Thursday, 2019-11-14

r0kk3rz@eugenio_g7 any troubles?02:09
xreactx[m]__mal: hey, deathmist is helping me make leaps and strides... Can you look at my recovery log
xreactx[m]__Installer doesn't recognize cache (should be fixed with his hybris installer) and data. Also, do you know of any special instructions for treble devices?03:25
xreactx[m]__Also, if you have time my repo's are up to date on
xreactx[m]__"/system/etc/ld.config.txt": Too many symbolic links encountered03:31
T42Shaka1409 was added by: Shaka140904:03
T42<Shaka1409> anybody port sailfish on xiaomi mi5?04:04
T42<birdzhang> zhxt04:08
rinigusThaodan: that should make it easier for you when you get to boot xperia 1 -
riniguspiggz: thanks for your help ^07:00
rinigusI would like to make a separate sub-project at OBS which will contain system and vendor image packages for sony tama (starting with xz2)07:03
rinigusThat sub-repo will have repository (maybe later repositories) corresponding to android releases used as a base: android-9.0.0_r46 for now. It will be shared by devel and testing (so better locate it at testing). Packages will be added as RPMs, as we do now with HAL07:03
rinigusAs vendor and system images are built from AOSP, I presume that it should be OK license wise. We have docs describing how its done and link to source.07:03
rinigusSuch arrangement will allow me to have the full build at OBS (see for current issues).07:03
rinigusis this OK?07:03
rinigusmal , lbt : ^07:04
T42<birdzhang> @adampigg how did you fix sfos 3.2 reboot to recovery on mido? just set CONFIG_SECURITY_SELINUX_DEVELOP=n ?07:15
T42<adampigg> No, just added selinux=0 to cmdline07:16
T42<adampigg> Rinigus, my pleasure07:16
T42<birdzhang> okay, i have selinux=0 , but still reboot to recovery, maybe i need rollback something, thanks07:18
T42<adampigg> My clues were in pstore..from recovery, use adb shell to look at logs07:19
T42<birdzhang> init.log
T42<adampigg> Kernel log07:23
T42<adampigg> Are the new github actions equivalent to gitlab ci?07:25
T42<adampigg> Atleast, i think its a new feature07:25
r0kk3rznot as powerful i dont think07:26
T42<eugenio_g7> r0kk3rz: no... @adampigg did everything :D07:26
r0kk3rzah, the tried and true 'wait for piggz to fix it' method of development07:27
T42<adampigg> lol, i that the repotation im getting?07:28
T42<adampigg> @eugenio_g7 what is special about your new mesa?07:28
T42<adampigg> rinigus, great that youve upstreamed the bootctl fix, should it go higher, as affects multiple devices, tho, i ack that is probably difficult07:31
T42<eugenio_g7> r0kk3rz: :D07:35
T42<eugenio_g7> @adampigg: that it doesn't work currently! :D07:35
T42<adampigg> huh?07:35
T42<adampigg> u mean your new one isnt working i guess07:36
T42<adampigg> u r fixing tearing?07:36
T42<eugenio_g7> Yea my new one doesn't work...trying to fix the browser actually07:39
Mister_Magister@adampigg good for you~~07:39
rinigusadampigg: in general, this part of gpt-parted is a big guesswork and is not universal. right now it assumes that partition differs from its disk by having numbers in the end of the name. so, no, I don't plan to upstream it further...07:39
r0kk3rzdoes the nfc stuff require binderfu?07:40
T42<adampigg> @eugenio_g7 neat07:41
T42<adampigg> r0kk3rz, nfc on pro1 works with binder, not sure of another way07:41
r0kk3rzah ok07:41
T42<adampigg> i used it to write my email address to a tag07:42
Thaodanr0kk3rz: if we had the userspace part of the nxp stuff without binder I think not but I think that comes second07:46
Thaodanrinigus: thanks for the pr. I also found some issues in my boot image creation, some parameters aren't really visible in the build system.07:49
