Friday, 2019-07-05

mccrearyelros34: Thanks for working on SFOS 3 for moto_msm8960_jbbl. I had a problem using your installer as nadlabak's twrp-3.1.0 doesn't support bzip2. I repackaged the installer ZIP to use gzip instead, and left it here:
mccrearyAlso, the xt897 hardware 'A' touchscreen doesn't work with your kernel, only the hardware 'B' version. I'll try to dig up the relevant kernel patches00:05
T42<elros34> mccreary: try twrp 3.0.2-0, I never have any install problems with that version. For touchscreen pastebin evdev_trace -I05:43
piggzmccreary: morning, where is your fxtec work .... is twrp working well?07:46
piggzdroidmedia down to 16000 targets....lets hope it works!08:02
piggzactually, this is a hybris-hal build...finger crossed for the droidmedia build after08:17
Mister_Magistermal: great droidmedia has problems with code08:21
Mister_Magisterpiggz: false08:21
Mister_Magisterthe thing is08:22
Mister_Magisteryou are not entering the build chroot08:22
Mister_Magisterfrom hadk08:22
Mister_Magisteryou miss parameters like PORT_ARCH08:22
Mister_Magisterpiggz: run make -j$(nproc --all) libdroidmedia minimediaservice minisfservice libminisf08:22
Mister_Magister anyway droidmedia is not building08:23
Mister_Magisterpiggz: let me handle that :P08:23
Mister_Magisterpiggz: without $PORT_ARCH detect targets script fails and you end up with "make -j12 "08:27
Mister_Magisterso you build everything08:27
Mister_Magisterpiggz: dieded08:28
piggzEnv setup for t508:28
piggzHABUILD_SDK [t5] piggz@linux-f1uu:~$ echo $PORT_ARCH08:28
Mister_Magistertry my command tho08:28
piggzim in the middle of a build atm08:29
piggzits slooooooow08:29
Mister_Magisterwell k08:29
Mister_Magisterbuit while building droidmedia use my command :P08:29
Mister_Magisterframeworks/av/services/camera/libcameraservice/device1/CameraHardwareInterface.h:32:10: fatal error: 'vendor/qti/hardware/camera/device/1.0/IQCameraDeviceCallback.h' file not found08:44
Mister_Magisterpiggz: clearly jolla is behind us :P08:51
malMister_Magister: nope09:08
malMister_Magister: do you have vendor repo?09:08
abransonboth of those are building fine for the xperia 10 on android 9, so it's probably a problem with specific droid sources09:10
Mister_Magistermal: ofcourse i have09:10
Mister_Magisterabranson: could be, we are working on stock sources so things are different from aosp09:10
abransonespecially odd that the cameraservice is depending on a header that doesn't exist. is there a choice of vendor repos at all?09:11
Mister_Magisteri'm thinking of replacieng frameworks_av with aosp one09:11
malMister_Magister: do you see IQCameraDeviceCallback.h anywhere in source tree09:11
abransonand treble was meant to standardise all those frameworks :D09:11
Mister_Magistermal: nope i checked that already09:11
malthat could also be automatically generated09:11
Mister_Magisterdoubt that09:12
malmany things are automatically generated now in android build09:12
Mister_Magisterbut this isn't09:13
malmaybe it should be but is not?09:13
malabranson: that build log is odd, the first error of some log format, why would that not happen in other android 9 builds09:13
Mister_Magister(someone can't write error free code?)09:13
Mister_Magisteri added simple casts and it fixed problem09:14
abransonhmm that shouldn't be necessary09:14
Mister_Magisterbut it is09:14
abransonunless there's a proper way of handling that left to be discovered09:15
malabranson: has size_t always been unsigned long?09:16
Mister_Magisterit can be because of my sources tho09:16
abransonmight be related to the bitness?09:17
malabranson: could be, the logging code is indeed a bit wrong, based on quick search size_t should be printed using %zu09:18
abransonthat sounds like the right fix09:19
malabranson: we could check if there are other warnings like that also09:20
abransonthe second one is actually the wrong size though09:21
