will build again12:58
ghosalmartinam curious if its just legacy thats its in there now12:59
ghosalmartinmy chiron build atm 6933/54470 not sure how much it needs to build :P13:00
ghosalmartinthe marlin was only 1500013:00
guhlahh stupid! thats what happened: the kernel i booted (with VT enabled) immediately reboots and it boots again with the kernel i flashed to /boot13:08
guhlso enabling VT makes the kernel reboot13:09
ghosalmartinahh yeah13:09
guhlI could go back the stony to debug this in kernel13:12
guhlstony (aka hard) way13:12
kimmolinearby bup13:12
guhlbut now i got to see a doctor - later13:12
ghosalmartina pint a day keeps the dr at bay ;)13:13
ghosalmartinmal: if a cm14 devices focus doesnt work. am i missing some config?13:29
echo_salikhi all...14:04
guhlghosalmartin, that's the problem living a metric country :-)14:31
ghosalmartinwhat the uk? :P14:32
ghosalmartinguhl, btw how long did you hybris-hal take? mines been going since you left14:34
ghosalmartinand its only 64% done :P14:34
r0kk3rzuk? metric? since when?14:38
guhlno here (austria). No pints so i have to see a doctor.14:39
guhlghosalmartin, yes i think first build without ccache took half a day14:39
ghosalmartinoh damn14:39
ghosalmartinwell atleast i have a zip built for marlin :D14:40
guhlbut i am building on a 2012 mobile i714:40
ghosalmartintime to see what happens14:40
ghosalmartini think am on a 2015 mobile i7 :P14:40
ghosalmartinaside from dhi failing am in :D14:42
ghosalmartinguhl, ive finally hit your error  :P14:43
ghosalmartin /bin/bash -c "POLICYVERS=30 BOARD_SEPOLICY_VERS=27.0 FRAMEWORK_VBMETA_VERSION=0.0 /home/mgh01/hybris-15.1/out/host/linux-x86/bin/assemble_vintf -i hardware/interfaces/compatibility_matrix.27.xml -o /home/mgh01/hybris-15.1/out/target/product/chiron/gen/ETC/system_compatibility_matrix.xml_intermediates/compatibility_matrix.xml  -c \"/home/mgh01/hybris-15.1/out/target/product/chiron/obj/ETC/manifest.xml_intermediates/manifest.xml\"14:43
ghosalmartin--kernel=3.18:  --kernel=4.4:  --kernel=4.9:"14:43
malghosalmartin: not sure what you meant by focus on your previous question14:44
ghosalmartinmal camera focus14:44
ghosalmartinive ran the appropriate scripts but nada14:44
ghosalmartinusb_moded is causing issues :P14:45
malghosalmartin: usb-moded changes got merged today, requires also updated buteo-mtp14:48
ghosalmartinmal: i take it, a simple mask will stop it?14:48
malghosalmartin: also usb-moded build also requires new rpm version or manual editing of spec14:49
ghosalmartinhmm disabling usb-moded broke my telnet :P14:49
ghosalmartinmal: how would surfaceflinger be started?15:11
ghosalmartindo I need to add a task to manually start it?15:11
malghosalmartin: this and the related stuff which should be easily found, and the droid-hal-device.conf changes of course15:13
ghosalmartini have the first and last thing15:13
ghosalmartinnot sure about related stuff :P am digging in the sony-nile repos, that sound about right?15:13
malghosalmartin: the script that uses and then the symlink to start the service15:13
ghosalmartini feel like am cheating by skipping stable usb to go straight to ui :P15:14
malghosalmartin: did that service start on the device?15:14
ghosalmartinprobably won't anyways, my droid-hal-init was failing15:15
malhow did it fail?15:15
ghosalmartinso i better fix the usb issue first15:15
maldhi is indeed the first thing to fix15:15
malghosalmartin: did you remember to add the stuff I mentioned yesterday?15:15
malto config repo15:16
ghosalmartinmal: yeahh thats all added in15:16
malghosalmartin: could you push the repo to github so I can check15:16
ghosalmartini need to add the systemd task still15:17
ghosalmartinand the conf file15:17
ghosalmartinbut i was just going to do it on device to test15:17
ghosalmartini take it this is failing due to older rpm
malghosalmartin: is detritus really that small?15:18
ghosalmartinseems like it15:18
guhlghosalmartin, i did the next nasty hack to come by that error15:19
guhlln -s /home/guhl/ws_and/chiron/test/vts-testcase/kernel test/vts-testcase/kernel15:19
ghosalmartinguhl, so are we missing another repo it requires?15:20
