Friday, 2020-11-27

T42<elros34> also when in logs do you connect display?00:00
T42<elros34> I can't find hwc11_callback_hotplug report00:00
T42<simonschmeisser> xmn: the sfos image runs latest megi kernel, however some config options where missing and crust is not installed by default yet. config options where added yesterday evening, for crust see my hacky PR for flashit06:40
T42<simonschmeisser> also it sounds like you might want to update the modem firmware, see if you get a warning about that in dmesg06:41
xmnoh cool, good to hear!06:41
T42<simonschmeisser> also wifi is broken since yesterday evening :D06:41
xmnon your pr?06:42
xmnwell the latest vers wifi is broken06:42
xmnalso which sfos image are you referring to?06:43
xmnah I see for the flash-it upgrade gotcha06:45
T42Michael %lastname% was added by: Michael %lastname%09:47
T42Katherina %lastname% was added by: Katherina %lastname%09:48
T42Lillian %lastname% was added by: Lillian %lastname%09:48
T42Lewis %lastname% was added by: Lewis %lastname%09:49
T42Magnus %lastname% was added by: Magnus %lastname%09:50
T42Rawlings %lastname% was added by: Rawlings %lastname%09:50
T42my90ds was added by: my90ds09:54
T42Litz %lastname% was added by: Litz %lastname%12:25
T42<ankaos> where is the problem? and how fix?
mal@ankaos you should tell what problem you are trying to fix, it's not very simple to guess that from some random log14:44
T42<ankaos> device not open. and read initi rootfs not mounted15:36
T42<adampigg> @elros34 [maybe check whether your wlan modules support …], I grepped the source....15:38
T42<adampigg> #ifdef CONFIG_PLATFORM_ANDROID15:38
T42<adampigg>   char *if2name = "p2p%d";15:38
T42<adampigg> #else /* CONFIG_PLATFORM_ANDROID */15:38
T42<adampigg>   char *if2name = "wlan%d";15:38
T42<adampigg> #endif /* CONFIG_PLATFORM_ANDROID */15:38
T42<adampigg> oh, and this...15:39
T42<adampigg> module_param(if2name, charp, 0644);15:39
T42<adampigg> MODULE_PARM_DESC(if2name, "The default name to allocate for second interface");15:39
T42<elros34> or you could use modinfo wlan :)15:39
T42<adampigg> its so long since i made  a modprobe conf ... need to google how!15:40
T42<elros34> @ankaos "mount: /system: special device /system_root/system does not exist." Does your device use system as root? Check what  you have in: ls -al /system_root15:44
T42<ankaos> ls -al /system_root ->
T42<elros34> so looks like regular system not system-as-root so now check in droid-hal-device.log whether it use correct fstab15:47
T42<ankaos> this is true?15:52
T42<elros34> looks correct15:54
T42<adampigg> @elros34 im not sure there are enough options....
T42<elros34> indeed a lot of15:57
mal@adampigg so some module parameter issue?16:40
T42floraeo was added by: floraeo16:51
T42<adampigg> mal: well, maybe also a connman/wpa_supplicant issue worked around...... driver by default exposes wlan0 and wlan1, but should only use wlan0 for normal, non p2p use, but wpa_supplicant was trying to use wlan116:51
T42<adampigg> by using the module option, i was supposed to name wlan1 to p2p0, but, it actually dissapeared from ifconfig, so, only wlan0 now and connects ok!16:52
mal@adampigg I was told it might be possible to use also wlan0 and wlan1 what capabilities those report in wpa_supplicant16:54
T42<adampigg> just walking 🐶, will report back!16:55
piggzmal: you can see it trying to associate with pgznet on wlan1, not wlan017:13
T42<simonschmeisser> after enabling autosleep/suspend on pinephone it now suspends even during calls ... should mce do a wakelock for that automatically or would I need to add/write another module for it?17:22
malpiggz: check dbus-send --system --print-reply --type=method_call --dest=fi.w1.wpa_supplicant1 /fi/w1/wpa_supplicant1/Interfaces/0 org.freedesktop.DBus.Properties.GetAll string:fi.w1.wpa_supplicant1.Interface17:26
malthen same for interface 117:27
malhmm, both return same caps17:31
mal@simonschmeisser is wakelock support enabled in kernel?17:33
T42<simonschmeisser> mal: megi showed the same yesterday17:33