Mister_Magisteri will try replacing frameworks av09:25
Mister_Magisterwhat could go wrong?09:25
malabranson: do you want to fix those or should I?09:27
abransonmal: do you know how to fix the second one?09:28
Mister_Magisteri fixed it already09:28
* Mister_Magister runs09:28
abransonMister_Magister: for all builds?09:28
Mister_Magistermal: i have answer for you why fxtec has f5 and f6 not SW_SLIDE09:29
Mister_MagisterRegarding keyboard slide with Hall sensor, I was told the head engineer wanted to do with SW_LID, but at the time the upper part has finished, so they did not change09:29
*** justgerd98 is now known as JustGerd09:36
Mister_Magistermal: i think that file might be generated after all09:45
piggzugh, 80000 targets for droidmedia09:48
piggzstill, 30000 better than befoe!09:48
Mister_Magisterpiggz: did you use my command?09:48
Mister_Magisterthen do it09:49
Mister_Magisterthere is like 1000 targets09:49
malMister_Magister: so you need to figure out what should generate it09:53
piggzok, im at the stage with the logging errors....10:15
piggzmade changes10:15
Mister_Magisterwhat logging errors?10:16
piggzthe size specifiers10:16
Mister_Magisterah the casts10:16
Mister_Magisterwelcome we are at same stage now10:16
Mister_Magisternow just wait with me10:16
piggzabranson: PRId64 for int64_t ?10:25
piggzaccording to stackoverflow :D10:26
malpiggz: hopefully the timestamp has same type in all android bases10:33
deathmisthey! aren't pretty much all dirs under /dev/cpuset meant to be owned by system:system because mine are mostly owned by root and causing permission issues. also dhi has references to many missing "mems" and "cpus" files but I'm only seeing "cpuset.cpus" and "cpuset.mems" instead in all the dirs10:47
Mister_Magisterdeathmist: you are missing udev rules then10:49
deathmistseems so as there's no files mentioning "cpuset" in /lib/udev/rules.d/ on the device10:56
Mister_Magisterthat's what i just said10:58
piggzfrom my understanding, we need to run hidl-gen on
piggzto get a .h from the .hal file?11:07
piggzabranson: mal ^^ ?11:07
piggznot sure on the exact parameters yt tho11:07
malpiggz: something like that, maybe just check output of make modules and see it has something like that11:19
Mister_Magistermal: i have now the file in out/soong/.intermediates/vendor/qcom/opensource/interfaces/camera/device/1.0/\@1.0_genc++_headers/gen/vendor/qti/hardware/camera/device/1.0/IQCameraDeviceCallback.h11:30
Mister_Magisterbut dunno how to use it11:31
piggzmal: hopefully its just11:35
Mister_Magisterpiggz: i already did it lol11:35
Mister_Magisterbut that ain't all11:35
Mister_Magistercheck what i just wrote11:36
abransonsounds like that should be a dep in the cameraservice mk file11:36
Mister_Magisterabranson: because it is11:36
abransonbut it's not getting brought in for your build?11:37
Mister_Magisterabranson: in LOCAL_SHARED_LIBRARIES:=11:37
Mister_Magisterit is11:37
Mister_Magisterdunno if that wil make it build11:37
Mister_Magisterbut if it's on out that's not all11:37
abransonmight be worth checking how it's included in the sony aosp, as everything builds fine there11:38
Mister_Magisterabranson: it doesn't exist in aosp11:38
malMister_Magister: usually those should be automatically found there, not sure why it fails to do that11:38
abransonthat's the problem to solve11:38
Mister_Magisteraosp doesn't include that file at all so sony ain't helpful there11:40
Mister_Magisterabranson: mal:
Mister_Magisterabranson: can i add to minimedia
Mister_Magisterlets see if that helps11:47
abransonthis looks relevant:
Mister_Magisteryeah we have that11:49
Mister_Magisterbut frameworks_av is from aosp11:49
Mister_Magisterand aosp doesn't include it11:49
Mister_Magisterabranson: mal: that did the trick!11:54
Mister_Magisterpiggz: oh you too read!11:54
Mister_Magisternow building minimedia includes it11:54