Thaodanrinigus: which guesswork,  where?07:53
rinigusThaodan: when you assume that sda123 belongs to sda. not sure how universal is that07:54
Thaodanwhy shouldn't it? I thought this was how device letters work.07:55
deathmist@nitin_3 yes, I fixed audio on 16.0 already. needed like and media/codec/mixer/policy configs in etc/ (some prob unneeded). then final step (for me at least) was to fix failing hw module audio.primary07:56
rinigusThaodan: example: /dev/mmcblk1p1 -> /dev/mmcblk . so, not always. its not the same type of storage, but still illustrates that its not as universal as one may think08:02
rinigussorry /dev/mmcblk1p1 -> /dev/mmcblk108:02
rinigusat least, I don't know what are the limitations and assumptions in device lettering to claim that its universal approach08:03
T42<adampigg> there is bound to be a way using /sys to find the parent device08:04
rinigusadampigg: that I believe. but then, we all prefer to cut the corners and get to the working device as soon as we can. don't we?08:05
rinigusif there is someone wishing to do that properly and thinking that its a good investment of time, please do so08:06
maldeathmist: can you check where your device has those xml files are in android side, probably under vendor/etc somewhere08:09
maldeathmist: for that audio symlink, maybe try this patch to libhybris08:11
T42<adampigg> rinigus, true, esp as you improved upon my version!08:17
deathmistmal: about your first message, they are symlinks, not sure exactly what you mean. there are probably many useless ones that get picked up from /vendor/etc/ anyway. about that libhybris commit, looks interesting. will definitely try when I get back home and see if it works without my "hacky" symlink :)08:22
deathmistalso, any clue about the dhd submodule update breaking my mic build I talked about yesterday?
maldeathmist: I mean check if the original androi xml files for codecs and stuff are in /system/etc or in /vendor/etc or /vendor/etc/audio08:37
T42<nitin_3> deathmist,mal08:38
T42<nitin_3> Audio works for me08:38
T42<nitin_3> if I flash sfos with xiaomi rom08:38
T42<nitin_3> instead of lineage08:38
T42<nitin_3> w/o changing anything on sfos side08:38
maldeathmist: I need to check the mic build, my own build machine is now down so I need to setup some environment to test local mic builds08:38
maldeathmist: usually android should load all those media xmls from /vendor08:39
maldeathmist: check if you have /vendor/etc/audio folder on the device08:40
maldeathmist: where do you have audio_policy_configuration.xml on your device08:41
deathmistmal: nope, no separate audio/ folder in etc/ of /vendor or /system. that file is on /vendor/etc/audio_policy_configuration.xml for me08:42
deathmistand I've looked into it before, it has includes with paths to /vendor/etc/... audio configs08:43
maldeathmist: there should not be need to link those to /etc08:50
malif everything is in /vendor/etc08:50
sledgesPSA: SFOS community collab meeting at #mer-meeting starting right about now:)09:02
Thaodanrinigus: Yes thats because the naming scheme is different for mmc and nvme devices.09:05
Thaodanits <device type><num> -> <device type><num>p<partnum>09:07
T42<adampigg> Mal, 1 accepted pr on dhv not merged09:32
mal@adampigg, you mean that native thing?09:33
T42<adampigg> Yes09:33
malI should check that again09:33
T42<布琪番 %lastname%> (Photo, 1219x693) "精准币圈菠菜棋牌粉" "非小号/mytoken惊爆价"1、币圈电报群/微信群炸群推广 2、twitter、facebook关注 3.国内外KYC身份材料出售  4、电报群/微信群/potato拉币圈活人 5、国内外交易所APP实名注册认证 6、电报群客服活聊天  7、批量注册国内外交易所实名账户,,出twitter、facebook、微信带粉10:44
Mister_Magisteriiis that a spam?10:45
T42<birdzhang> yup10:47
T42<dontatmepls> yeah thats a classic telegram spam account10:49
