Friday, 2019-10-18

xreactx[m]__is this a problem? "Skipping repository 'adaptation0' because of the above error."01:58
T42<birdzhang> it's normal, ignore it02:03
xreactx[m]__I thought that was the root of another problem...02:06
xreactx[m]__is that a "straggler file"?02:10
T42<birdzhang> cd hybris/mw/nfcd-binder-plugin && git reset --hard 1.0.402:26
xreactx[m]__do I need to execute git pull after that?02:28
T42<birdzhang> nope02:32
xreactx[m]__ problem still there03:00
T42<birdzhang> oh, git Fast-forward03:09
T42<birdzhang> you can manual build it after git reset --hard 1.0.4, and comment nfcd-binder-plugin in rpm/dhd/helpers/build_packages.sh03:11
T42<birdzhang> rpm/dhd/helpers/ -b hybris/mw/nfcd-binder-plugin03:12
T42<birdzhang> for example:
xreactx[m]__Thanks, I'll have to look into it tomorrow.03:39
xreactx[m]__birdzhang: thanks03:39
T42<adampigg> mal: looking at res bug, easyish to debug, added the imageCapture resolution to the label06:25
T42<adampigg> only happens on back camera!06:25
T42<adampigg> a good reason for my last idea for the ui to present the actrual state and not the setting value?06:25
Mister_Magisterdoot doot08:03
deathmistmorning, mal: so after the "something to test" stuff was ran I built a new zip listed and now systemd hangs on "starting" state for quite a while but in the end DHI still fails & *servicemanager services seem to get ended by a SIGKILL, another journal update
malhmm, something went wrong then08:15
malcheck what the first few lines of command git log in system/core shows08:16
deathmistmal: is what the first 4 shown commits say08:19
maldeathmist: are you sure you correctly disabled selinux on the device? via defconfig + kernel commandline?08:22
malok, based on the log that should be ok08:23
deathmistmal: I should've, this is how it also booted perfectly fine on hybris-15.1 w/ cmdline defconfig
malverify selinux_stubs is in the servicemanager rc files in /usr/libexec/droid-hybris/system/etc/init/08:28
deathmistmal: the files look ok to me and /usr/libexec/droid-hybris/system/lib64/ definitely exists on the device08:42
deathmistI still have a feeling my disabled_services.rc disables too much stuff and I'm going to experiment around a bit with that08:58
malbut if servicemanagers fail then the issue is something else09:00
malwait, you have qti disabled? that looks so wrong09:00
deathmistyeah it used to just fail and get restarted on 15.1 constantly so caused battery drain, it booted fine without it tho but I'll enable it again09:01
deathmistmal: nope, still not booting even with only disabled (and ofc I import init.qcom.rc manually on top becuause I disabled it in main init.rc)09:28
deathmistis it bad to have this in journal "selinux: SELinux: Loaded file_contexts" ? is comes up after "droid-hal-init: init second stage started!" after which I also see a few lines like "droid-hal-init: Forked subcontext for 'u:r:vendor_init:s0' with pid 3368"09:34
T42<birdzhang> deathmist: how do you get telnet work? i have same issue with you. "/etc/udhcpd.conf: No such file or directory"09:40
T42<birdzhang> init.log:
deathmist@birdzhang it is very finicky for me still even if I manage to get it, seems to work 100% of the time on first boot, other boots are basically a roll of the dice. not sure if it's an issue is the cable, device or my host. about those udhcpd.conf init messages I still get them even when telnet works fine09:42
T42<birdzhang> okay09:44
deathmist are my current defconfig options with which I can get telnet on the OnePlus 5 (cheeseburger)09:44
T42<birdzhang> thank you09:45
deathmistoh I also made a "blind build" for OnePlus 5T (dumpling) on hybris-15.1 from mostly the same sources, now someone tested it on XDA and apparently it "worked perfectly" :D  (I don't own the 5T) so could be worth looking into Jolla Store etc after verifying MAC addresses & IMEI's stick with reboots etc09:56
deathmistmal: have you found anything yet that could cause servicemanager to fail? I could strace it too in a bit if that helps13:22
maldeathmist: sorry, I haven't even looked at that yet13:23
malstill at work13:23
deathmistoh ok, no problem13:23
nitin03I need some help porting sailfishOS to my device...I was able to build sailfish os package.13:43
nitin03and flashed it after lineage os...But was unable to boot into sailfish os ...I tried to debug but.not able to as I can't telnet into my device13:44