Mister_Magisterbut i think it's specific to our sources only11:55
Mister_Magister i fixed something myself11:55
* Mister_Magister is shocked and proud11:55
Mister_Magister#### build completed successfully (10:12 (mm:ss)) ####11:56
malthe question remains why the sources were broken11:56
Mister_Magistermal: we don't have pure aosp they are stock sources11:57
Mister_Magisterin short: blame fxtec11:57
Mister_Magisterthey added it and minimedia didn't include it11:58
Mister_Magisterif you would build libcameraservice standalone it would work11:58
piggzwhat was the fix?11:58
Mister_Magisterpiggz: read please what i wrote11:58
piggzah, i c11:59
Mister_Magisterpiggz: y u not reading my messages ;-;11:59
Mister_Magisterpackaging droidmedia now12:00
Mister_MagisterBuilding of droidmedia-localbuild finished successfully12:01
Mister_Magisterlet's see if sensors work12:03
malMister_Magister: which sensor backend did you build, binder or hal?12:04
Mister_Magisteri dunno12:05
Mister_Magisterare there options?12:05
Mister_Magisterdroid-hal-init: Could not find service hosting interface android.frameworks.sensorservice@1.0::ISensorManager/default nope still piggz12:05
Mister_Magisterpiggz: droidmedia didn't help12:06
malMister_Magister: which sensor package do you have hybris or binder12:06
malMister_Magister: are mini* services running?12:07
Mister_Magisterthey aren't12:07
Mister_Magistermal: sensorfw12:07
Mister_Magisteri don't understand what do you mean hybris or binder12:08
piggzMister_Magister: where in mimmedia
Mister_Magisterbluebinder/            droid-configs/  droid-hal-version-t5/   geoclue-providers-hybris/  libgbinder-radio/  libhybris/             ngfd-plugin-droid-vibrator/  pulseaudio-modules-droid/             qt5-qpa-hwcomposer-plugin/      qtscenegraph-adaptation/  sensorfw/ community-adaptation/  droid-hal-t5/   droidmedia-localbuild/  libgbinder/                libglibutil/       mce-plugin-libhybris/  ofono-ril-binder-plugin/     qt5-12:09
Mister_Magisterfeedback-haptics-droid-vibrator/  qt5-qpa-surfaceflinger-plugin/  repodata/12:09
malMister_Magister: zypper se sensor12:09
Mister_Magisteryeah they should be native but that's minor thing12:09
malon device12:09
Mister_Magisteryeah surely on device :P12:10
* Mister_Magister is waiting for all repos to timeout cause wifi ain't working yet12:10
piggzMister_Magister: it should be working12:10
Mister_Magisterbut it isnt'12:10
piggzworks a charm here!12:10
Mister_Magisterafter reboot?12:11
Mister_Magistermal: hybris-libsensorfw-qt5-hal-0.10.13-0.armv7hl12:11
maltry building binder backend12:11
malit should be default for android 9 if you have new enough dhd submodule12:12
Mister_Magistermal: dhd is at bd43b2cf63512cfc9789511e6c27a50504b4f44812:12
malso too old12:13
malmake sure you remove the hal rpms from droid-local-repo before building12:13
Mister_Magistershould i just clean droid-local-repo and rebuild everything?12:14
Mister_Magisterill do it12:16
piggzMister_Magister: where in mimmedia
Mister_Magistersure gib second12:18
Mister_Magistersent on telegram sir12:19
Mister_Magistermal: whichpackage should i build? sensorfw?12:23
malthe commit I linked tells everything you need to know12:24
Mister_Magisterah sorry sorry12:24
Mister_Magisteri'm retarded12:24
Mister_Magistermal: yay test_sensors works12:35
piggzabranson: mal: if we can build the .h file in out/.... how can we make the build find it?12:37
piggzwithout copying to vendor/....12:37
abransonpiggz: a lot of things are generated in out/. i think the dependency declarations in the files make sure that they are generated and found at the right time12:39
piggzabranson: yeah, but it doesnt :)12:40
piggzfile is generated, but not getting picked up12:40
abransonmaybe the dependency type is wrong?12:40
piggzactually, probably need to change frameworks/av/services/camera/libcameraservice/device1/12:40
