Tuesday, 2018-09-04

Thaodancan I replace packages that are on the SDK already when running build packages mw ? I need a newer version of ofono.05:55
mgrovermal: for android8, is it still using the audioflingerglue?07:04
mgroverabranson, if you have time, can you look at this please https://github.com/sailfishos/gst-omx/pull/407:28
abransonmgrover: are you sure you need to use gst-omx? Droidmedia should be better.07:51
mgroverabranson, i needed it for my nexus 5x and it worked perfectly :P07:51
abransonIt doesn't support passing the video buffers around like droidmedia can, so it'll never perform as efficiently as droidmedia, which passes the decoded video stream directly to the renderer.07:54
mgroverabranson, yeah but droidmedia didnt work :P07:54
*** mgrover is now known as ghosalmartin07:55
abransonThat should be workable all the way to 8 now I think.07:55
abransonBut i guess omx is an ok fallback.07:56
ghosalmartinabranson, okays ill give it a shot, but that pr still fixes the issue with it not linking07:56
Thaodancan I skip a pkg of a pkg when building a pkg with build_packages.sh/mb2?07:56
abransonghosalmartin: yeah I don't see why that shouldn't go in. I just strongly discourage omx use as it won't be maintained.07:58
ghosalmartinabranson, thats fine, i think I have the latest droidmedia anyways, so if I remove omx it should just work07:58
abransonI was in the middle of playing with allocators for it, the mtk came along and was far too different to qcom to even try :P07:59
ghosalmartinabranson, ahhh :P is mtk lost cause atm?08:09
ghosalmartinabranson, turns out i didnt even have ti installed anymore on this latest zip, gg :P08:22
abransonghosalmartin: it is on gst-omx. Droidmedia works fine though.08:23
ghosalmartini mean the bullhead image i built recently doesn't have gst-omx for whatever reason, so must be working off just droidmedia08:24
ghosalmartinguhl flashing a chiron zip atm08:53
ghosalmartinlbt: can I ask for two extra dirs in devaamo please08:53
ghosalmartina marlin and a chiron08:54
abransonghosalmartin: that's good news. I'll merge that.pr anyway - good to keep around as a fallback. For some reason the merge button doesn't work on my phone though, and my plane's about to take off, so I'll try to remember later :D09:01
ghosalmartinabranson, awesome, ta :)09:06
demonaxshi have a strange behavior, when i add wi-fi connection in wlan (in settings) all works, but after rebooting device saved connection not enabled , disable and enable again wlan not helped, if I delete connection and with create it again, all works... Essence of problem is that after saving connection and restarting, for some reason it does not turn on.09:29
malghosalmartin: tbr is the one maintaining devaamo09:34
r0kk3rzdemonaxsh: mac address changing?09:34
maldemonaxsh: do you have wlan as built in drivers in kernel or as a module?09:34
ghosalmartinmal: ahh thanks :) ever had a weird issue when zip is installing that it fails with error 7?09:34
ghosalmartinbut when I boot to first stage telnet and extract it mnaually it works fine09:35
demonaxshNo I did not change anything. On  same device of  same type there was no such problem.09:35
demonaxshI have no idea what could happen09:37
demonaxshno, wifi driver build not as module, and wifi works, not worked after reboot09:37
kimmolicheck mac address now and after reboot.09:37
kimmoliconnections are stored per device mac, so if it changes, they are not enabled. was this your first reboot? or does it do it repeatedly?09:38
kimmoliwhich device?09:38
malghosalmartin: not sure what error 7 means but some people have had issues using new twrp, the tar unpacking fails due too oom killer being too eager to kill it09:41
ghosalmartinmal: hmm could be related i guess09:42
maldemonaxsh: I would suggest trying using wlan as module, maybe the driver starts fine otherwise but some firmware related things are missing and it gets a random mac address09:42
ghosalmartinhmm when I attempt to boot into real rootfs it reboots. no logs though09:43
demonaxsh"wlan0     Link encap:Ethernet  HWaddr 00:08:22:54:CE:FB"09:44
demonaxsh"wlan0     Link encap:Ethernet  HWaddr 00:08:22:D6:CE:FB"09:44
demonaxshok... i try build as module09:44
kimmoli00:08:22 InproCom InPro Comm09:49
kimmoliit should be same what you see on Android side09:49
ghosalmartinis there anyway to debug why rootfs is breaking everything? :{10:40
demonaxshturn off randomization as a centos, can not?10:48
r0kk3rzghosalmartin: whatcha talkin bout willis?12:15
ghosalmartinr0kk3rz, i can boot to first stage telnet, ie access on port 23, but when I continue it reboots :/12:16
r0kk3rzmounts ok?12:16
malghosalmartin: have you tried disabling droid-hal-init etc?12:16
ghosalmartinmounts seem fine12:16
malghosalmartin: and is the rootfs properly mounted12:17
ghosalmartini can access .stowaways12:17
r0kk3rzthats a start12:17
ghosalmartini do actualyl have that12:17
ghosalmartinsome obsecure error12:17
ghosalmartinwith swapper12:17
r0kk3rzobsecure, nice word12:19
r0kk3rzsecurity by obscurity, or obsecure12:20