nitin03as in guide it says..If you connect your device and run dmesg on host it should say mer bootloader but in my case it says google device13:45
malsome basic information needed first, which device? which android base? does the device reboot after some time?13:45
nitin03xiaomi poco f113:45
nitin03lineage 16 base...Yeah it boots up gets stuck on manufacturer logo...and stays like that and then reboots again and gets stuck on same logo...this goes on and on13:47
malhow long does it take for it to reboot?13:47
nitin03I cant say exactly but about 3 to 5minutes13:48
nitin03Is there any other way to know what is failing13:55
maldo you see anything in host dmesg about usb during that time?13:58
nitin03yeah i checked It should show mer bootloader..but it shows as google device...also i checked shows nothing,13:59
nitin03It shows things that I should See if I am booting android14:01
malbut is only one message or does it show any changes during that time14:04
nitin03on host shows things like vendor etc...14:07
malI'm interesting if something changes there during the time the device is on and also when the reboot happens14:08
nitin03i did pull dmesg from my shows avc denied errors...but i disabled it in by putting selinux=0 in kernel cmdline14:09
nitin03ohh..i will pay attention to it next time14:10
maldmesg from device? how?14:10
nitin03booted into twrp..and it has an option to store dmesg and recovery to sdcard ..then i pulled it from there using adb pull14:12
malbut isn't that the dmesg from recovery, not from sfos14:12
malnitin03: you might want to check the sfos rootfs in /data/.stowaways/sailfishos in recovery and check if that had init.log14:13
T42<BusterBg_18> nitin03: Let me see your sailfish defconfig14:13
nitin03ok thank you14:13
nitin03wait a minute14:13
malthat rootfs would tell if it even got to sfos14:13
malalso better show the changes you did to fixup-mountpoints file14:14
T42<y_so_seriouss> @deathmist [oh I also made a "blind build" for OnePlus 5T …], I would like to test on my OnePlus 5T, please share the download link of files14:19
maldeathmist: didn't the other servicemanager work before the changes I asked you to make? which makes very little sense though14:22
nitin03sorry i got disconncted14:31
deathmist@y_so_seriouss check follow flashing guide etc carefully14:32
nitin03did you check the mountpoints?14:35
T42<y_so_seriouss> @deathmist [@y_so_seriouss check foll …], Thanks a lot for bringing Sailfish OS for OnePlus 5T :)14:37
T42<adampigg> mal: looks like ive fixed the resolution bug15:01
T42<adampigg> via code removal15:01
mal@adampigg nice15:01
T42<adampigg> mal: i will PR later and we can be sure15:02
T42<adampigg> Vknecht:15:04
T42<adampigg> (Photo, 1080x1920)
T42<adampigg> Comment 215:04
T42<adampigg> 1 has a pr, 3 is likely super difficult!15:05
vknechtah ! there are always interesting ways to make things fail :-)15:05
vknechtwill check, on the videoprobe atm15:07
T42<adampigg> sucks that the propery way of frame grabbing doesnt work, and the PR looks dead15:08
vknechtthe videofilter ? well, in fact the probe seems more appropriate since no need to edit the frame15:09
vknechtbtw if I understood correctly, seems the frame would no be editable due to missing support in gstnemotexture or something15:10
vknecht*editable anyway15:11
vknechtnot that we need it right now I guess, but still good to know15:11
T42<adampigg> yeah, it seems none of that is properly implemented15:11
T42<adampigg> on a demo appi have, i draw on the frame using qpainter15:12
T42<adampigg> on a desktop tho, not mobile15:12
T42<adampigg> mal, vknecht, in portrait, how about having the res/timer above the shutter?  reason not to have it at the top, is that area auto fills with the menu buttons, so different devices have different areas15:15
T42<adampigg> above shutter would be consistent15:16
mal@adampigg that will be a pain because of how the rotation is currently done15:17
vknechtconsidered keeping at the same place, but written vertically ?15:18
malI think that would be terrible to read15:18
malvknecht: especially in video mode when it has also the time15:18
vknechtmaybe, but it's not so critical information ; but yes, with time it's worse15:19
vknechtbtw shouldn't time be reset as soon as recording stops ?15:19
malwe need to rethink the UI at some point15:19
mal@adampigg do you think it's ok to replace the "Checking supported:" with "Found support for" to minimize debug spam15:31