piggzas, we added it to a dep for minimedia, but thats not where the build fails12:41
malMister_Magister: what fixed sensors?12:41
Mister_Magistersensors aren't working yet12:48
Mister_Magisteronly test_sensors12:48
Mister_Magistermal: test_sensors says Got 0 sensors12:49
malMister_Magister: did you already build binder package?12:50
Mister_Magisterand installed it12:50
malwhat does journal tell?12:50
Mister_Magisterthat fixed test_sensors but not…12:50
malso logs logs logs12:50
Mister_MagisterCould not find service hosting interface android.frameworks.sensorservice@1.0::ISensorManager/default still counts?12:50
malMister_Magister: you never told which service complains about ISensorManager12:52
Mister_Magisterwait no12:52
malwhat does logcat have?12:52
malI mean give whole logcat12:53
malwhat have I told you about giving only selected parts of logs12:53
Mister_Magisterdunno i give only meaningfull ones12:53
Mister_Magisterlemme reboot12:53
malwhat you think is meaningful might not really be the important part of the log12:55
malyou might only give the part relevant for that service when the real issue can be in other services12:55
Mister_Magisterwell k12:55
malso many errors13:03
malMister_Magister: show output of mount and also "ls -l /"13:04
Mister_Magister /persist doesn't exist if thats what you mean but i will13:04
malMister_Magister: so you haven't added the symlinks I told to add for the other device13:05
malno wonder everything fails13:06
Mister_Magisterwhat symlink13:06
malask piggz for the changes done to the other device13:06
Mister_Magisterit's same device13:07
malyou should probably have access to the repos of thayt13:07
malI really have no idea what device is what13:07
Mister_Magisterfxtec pro113:07
malso ask piggz13:07
malhe fixed the issues13:07
Mister_Magisterpiggz didn't add any symlink in root13:07
malthen he has been lazy13:08
malbecause those are needed13:08
malunless those come from straggler files now13:08
malyou can check that13:08
Mister_Magistersry for causing trouble13:08
mallike I guessed you were focusing on wrong error messages13:09
Mister_Magisterwhat did you focus on?13:09
malthe real issue was in lower level13:09
Mister_Magisterwhere did you saw it?13:10
malthose many messages about not found files and permission issues13:10
Mister_Magistercheck_sensors_enabled: open error: settings file "/persist/sensors/sensors_settings" ?13:10
mallike that13:10
Mister_Magisterwhen i saw it i mentioned persist13:11
Mister_Magisteri just didn't check logcat13:11
T42<adampigg> Sorry, i have an electrician here, no power)13:11
malwhich should have told you that /persist is needed13:11
Mister_Magisteryeah just like i said i didn't look into logcat13:11
T42<adampigg> I fixed some files in /persist that the sensor sercice lools for13:11
maland if you would have looked at the new sony adaptation repos you would have noticed that is13:11
malMister_Magister: why not, if you have failing android service then logcat is obvious place to look13:12
Mister_Magister@adampigg u diidn't push it tho13:12
Mister_Magistermal: because i'm a retard13:12
mal@adampigg isn't it just the symlink13:12
T42<adampigg> Not yet..13:12
T42<adampigg> No13:12
mal@adampigg what else then13:13
Mister_Magistersony doesn't have symlink either13:13
T42<adampigg> There is a file that must contain a 1 to enable sensors13:13
malMister_Magister: really
T42<adampigg> And a folder must exist to allow a sensor.reg to be created13:13
T42<adampigg> I will post as soon as i get ppwer amd my laptop back13:13
Mister_Magistermal: how
malMister_Magister: read the commit I linked13:14
Mister_Magistersure i mean13:14
Mister_Magisteri checked the source and they aren't there13:15
Mister_Magisterwhy would i search in commits history13:15
malwhy wouldn't you?13:15
Mister_Magisterif it had them then just check source?13:15
Mister_Magisterhmm /mnt/vendor/persist doesn't exist13:16