T42<dontatmepls> can someone take a look and tell me how i can fix this fscrypt10:50
deathmist@dontatmepls didn't you already have your kernel compile? are you trying another kernel? if so, it's broken and you'll have to fix the sources manually yourself, and even so it may still not boot if you use LineageOS sources etc13:50
T42<dontatmepls> I didnt change the kernel, but now i swapped it with another one and it seems to work fine13:50
deathmistmal: found the offending dhd submodule part seems you can't use "{$ONETHING,$ANOTHERTHING}" in grep like that, pattern_lookup is set correctly before too14:21
deathmistand the issue is indeed it trying to build a loop image instead of an fs one like it used to always before, seeing if I can come up with a solution14:23
maldeathmist: hmm, need to look into that14:23
deathmistmal: well that was quick, guess 'egrep "^- droid-hal-($DEVICE|$HABUILD_DEVICE)" $pattern_lookup' works :D14:26
deathmistyep and build works too14:27
*** birdzhang1 is now known as birdzhang14:27
maldeicide-: try that it works the other way also14:28
malautofill mistake14:29
deathmistmal: works too, sorry got disconnected cuz web client is bad :p should I make a PR or let someone else handle it?14:31
maldeathmist: make PR14:54
rinigus_lbt: reminder - please patternise nemo:devel:hw:sony:h821615:20
deathmistmal: done :) feel free to tell me if I did something wrong15:27
malrinigus_: should we have renamed that?15:35
maldeathmist: thanks, that was already merged :)15:42
lbtrinigus_: done15:44
rinigus_mal: yes, we should. but I wanted to get the first complete build and then see how to partition packages to support more variants. that would result in appropriate name as well15:52
rinigus_lbt: thank you very much!15:52
rinigus_lbt: shouldn't there be _pattern now at ? missing it...15:54
rinigus_(have to run now)15:54
malrinigus_: did you trigger some rebuild like config package?15:54
rinigus_mal: no, I didn't. sorry15:54
rinigus_thanks, will do15:55
T42<adampigg> Rinigus, patterns are so 201815:55
* lbt sees mal had that "hmm that's too quick for him to have triggered a build and it to have failed" thought too :D15:55
T42<dontatmepls> so now ive made a sailfishos zip but when i install it the fricking LineageOS base image boots, w h a t16:29
deathmist@dontatmepls then proper hybris modified kernel isn't booting16:38
deathmistflash hybris-boot manually if nothing else helps, you're booting the base LineageOS boot image for some reason16:39
T42<dontatmepls> i dont know anymore16:43
T42<dontatmepls> it still boots to LOS, and i dont know why, what do i do now16:43
T42<elros34> check in LOS (not in twrp) whether you have in boot params (cat /proc/cmdline or in dmesg)  skip_initramfs added.16:46
rodrigosolarii have a question16:51
rodrigosolariim building in platformsdk in arm arch, is ok16:51
rodrigosolaribut in x86 +same errors16:52
rodrigosolari relocations in generic ELF (EM: 40) this error16:53
rodrigosolariin any library for building x8616:53
T42<dontatmepls> @elros34 there is skipinitramfs inside dmesg boot params, now what16:53
T42<elros34> you need to revert commit which adds support for it in kernel. Show your kernel repo if you can't find it16:54
T42<dontatmepls> didnt find it16:55
T42<elros34> here: and maybe next one after that in same file if it fails16:57
*** OhYash1 is now known as ohyash17:00
rinigus_lbt & mal: yep, the _pattern is there - thanks!17:02
T42<dontatmepls> it cant be reverted because of conflicts17:02
T42<elros34> @dontatmepls maybe revert this one also
T42<dontatmepls> it also can't be reverted @elros3417:04
T42<elros34> I don't believe :P17:04
