Saturday, 2026-04-11

Mister_Magisterfixed led (just needed path)00:00
Mister_Magistervendor.dataqti is bootlooping, i wonder if i need it00:01
malis mobile data working?00:03
Mister_Magisteryeah00:03
Mister_Magistergot 40Mb/s on 5G00:03
Mister_Magisterehh i probably do need it00:04
Mister_Magisteraaand no suspend00:04
Mister_Magisterthere doesn't seem to be autosuspend at all00:05
maldo you have then autosleep kernel option set?00:05
Mister_Magisterno00:06
Mister_Magisterby the looks of it00:06
Mister_Magisterah it's autosleep yeah no it's not enabled00:07
malhttps://github.com/mlehtima/android_kernel_fairphone_qcm6490/blob/hybris-23.2/arch/arm64/configs/vendor/fp5-qgki_defconfig#L103600:07
Mister_Magisterye ye i know i know00:07
Mister_Magisterit's not my first device i had to enable it on :)00:07
Mister_Magisteri lost count but it's probably nearing 20th device00:08
malI think android doesn't use autosleep anymore00:08
Mister_Magisterbrother it doesn't use autosuspend probably since like 8 years ago00:08
Mister_Magisterwhen i mentioned autosuspend to android dev they were like "what?"00:09
T42<the_hiktor> mal , what did i miss ?🙃 https://hastebin.com/share/edufuzonot.bash00:13
Mister_Magisteri had same issue and gave up on building locally instantly xd00:13
T42<the_hiktor> what would be the difference if i used a server regardless of building speed00:20
T42<the_hiktor> and other cool shit00:20
mal@the_hiktor too new droid-config submodule00:21
T42<the_hiktor> mal, so.... what could be done  here ?00:22
Mister_Magistergo back couple commits00:23
malgo to hybris/droid-configs/droid-configs-device and run "git reset --hard c67fe5b47b81714c1338c39d5d2547f528933b15" then back to $ANDROID_ROOT and run build_packages.sh -c00:23
Mister_Magistermal: the phone, despite having 120Hz screen, runs in 60Hz :( fps display says 60. ofc not debugging it now just saying00:25
T42<the_hiktor> mal, dude https://hastebin.com/share/jixebahuhe.less00:33
T42<the_hiktor> wrong commit ig00:34
maldifferent error00:39
maladd lines 14-17 and maybe also 22-24 to your droid-config spec https://github.com/mlehtima/droid-config-fp5/blob/master/rpm/droid-config-fp5.spec#L1400:40
malthen again build_packages.sh -c00:43
T42<the_hiktor> 😐 https://hastebin.com/share/kurogacowa.yaml00:51
malsb2 -t $VENDOR-$DEVICE-aarch64 -R zypper in droid-local-repo/$DEVICE/droid-config-surya-1*.rpm -ofono-configs-binder00:54
Mister_Magistermal: do you remember the config to tell jolla camera i got two back cameras or at least label them? i lost hat somehow00:54
malMister_Magister: what did I say earlier :)00:55
Mister_Magisterfp5 :)00:55
malhttps://github.com/mlehtima/droid-config-fp5/blob/master/sparse/etc/dconf/db/vendor.d/jolla-camera-hw.txt00:55
T42<the_hiktor> No update candidate for 'droid-config-surya-1-202604110047.aarch64'. The highest available version is already installed00:56
malok, so the error is something else then00:57
malsb2 -t $VENDOR-$DEVICE-aarch64 -R zypper in droid-local-repo/$DEVICE/droid-config-surya-sailfish-1*.rpm -ofono-configs-binder00:59
T42<the_hiktor> yup that was the one :D01:02
Mister_Magistermal: where the hell did you get ofono-binder-plugin-ext-qti from01:19
Mister_Magisterayy vibra working01:21
Mister_MagisterOkay i think that's bulk of the fixes01:28
Mister_Magisteroh you're building it01:31
Mister_Magisteroh wait it's not the jolla's01:31
Mister_Magisteryou're using ubports one on fp501:31
Mister_Magistercarry on01:31
Mister_Magistermal: turns out neither stock camera nor advanced camera sees second back camera o_O but video recording works02:15
Mister_Magistervolte registered instantly after adding ext-qti :)02:17
Mister_Magisterokay so what's left not working is: wide angle camera, mic in calls, charging indicator, sleep, fingerprint02:23
Mister_Magisterif i'm not mistaken that's literally it02:23
Mister_Magistermaybe nfc but idc, same as fingerprint i'm not doing in-screen fp02:23
Mister_Magisterohhhhhhhhhhh usb networking i forgot02:24
Mister_Magisterusb netwroking, charging limit, ui sounds volume02:25
Mister_Magisterbut like the big features work flawlessly already02:25
Mister_Magisterbro in one evening the port is literally 90% done02:26
Mister_Magisterthere's only one service bootlooping and it's vendor.dataqti02:28