T42<adampigg> Sure15:32
malI'm doing some code cleanup15:32
mal@adampigg also found a new improvement, hiding the menu icons if there are no such modes, iso modes are often missing on new android versions15:45
deathmistmal: I don't remember and I don't have the old build anymore :/ should I revert the changes and how? I'll definitely keep this current build too if we need it again15:49
maldeathmist: I think the previous code was wrong, not sure what the issue is now15:50
malThaodan: ping16:08
SimdolHey, has there been an attempt at porting SOS to oneplus 6?16:31
SimdolIn particular, with the new A/B partition scheme and project treble, is porting possible?16:31
Thaodanmal:  yes?16:33
ThaodanSimdol: just ignore ab layout or use it to extend the system partition.16:34
SimdolThaodan: Okay, that sounds feasible. For the 'base' of Android, any AOSP based rom is supported right?16:37
Simdoli.e Android Pie works with HADK?16:37
malThaodan: about the droid-hal-device PR, not sure I understand what issue you have with -b16:45
malSimdol: lineage os or aosp are fine, up to android 916:46
malandroid 9/ lineage 16.0 needs an additional command before build but otherwise the same16:46
Simdolmal: is that command documented somewhere?16:54
malwell it is in channel logs17:05
Simdolmal: well, what command is it?17:06
Simdoli don't think i know where to look for that17:06
mal"hybris-patches/ --mb"17:06
malbefore doing make hybris-hal17:06
Simdolmal: ah, so just gotta apply that patch?17:07
malit will apply many patches17:07
Simdolalso for some odd reason, i can't do  'sudo zypper ref' in a chroot of platform_sdk17:08
Simdolit hangs, one of the core is at 100%17:08
Simdolcan't terminate it either :/17:08
maldoes it output anything?17:10
Simdolmal: nope17:11
Simdolmaybe i should run with -v to see if it helps?17:11
Simdolsudo zypper -v ref17:12
SimdolVerbosity: 117:12
SimdolInitializing Target17:12
Simdolwarning: rpmdb: Freeing read locks for locker 0x7: 3821/413720000017:12
Simdoland it hangs there17:12
malwhat step in build process are you doing?17:12
Simdolmal: just trynig to setup the build environment17:13
maldid you try exiting the sdk and going back17:16
Simdolmal: actually done that multiple times17:17
Simdoli think double chrooting might be the cause of this17:17
malyou are in two chroots?17:22
Simdolmal: yes.17:22
Simdolmal: i needed a clean build environment, it's ran in lxc container17:23
Simdolwell, i will try one more time outside the container to see if that works17:23
malsfos build happens in chroot already so the host doesn't matter really17:24
Simdolmal: i figured, i just like having them separated so that i can nuke and remake easily if i have to17:26
Simdolalso allows for really easy cloning17:26
deathmistSimdol: if you want a clean environment I recommend just making a new user on your actual host specifically for building stuff w/ bash as the shell17:26
deathmistthen you can just e.g. "su - builder" and off you go17:27
Simdoldeathmist: yeah, that's what i just did17:27
Simdolhmmm Repository 'adaptation0' is invalid.17:27
Simdol - Store credentials not received. Not connected to D-Bus server17:27
deathmistyou can ignore most "errors" unless builds fail :p17:28
Simdoldeathmist: i figured :P17:28
Simdolokay, which repo do i need for my device? i see hybris-16.0 (which i assume is los based), but if i want to use an aosp rom, do i clone that instead?17:34
Simdolcan't i just git clone
malSimdol: aosp porting is not as straightforward as lineage porting, mainly because there is no ready manifest for it, is there something wrong with the lineage os for that device?17:39
T42<Akatsu %lastname%> How make works under HADK? Do hybris-hal execute hybris-boot aswell?17:39
malhybris-hal build also hybris-boot and other things17:40
Simdolmal: the support for los for my device is not prefect17:40
Simdolmal: omnirom on the otherhand, has much better supprot17:40
malSimdol: well if you can find the android manifest used for omnirom then it might be possible17:40
T42<Akatsu %lastname%> Guide for SFOS isn;t very idiot friendly, halium have it done much better17:40
malimprovement ideas are welcome17:41
malI haven't looked at halium porting instructions17:41
malwhere are those17:41
T42<Akatsu %lastname%>
T42<Akatsu %lastname%> there is no need for sdk,chroot etc17:42
Simdolmal: is this it?