guhlnow i have a compatibility_matrix.xml in /home/guhl/mer/chiron.los15.1/droid/out/target/product/chiron/obj/ETC/system_compatibility_matrix.xml_intermediates15:20
guhland it goes on15:21
ghosalmartin   <project path="test/vts-testcase/kernel" name="platform/test/vts-testcase/kernel" groups="vts,pdk" />15:21
ghosalmartinam curious whats going to happen since I dont have the cache enabled :P not enough disc space15:23
ghosalmartinits started from scratch...15:24
guhlwell that gives you enough time for that pint :P15:25
ghosalmartinmaybe :P15:25
ghosalmartini guess i can crack on with the marlin build then :P15:25
ghosalmartinand a fresh bullhead build15:25
guhlmight get warm but i heard you like it that way15:25
ghosalmartinright now a cold pint would be welcome :P15:26
ghosalmartinmal: if i want to build the updated usb-modded on obs, will it require someone updating the rpm? or can I do it?15:31
vknechtis it a common problem that UI doesn't show up eg. on incoming call, alarm...15:35
* vknecht suspects proximity-sensor flip-flop15:36
ghosalmartini wonder have there been some fixes in mce since recently on my bullhead the sensors have been very consistent15:36
* vknecht doesn't know exactly, and there was that sensor CSD config change recently15:40
malghosalmartin: I think it needs to be updated in the lower level15:42
ghosalmartinmal: damn guess ill have to do it manually15:42
ghosalmartinactually if I just remove the Recommends should be fine15:44
malyes, that is enough, and the buteo-mtp dependency needs to be built15:46
ghosalmartinthats done15:46
ghosalmartinhmm I cant fork it15:46
ghosalmartinah reached my project limit :P15:46
malghosalmartin: gitlab has quite low project limit15:48
ghosalmartinah tis fine their old projects anywya15:48
echo_salikhi mal15:53
echo_salikdid you get that HDD issue resolved by the shop?15:53
malyes, got a new disk without issues15:54
echo_salikany data loss?15:56
malno because it was a raid5 setup on a NAS so it fixes itself automatically15:57
echo_salikcool... raid 5 is the one with parity right?15:58
ghosalmartinmal: how do I disable telnet access?15:59
kimmoliisn't it disabled when you get from devel: to testing: (?)16:00
malkimmoli: yes, it is16:01
malecho_salik: yes16:01
ghosalmartinahh i think its done it automatically :D awesome :)16:01
ghosalmartinguhl, was this the next problem? /bin/bash -c "(python /home/mgh01/hybris-15.1/out/host/linux-x86/bin/ -l /home/mgh01/hybris-15.1/out/host/linux-x86/lib64 -f /home/mgh01/hybris-15.1/out/target/product/chiron/obj/ETC/plat_file_contexts_intermediates/plat_file_contexts -f /home/mgh01/hybris-15.1/out/target/product/chiron/obj/ETC/nonplat_file_contexts_intermediates/nonplat_file_contexts -p /home/mgh01/hybris-15.1/out/target16:05
ghosalmartin/product/chiron/obj/ETC/sepolicy_intermediates/sepolicy ) && (touch /home/mgh01/hybris-15.1/out/target/product/chiron/obj/ETC/sepolicy_tests_intermediates/sepolicy_tests )"16:05
echo_salikmal: any idea regarding my build?16:13
malecho_salik: can't remember anymore what the situation was16:14
echo_saliklol ok... new logs?16:14
echo_salikservicemanger wasa failing btw16:14
r0kk3rzthats a lot of libc errors :/16:23
malyes, the binder errors are really bad16:23
r0kk3rzis this a treble device?16:29
echo_salikyeah i think so16:31
echo_salikvendor is partitopned16:32
ghosalmartinvendor has been partitioned since android 616:32
ghosalmartinwas the device released with android 8?16:33
r0kk3rzi just saw some records there looking for an odm partition16:33
echo_salikno it was a marshmellow device that has got an official nougat and oreo (beta) update16:33
ghosalmartinso probably not trebled. but ya never know the porters might of taken the time16:34
malghosalmartin: wondering if the device has odm parition16:35
echo_saliki ahve seen treble support for my device though16:35
echo_salikhow can i check that?16:35
malor oem partition16:35
malls -lR /dev/block16:36
echo_salikfrom recovery i think?16:36
ghosalmartinisnt odm more of a sony thing not a treble thing?16:36
malghosalmartin: not sure how many devices have oem/odm partition16:37
echo_salikmal: lrwxrwxrwx 1 root root 22 Jun 13  1971 oem -> ../../../../mmcblk0p3016:37