T42<simonschmeisser> mal: yes, enabled17:34
piggzmal: ifaict, wlan1/p2p is for wifidirect type stuff17:35
malpiggz: probably, there should also be option somewhere to make connman ignore some interface based on name17:37
malpiggz: NetworkInterfaceBlacklist in /etc/connman/main.conf maybe you edit that or optional give connman -I parameter to ignore a specific interface17:39
piggzmal: im ahead of you, done that and its working17:40
T42<simonschmeisser> mal: I can succesfully inhibit suspend with eg pure maps17:40
malpiggz: which wya did you use?17:41
piggzmodified main-native.conf (what we use on pinephone)17:41
malpiggz: what other customizations do you have in the config?17:42
T42<simonschmeisser> hmm, ssh does not inhibit but should also do it?17:42
piggzmal: remove usb from the same line, as the usb net is usb017:44
piggzi think android uses rndis0 ?17:45
wdehoog@mal what would be the best way for me to look for the cause of all the EGL errors I see in my logcat (when using latest libhybris)?18:22
riniguswhen updating port to 3.4.0, is there any package that should be not updated to the latest? like nfcd was earlier, for example?18:37
piggzrinigus: sensorfwd plugin, use 0.11.918:44
riniguspiggz: but that doesn't have my compass fix. the one we are using on tama for a while already18:46
piggzrinigus: ask mal then, thats what he told me :)18:47
riniguspiggz: yes, I remember, he did. will have to grep history and see what was the reason18:51
riniguslooking at commits, I may as well stay with 0.11.8 + my patches for tama (compass and pickup sensor)18:55
T42<floraeo> Fwd from Recap:
rinigusfrom scanning logs, no idea what is broken newer sensorfwd. maybe mal can explain...19:00
malrinigus: sensorfw should be 0.11.919:04
malrinigus: the hybris/binder plugin19:04
mal.8 should be fine also19:05
malrinigus: anything in logs about sensorfw19:05
rinigusmal: no logs yet, I am going through packages at OBS and preparing for update. on 3.3, I am running .8 + rotation vector (merged to upstream now) + pickup sensor (tama specific) just fine. compared to 0.12.1, that just doesn't have one dbus commit19:07
rinigus... commit from upstream. and your fixes for older androids19:07
rinigusbut ok, I will stick with my current version then19:08
malrinigus: what fixes to older androids?19:13
malah, that one19:14
malrinigus: maybe you should then make a custom branch with 0.11.9 + the rotation vector19:14
rinigusmal: that would be needed only when switching to aarch64. the diff is just this:
malyeah, 0.11.8 should be fine also19:18
rinigusbut surely I will be interested to try to switch over to aarch64. just not this weekend :)19:18
malrinigus: aarch64 is not yet perfect, I have issues with video recording, I know in theory how to fix it probably but I have some issues with that on my device19:19
rinigusmal: good luck with fixing it!19:20
malotherwise it works quite well19:20
rinigussounds promising. I will keep it in mind and try to join aarch64 in foreseeable future19:25
rinigusThaodan: you switched to your own dhd fork in, is it still needed?20:14
rinigusnote that tama is so far configured with its own "meta" packages support and I wonder whether/how those mods and upstream will clash20:15
T42<adampigg> rinigus: i switched pro1 and volla to the new meta-packages,  it wasnt too hard, an lets you delete a few packages from obs20:17
rinigusmust say that I hate to redo something that was already working. meta is there already for tama, would have to see what's the difference. maybe later tonight...20:19
T42<adampigg> rinigus: basically, move your package .specs to pattern/ and rename .inc ... include both in the .spec, and some edits to change some names in the files20:26
T42<adampigg> because it gets included, you dont have to build the other specs individually20:26
malwdehoog: do you see any other erros than egl errors in logcat, do things crash?20:42
wdehoogsome 'unrecognized pixel format: 0xb19909c9'. 10 min ago I reinstalled it. lots of egl errors, then lipstick crashed and restarted and now no more egl errors. huh ????20:44