T42<Akatsu %lastname%> pray for me to not fail on building17:44
T42<Akatsu %lastname%> Are tooling SDK meant for app development? I remember seeing in HADK to not install them17:45
T42<BusterBg_18> I find the HADK fine as it is,  besides halium only builds images or I am missing something?17:47
T42<Akatsu %lastname%> yes17:47
*** OhYash1 is now known as ohyash17:47
T42<Akatsu %lastname%> its only for building images. systemimage and hybris-boot/hallium-boot17:48
T42<Akatsu %lastname%> sadly its stuck on 7.1 so before android got some unified stuff17:48
T42<Akatsu %lastname%> will it ever be possible to build Treble SailfishOS and just have to manualy build custom kernel image?17:50
malSimdol: looks like it17:52
malsailfish already uses treble in many places but devices aren't that simple, a lot of configuration changes etc17:53
mal@adampigg starting to think of running automatic code formatting to qml code is a good idea, we have too many things that do not match Qt style guide17:54
T42<adampigg> mal: you tell me off when i do that!!!17:55
Simdolmal: can you help me decipher what .local_manifests/$DEVICE.xml values are?17:57
mal@adampigg so what should we do, makes git blame more difficult but would make code more consistent, I already fixed c++ files17:57
malSimdol: you find those in omni.dependencies files in your device, you should also check the projects linked there for additional dependencies files17:58
Simdolmal: where is the omni.dependencies located?17:59
Simdolmal: ah of course. *fale palm* thank you so much18:00
Simdolmal: wait, that's not an xml file. so do i just simply repo sync?18:01
vknechtthat qml formatting, maybe it's doable from a commit hook ?18:09
malSimdol: no, you create the xml file based on the information in dependencies file18:19
malSimdol: hadk pdf should have an example which you should be able to use to determine what goes where18:19
Simdolmal: so repository = name; revision = revision and the rest i can ignore?18:24
Simdolpath = remote?18:25
malpath = target_path obviously18:28
T42<Akatsu %lastname%> not really sure but is treble about these binder interfaces?18:30
T42<Akatsu %lastname%> so in short vendors are pushed to make universal binaries which will like expose devices in specific format to be accessible by standarized api?18:31
mal@adampigg are the CameraImageProcessing.ColorFilterEmboss and related additions already in qtmultimedia?18:31
rinigusas I have to make (or at least it looks like it) vbmeta image, I am stumbling into error while making vbmetaimage in hadk (
mal@adampigg looks like 3.1.0 has those, do we want to support older sfos version in cam app?18:33
rinigussame will happen if I try to make systemimage vendorimage18:33
riniguslooks like hybris patches don't allow it...18:33
T42<Akatsu %lastname%> im idiot, can anyone tell me how im supposed to get flashable zips? from make  hybris-hal i got hybris-boot.img and folders with system and root, what im supposed to do later?18:36
malrinigus: was that android 9 base? did you apply patches?18:36
malcontinue reading the hadk pdf, make hybris-hal is just the first step in the build18:37
T42<Akatsu %lastname%> oh scratchbox sorry18:37
rinigusmal: sorry for skipping specs. yes, its aosp9 with applied hybris patches.18:38
piggzmal: yes, 3.1 only, im happy to support 3.1+18:39
piggzalothough, harbour doenst let me choose that as an option18:39
Simdolmal: having a hard time setting up that .xml file "fatal: remote error: Git repository not found"18:41
Simdolit tries to clone from when i need it to clone from github repo18:41
malSimdol: so you didn't set proper remote for those18:54
maladd remote="omnirom"18:55
Simdolmal: got it! i also had quite a few errors, someone should consider adding as documentation19:00
Simdol*to the documentation pdf19:00
malvknecht: the automatic formatter does some a bit strange formatting when ternary operators are used19:07