Mister_Magisterwait it might not be sleeping because… because it thinks it's charging02:30
T42<the_hiktor> mal, Failed to execute %pack script with /bin/bash07:20
T42<the_hiktor> wait i forgot this07:51
T42<the_hiktor> Info[04/11 07:07:02] : Running pack scripts ...07:51
T42<the_hiktor> /parentroot/mnt/Imp/hadk/SailfishOScommunity-release-5.0.0.62-surya-my1 /parentroot/mnt/Imp/hadk07:51
T42<the_hiktor> mv: cannot stat 'update-binary': No such file or directory07:51
T42<the_hiktor> Info[04/11 07:07:02] : Script returned with non zero status, failing.07:51
T42<elros34> are you sure make hybris-hal doesn't show errors especially related to hybris-boot part?09:20
T42<elros34> why did you comment out installable_zip?09:36
T42<the_hiktor> 🤦‍♂️10:16
T42<the_hiktor> Yeah no errors related to hybris boot10:16
T42<the_hiktor> Lemme try again10:17
T42<the_hiktor> well that was dumb 🥲10:56
T42<the_hiktor> but it finished so well😄10:57
Mister_Magistermal: are you up for some more debugging?12:28
Mister_MagisterKeto: IIRC you're the guy to ask about jolla store access no?12:31
Ketoyes12:57
Ketohttps://github.com/mer-hybris/hadk-faq?tab=readme-ov-file#jolla-store-access13:00
Mister_Magistersomehow missed this13:07
Mister_MagisterKeto: https://paste.opensuse.org/pastes/5538d828d586 :)13:09
Ketoadded13:12
Mister_Magisterthank you kindliy :)13:17
Mister_Magistermal: i tried building halium thingy but it doesn't seem to compile, maybe it doesn't have a15 compatibility yet13:54
Mister_Magisterhttps://paste.opensuse.org/pastes/11ca98fcf37a13:57
malah, I haven't pushed my changes anywhere yet13:59
Mister_Magisterand i was reading into the c code14:01
Mister_Magisterseems like I don't have to? :P14:01
Mister_Magisteroh mal about volte calls, it works but funny issue, I can't hang up! is that somehow normal or it's your first time seeing that?14:01
maltry this https://github.com/mlehtima/android_vendor_halium_hardware/tree/halium-15.014:02
Mister_Magistertrying14:05
Mister_Magistermal: call mic fixed! but there are two issues remaining: can't hang up, and even if other person hangs up call never ever ends. And proximity sensor seems bugged during call, it keeps turning my screen off even if i press power button, i tried on demand and i thought it fixed it but it didn't14:34
malhmm, found some odd thing in the volte plugin14:40
Mister_Magisterhm?14:41
malhttps://gitlab.com/ubports/development/core/hybris-support/ofono-binder-plugin-ext-qti/-/blob/main/src/qti_ims_call.c?ref_type=heads#L33214:41
Mister_MagisterI'm sad to admit that doesn't tell me much14:41
Mister_Magisterit completes only if it's complete?14:41
malthe code doesn't use the volte hangup14:42
Mister_Magisterhuh14:43
Mister_Magisteridk what's that14:43
Mister_Magisterbut like on oneplus 6 it worked fine14:43
Mister_Magisterah14:43
Mister_Magisterso we need to implement it <eventually>14:43
malso now it seems your device needs that14:44
Mister_Magisteruh-oh14:44
Mister_Magisterwell, i have no idea how to go about implementing it14:44
malmost of the code is there, just might have some bug14:45
Mister_Magisterwhat are we doing then chief?14:45
malmaybe try removing that return 0 and see what happens14:46
Mister_Magisteroh the dcd already provides ondemand proximity sensor config14:47
Mister_Magisteri shall do that then14:47
Mister_Magisterbut do you have idea as to why it blanks the screen?14:47
Mister_Magisterwould be easier to try to hang up if i can press hang up14:47
maldoes csd show proximity sensor as working?14:48
Mister_Magisteryes14:48
malmaybe check journal during a call and see if you see anything there14:50
Mister_Magisterwill do14:51
Mister_Magisterext-qti is building14:51
Mister_Magisterhttps://github.com/VerdandiTeam/ofono-binder-plugin-ext-qti/commit/20d36eb6f745b1ba2e1414a863356e3fbd96516a14:51
Mister_Magisternope, didn't fix it14:53
Mister_Magisterhttps://paste.opensuse.org/pastes/91b75c9af3c514:54
malcheck logcat for binder errors14:54
Mister_Magisterhere's journal, with the proximity stuff and hanging up14:54
malI wonder that is happening, you see that ims:imsradio0: failed to parse incoming SMS data14:54
Mister_Magisteralso the vendor.dataqti seemingly wants /dev/rmnet which it doesn't find and then exits but that's beside the issue14:55
Mister_Magisterhttps://paste.opensuse.org/pastes/0c0700830a6d logcat14:57
