Monday, 2022-02-21

T42<HengYeDev> possibly i need ?00:39
T42<elros34> this is different android base so unlikely01:12
T42<HengYeDev> could it be the vintf build problem? i commented out the line that verifies system compayibiliy matrix of vintf in Makefile @elros3401:18
T42<elros34> no idea what's that01:19
T42<elros34> do you have up-to-date repos uploaded?01:21
T42<HengYeDev> not quite, i it tomorrow (re @elros34: do you have up-to-da...)01:23
T42<elros34> in the meantime you can starts with something simple like mask user@100000, reboot then start /system/*/surfaceflinger01:24
T42<edp_17> @elros34 : Hello. I think need your help with the BT again.09:00
T42<edp_17> Note4 (sfos4.2, broadcom-bluetooth and bluetooth-rfkill-event updated to the latest) :
T42<edp_17> If I start /usr/sbin/brcm_patchram_plus manually, BT work, otherwise it doesn't turn on.09:02
T42<edp_17> in the log it says, cannot execute brcm_patchram_plus.09:03
T42<elros34> what about "Can't open Bluetooth address file /factory/bluetooth_address"? Missing bdaddr in conf?09:13
T42<edp_17> In conf I have:  daddr = /efs/bluetooth/bt_addr09:14
T42<edp_17> I'm trying to symlink this to /factory/bluetooth_address to see whether that works09:14
T42<edp_17> Yep, that woked.09:15
T42<elros34> Do you have that bdaddr = in General section?09:16
T42<edp_17> How do you mean General section? There are no sections in there. I'll amend it according to the example in rfkill package and will retry. Thanks.09:18
T42<edp_17> That wasn't enough to symlink bt address, I also needed to rename the config file in /etc/firmware from bcm4358a1.conf to bcm43341.conf.09:22
T42<edp_17> Now, I try to remove the symlink to see whether the renamed conf file solves the issue.09:23
T42<elros34> name of conf file you should have defined in /etc/sysconfig/, probably in droid-config sparse09:25
T42<edp_17> Yes. The symlink is not necessary. The conf file must have the new name.09:26
T42<elros34> so if you changed nothing and you need different name for config file then probably you are missing mentioned sysconfig09:28
T42<edp_17> No, what I've done was only renaming that file bcm4358a1.conf to bcm43341.conf09:47
T42<edp_17> What is the "mentioned sysconfig" you were referring to?09:48
T42<edp_17> Can I get some help with GPS as well. In csd it says 'GPS in error state'. Journal:  and logcat:
T42<edp_17> strace of gpsd:
T42<elros34> this one
ThaodanJust note this issue here is that you have to specific the firmware, on newer kernels the kernel loads the firmware and the bluetooth module patches the ram.10:29
Thaodan(ram of the bluetooth/wifi chip)10:29
T42<edp_17> @elros34 : Yeah, that file is empty (except some comment lines) I try to put the original content back and removing /etc/firmware/bcm43341.conf10:37
T42<edp_17> Yep, BT still works.10:39
T42<edp_17> Thanks Thaodan.10:40
T42<adampigg> lbt: stuck dod :)10:44
T42<edp_17> Can the 'failed to create flashlight gstreamer source element' error be related to a camera issue? (Flashlight doesn't work and camera just gives a black screen sometimes flashing up an old tv style image where there was no signal.) Strace of camera is here:
T42<edp_17> I've also tried " gst-inspect-1.0 | grep -Ei 'droid|camsrc' " but gave me nothing back.11:12
T42<elros34> AFAIK flashlight is controlled by gstreamer so working camera is required. There are apps which controla flash using sysfs. Remove ~/.cache/gstreamer* and try again gst-inspect,11:44
T42<edp_17> @elros34 : I got this lones twice: gst-plugin-scanner: ../hybris.c:75: __resolve_sym: Assertion `ptr != NULL' failed.11:47
T42<elros34> read hot-hadk, 4.3 section, there was bug ignored for a while and recently fixed in bugfix release
T42<elros34> if that is not the case then maybe you have version mismatch for droidmedia/gst-droid11:53
T42<edp_17> I haven't updated the chroot. I'm still using '14.04.5 LTS, Trusty Tahr'11:58
T42<edp_17> This tar.bz2: ubuntu-trusty-20180613-android-rootfs.tar.bz211:59
T42<elros34> ?grep for your error12:02
T42<elros34> it's there12:02