echo_saliki got this in telnet16:37
malwondering if that should be mounted to /odm or not, sony has those mounted16:40
r0kk3rztheres also this hidl_memory service too16:40
malecho_salik: it would be interesting to see what is in that oem partition16:41
echo_salikshould i mount it?16:41
wiktorek140sledges: ping16:41
malecho_salik: I would try it16:42
echo_salikmount: /oem: unknown filesystem type.16:44
r0kk3rzecho_salik: if its originally a marshmellow device, why are you using 15.1?16:44
r0kk3rzis there no LOS 14 for it?16:44
echo_saliki havent seen a stable los14 with sources16:45
echo_saliki was using oero on this device before this16:45
echo_salikan i have seen more sources for oreo then nougat16:45
malecho_salik: ok, then the issue might be something else16:45
echo_salikbecause by the time this device's kernel was out, the device had official nougat...16:46
echo_salikso i guess porters preferred to work on oreo?16:46
r0kk3rzyeah makes sense, android people tend to want the latest android16:47
r0kk3rzyou just might have an easier time with a 14.1 base, if a stable one existed16:47
echo_saliki have seen an (only one) lineage 14 but even the modder has i think abandoned the thread16:48
echo_salikps: can i build a lineage 14.1 rom?16:49
r0kk3rzyou could16:49
guhlghosalmartin, sorry been a bit distracted16:51
ghosalmartinguhl, ah its alright, am juggling devices atm :P bullhead has a nice release that feeds of OBS now16:52
ghosalmartinmarlin has an issue iwth usb storage mode16:52
guhli am running into16:52
guhlbuild /home/guhl/mer/chi..._tests_intermediates/treble_sepolicy_tests16:52
guhlFAILED: /home/guhl/mer/chiron.los15.1/droid/out/target/product/chiron/obj/ETC/treble_sepolicy_tests_intermediates/treble_sepolicy_tests16:52
guhlTraceback (most recent call last):16:52
guhl  File "/home/guhl/mer/chiron.los15.1/droid/out/host/linux-x86/bin/", line 4, in <module>16:53
ghosalmartinkeeps doing this
guhl    import mini_parser16:53
guhlImportError: No module named mini_parser16:53
guhlsorry for pasting16:53
ghosalmartinhow did you provide the module policy?16:53
guhlnot that i would know16:53
guhlmight just be that the parallel execution failed with something different for me first16:53
ghosalmartinyeah could be16:54
ghosalmartinyeah am rocking
echo_saliki do have the original kernel from xiaomi... its based on nougat...16:54
echo_salikwait... would that mean i'd have to redownload all the repos?16:54
echo_salikit took me a week for that... plus this happens...16:55
guhlare these python modules supposed to be installed on the host?16:55
ghosalmartinguhl, i tried that with no luck16:55
ghosalmartinbut then again maybe I just need to find the right one16:55
malecho_salik: if the android version works then sailfish should work on top of it16:56
malchanging bases without a good reason is questionable in my opinion16:56
guhlwell i am using the same ubu-chroot for building the hybris-hal that i used to build LOS 15.116:57
echo_salikmal: i did try out this rom though... if this fails i'd use that16:57
ghosalmartinguhl, yeah should be fine just a sudo apt-get install16:58
guhlit is in system/sepolicy/tests16:59
Mister_Magisterhi guhl! hi mal hi ghosalmartin16:59
malecho_salik: what did you mean by that?16:59
guhlthe Magister! nice seeing you16:59
Mister_Magisterguhl: your thea is doing great under my lead :P16:59
guhlit's doing perfect!17:00
Mister_Magistermy mom is using it :P17:00
ghosalmartinMister_Magister, hey up :)17:00
* Mister_Magister is adding hw-settings.ini to every device17:01
echo_salikmal: sorry... i meant i had thried the lineageOS rom that i am using with sailfish... really liked it... simple... not much clutter17:01
ghosalmartinMister_Magister, whats that do?17:01
Mister_Magisterghosalmartin: fixes brightness settings and sensors17:01
Mister_Magistersee in hadk-faq17:02
guhlghosalmartin, the question is why it does not find the modules there17:02
ghosalmartinguhl, ...missing another repo? :P17:02
* Mister_Magister hopes he won't blew up merobs some day by building all devices at once17:18
Mister_Magistermal: how ya feelin today?17:22
ghosalmartinmal: it was this i needed