Mister_Magisterit has some "undelivered transaction 47841, process died" but idk if that counts as "binder errors"15:00
Mister_Magisterit could be the vendor.dataqti15:01
Mister_Magisterthe vendor.dataqti is doing this stuff as I mentioned https://paste.opensuse.org/pastes/e0344184e21e15:02
Mister_Magisterit really wants /dev/rmnet_ctrl15:02
maland there is no such device?15:03
Mister_Magisternope15:03
Mister_Magisterbut there is on android15:03
malsome module not loaded?15:04
malor some init not done?15:04
Mister_Magisteri thouhgt we loaded all of them hold on i'll verify agianst android15:04
malany systemd services failing? did the libhybris fix the getprop/setprop things?15:04
Mister_Magisterno services failing, yes it did i think15:05
Mister_Magisteryeah rmnet modules are clearly missing15:05
malit could cause all kinds of issues15:05
malcheck all other modules also15:05
Mister_Magisteri shall15:06
Mister_Magisteri'll grep for the rmnet module and maybe i'll find other modules.list or whatever it was called15:06
malmodules.load I think15:07
Mister_Magisterout/target/product/miami/vendor/lib/modules/modules.load has the rmnet stuff15:07
Mister_Magisterbut my modules-load already has the rmnet modules15:08
Mister_Magisteron the list15:08
malbut don't get loaded?15:08
Mister_Magisteri think so yes15:08
malbut module loading system service succeeds?15:09
mal*systemd service15:09
Mister_Magisterrmnet_core and rmnet_ctl loaded15:09
Mister_Magisterbut rmnet_offload did not15:09
Mister_Magisterbut i can load them manulally15:09
Mister_Magisterbut i can load them manually15:09
malbut still no device node?15:09
Mister_Magisternope15:10
malanything in dmesg?15:10
Mister_Magisterit's hard to tell which ones are msising15:10
Mister_Magisterthe modules i mean15:10
malabout symbols or modules15:10
Mister_Magisterhttps://paste.opensuse.org/pastes/f3a85c10f89415:10
malmaybe grep kernel sources for rmnet_ctrl15:11
Mister_Magisteri'll ask guy with same phone on android about dmesg and maybe i'll see something15:11
malwhere is the kernel source on github?15:11
Mister_Magisterokay i compared the lsmod and now only lzo_compress and lzo_decompress are missing15:13
Mister_Magisterso i think we are good in term of modules15:13
Mister_Magisterhttps://github.com/VerdandiTeam/android_kernel_motorola_sm6375/tree/lineage-22.215:14
mallooking at the code it's usb f_gsi.c which should be creating the missing device node15:16
malhttps://github.com/LineageOS/android_kernel_motorola_sm6375/blob/lineage-22.2/drivers/usb/gadget/function/f_gsi.c#L179315:17
malmaybe gsi function needs to be loaded in configfs15:17
Mister_Magister# CONFIG_USB_CONFIGFS_F_GSI is not set15:18
Mister_Magisterwhich would be funny as it wouldn't be on android either15:19
malhmm15:19
malok15:19
Mister_Magisterbut i can just =y it15:19
malmaybe not if android doesn't have it15:19
malthen it's something else15:19
Mister_Magisteri'll verify still15:19
Mister_Magisteroh15:20
Mister_Magisterandroid HAS it enabled15:20
malhow did it get disabled?15:20
malyou disabled it15:21
Mister_Magisteri… i did?15:21
Mister_MagisterI did15:21
Mister_Magisteri copied usb stuff from op6 without looking15:21
malhttps://github.com/VerdandiTeam/android_kernel_motorola_sm6375/blob/lineage-22.2/arch/arm64/configs/vendor/ext_config/moto-holi-miami.config#L16115:21
Mister_Magisteryeap15:21
malnot sure about the other changes also15:21
Mister_MagisterI left only the CONFIG_USB_F_RNDIS=y CONFIG_USB_CONFIGFS_RNDIS=y15:21
Mister_Magisteryeeted the rest15:22
Mister_Magisterin that code block15:22
Mister_Magisterokay next on the list of annoying stuff will be, charging indicator is broken15:23
Mister_Magisterit always thinks it's charging15:23
malany of the power_supply nodes showing the correct state?15:23
malin sysfs15:23
malmce has support for blacklisting broken ones15:24
Mister_Magisterwhat do you mean by state? which file is that15:24
malstatus maybe15:24
malsomething that says Charging or something15:24
Mister_Magisteroh wait i cat'd all the files and see two discharging and one charging15:25
Mister_Magisterlet me see which ones are those15:25
maljust fyi the blacklisting is done with [BatteryUDevDeviceBlacklist] and then next line is name like this based on the last part of power_supply path "something = true"15:26