deathmistmal: seems my dumpling (OnePlus 5T) builds work fine, sources are like 95% identical to my cheeseburger (OnePlus 5 non-T) ones. what would be the best way of doing OTA with the 2 almost same models? the only packaging differences are ofc the dcd, dhd and dvd packaging19:14
Simdoldeathmist: huh, that's pretty cool19:22
Simdolyou managed to port sos 3 to 5/5t?19:23
Simdoldeathmist: apart from hwdk's instructions, did you have to do anything special to get stuff to work? loaded question ik, just curious19:24
piggzmal: argh, what is going on!19:27
deathmistSimdol: I looked at other device (droid-config-*) repos and picked up stuff that I needed e.g. media and audio configs etc, some you'll have to get working manually with device/SOC specific symlinks. either way here's a scuffed screenshot of OP5T running SFOS :)
deathmistbranek would be proud lol19:31
Simdoldeathmist: ah, cool. i'm trying to sos to op6, so this should be interesting...19:33
maldeathmist: are dumpling builds also 16.0?19:34
malpiggz: ?19:34
piggzmal: onCameraStatusChanged ....19:35
piggzapplies settings only when in ActiveStatus19:35
piggzexcept, it doesnt19:35
piggzit calls the function in the StartingStatus, and it def shouldnt19:35
deathmistmal: basically when I get them to work on cheeseburger too lol, for now it's working perfectly fine on 15.1 as is cheeseburger. the kernel sources etc are the exact same! and I can just sync and adapt droid-config and such sources from 5 non-T19:35
piggzmal: no, im wrong...19:37
maldeathmist: you said the dumpling build works but you don't have working cheeseburger build19:37
piggzits just lost in all the debug19:37
deathmistno I do, both work fine on 15.1 but 16.0 is equally broken on both (haven't bothered to test dumpling on 16.0 as I don't own the device)19:37
maldeathmist: I got confused which build was based on which version19:39
deathmistthe sources are in sync basically, once I get hybris-16.0 working on cheeseburger it will also work on dumpling which is kinda awesome :)19:40
T42<BusterBg_18> hey mal: guess what? I now have telnet on xiaomi_sirius (hybris-16 base)19:55
T42<BusterBg_18> And as expected I dont have GUI19:56
deathmistBusterBg_18: DHI ( droid-hal-init) most likely fails for you too as it does for me currently on 16.0 (check systemctl status)19:56
deathmisttho I think you'll also need mount units for /system and /vendor if they are not mounted (see mount command or cd in there and ls to check)19:57
malmaybe I should actually try using hybris-16.0 myself :)19:57
maldeathmist: can you recheck the mounts are ok19:58
maldeathmist: in case you rebuilt the image and didn't add those to config repo yet19:58
mallooks like you have those on github20:00
T42<BusterBg_18> I just checked init.log seems like my fixups are missing20:05
T42<BusterBg_18> Now that I remember I did a repo sync before saving them20:05
deathmistmal: yeah I do and they work fine on current 16.0 build
mal@adampigg which android base does fxtec use?20:07
maldeathmist: try adding this file to your device
maland this
malalthough that might have some issue as that is from aosp20:11
maldeathmist: instead of that last link copy the file from /system/etc/ld.config.28.txt and add the libexec parts the same as in the link I gave20:12
T42<Akatsu %lastname%> How long does "Building rpm/libhybris.spec" takes?20:14
T42<Akatsu %lastname%> i said y to build libhybris after build_packages.sh20:14
T42<BusterBg_18> It takes a little bit longer than the other middleware packages in my experience20:15
T42<Akatsu %lastname%> so should i use all on prompt?20:16
T42<BusterBg_18> No20:16
T42<Akatsu %lastname%> i only accepted libhybris and now i get libglibutil20:16
T42<adampigg> mal: pie20:17
T42<BusterBg_18> Check out my messages asking about the middleware that fails/needs git reset on the channel logs20:17
vknechtpiggz, sent a fix for gallery delete, hope it's fine with you ; there's an imperfection in that when deletion happens when previewing another media, the preview switches to another media, but that's a different (unseen previously) problem (gallery index reordering I guess)20:26