malghosalmartin: can you get to real rootfs by adding init_enter_debug2 to sailfish rootfs, it should stop the boot very early12:21
ghosalmartinmal: nah tried that with no luck12:22
ghosalmartinthe only way I can make it stop at telnet 23 is by renaming the sailfishos folder to sailfishosbak12:22
malhmm, can you double check init-debug script in rootfs12:25
ghosalmartinthe data partition or?12:25
malsailfish roootfs i.e. .stowaways/sailfishos or something like that12:26
r0kk3rzghosalmartin: boot with debug flag?12:34
ghosalmartinhmm maybe I should do that :P12:36
r0kk3rzthat should stop at first telnet without any shenanigans12:36
ghosalmartinyeah first telnet is fine12:37
ghosalmartinbut after that is the issue12:37
malghosalmartin: did you check init-debug12:37
maland maybe also init.log in rootfs12:37
ghosalmartininside /target is that meant to be rootfs?12:38
ghosalmartinah mines mounted the data partition directly12:39
malit's also there12:40
malthe mounts are confusing12:40
ghosalmartinso through target, my path to the rootfs is12:40
ghosalmartin /target/data/.stowaways/sailfishos12:40
ghosalmartinis that correct? or should it just be /target/12:41
malghosalmartin: check the link I just gave12:42
ghosalmartinso yeah its incorrect12:42
malshow the output of mount12:42
ghosalmartinso if I mount it under /target manually and continue will ti work?12:42
ghosalmartinno device12:43
ghosalmartinno dice even12:43
ghosalmartinlemme recompile kernel with debug flag12:43
maljust give the output :)12:44
ghosalmartinokays :)12:44
ghosalmartinmal: the mount works fine12:50
ghosalmartinnow /target has the rootfs12:50
malI can't help if I don't know exactly what is happening12:51
ghosalmartinokay these are the steps12:51
ghosalmartinI reboot to recovery, rename sailfishos to sailfishosbak to gain access to telnet on 2312:51
ghosalmartinthen when I reboot into telnet 23 i rename folder back to sailfishos12:51
ghosalmartinand then I call   echo "mount_stowaways" >/init-ctl/stdin12:52
ghosalmartinso sfos rootfs will be under /target12:52
ghosalmartinthen I run   echo "ls -l /" >/init-ctl/stdin12:52
ghosalmartinand reboot12:52
malstill waiting for the damn output of mount or even better init.log12:52
malbased on init.log it cannot find .stowaways/sailfishos in your data partition12:54
ghosalmartinyeah, when I fix that, reboots12:55
malso where is .stowaways/sailfishos if it's not in data partition?12:56
ghosalmartinthats the latest mount stowaways12:56
ghosalmartinis that not finding it and mounting it?12:56
malcan you answer the question, run find / -name .stowaways or something and se where in the hell the rootfs is12:57
ghosalmartinits there under the name sailfishosbak12:57
ghosalmartinotherwise bootloader won't stop12:57
maldammit, then say that clearly when giving the pastes12:58
malI thought you had some issue of wrong partition because it didn't find the rootfs12:58
ghosalmartinive been trying to say that 3 times now if you read up12:58
ghosalmartinand for whatever reason bootmode=debug doesnt work. it continues to boot12:59
malhow do expect me to debug any issues if the boot is intentionally broken by you12:59
ghosalmartinno other way to stop at telnet 2313:00
malso did you ever check if there is init.log in sailfish rootfs, not the telnet rootfs13:01
maland also the init-debug13:01
demonaxshfucking mediatek, can not be built wi-fi driver as module13:02
malsome devices have a /sys path to write MAC address13:04
ghosalmartinmal: no init.log, and what do I need to check inside init-debug? it exists13:07
ghosalmartinbut everything is a variable, not concrete?13:07
malghosalmartin: did you need any kernel hacks?13:08
ghosalmartinmal: not really, i just disabled DM-VERITY and CONFIG_VT13:21
ghosalmartinbut another device marlin booted into rootfs with the same hacks13:21
malI assume you disabled the kernel config check for CONFIG_VT in init-script?13:23
r0kk3rzi wonder if systemd is balking without that13:27
malbut init_enter_debug2 should stop before systemd13:29
ghosalmartinsystemd manages to start fine without CONFIG_VT on marlin13:30
Thaodanwhen I boot with hybris-intrid recovery image I get no usb telnet connection but with hybris-boot it works. Can I can change hybris-initrd someone to behave like hybris-boot on recovery?13:36
demonaxshor... patch kernel driver, I found where, though it's certainly radically13:39
Thaodancan someoe do ls /sys/class/android_usb/android0 for me on the xperia x?14:21
Thaodanit seems that this doesn't work with  4.4.x on sony-nile14:21
malsony nile needs newer usb-moded which is not yet included in latest sailfish release so you need to build it yourself14:26
malsame for all devices which use the new configfs based kernel usb drivers14:27
malusually devices which use 4.4 or newer kernel14:28
Thaodanah, good14:33
Thaodanhybris-boot supports configfs based usb-gadget would that be enough or is it used elsewhere?14:33
malit might be enough if it works well enough and cable is connected14:34
malbut that won't handle things completely14:35