T42<edp_17> Found it on collabedit. That says for 4.3.x but this is 4.2 build12:05
T42<elros34> so do you use correct versions for droidmedia/gst-droid?12:08
T42<edp_17> I am using the ones mentioned on collabedit. droidmedia: 0.20210702.0, gts-droid: upgrade-4.2.0, and gmp-droid: b61b740987a9880fccf5401220644c7dcf184be2 (tag 0.2)12:13
T42<elros34> did you check it on device?12:19
T42<edp_17> droidmedia-0.20210702.0-1.armv7hl and gmp-droid-0.2-1.armv7hl. It says gts-droid is not installed.12:24
T42<edp_17> gstreamer1.0-1.18.4+git1-1.4.1.jolla.armv7hl12:25
T42<edp_17> gstreamer1.0-droid-0.20220207.0-1.armv7hl12:25
T42<elros34> so clerly wrong version of gst-droid12:29
T42<elros34> gst-droid repo => gstreamer1.0-droid package if that is not clear12:30
T42<adampigg> lbt: thinking out load, could we get a chumgui obs user, with limited privs that could make api get calls only?12:36
T42<edp_17> @elros34 : I see now. Deleted and cloned upgrade-4.2.0 branch and now I see the difference in git log.12:39
T42<edp_17> @elros34 : is it enough to rebuild/install gts-droid?12:39
T42<elros34> should be12:41
T42<edp_17> Failed to build:
T42<adampigg> You shouldnt need to build gst droid these days12:43
T42<edp_17> @adampigg : I am building 4.2.12:44
T42<edp_17> As 4.2 is a stop release :( I need it for OTA from 4.1 to 4.3.12:45
T42<adampigg> K12:46
T42<edp_17> @adampigg : What's that?12:49
T42<adampigg> Ok12:50
T42<edp_17> Oh, :) Thanks.12:51
T42<elros34> did you really have correct droidmedia installed? What zypper se -s droidmedia returns for both packages?12:52
T42<elros34> and if you want to OTA then why do you even bother to build everything locally?12:52
T42<edp_17> On device? This:
T42<edp_17> I am building locally because I wan't be sure that once I OTAd, all stuff works fine.12:56
T42<edp_17> Although, my aim is 4.3 so you are right, I shouldn't bother to fix stuff for 4.2.12:57
T42<elros34> I can't see how using obs will make harder than local build. About error: is in sdk so check zypper output in target13:00
T42<edp_17> zypper output in target:
T42<edp_17> @elros34 : On OBS, all packages for 4.2 have built including gmp-droid and gts-droid. Should I try to install those packages on device?13:11
T42<edp_17> @elros34 : I've managed to install gstreamer1.0 from the OBS repo and main camera and flash do work now. 👍13:37
T42<elros34> I guess adaptation-community-common mess with your local build. Is that devel repo?13:37
T42<edp_17> Now only the GPS should be fixed. 😊13:38
T42<edp_17> @elros34 : Yes, it did mess with the local build. I've downloaded the packages from OBS and then installed manually. Nope, that is a testing repo.13:39
T42<elros34> and you sure target is 4.2?13:39
T42<edp_17> Yes.13:40
T42<LSolrac> So, like 3-4 repo resyncs later, Im unfortunately stuck with this. Ive no vendor repos that I know of, the only one being the actual image :
T42<elros34> @edp_17  hm but there is no droidmedia-devel in 4.2.0 repo, and your zypper output shows it but good it works13:46
T42<elros34> so why didn't you add proprietary repo like we wwre talking about?13:47
T42<elros34> @LSolrac ^13:47
T42<edp_17> @elros34 : Now zypper shows this:
T42<edp_17> In my treltexx (testing 4.2) repo has these 5 gstreamer1.0* packages:
T42<edp_17> I've only installed gstreamer1.0-droid* and tool.13:51
T42<HengYeDev> @elros34 I have updated the repos
T42<HengYeDev> some parts of sparse i didn't add yet but applied on my device to test13:58
T42<HengYeDev> and init script isn't pushed yet to hybris-boot fork13:58
T42<HengYeDev> surfaceflinger says failed to get hwcomposer service than segfaults14:02
deathmistwell yeah because hwservicemanager dies -> hwcomposer can't run -> surfaceflinger can't work14:05
T42<HengYeDev> since the backtrace mentions /system/lib64/ (android::vintf::HalManifest::getInstances as its first entry, does that correlate with the build errors i had about vintf compatibility matrices14:09
T42<LSolrac> Similar reason, I haven't found one as of yet. (re @elros34: so why didn't you ad...)14:09
T42<LSolrac> Not for Android 10/Lineage 17.1 anyway14:10