Mister_Magister /mnt was mounted after /mnt/vendor/persist13:17
Mister_Magistermeh i'm failure13:19
malMister_Magister: didn't piggz also sort that out13:32
Mister_Magisteri don't know we are not same person13:34
* Mister_Magister all happiness faded away13:38
malMister_Magister: don't you communicate at all since you are doing the same device13:41
Mister_Magisternot really13:41
malI really don't want to fix the same issues many times13:41
Mister_Magisteri know13:41
Mister_Magisteri'm just sad because i suck at porting13:42
Mister_Magisteri will wait for piggz13:42
piggzI have power!14:03
piggzMister_Magister: logcat pls14:04
mccrearyelros24: TBC, the touchscreen on the HW A xt897 does not work in cm-11, but here is the evdev_trace output you wanted:
mccrearyI've included output for both HW A and HW B variants14:47
KALUbEHi all, I'm having issues with WiFi, this phone is using a Spreadtrum chip and I haven't been able to find any others that use it. I can load the sprdwl module which is what is used I believe. The wlan0 socket is in /data/misc/wifi/sockets/ but I don't think there's a driver on the other end. There is a file called wpa_supplicant.rc in my device tree which contains this
mccrearypiggz: I haven't pushed anything to github, but tdm's kernel works well for TWRP if you add this patch:
mccrearyMy device tree is currently a hacky mess, but started as a fork of
malKALUbE: did you get UI working?14:51
malKALUbE: also you should not care about android side things for wlan14:51
KALUbEmal: Yes, adding property_contexts to straggler files made it all work15:02
KALUbEThanks to elros34 for that15:04
T42<elros34> mccreary: I didn't even know there are 2 versions. Driver is loaded (atmxt) so I guess it doesn't generate anything. You can read raw events using evdev_trace -t 7. It's about cm version or kernel?15:04
malKALUbE: wondering how those were missing from the originally? shouldn't those have been needed to finish the build successfully15:05
KALUbEI have no idea, I guess my device tree isn't brilliant. I have rerun build_packages and rebuilt a new rootfs that boots fine so it should be good now15:06
malKALUbE: most likely reasons for wlan issues are that either it's not finding the firmware files or it's missing some initialization15:06
malKALUbE: show dmesg15:07
KALUbEmal: Will grab that. My kernel defconfig references firmware files that don't exist:15:09
malKALUbE: what wlan chip does that really have "CONFIG_SPRDWL is not set" earlier you said it has sprdwl module15:10
KALUbEGiven it specifies the location of BCM fw I would guess it's actually going to be broadcom?15:12
KALUbEHowever, when building the kernel it also compiles SPRD_MODULES which includes the sprdwl.ko module15:13
KALUbEWould that be why CONFIG_SPRDWL is not set?15:15
malKALUbE: can you show output of command "mount"15:18
malKALUbE: now show the original fstab in your device repo15:20
malKALUbE: hmm, now "ls -l /dev/block/platform/sdio_emmc/by-name/"15:22
mccrearyelros34: The touchscreen isn't generating events on the HW A xt897. The fix for this predates my involvement w/ LineageOS on the xt897, but I remember talking to nadlabak about it ~2.5 yrs ago15:24
malKALUbE: does audio work?15:35
KALUbENo, no audio15:37
malok, log has an error about audio firmware15:38
deathmistpketo: could you bring Store to the OnePlus 5? Device model: 5 (cheeseburger / cheeseburger)15:40
KALUbEmal: Haven't done much for audio yet, so that's fine.15:40
malKALUbE: I was just thinking if the reason is the same, some missing firmware symlink or something15:40
KALUbEmal: I found the firmware file referenced by my defconfig in /hardware/broadcom on my build and pushed it to the right location in /system, no change.15:42
malKALUbE: I think you might really have sprdwl15:43
malcheck line 240 in dmesg15:43
malit mentions sprdwl_mem_alloc15:44