T42<ankaos> elros and mal: thank you20:47
T42<ankaos> but not use toucpad20:47
wdehoog@mal sorry. see ^20:48
wdehoogis it good/required/bad to  install parts of sailfishsecrets in a port?20:49
malwdehoog: so it fails only on first boot?20:52
wdehoogOne moment.....20:53
malwdehoog: did startup wizard work?20:53
wdehoogstartup wizard worked. after reboot egl errors are back20:55
T42<ankaos> which ı use? sh-3.2# ls /dev/input/20:57
T42<ankaos> by-path  event0   event1   event2   event3   event4   mice     mouse020:57
wdehoogat boot the 3 bottom icons on topmenu are invisible. after 'systemctl --user restart lipstick' they become  visible20:58
wdehoogand no more egl errors20:59
mal@ankaos you should not need input device in compositor config
malthat should be enough21:02
malwdehoog: quite odd issue, can you get logs from such failing situation, journal and logcat, as soon as possible after boot so we can any early messages21:03
T42<ankaos> mal: thanks ı testing21:04
T42<ankaos> mal: not work touchscreen21:05
malso then set it manually, you need to check either manually using cat to see which input device outputs something when you touch the screen or use evdev_trace -t21:06
malthat evdev_trace is from mce-tools package21:06
wdehoog@mal don't know if this is usefull. part of the log I saved  when lipstick crashed:
T42<ankaos> mal:
T42<ankaos> touchscreen not work21:08
wdehoog@mal logcat: and journal:
mal@ankaos then you need to provide full journal log21:12
T42<ankaos> mal: journalctl output:
malwdehoog: so does it fail right away or after a moment?21:23
T42<ankaos> mal:
wdehoog@mal after reboot topmenu does not display the 3 bottom icons, system feels sluggish and egl errors appear in logcat (not always from the start but in the end the always appear)21:33
T42<ankaos> ı dont found error21:33
mal@ankaos where is the kernel source repos for your device21:35
malwdehoog: how does it work after you restart lipstick, does it stay stable for long time?21:36
T42<ankaos> mal:
wdehoog@mal yes it seems to be quit stable. launched lots of stuff and nothing bad happens . you know some stress test?21:38
wdehoogonly these appear: qdmemalloc(15211): unrecognized pixel format: 0xb291f9c921:40
UmeaboyHopefully someone here can answer my question I asked yesterday.21:40
T42<ankaos> mal: problem is kernel source21:40
UmeaboyIF my phone (it currently runs an unofficial build of LOS 17.1 and the HADK is written to use AOSP as base. Problem is, there is no Android 10 base for my phone on github so how can I instead use the HADK with LOS and get the patches working?21:42
malwdehoog: I was just wondering about if there was some fence leak or something but doesn't seem like it21:42
UmeaboyIt's the Samsung Galaxy S10 Plus.21:42
UmeaboyNot a Sony phone.21:42
T42允棕号 %lastname% was added by: 允棕号 %lastname%21:43
UmeaboyHere's the repo used for the S10 Plus:
malUmeaboy: why do you keep claiming hadk is based on aosp, it's not, hadk pdf is based on lineage/cm21:44
UmeaboyWell, do I just need the device and kernel repo for the patches to work or do I also include the rest of them in that roomservice.xml that whatawurst made?21:45
malthe hybris-17.1 patches are not public yet, there are aosp-10 patches though, I haven't pushed my hybris-17.1 stuff anywhere21:46
UmeaboyOK. aosp 10 is made for Sony devices, right?21:46
malmaybe I should try to rebase my hybris-17.1 this weekend21:47
UmeaboyMy phone in this case is a Samsung Galaxy S10.21:47
malneed to add more disks to my build machine to make space21:47
malI don't have enough space to build anything now21:47
Umeaboymal: Have you heard of this company?
UmeaboyYou did mention you lacked of space.21:52
malwell I have 2x10 TB disks waiting for installation21:53
malrinigus: ping, I noticed a typo in your sensorfw PR, you wrote pkconfig instead of pkgconfig in one of the commits couple of times22:21
malrinigus: did you ever test the built-in compass in sensorfw?23:11

Generated by 2.17.1 by Marius Gedminas - find it at!