malhttps://github.com/mer-hybris/s19mps-droid-configs/blob/master/sparse/etc/mce/70-battery-s19mps.ini15:27
Mister_Magisterbattery: discharging, bms charging, charger discharging, mmi_battery discharging,15:27
Mister_Magisterso we blacklist bms15:27
malbtw, maybe check for the charging control if you have something like in that15:28
Mister_Magisterwe were already checking for it yesterday no?15:28
Mister_Magisterand we didn't find anything15:28
malbut look at the path in that example15:28
malit's probably somewhere but it's just a bit hidden15:29
Mister_Magisteri don't have stop_charge or anything like that15:29
Mister_Magisterbut first blacklisting15:29
Mister_Magisteri'm not sure what to write there15:30
malwhat is there in charger.0 if you have it15:30
Mister_Magisteri don't have it15:30
malbms = true15:30
malmaybe15:30
malcheck the uevent file in bms15:30
malor what was it15:30
Mister_Magisterhttps://paste.opensuse.org/pastes/112dad4b45a715:31
Mister_Magisteri added bms = true and restarted mce and it's still charging15:31
malshow the whole config file15:32
malalso is usb connection detection working correctly?15:32
Mister_Magisterhttps://paste.opensuse.org/pastes/02841660af7b15:32
maldoes it use extcon or android_usb thing15:32
Mister_Magisterusb probably doesn't work at all as i masked usb-moded15:33
Mister_Magisteri didn't yet get to usb-moded15:33
Mister_Magisteri think that should be next step then15:33
Mister_Magisteri think elros was mentioning some different solution to networkign than rndisk last time i mentioned it15:33
Mister_Magisteri need to dig that up15:33
Mister_Magisterah it was adam15:34
Mister_MagisterNCM is the thing we want15:34
Mister_Magisterlet me try15:34
Mister_Magisterkernel built15:38
Mister_Magisterhmm i don't have /sys/kernel/config/usb_gadget/g1/configs/b.1 only /sys/kernel/config/usb_gadget/g1/configs/c.115:39
Mister_Magisterand the usb-moded config wants b.115:39
malso you change it15:39
Mister_Magisterokay15:39
Mister_Magisterbut c.1 has rndis.usb15:39
malhttps://github.com/mlehtima/droid-config-fp5/blob/master/sparse/etc/usb-moded/usb-moded-fp5.ini#L315:40
mallike that but with c.115:40
malb.1 is used on lvm based builds, c.1 in community style builds15:40
Mister_Magister /sys/kernel/config/usb_gadget/g1/UDC: write failure: No such device15:41
Mister_Magisterfor reference i literally copy pasted this https://github.com/HelloVolla/droid-config-halium-mimir/blob/master/sparse/usr/bin/droid/setup-configfs.sh https://github.com/HelloVolla/droid-config-halium-mimir/blob/master/sparse/etc/usb-moded/99-usb-moded-config-mtk.ini15:41
maldo you have anything under /sys/class/udc/15:42
Mister_Magisteri will tell you in 20 seconds once it reboots15:43
Mister_Magisteri saw setup-configfs.sh also has b.115:43
Mister_Magisterso i changed it and rebooted15:43
Mister_Magisteryes it does have /sys/class/udc/ and it's non-empty15:44
malcheck the usc things in that script15:45
maland see if the paths exist15:45
Mister_Magisterwill do15:45
Mister_Magisterthe /sys/kernel/config/usb_gadget/g1/UDC is empty file and i suspect it's not supposed to be15:47
malthe script or usb-moded is supposed to write there15:48
Mister_Magisterso "ls /sys/class/udc | grep -v dummy | head -n 1" returns 4e00000.dwc3, and writing that to /sys/kernel/config/usb_gadget/g1/UDC returns bash: write error: No such device15:49
Mister_Magisterso script does it just fails i think15:49
malare there others than 4e00000.dwc3 in the folder?15:49
Mister_Magisteryes15:49
Mister_Magisterthere's dummy_udc.015:49
maldoes getprop androidboot.usbcontroller return anything?15:50
Mister_Magisterno15:50
malanything in dmesg when you do that echo?15:50
Mister_Magisterudc 4e00000.dwc3: failed to start g1: -1915:51
malcheck getprop for anything else related to usb, and kernel cmdline also15:51
Mister_Magisterandroidboot.usbcontroller=4e00000.dwc315:51
malmaybe the usb kernel config changes broke it?15:51
Mister_Magisteri mean not sure it was ever working xd it's first time im trying to do udc not rndis15:51
Mister_Magisteror i mean15:51
Mister_MagisterNCM15:51
maludc is a different thing15:52
Mister_Magistergetprop: https://paste.opensuse.org/pastes/6d6c215c4de215:52
Mister_Magisteryeah sorry15:52