KALUbEmal: It was being built but not loaded, I have added it to /etc/modules-load.d/ and it now gets loaded on boot. But it doesn't get used by any processes and doesn't seem to create any devices, although I'm not sure how to check15:45
KALUbEI've been messing with the wpa_supplicant.service file and .conf file, changing the ctrl_interface to /data/misc/wifi/sockets. There is no other wlan that I can find15:47
malKALUbE: usually we don't use /etc/modules-load.d/ in sfos because some drivers might want firmware to be present when loading the module15:47
KALUbEmal: IS there a better method I should use instead?15:47
malmodprobe in a systemd service15:47
malsomething like that but of course adapt it15:48
malKALUbE: but before doing any service like that we need to get it working manually15:48
malKALUbE: anything in dmesg after modprobing the module?15:48
KALUbENothing, just [    0.421508] [c0] [sprdwl_mem_alloc]15:52
KALUbEmal: [   69.455902] [c0] [SC2331]######## wlan_module_init sdio is not ready  ##########15:55
KALUbEShould there be a driver that loads on boot?15:55
mccrearyelros34: FYI, relevant info from dmesg:
Mister_Magistermal: /mnt/vendor/persist was never fixedf16:20
mccrearyThe cm-14.1 kernel says, "[    4.352723,1] input: atmxt-i2c as /devices/virtual/input/input6", while the cm-11/SFOS kernel says, "[   28.343995]   28.343995,1] input: atmxt-i2c as /devices/virtual/input/input7"16:20
Mister_Magistersomething in dhi is mounting /mnt as tmpfs16:20
T42<elros34> mccreary: maybe atmxt-r2.tdat  have different content for cm11 and cm1416:41
piggzmal: how can a customise a mount unit created by dhd, so that it mounts after DHI? im thinking maybe masking that one from creation, then just add the unit in config/sparse... ?16:53
Mister_Magistermal: btw friendly reminder about obs projects16:56
Mister_Magisteralso mal we got sensors woohoo whatever16:57
malpiggz: add the mount point to and then add it to sparse16:58
Mister_Magisterskip mountpoints?16:58
piggzmal: yup, thats whay ive done16:58
malpiggz: I don't know any better way16:59
piggzyeah, that was my first thought16:59
malpiggz: what is in the /vendor normally, other than the other mounts17:00
piggzlibs n stuff17:00
piggzmnt vendor persist has config17:00
Mister_Magisterrc files17:00
piggzneed mnt vendor persist to be mount after dhi17:01
Mister_Magisterfstab and uevent files17:01
mccrearyelros34: There are multiple versions of atmxt-r2.tdat in /system, and replacing /system/etc/firmware/atmxt-r2.tdat with /system/xt897/etc/firmware/atmxt-r2.tdat fixes my HW A touchscreen problem17:03
mccrearyFWIW, the MD5 checksum for /system/xt897/etc/firmware/atmxt-r2.tdat matches the file in this commit from nadlabak:
piggzMister_Magister: i also fixed browser pink17:04
Mister_Magisterwell k17:04
Mister_Magisterwho will fix my hwcomposer?17:04
piggzu will17:07
KALUbEMister_Magister: I think I have a wrench somewhere17:07
Mister_Magisterpiggz: i won't17:07
piggzgrep the logs luke
Mister_Magisterhave nothing left to debug17:07
Mister_Magisterpiggz: >noone has this new device17:08
Mister_Magister>grep logs17:08
Mister_Magistergood idea17:08
KALUbEWhat's the error though?17:08
Mister_Magisterfailing to get main display17:09
KALUbEThat sounds rough...17:10
deathmistpketo: also the device is uniquely identified via it's IMEI1 from the back of the packaging box (and about screen in settings) and builds are not yet on OBS17:31
vknechtthe ssu board mappings, are they used by mic, or only on device ?17:34
vknechtor by obs ?17:40
malvknecht: on device18:03
vknechtok, related to the graphical update client I guess ?18:05
malvknecht: what is the problem you have?18:06
vknecht"not installable provider" when running mic for droid-config-$DEVICE-sailfish/bluez5 when they're enabled in patterns18:08
vknecht(different from uninstallable provider from hadk)18:08
vknechtwas wondering if board mapping could be involved, but I guess that's not the case18:09