T42<dontatmepls> ill dm a screenshot so you do xD17:05
T42<elros34> it's latest 4 lines change in that file...17:05
T42<elros34> latest commit*17:06
rinigus_as my morning plan on images did not raise any objections, I'll proceed with it. let me know if it needs to be deleted17:09
T42<dontatmepls> @elros34 i managed to do it nvm17:11
deathmistmal: about that libhybris commit from earlier today, audio works after applying it and removing the symlink to :)17:13
maldeathmist: nice17:13
malkrnlyng: ^17:13
vknechtrinigus_, these few lines should be reassuring ;-)
T42<elros34> @dontatmepls you will also need this %define makefstab_skip_entries / in droid-hal-$DEVICE.spec17:17
rinigus_vknecht: so, packages as in system and vendor should be fine then. /odm is this binary package that we are expected to download from sony site, right?17:18
rinigus_vknecht: thanks! I will take care to minimize storage space as much as I can with them.17:20
rinigus_on obs17:20
deathmiststill no luck with cellular data on 16.0 having audit enabled too :/ not sure if something else is still broken. for sure RTC & proximity sensor too (works for 3s after unlock & 8s after blanking, then state goes to UNDEF as reported by mce)17:22
T42<elros34> I think this is normal if you use on_demand mode for proximity, still messwerk should report it correctly17:31
T42<dontatmepls> still booting to los after this17:39
T42<dontatmepls> ill try more tomorrow, pls send suggestions on how i could fix it17:39
deathmist@elros34 you're right! proximity seems to be ok then, that's great ^^ I just use the mce status from journal in my very bad attempt at a gesture handling "daemon" I honestly need to rewrite it in C / C++ or something but I still don't understand how MCE works with gestures like used here
T42<Lukapanio> just flash hybris-boot17:39
T42<Lukapanio> and it wount boot to los17:39
T42<elros34> @dontatmepls are you sure you use hybris-boot.img not boot.img?17:40
T42<dontatmepls> @elros34 [@dontatmepls are you sure you use hybris-boot. …], yees17:41
T42<elros34> so no way it boots los unless you forgot to make hybris-hal after reverting commits17:42
T42<dontatmepls> @elros34 [so no way it boots los unless you forgot to ma …], i did make it17:42
T42<dontatmepls> i can do a clean build tho17:42
T42<elros34> hold recovery button so bootloader will not add skip_initramfs to kernel params17:43
T42<dontatmepls> ahhhhh i did something idk what17:43
T42<dontatmepls> i made progress i think17:43
T42<dontatmepls> blackscreen and pc recognized some device17:44
T42<dontatmepls> ill try ssh and stuff now17:44
T42<elros34> it should show something about mer boat loader and available tlenet port17:44
T42<elros34> telnet*17:44
T42<dontatmepls> 3 Mer Debug telnet on port 23 on usb0 - also running udhcpd17:47
T42<dontatmepls> yeeees17:47
T42<dontatmepls> okay im connected via telnet now17:51
T42<elros34> if it's 23 port then something is wrong. I guess fixup-mountpoints17:53
T42<dontatmepls> i was able to connect17:54
T42<dontatmepls> no worrys17:54
T42<dontatmepls> im in it now17:54
T42<dontatmepls> ig ill send logs like dmesg?17:54
T42<elros34> check /init.log (not sure if it's available in initramfs)17:54
T42<dontatmepls> i got the log17:55
T42<dontatmepls> ill upload it and sedn17:55
T42<dontatmepls> thats the output17:56
T42<elros34> ok so it's 2323 not 23, so now dmesg and journalctl logs are important17:58
T42<dontatmepls> heres dmesg17:59
T42<elros34> standard: you forgot about selinux_bootparam in defconfig18:00
T42<dontatmepls> uhhh im 99% sure its there18:00
T42<dontatmepls> one sec18:00
T42<dontatmepls> okay i really forgot18:01
T42<dontatmepls> ill rebuild hybris boot with it, aight?18:01