malbut did you revert the kernel config changes and booted that already?15:52
malrelated to usb15:52
Mister_Magisterno not yet15:53
Mister_Magisterit's building on obs15:53
malcontinue debugging after you have done that15:53
Mister_Magistertho i don't have the scripts for flashing kernel15:53
malI'll be away for an hour or more15:53
Mister_Magisteroki15:53
Mister_Magisterill fix the OTA15:54
Mister_Magisterdo you have any tip to leave before you leave about the ui sounds?15:55
malnot at the moment15:55
Mister_Magisteroki odki15:56
malmaybe check logs for ngf related things15:56
maland pulseaudio15:56
Mister_Magistershall do15:56
Mister_Magisterso i'll fix ota, and debug usb networking/charging15:56
Mister_Magisterand the vendor.dataqti15:56
Mister_Magisterafter that we'll see if suspend works15:56
Mister_Magisteri'm not fixing fingerprint cause that's in-screen15:56
Mister_Magisterand beside that the only thing left will be wide angle camera not showing up but i want camera2 api anyway so maybe that'll fix it15:57
Mister_Magisterand ye ui sounds15:57
Mister_Magisterthat's the plan15:57
Mister_Magisteruh-oh /proc/cpuinfo doesn't have Hardware line16:01
Mister_Magisterahh it's also using /sys/firmware/devicetree/base/model which returns miami16:04
Mister_MagisterI shall use that16:04
Mister_Magisterso gsi is =y now but still no rmnet16:18
Mister_Magisterhttps://paste.opensuse.org/pastes/a57796d1330b Hmm16:21
Mister_Magistervendor does stuff like /vendor/bin/modprobe -a -d /vendor/lib/modules rmnet_ctl but that doesn't work16:32
Mister_Magisterokay so the "disagrees about version of symbol module_layout" is due to loading of the vendor module which doesn't want to load16:35
Mister_Magisterso i think it can be ignored16:35
Mister_Magisteridk man16:45
Mister_Magisterapparently in other devices that service is just disabled, since mobile data works fine without it i also think we can disable it https://gitlab.com/ubports/porting/community-ports/android10/xiaomi-poco-x3/xiaomi-surya/-/merge_requests/2116:51
Mister_Magisterfinally silence in logs16:55
Mister_Magisternow, suspend doesn't work, it doesn't have single wakelock and all I can see is this https://paste.opensuse.org/pastes/279c67094b32 (i've fixed charging indicator)17:24
malhow did you fix charging indicator?17:31
Mister_Magisterblacklisted 2 more devices :)17:33
Mister_Magisterwelcome back17:33
Mister_Magisteri've blacklisted bq25960-slave and bq25960-master17:33
Mister_Magisterand it works now17:34
T42<Mister_Magister> @b100dian yooooo do you have a second to talk about out lord and saviour in-screen fp?17:34
Mister_Magistermal: if you're back i got couple burning questions :D17:34
malhow about you just ask17:35
Mister_Magisteri just dont' know if you're back yet or just popped in17:35
Mister_Magisteri'll just assume you're back then17:35
Mister_Magisterso i disabled vendor.dataqti service as seemingly other ports done the same so we do not worry about that anyomre, fixed charging indicator by blacklisting those 2 devices, no clue why suspend doesn't wor, nothing seems to be blocking it, ncm still doesn't work oh and we got double tap to wake which makes /sys/class/touchscreen/primary/double_tap_pressed go from 0 to 1 when you double tap and then resets to 0 once screen17:37
Mister_Magisterturns on, could we handle that in mce?17:37
Mister_Magisteralso you didn't push the patches for the suspend service, tho i'm not sure that will fix suspend despite the name? also turning wifi off doesn't help either so it's not wifi17:38
mal se17:39
maloops17:39
malcheck wakeup_sources17:39
maland journal for suspend or wakeup related messages17:39
Mister_Magisterno such file in sysfs17:40
malshould be/sys/kernel/debug/wakeup_sources17:40
Mister_Magisterin journal, beside the dmesg messages i mentioned already, nothing else17:40
Mister_Magisterhttps://paste.opensuse.org/pastes/b197451eed0b it just does this17:40
Mister_Magister /sys/kernel/debug/wakeup_sources does not exist, /sys/kernel/debug doesn't exist even17:41
Mister_Magisterthere's /sys/kernel/wakeup_reasons/last_resume_reason but that's empty17:42
Mister_Magisteri don't think it even attempts suspend which's worse17:42
malsuspend service patch pushed17:43
Mister_Magisterthanks17:43
Mister_Magisterwait where was the path for count of failed suspends?17:43