T42<elros34> How can  you not found it proprietary_vendor_samsung in google gives you correct repo in first result:
T42<elros34> @HengYeDev but hwcomposer still fails with masked user@100000?14:18
T42<HengYeDev> Yes14:19
T42<elros34> have you tried to strace hwservicemanager? What about /dev/ permissions?14:20
T42<HengYeDev> what exactly do you mean by /dev permissions? should i send a paste of all files in /dev in ls-l?14:22
T42<HengYeDev> also, not sure if relevant, but /odm/etc points to /vendor/odm/etc, but /vendor/odm points to /odm @elros3414:23
T42<elros34> according to your repo, you do not have mounted /odm14:24
T42<HengYeDev> @elros34 i forgot to add my mount-dynparts script to sparse14:26
T42<elros34> same thing, there is no /odm, only /odm214:26
T42<elros34> About permissions, check /dev/mali* /dev/ion14:28
T42<HengYeDev> crw-rw-rw-    1 system   system     10,  95 Nov  1 08:24 /dev/ion14:28
T42<HengYeDev> crw-rw-rw-    1 system   system     10,  92 Nov  1 08:24 /dev/mali014:28
T42<HengYeDev> @elros3414:28
T42<elros34> check also whether you have some uevent*rc file in /system or /vendor and compare with your /dev/14:29
T42<elros34> "/odm/etc/vintf/manifest.xml: Too many symbolic links encountered" also suggest something wrong with your /odm14:36
T42<HengYeDev> @elros34 perms seem to match- i get 0666 in the ueventd.rc14:37
T42<HengYeDev> where are you getting that odm from, i don't see that on my device14:37
T42<LSolrac> That repo didn't include it. But I was able to find one, thanks to someone in another group. This goes into the .repo/local_manifests/$devce.xml correct? (re @elros34: How can  you not fou...)14:37
T42<HengYeDev> @elros34 perms seem to match- i get 0666 in the ueventd.rc14:37
T42<HengYeDev> where are you getting that odm manifest.xml from, i don't see that on my device (edited)14:37
T42<elros34> @LSolrac it does:
T42<elros34> @HengYeDev from one of your logs and lack of mounted /odm from your repo14:39
T42<LSolrac> In that case how does one go about referencing a subdirectory in a repository? (re @elros34: @LSolrac it does: ht...)14:50
T42<elros34> set path to vendor/samsung, that is all14:53
T42<elros34> same as for android_device_samsung_star2lte you set path to device/samsung/star2lte14:55
T42<HengYeDev> @elros34 why is odm mounted at /odm2 in the script? what should i try15:01
T42<elros34> you should not copy scripts from random users15:02
T42<elros34> I guess its in /odm2 to created libandroidicus symlink and then content of /odm2/etc or something should be bind mounted in /odm/etc15:03
T42<elros34> but for some reason bind mount wasn't included15:04
T42<HengYeDev> after modifying the script to match for odm, droid-hal-init no longer crashes15:07
T42<HengYeDev> but now the libandroidicu is complaining again,  though not rebooting15:08
T42<elros34> so what you have in /odm/? Only etc directory?15:09
T42<HengYeDev> @elros34 yes, only etc and inside etc, build.prop,fs_config_dirs,fs_config_files,and selinux dir15:10
T42<elros34> so mount odm partition to for example /.odm_root, make in sparse /odm/etc then bind mount /.odm_root/etc to /odm/etc. Then you can add libandroidicu symlinks do rw /odm15:11
T42<elros34> you can see /odm it's handled also in vayu but with slightly different way:
T42<edp_17> @elros34 : I've built an image using OBS. That looks like fixed the GPS issue I had. Now, almost everything works on my treltexx 4.2. The only remaining issus is video recording hangs the camera app and it cannot be started again without rebooting the device. I've seen this is mentioned on collabedit.15:39
T42<edp_17> Thanks for the advice on building image using OBS! 😃15:40
T42<edp_17> What should I need to modify locally to build an image that installs boot.img ranther than hybris-boot.img?15:49
T42<edp_17> I've updated the hybris-updater-script but I need to copy the boot.img into ~/mer/android/droid/out/target/product/i9100/obj/ROOT/hybris-boot_intermediates to be ready for packaging. (I think.)15:51
T42<elros34> are you sure you want boot.img? IIRC it's android abootimg. Didn't you want some ZImage?16:26