vknechtabout only reference I found about this error is
vknechtguess I can live without bluez5 for now, but maybe the missing sailfish configs are what prevent getting to gui... or would those missing only imply wrong scaling/positionning of buttons/widgets/etc. ?18:13
vknechtalso, is there something with mic caching or droid-local-repo so that we might have to remove stuff before building/mic'inc ? looks like rebuilt packages are not taken in account18:19
malvknecht: show your repos18:39
vknechtnot in sync atm (switched to community-way), but here they are :
vknechtI think I'd better get all these straight first, so you don't lose time18:46
pketodeathmist: could you remind me again on monday :)18:49
deathmistpketo: sure thing!18:51
vknechtcurrent journal and dmesg just in case19:02
vknechtno dhi running yet it seems ; and no telnet/usb connectivity past the switch_root, gotta get them from sdcard/raw journal files19:03
piggz_abranson: referring to you knowledge deomstrated here .....19:04
piggz_how can i write a custom keyboard layout for the fxtec which sends correct keys when modifiers are pressed?19:05
vknechtbut it's only been one day since I get journalctl logs, started almost a month ago :-)19:05
piggz_or, any good guides to writing keympas, the only bit i nned to know then is how to apply them in sailfish19:07
piggz_thi s might help
Mister_Magisterwhy isn't preloading selinux_stubs solving every my problem19:17
Mister_Magistermal: want to help me with hwc2 or did i piss off you enough?19:17
Mister_Magister assuming the latter one19:19
Mister_Magisterblob is dying19:22
Mister_Magisterdebug blob19:22
Mister_Magisterhwc blob service dies when registering callback19:37
Mister_Magisterwithout anything obvious in strace19:37
Mister_Magisteris tombstoned supposed to work on sfos?20:03
malnot sure, but those could be in /data/tombstones20:04
Mister_Magisterit's segfaulting20:05
Mister_Magisterwell not really segfaulting20:05
Mister_Magisterit's commiting suicide20:05
Mister_Magister"failed to get socket from init"20:05
malpiggz_: ping20:07
piggz_mal: pong20:07
Mister_Magistermal: so… no idea how to debug blob? though you might have some idea20:07
Mister_Magisteri'm causing too much trouble20:07
malpiggz_: could you fix the commit message in you are not enabling it only for x8620:08
malMister_Magister: have you verified the usual suspects, correct mounts, udev rules and that all needed firmware is correct20:09
piggz_mal: sure20:09
Mister_Magisteri don't know what does it mean to be correct20:10
Mister_Magisterhow do i know if my udev rules are correct or not20:10
Mister_Magisteror mouts20:10
piggz_mal: looking at my pulseaudio log, it seems to be using null sink?20:11
malMister_Magister: well if not all needed ueventd*.rc files are not created to out/ then some udev rules are not created which generate all sorts of interesting issues20:11
Mister_Magistermounts seem to be okay, udev rules have files from vendor so should be okay aswell20:11
Mister_Magisteri copied ueventd files from /vendor20:12
Mister_Magisterto out/../root/20:12
Mister_Magisterall rules should be there20:12
Mister_Magistercan i just chmod 777 /dev?20:12
Mister_Magisterholy moly20:13
Mister_Magistertest_hwc2 is not dying20:13
Mister_Magisterit did something20:13
Mister_Magisterno screen but not dying20:13
Mister_Magistermccreary: ^20:14
malshow your udev rule file from the device,/lib/udev/rules.d/999-android-system.rules20:14
malMister_Magister: is that the generated file or have you modified it in any way?20:15
Mister_Magisterexcept for binder files20:15
piggz_E: [pulseaudio] droid-util.c: Failed to load audio hw module audio.primary : Invalid argument (22)20:16
piggz_E: [pulseaudio] module.c: Failed to load module "module-droid-card" (argument: "rate=48000"): initialization failed.20:16