malI wonder why wakeup_sources is not there, maybe some kernel option17:43
malgrep kernel sources17:43
Mister_Magisterlike system monitor doesn't even count failed ones so i don't think it attempts suspend17:43
Mister_Magisteri'll apply just that one patch17:44
Mister_Magisterpatched17:45
Mister_Magistermal: is handling of the dt2w possible in mce?17:46
malyes17:46
malprobably17:46
Mister_Magisterxd17:46
Mister_Magisteri mean SHOULD be possible but can it be done currently or does it require changes in mce?17:46
Mister_Magisterlike some magical config17:46
malwell what kind of event does it send17:47
maldoes it show in evdev_trace -t17:47
Mister_Magisterit doesn't send evdev events17:47
Mister_Magisterit only sets that sysfs file to 117:47
Mister_Magisterthat's all it does17:47
malmaybe look at kernel driver17:48
Mister_Magisteri think CONFIG_PM_DEBUG is the one we want17:49
maljust wondering if we need to modify it or if there is some configuration17:49
Mister_Magisteri think we need to modify it17:50
Mister_MagisterInstall: out/target/product/miami/system/bin/hw/android.system.suspend-service :)17:50
Mister_Magisterlet me rebuild kernel with CONFIG_PM_DEBUG and install both17:51
Mister_Magisterwill dhd pick the suspend service?17:51
malafaik it should pick it https://github.com/mer-hybris/droid-hal-device/blob/master/droid-hal-device.inc#L71917:53
Mister_Magistermeanwhile mal any ideas about NCM?18:00
Mister_Magisteror if you're up to me, could you guide me through camera2 api18:00
malis the udc still a problem?18:06
Mister_Magisteryes18:06
maleven after the reverts18:06
Mister_Magisterreverts?18:06
malthe config changes18:06
Mister_Magisteryes18:06
Mister_Magisteroh wait no it just worked18:07
Mister_Magisterhold up18:07
Mister_Magisterno i'm not getting anything on the pc18:07
malbut udc seemed ok now?18:08
Mister_Magisteryeah now errors out18:08
malis usb-moded configured correctly for ncm18:08
T42<adampigg> finally, people onboard with using NCM over rndis18:09
Mister_Magisteroh18:09
Mister_Magisterthere's typo in the script18:09
Mister_Magister$USB_DRVER18:09
Mister_Magister@adampigg fix it18:09
Mister_Magisterand since you're to blame for this mess, help me fix the ncm18:09
T42<adampigg> is that in my configfs scipt?18:09
Mister_Magisteryes18:10
Mister_Magisterwriting to UDC says sh: write error: No such device18:10
T42<adampigg> its correct here https://github.com/HelloVolla/droid-config-halium-mimir/blob/master/sparse/usr/bin/droid/setup-configfs.sh18:10
Mister_Magisterno it's not18:10
malhttps://github.com/HelloVolla/droid-config-halium-mimir/blob/master/sparse/usr/bin/droid/setup-configfs.sh#L2918:10
T42<adampigg> oh, i see it18:11
Mister_Magisterhttps://github.com/HelloVolla/droid-config-halium-mimir/blob/master/sparse/usr/bin/droid/setup-configfs.sh#L29 DRVER18:11
Mister_Magisterand the 33 and 34 are failing18:11
malbut you just said a moment ago that udc worked?18:12
Mister_Magistermal: yes usb-moded has the configuration files18:12
Mister_Magistermal: yeah it's weird18:12
Mister_Magisterthis is what usb-moded says when i plug the phone in https://paste.opensuse.org/pastes/15549b87772318:12
malwhat does it try to write there? do you udc configured in usb moded config?18:13
malthe dummy one could be messing up things if you don't tell it what to use18:13
malcheck fp5 how to do it18:14
Mister_Magisteri don't know what does it try to write, i'm using adam's config files https://github.com/HelloVolla/droid-config-halium-mimir/tree/master/sparse/etc/usb-moded18:14
mallook at the udc in that18:14
malit's not even close to what you should have18:15
Mister_Magisteryou mean the gadget_udc_device?18:15
malyes18:15
malI asked about it earlier today18:15
maland you told what you should have18:16
Mister_Magisterwell to be honest i have no idea what i'm doing18:17
Mister_Magistersorry18:17
Mister_Magisterokay now when i connect usb it stars charging and usb-moded doesn't say anything18:18
Mister_Magisternothing in host dmesg18:18
Mister_Magisterand when I disconnect the usb it fails to stop charging18:19
malI wonder why you use different gadget_base_directory18:19
Mister_Magisteri copied configs from adam thats why18:19
malI also told you to check if extcon and android_usb report state correctly18:20