vknechton another subject, wondering if QVideoProbe is also faulty, I get signal ( ) but can't get QVideoFrame to QImage saving for now20:36
vknechtugly, non rebased patch for videoprobe attempt :
vknechtvoid FrameGrabber::emitVideoFrameProbed() being the critical part20:43
vknechthope you all enjoyed this historical meme video XD20:45
vknechtanother one, an hommage to all developers, with love from MS XD :
deathmistmal: help, it's doesn't make any sense xD dcd commit result on system
deathmistand just to be clear, DHI shows "Process: 5200 ExecStartPre=/bin/sh /usr/bin/droid/ (code=exited, status=0/SUCCESS)"20:51
deathmistso the mount apparently "works" but it doesn't do anything :/20:52
deathmistand the shown ld.config.28.txt was originally in my /system/etc/ dir, I just added the 2 libexec lines to it like on ganges-pie repo21:01
maldeathmist: can you check if the mount worked21:06
maldeathmist: no difference in logs?21:07
deathmist:o droid-hal-init is up \o/21:09
deathmistnvm it failed now .-.21:10
maldeathmist: same error?21:12
deathmistmal: guess so, servicemanager services still die. doesn't look like the mount works from (even after letting it sit for minutes after powering on) but if I run it manually after it works fine, why could this be? I can give new journal21:14
T42<BusterBg_18> After adding once more my changes to fixup-mounpoints, do I need to rebuild everything or only hybris-hal?21:14
mal@BusterBg_18 depends on what you mean by everything, hybris-hal and -d are needed21:16
T42<BusterBg_18> I was refering to packaging it once again21:17
maldeathmist: maybe try the ugly way, backup the original file from /system, mount /system with write permissions and copy the modified file21:17
deathmistmal: that definitely did something good, *servicemanager services don't die now it seems and DHI is actually alive, now only I still get no graphics, systemd is left in a degraded state (dev-cpuctl.mount dev-stune.mount sys-fs-bpf.mount & systemd-tmpfiles-setup.service fail)21:22
deathmistfollowing journal I see some RTC errors spammed and vendor.dataqti seems to die and get restarted constantly21:25
deathmistoh yeah date shows "Fri Jul 26 06:38:48 EEST 2019" oof21:25
malthose failures are ok21:26
malI mean the earlier ones21:27
malnot sure about vendor.dataqti21:27
maldeathmist: check if logcat works21:27
deathmistseems to work fine21:28
malanything about graphics in there21:28
malhwcomposer or stuff, also check journalctl21:28
T42<BusterBg_18> mal: how can I check if I pointed fixup to correct userdata partition?21:57
deathmistBusterBg_18: pretty sure if you get a "/data/.stowaways/sailfishos/init.log" then that is taken care of21:59
malone way is to check the ramdisk init somewhere under out/22:00
T42<BusterBg_18> I get:22:00
T42<BusterBg_18> mount: mounting /dev/block/bootdevice/by-name/userdata on /data failed: No such file or directory22:00
T42<BusterBg_18> And a pretty ominous Debug session : Failed to boot init in real rootfs22:00
malsomething like out/target/product/$DEVICE/obj/ROOT/hybris-boot_intermediates/initramfs/init22:01
mal@BusterBg_18 why not show your fixup-mountpoints changes and also give link to the fstab of your device repo22:02
mal@BusterBg_18 where do you see that error?22:02
T42<BusterBg_18> Dont worry about the missing "@" symbol I already have that
T42<BusterBg_18> On init.log in telnet22:03
malyou should already see an error in the fixup22:04
mallook at the the second line in that diff and then look at the end of what you added22:05
deathmistcould it be failing because userdata is F2FS by default in that fstab? I've only used ext4 on SFOS22:08
deathmistand so is cache too22:08
T42<BusterBg_18> The "$@"?22:08
mal@BusterBg_18 yes, make sure the end of your additions looks exactly the same as the previous one22:09
T42<BusterBg_18> I already have that, just not commited :p22:09
mal@BusterBg_18 come on, why would you show me obsolete code22:09