ghosalmartinwell just the last file edit sorted my problem17:24
ghosalmartinguhl, ^^ may need it as well17:24
ghosalmartinmal: do you know how the /lib/systemd/system mounts get genearted17:26
malghosalmartin: from fstab/*.rc files with help of fixup-mountpoints in rpm/dhd17:30
ghosalmartinmal: ah...dont got one of those :P17:31
ghosalmartinwell not with them all defined17:31
guhlghosalmartin, i worked around the error going into system/sepolicy/tests and building/installing it with mm17:35
ghosalmartinguhl, yay17:36
guhlthat's gonna need some cleanup afterwards17:37
ghosalmartini think it all does :P17:38
guhlback to the asound.h error in the kernel-header files18:05
guhlthat has been fixed in the kernel sources (by including time.h) but not in the kernel-header sources18:06
ghosalmartinama head home and then run that build and see whats what with the kernel sources18:07
ghosalmartinguhl ive moved onto illegal argument exceptions now18:44
Mister_Magisterghosalmartin: were you working on roma chip?19:23
malroma is quite difficult it seems19:26
Mister_Magisteryeah so i just wanted to ask if he got something19:28
* Mister_Magister needs to move g2 and g2lte to bluez519:28
ghosalmartini got knowhere19:30
guhlghosalmartin, i fail here now:
guhlwhere is your illegal argument exceptions now19:30
malI also tried to get rome working but couldn't yet19:31
ghosalmartinso i went inside conscrypt and tried to build with with mm, which needs some libcore etc etc19:31
ghosalmartinam just following the chain up19:31
guhlghosalmartin, i need to stop for today, have a flight to catch tomorrow morning at 519:36
ghosalmartinhopefully somewhere nice :P19:36
guhlif you like frankfurt and two days of finance meetings19:37
ghosalmartinfrankfurt is lovely, meetings not so much :P19:37
guhlhave fun - night!19:38
wiktorek140ghosalmartin: who should i ping to repaternise my repo on obs?20:15
r0kk3rzlbt is the patterniser20:17
wiktorek140lbt: ping20:20
wiktorek140lbt: and if u just read it later, plase repatternise this repo:
malwiktorek140: that already has patterns?20:22
malor did you add those manually?20:22
wiktorek140mal:  yes, but show me that is disabled any package20:22
wiktorek140it was paterrnised but somehow broken20:23
malyou just enable the packages again20:23
wiktorek140and then i try to fix by hand but dont help20:23
malfrom repositories subpage20:23
wiktorek140mal: where if i can ask?20:23
malyou see the target is disabled20:24
malthe build flag part20:24
wiktorek140yes, i see20:24
wiktorek140but how to enable it? i didnt see any switch or something similiar20:24
malso ust toggle that by pressing the icon20:25
malit should show a menu20:25
malor something like that20:25
wiktorek140oh, so text is link but icon not, clever xD20:26
malthe icon is also, it should show a hover-on menu20:27
malwiktorek140: it seems you haven't updated packages in a while, libhybris is many versions behing20:28
wiktorek140mal: probably, just fixing things to up to date20:32
wiktorek140and ngfd is broken, maybe u know how to fix it?20:33
wiktorek140mal: fixed, just renamed20:34
malrenamed what?20:34
wiktorek140ngfd-plugin-droid-vibrator to ngfd-droid-native-vibrator20:36
malbt o didn't change that on OBS20:40
mal*but you20:41
wiktorek140yes, now i change it on obs20:41
wiktorek140mal: but pulseaudio-modules-droid fail20:42
malvery odd error20:47
wiktorek140can u tell hot to set tag on repo in obs?20:48
maljust set the revision in _service20:48
Mister_Magistersome really old problems i see :P20:48
malwiktorek140: try building 11.1.6820:51
wiktorek140mal: work20:58
malwiktorek140: I think you should define this
malands update the submodule of your droid-hal-version-$DEVICE repos20:59
wiktorek140mal i have defined have_vibretor_native21:00
wiktorek140ohh, i see what can by wrong21:01
malyou didn't update submodule21:01
wiktorek140mal: yes, i know, but it not for today ;)21:02
wiktorek140see you later21:02
maldammit, why don't people just do what I say21:02
malthat define is not even there if the submodule is too old21:03
r0kk3rzgod emperor mal demands obedience!21:07
Mister_Magistermal: lol21:31

Generated by 2.17.1 by Marius Gedminas - find it at!