malif one is wrong you need to disable it18:20
Mister_Magistersure but that doesn't tell me anything18:20
Mister_Magisterwhat's extcon18:20
malcome on, obviously /sys/class/extcon18:20
Mister_Magisterwell it's not obvious to me, seriously18:21
Mister_Magisterthere's no /sys/class/android_usb/android0/enable but there's /sys/class/android_usb/android0/state18:21
malI did ask if state is correct or not18:21
Mister_Magisterstate is correct18:21
Mister_Magistersays DISCONNECTED18:21
malhow about extcon18:21
Mister_Magisterthere's extcons 0 2 418:21
malcheck all state files there18:22
Mister_Magisterhttps://paste.opensuse.org/pastes/6c327eb833d918:22
maldo those change correctly if you plug in usb18:22
Mister_Magisterhttps://github.com/HelloVolla/droid-config-halium-mimir/blob/master/sparse/etc/usb-moded/dyn-modes/99-developer_mode.ini#L12C20-L12C58 is this path correct though? it doesn't exist here18:22
malI have no idea about the stuff piggz has done18:23
Mister_MagisterI've connected usb yet extcons and the /sys/class/android_usb/android0/state are all 0 or disconnected18:23
malhmm18:23
malis there only android0?18:24
Mister_Magisteryes18:24
Mister_Magisterother files in dyn-modes also use the /enable path so ig its correct18:25
malreboot just in case18:25
Mister_Magisterwhen i rebooted with usb in the /sys/class/android_usb/android0/state is CONFIGURED18:27
Mister_Magisterextcons are 018:28
Mister_Magisterand when i unplugged usb it remains configured18:28
malok so usb detection is broken then somehow18:29
Mister_Magister:(18:29
malfp5 needed kernel patch to make it work18:29
Mister_Magisterspeaking of, the suspend service and new kernel are ready18:30
T42<TheVancedGamer> USB was awful on FP518:30
Mister_Magistermal: cant' we use the charging indicator as usb plugged in indicator?18:30
Mister_Magisterdumb question maybe18:30
Mister_Magisterlike the /sys/class/power_supply/usb/18:31
malit wouldn't be enough to detect which mode is wanted18:31
malafaik18:31
Mister_Magisterah yes18:31
malif it's a charger or computer18:31
Mister_Magistergotcha18:31
Mister_MagisterI think i'll yeet the NCM changes and try the same configs i used in pipa18:32
Mister_Magistermaybe @adampigg chekhov's gunned me18:32
malthis is not related to ncm18:32
Mister_Magisterif you say so18:33
Mister_Magisterso I added PM_DEBUG but still no /sys/kernel/debug/ :(18:33
T42<TheVancedGamer> you need CONFIG_DEBUG_FS18:33
T42<TheVancedGamer> which will probably break a lot of things in GKI kernel18:34
Mister_Magisterthanks18:34
malbut it's a problem since we rebuild vendor_boot also18:34
Mister_Magisterwhat do you mean?18:34
mal*not a problem18:34
Mister_Magisterah18:35
malmy multitasking is getting ridiculous, watching a movie, then I have a golf tournament in another display and irc on another18:35
T42<adampigg> mal: thx, i just improved my configs :)18:36
Mister_Magisterey, i said that typo18:36
T42<adampigg> that dyn-modes stuff is probably not needed, didnt know about [network] section in main conf18:36
T42<adampigg> mister-magister: no, other stuff simplified :)18:37
T42<adampigg> but also thx for typo18:37
Mister_Magister>:(18:37
mallet me check but I think overriding main.conf is not needed in 5.0.0 anymore, at least in newer ones18:38
Mister_Magistermal: what do we do with the suspendservice?18:39
mal5.0.0.73+ doesn't need main.conf override because it supports main.conf.d18:42
Mister_Magisterthere's /system/bin/hw/android.system.suspend-service running18:42
malhmm18:42
malis something missing, maybe18:42
Mister_Magisterwhat do you mean?18:42
Mister_Magisterthere's /usr/libexec/droid-hybris/system/bin/hw/android.system.suspend-service18:43
Mister_Magisterdo we want to start it somehow?18:43
Mister_Magisterlike the executable is there, what next? xd18:43
Mister_Magisterahhh there's .rc that needs to be added18:46
Mister_Magistershould've said so18:46
Mister_Magisterhmm, it did not start18:48
maloverride it in libexec rc18:49
Mister_Magisteri did that18:50
Mister_Magisteri added this https://github.com/sailfishos-sony-nagara/droid-config-sony-nagara/blob/main/sparse/usr/libexec/droid-hybris/system/etc/init/android.system.suspend-service.rc18:50
Mister_Magisterbut it still starts /system one18:50
malis the service name the same?18:51
Mister_Magisteryes18:51
malwhat are the permissions of the file?18:52