T42<HengYeDev> @elros34 After i bind mount /odm/etc it bootloops16:36
T42<HengYeDev> dmesg:
T42<HengYeDev> the device stopped responding and rebooted before it could print the call trace, it seems16:36
T42<HengYeDev> logcat:
T42<HengYeDev> Journal:
T42<HengYeDev> is it bad that in the journal it says that selinux is denying a lot of operations from android things?16:36
T42<edp_17> Actually I want the file named "kernel" but boot.img is the same. (re @elros34: are you sure you wan...)16:38
T42<elros34> not sure, use file command on that16:39
T42<elros34> @HengYeDev why you still didn't add audit=0 to kernel bootcmd16:39
T42<edp_17> @elros34 : file boot.img: boot.img: Linux kernel ARM boot executable zImage (little-endian)16:40
T42<edp_17> file kernel : kernel: Linux kernel ARM boot executable zImage (little-endian)16:40
T42<edp_17> And theirs sizes are the same.16:41
T42<elros34> so zImage , isn't the same packaged here:
T42<edp_17> This is for i9100 not for the treltexx. Sorry, should have mentioned.16:42
T42<elros34> what the difference, i9100 will also have it16:43
T42<edp_17> Yes, but on treltexx, I don't need this as installing hybris-boot.img will work. On the i9100 I need to flash kernel (or boot.img) to KERNEL partition to make it boot. I just wanted to build an image that flashes the correct kernel into the KERNEL partition and then I don't need to flash it manually.16:45
T42<elros34> I get it just answered part about packaging which is already done16:45
T42<edp_17> Okay, I see.16:46
T42<HengYeDev> @elros34 seems that kernel commandlines are not being applied, i added audit=0, unpackimg says its there, but cat /proc/cmdline does not include it16:47
T42<elros34> Could be that bootloader overwrites boot arguments, have you tried adding them in defconfig or maybe disabling audit here will work:
T42<HengYeDev> when you mention defconfig does that mean CONFIG_AUDIT=0? doesnt that disable selinux16:59
T42<elros34> no, more like CONFIG_CMDLINE_EXTEND16:59
T42<elros34> when we talk about audit I do not see /etc/selinux in your sparse17:07
T42<HengYeDev> i put it in the rootfs manually after test, i followed hadkfaq (re @elros34: when we talk about a...)17:41
T42<BenKerry> Is there a port for Redmi Note 10 Pro?18:00
piggz_PSA, is back18:00
Thaodanpiggz: thanks18:30
T42<b100dian> +1 thanks Adam18:46
T42<HengYeDev> @elros34 ok, i turned it off in the audit.c20:16
T42<HengYeDev> there is a lot of avc: denied { open } and other selinux problems, could that cause the bootloop?20:16
T42<HengYeDev> I have already made the changes for selinux recommended in hadk-faq on the device, not in sparse yet tho20:16
T42<b100dian> Do the avc denied lines end in permissive=1 (or true)?20:21
T42<HengYeDev> yes20:23
T42<elros34> if selinux is in permissive mode than it shouldn't cause issues except  for spam in logs. You refuse to upload up-to-date repos so are you sure that you copy correct files not symlinks in etc/selinux?20:34
T42<HengYeDev> whoops, forgot that20:38
T42<HengYeDev> i'll get the repos up soon20:38
T42<HengYeDev> @elros34 Updated repos with all selinux stuff in sparse:
T42<HengYeDev> here is full logcat:
T42<HengYeDev> device reboots after a few seconds on blank screen22:14
T42<HengYeDev> dmesg:
T42<HengYeDev> seems to be with decon?22:17
T42<HengYeDev> or possibly  [32m[   30.586345]  [0m[@][ERR]fimc_is_vender_hw_init:605:Do not init hw routine. Check sensor failed! ?22:17
T42<elros34> so that kernel change didn't help22:21
T42<HengYeDev> doesn't seem so, i will try the cmdline in config22:22
T42<elros34> grep for printk.devkmsg=on, I think this was added manually to your kernel cmdline, maybe in dts22:29
T42<HengYeDev> yeah its there @elros34
T42<elros34> you have even androidboot.selinux=enforcing22:33
T42<HengYeDev> @elros34 in dmesg, selinux starts in permissive tho22:34
T42<elros34> not sure if that will help for current issue (device still fails with masked user@100000?) but looks like your mali driver needs

Generated by 2.17.1 by Marius Gedminas - find it at!