Thaodanmtp support and so on needs it or not?14:35
Thaodanand hybris-intrid should use usb gadget mode instead when aviable?14:36
r0kk3rzhybris-boot probably wants patching for configfs, its not that hard14:37
r0kk3rzi wrote a configfs rndis script the other day14:37
malr0kk3rz: hybris-boot already has some configfs support14:38
r0kk3rzoh nice14:38
ThaodanI made a bug on this14:42
merbotMer bug 1950 in .Other "Hybris-initrd: support usb-gadget mode" [Normal,New]14:42
ThaodanI'll try to add the changes from hybris-boot to hybris-initrd later14:43
malThaodan: I thought that hybris-initrd doesn't even support telnet like the hybris-boot does14:44
maland it's not even supposed to because official device not need telnet in init14:45
mal*don't need14:45
malrecovery of course is different14:46
ThaodanI was talking about recovery. I was testing why I get a boot loop with the normal image and just found that rndis doesn't work at all14:53
ThaodanCan I use logcat or something a like to diagnose the bootloop?14:58
malhow soon does it reboot?15:09
ThaodanAfter Sony Logo15:23
malThaodan: how soon? seconds? tens of seconds or what?15:26
ThaodanSome seconds.15:27
malThaodan: you shouldn't disable audit in kernel, just check what was done to xa215:29
ThaodanOk,  is it used by some driver's or what?15:30
malThaodan: your PR has some not so nice things and obsolete stuff15:30
maland some things twice15:31
malhow about you try just applying the changes that were done to xa2 and see what happens, without any additional changes15:31
ThaodanSo the checks from the verify Script are somewhat outdated?15:31
malno, it just has things that not used on all versions, for example audit says that it's needed if you use selinux bootparam15:32
demonaxshto work bluetooth necessarily having rfkill? /dev/rfkill not exists in /dev, but:15:41
demonaxshwho controls bluetooth? connman?15:41
demonaxshyes, i know this, but usually with rfkill can enable/disabling bluetooth, and it's not part bluez15:49
ThaodanMal: ok what should I remove including the duplicates of course?16:00
Mister_Magistermal: hii do you have some time?16:49
malI was just about to start doing other things, making certain type of porridge which takes some time16:49
Mister_Magistermal: oh okay then16:58
Mister_Magisterhave fun lol16:58
vknechtbon ap├ętit :)16:58
malwell that is my breakfats for the next couple of days to I'm not going to eat it today, the active making is about 30 minutes or a bit more but I have to cool it at some point so the whole process takes at least 2-3 hours17:03
UmeaboyI'm willing to pay anyone in this room that can help me build a stable enough version of Lineage OS 15.1 (or 16.0 for that matter) that I can install and test on the Sony Xperia XZ Premium Dual G8142 $109,70 to a Paypal account. Who's up for it? I have CarbonROM already running on it, but that's encrypted as standard and I hate encryption on cell phones especially if they can't have the encryption disabled during the21:52
Umeaboyinstallation. The device is officially supported and gets nightlies.21:52
UmeaboyYou won't get it until I can confirm that it is indeed working and that it gets supported by Lineage. I've read the documention about using their review system to add new repos, but I prefer building the old fashioned way.21:54
malwouldn't building on top of sony aosp be a good idea?21:57
malthat device is part of open devices and still maintained21:58
Umeaboymal: Well, I thought CarbonROM is like having AOSP + Lineage in the same ROM.21:59
UmeaboyCarbonROM is an aftermarket firmware based on the Android Open Source Project created with the purpose of adding versatility and customization to stock Android.21:59
UmeaboyI guess it would be simpler and fast if libhybris could be built with CarbonROM if only I could get a decrypted build of CarbonROM.22:01
UmeaboyHaving encryption enabled seems to prevent me from installing Busybox to the right path.22:02
malwell I have no idea what codebase that rom uses22:02
UmeaboyIt just won't let me install to /system/xbin/22:02
UmeaboyI could be wrong, but it looks as if it's Android.22:04
malwell considering that device uses the same kernel as xa2 in sony aosp we could reuse the same repos almost completely, just need to do the mimimal patching of device repo and of course the defconfig in kernel22:07
maland adaptation repos could also reuse most of the stuff from xa222:09
Umeaboymal: Sorry, missed the last part.22:45
UmeaboyThe connection died.22:45
malcheck channel log linked in topic22:46
UmeaboyLineage doesn't have Sony's msm8998 kernel yet.22:50
UmeaboyAt least not officialy.22:50
malI consider aosp based build easier in this case22:52
UmeaboyIt's a bit outdated, but here's my git repo for maple to prepare for Sailfish.22:52
UmeaboyMy Lineage manifest.22:55
UmeaboyIt won't build, but it's a start.22:55
UmeaboyMy offer still stands.22:56
UmeaboyAnyway, I have to go to sleep. Take care.22:59
OrokuSaki$109,700 isn't a bad deal23:39

Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!