maldo you expect me to find what is wrong that way22:09
T42<BusterBg_18> Sorry22:11
T42<BusterBg_18> It slipped off my mind22:11
maldid you check that out/target/product/$DEVICE/obj/ROOT/hybris-boot_intermediates/initramfs/init22:12
malit should have the actual device node near the top22:12
maldeathmist: sfos installer script used in recovery has hardcoded ext4
maldeathmist: so we assume everyone formats data to ext422:15
deathmistah, out of curiosity would it work with F2FS if I got rid of the check?22:15
T42<BusterBg_18> Past the Running mer Boat Loader I have DATA_PARTITION=/dev/block/bootdevice/by-name/userdata22:16
maldeathmist: maybe22:18
mal@BusterBg_18 so fixup doesn't work then22:18
T42<BusterBg_18> So I suppose it should say /dev/block/mmcblk0p81?22:19
maltry running make hybris-hal again and see if thereis any difference, does the build report what it found from fstab, it tells it quite near the beginning22:19
T42<BusterBg_18> It says out/target/product/sirius/obj/ROOT/hybris boot_intermediates/init Fixing mount-points for device sirius22:23
deathmistmal: could line 2777 in journal be an issue for gfx? here's also logcat
mal@BusterBg_18 it should saud something about found data at somewhere22:28
malsimilar for boot22:28
malactually "/data appears to live on"22:29
T42<BusterBg_18> And after that its just kernel stuff22:30
T42<BusterBg_18> Should I make clean and see if it changes anything?22:30
malyou still have the wrong one in that init file?22:32
T42<BusterBg_18> Yep22:32
T42<BusterBg_18> I just found something, theres another sirius on fixup-mountpoints22:33
T42<BusterBg_18> Can that be the problem?22:33
malit uses the first one it finds22:34
malthat other one is for sony's sirius device22:35
malno idea why xiaomi decided to use a codename that is already being used22:35
maldeathmist: can you should output of "ls -l /" from device22:36
maldeathmist: also output of "mount"22:36
T42<BusterBg_18> Ok, I just deleted it currently rebuilding hybris-hal22:38
T42<BusterBg_18> Success!!!22:39
T42<BusterBg_18> It now shows DATA_PARTITION=/dev/mmcblk0p8122:39
maldeathmist: verify that /mnt/vendor/persist has some reasonable content23:02
deathmistmal: look... reasonable to me?23:03
mallooks ok23:04
T42<BusterBg_18> So now my device led lit up23:04
T42<BusterBg_18> It no longer complains about not finding /data23:04
deathmist@BusterBg_18 that's good... now can you get in with "telnet 2323" ?23:06
T42<BusterBg_18> Indeed23:06
T42<BusterBg_18> With only 23 It refuses connection23:06
deathmist23 is only used during very early stage debugging (i.e. not in SFOS rootfs)23:07
T42<BusterBg_18> Ahh, I didnt remember that23:07
maldeathmist: try EGL_PLATFORM=hwcomposer test_hwcomposer23:08
deathmistas nemo or?23:08
deathmistah, "library "" not found" and segfault23:08
maldoes the device that that anywhere23:09
deathmistin /vendor/lib64/ only, isn't this bad? :(23:10
malnot sure23:10
malin logcat " SDM     : HWCColorManager::CreateColorManager: Successfully loaded"23:10
deathmistmal: here's strace too for test_hwcomposer23:15
T42<BusterBg_18> Here is the latest init.log
mal@BusterBg_18 if you are in telnet 2323 then init.log is useless, you should be looking at dmesg, journalctl and sometime /usr/libexec/droid-hybris/system/bin/logcat23:18
T42<BusterBg_18> As I tought, well here is dmesg
maldeathmist: let's try a very ugly way, copy this file (click view raw to download it) to /usr/libexec/droid-hybris/system/lib/libsdm-disp-vndapis.so23:29
mal@BusterBg_18 was that also hybris-16.0 base? if so you might need to do similar things that deathmist did23:30
T42<BusterBg_18> Yep its also 16.023:30
malmost likely some issues with mounting and other things23:30
malbut it's quite late now so maybe tomorrow23:30
T42<BusterBg_18> No prob, my ass is starting to get sore from sitting so its also a good idea for me to take a break, thanks for your help as always mal c:23:33

Generated by 2.17.1 by Marius Gedminas - find it at!