T42<elros34> if you are changing kernel then you only need to rebuild kernel (hybris-boot)18:03
T42<dontatmepls> i know, doing it rn :)18:03
T42<dontatmepls> is it disabled now, for real?18:10
T42<dontatmepls> i added it to defconfig18:11
T42<elros34> nope: "SELinux:  Starting in permissive mode". Did you checked generated .config somwhere in out/18:12
T42<adampigg> @eugenio_g7 i rebuild d-c-l ... it was building the wrong branch18:14
T42<dontatmepls> @elros34 [nope: "SELinux:  Starting in permissive mode". …], i added these to the defconfig but they arent in the .config18:18
T42<dontatmepls> im lost18:18
T42<elros34> some config options have dependecies, show diff of your defconfig changes18:19
T42<dontatmepls> is this what ya wanted?18:24
T42<elros34> sure, -# CONFIG_SECURITY_SELINUX_BOOTPARAM is not set18:29
T42<elros34> +SELINUX_BOOTPARAM=y18:29
deathmist@dontatmepls every kernel config option starts with "CONFIG_", remove your incorrect "SELINUX_BOOTPARAM=y" line entirely and do "CONFIG_SECURITY_SELINUX_BOOTPARAM=y" instead18:32
deathmistthen make sure it is also "=y" in your .config => find out/ -name .config18:34
deathmistafter "mka hybris-boot" ofc18:34
T42<dontatmepls> [    0.002223] SELinux:  Disabled at boot.18:41
T42<dontatmepls> ayyo18:41
T42<dontatmepls> ok, whats next?18:41
deathmistjournal please :)18:43
T42<dontatmepls> aight18:43
T42<dontatmepls> it literally says no entrys, what am i doing wrong now18:44
deathmistthat's in telnet 2323?18:44
T42<dontatmepls> i need ssh, right18:45
T42<dontatmepls> gotta enable that18:45
T42<dontatmepls> i cant ssh after setting passwd18:46
T42<elros34> what are you doing? you don't need ssh, you have telnet18:46
T42<dontatmepls> "No journal files were found."18:48
deathmistsome parts of systemd are not running then, journald should be up. it can be manually started with "/lib/systemd/systemd-journald", but you probably have something still wrong in your defconfig18:51
T42<Lukapanio> set persist logging18:52
T42<elros34> wait for a while and grab full dmesg, you should have at least preinit message there according to your latest /init.log18:52
T42<dontatmepls> deathmist running your command outputs this: "Failed to get machine id: Input/output error"18:53
deathmistdo what elros34 said, your systemd prob just hangs and we can't do much with it now18:53
T42<dontatmepls> ill wait 2 mins18:55
T42<dontatmepls> heres a dmesg after waiting some time18:58
T42<dontatmepls> anyways ill continue tomorrow, id appreciate some suggestions as always :)19:06
T42<elros34> there is no preinit msgs but if sda31 is your data partition then at least it's /sbin/preinit is executed19:07
T42<elros34> you can add few echo_g  at the beginning of  main() part in /sbin/preinit19:09
T42<Lukapanio> hmmmm for treble device must i insatall generated vendor.img19:22
T42<Lukapanio> ?19:22
*** electro is now known as Guest859419:43
T42<Lukapanio> @hacker1245519:49
T42<hacker12455> No idea19:50
rinigus_as xz2 port has lvm, I do wonder whether storage is best split by ~2.4GB for / and 44GB for /home. I am considering increasing / to 4, as its only 1gb free after fresh install in current config20:00
rinigus_what's experience with official devices? is such "small" / sufficient?20:01
rinigus_like 2.5gb20:01
vknechtsurely it would do no harm (or even prevent harm) to pump it up20:19
vknechtespecially with that much room for home20:19
vknechtseeing the same 2.5Gb for /, the rest for home on unofficial X lvm build so it's just (tight) default I guess20:26
rinigus_vknecht: also /home can be somewhat "helped out" by sdcard as well.20:31

Generated by 2.17.1 by Marius Gedminas - find it at!