malMister_Magister: if binder stuff is missing then something is missing from out afaik20:16
Mister_MagisterHWCSession::CreatePrimaryDisplay: Primary display created.20:16
Mister_Magistermal: i think that too20:17
malMister_Magister: find eny uevent* files in device and vendor repos20:17
Mister_Magisternone in device20:17
Mister_Magisteri will show you20:17
malany .rc files20:17
Mister_Magister /system/bin/tombstoned: Tombstone written to: /data/tombstones/tombstone_0420:17
Mister_Magisterso it's working after all20:17
malfirst thing is to figure out udev rules20:18
Mister_Magisterjust oftopicking20:18
Mister_Magisteron device tho things are different20:19
Mister_Magisteri have in / and in /vendor uevent.rc and uevent.asus.rc20:20
Mister_Magisterno other uevent .rc files20:20
malMister_Magister: do you have a link to device repo in github or somewhere?20:21
malor is it not public20:21
Mister_Magisteryess… yes20:21
Mister_Magisteri was waiting for it to become public20:21
Mister_Magisterdevice is 5z not pro120:21
Mister_Magisterso not the one piggz and thekit has20:22
Mister_Magistersorry for confusion20:22
malMister_Magister: very odd, the binder rules should be in system/core/rootdir/ueventd.rc20:27
malwhich should automatically end up in out/20:27
Mister_Magisterbut they aren't?20:28
malMister_Magister: well I don't know, you tell me20:28
Mister_Magisterthey are here20:28
Mister_Magisterrepo is standard lineage20:29
malbut is the same file in out/20:29
Mister_Magisteri copied ueventd.rc from device20:29
Mister_Magister99% nope20:29
malpiggz_: change still merge request title, it's the old one20:30
malMister_Magister: hmm, so how can the one on device be so different20:31
Mister_Magisteri will use the ueventd.rc from sources20:31
piggz_mal: done20:31
Mister_Magisterand ueventd.asus.rc from device20:31
Mister_Magisteris that ok?20:31
malpiggz_: thanks20:31
malMister_Magister: based on the rules you gave that is quite empty20:32
Mister_Magisterin source?20:32
Mister_Magisteri can give you files if you want20:32
malMister_Magister: look here after "From ueventd.asus.rc"20:32
Mister_Magisterhere is sauce if you want20:33
Mister_Magisterwell yeah they are empty20:33
malMister_Magister: I still cannot understand where the ueventd.rc on device came from, the rules from it don't match that20:33
Mister_Magisteri know i have same feeling20:34
malMister_Magister: have you tried to see what happens if you just copy that to out/ and build droid-hal rpms again20:34
Mister_Magisteri will copy ueventd from system_core and rebuild udev rules okay? or shouldn't i20:34
Mister_Magisterthat you mean from device?20:35
Mister_Magisterthat's what i'm actually on20:35
malno, from system/core20:35
Mister_Magisterokay so the thing i said above20:35
Mister_Magisteri'm doing it20:35
malyou already tried the other one20:35
malyes, we said the same thing at the same time20:35
Mister_Magistermissing kgsl rules now20:38
Mister_Magisterwhat if i put both of them?20:39
Mister_Magisteri can put the one from device as ueventd.device.rc20:39
Mister_Magisterdroid-hal won't care20:39
Mister_Magisterit will merge them20:39
Mister_Magistermal: ?20:44
Mister_Magisteru okay?20:44
malI'm only randomly on irc rest of tonight20:59
Mister_Magisterokay yeah its late so i'm going to sleep myself20:59
Mister_Magisterif i'm not bothering you i will ping you another day20:59
Mister_Magisterif you don't hate me enough yet20:59
T42<elros34> mccreary: so it's a matter of correct atmxt-r2.tdat firmware? Do you know any sysfs entry to distinguish between A and B hardware?22:07
T42<elros34> mccreary: maybe it's too late... In cm11*zip /system/etc/firmware/atmxt-r2.tdat has SCORPION header but after installing on device it has ASANTI23:00
mccrearyelros34: I'll bet my recovery doesn't assert 'asanti'23:41
mccrearyI will check that23:41

Generated by 2.17.1 by Marius Gedminas - find it at!