malhas to be 64418:52
Mister_Magisterit wasn't18:53
malI think android is strict about that now18:53
Mister_Magisterkernel finished rebuilding18:53
Mister_Magisternow it worked18:54
Mister_Magisterso ig android can now block suspend18:54
Mister_Magisternow in /sys/power/wake_lock we have rich-core-dumper all day every day18:56
Mister_Magistermaybe that's our little culprit?18:56
Mister_Magisteroh its gone18:56
Mister_Magisterstill no suspend though18:57
Mister_Magisterwhat's interesting it doesn't wake up the display when you connect charger18:57
malbecause usb detection is broken18:58
Mister_Magisterahh gotcha18:58
malmaybe18:58
Mister_Magisternow droid-hal is rebuilding and give me 15 minutes i have to go to the shop before they close it18:58
Mister_Magisteruh-oh mal with the latest kernel device is bootlooping19:25
Mister_Magisterso maybe DEBUG_FS wasn't such a good idea19:25
malyou flashed also vendor_boot?19:26
T42<Mister_Magister> i assume sfos flashed it yes19:27
malrecheck the flash script etc19:28
T42<TheVancedGamer> mal: FP5 also crashes with debug_fs and all modules packaged, but I never debugged it :P19:28
maloh19:28
Mister_Magisteri just flashed them manually19:28
T42<TheVancedGamer> it violates some GKI thing apparently19:29
Mister_Magisteryeah no it bootloops19:30
T42<adampigg> heh, i couldnt find my sonos speakers on the lan, realised im still tethering over usb to my phone as i was testing that out19:31
T42<adampigg> i bet not many ports check that19:31
malI have debugfs on a16-based device19:36
Mister_Magisteridk what to tell you19:36
Mister_Magisterlooks like i'll have to get REALLY familiar with the kernel :(19:54
Mister_Magistermal: okay so ig currently the biggest issue is lack of suspend, any tips on how to even go about debugging it?20:13
Mister_Magistermal: does jolla's volte plugin support hanging up? can we try that?20:18
malit does20:56
Mister_Magistercan we try that? i forgot the package name21:15
Mister_Magisteralso i read a bit of kernel code and the double tap DOES have evdev device and it DOES send event before setting sysfs file but for whatever reason nothing is emitted on the event21:15
malis there some path for controlling the double tap?21:21
Mister_Magisteryes but that doesn't enable the evdev21:21
Mister_Magisterholdon21:21
Mister_Magisterso21:22
Mister_Magisterwe literally need this https://github.com/VerdandiTeam/android_kernel_motorola_sm6375/blob/lineage-22.2/drivers/input/touchscreen/touchscreen_mmi/touchscreen_mmi_gesture.c#L30621:22
Mister_Magisterit's picture perfect of what we need, the evdev device even exists21:22
Mister_Magisterbut this is my driver https://github.com/VerdandiTeam/android_kernel_motorola_sm6375/tree/lineage-22.2/drivers/input/touchscreen/goodix_berlin_mmi21:22
Mister_Magisternow i gotta figure out the diff21:22
Mister_Magisteram i thinking correctly here?21:24
Mister_Magisterye i have INPUT_TOUCHSCREEN_MMI=y too21:24
Mister_Magisterbut it does'nt seem to send the evdev stuff21:24
maland evdev_trace -i doesn't show that?21:29
Mister_Magisternope21:30
Mister_Magisternot even cat event2 shows anything (event2 being the double-tap)21:30
Mister_Magisteri have no idea why at this point21:30
Mister_Magisteryeah reading through this there's 0 reason why evdev shouldn't be sent21:32
malcheck sysfs input device21:32
Mister_Magisterwhat do you mean?21:33
malevent2 in sysfs21:33
malwhat files does it have21:33
malmaybe there is some enable file or something21:34
malcould be input2 or whatever21:34
maljust find the correct one21:34
Mister_Magisterhttps://paste.opensuse.org/pastes/b0f7fc83dded21:34
Mister_Magisterevent3 is the touchscreen21:34
Mister_Magisterall the event directories contain the same files21:34
malinput2?21:34
malshow the evdev_trace -i entry for double tap21:35
Mister_Magisterhttps://paste.opensuse.org/pastes/e6af37e71a0621:35
Mister_Magisterit should send BTN_TRIGGER_HAPPY6 on DT21:36
malso what does the sysfs for input2 contain?21:36
Mister_Magisterhttps://paste.opensuse.org/pastes/551a71627bc221:37
Mister_Magistercat /sys/class/input/input2/dt-gesture/enable is 021:37
malso what happens if you write 1 there?21:38
Mister_Magisterhttps://paste.opensuse.org/pastes/db6f68ed7c5a21:38
Mister_Magisternothing still21:38
malor enable_wakeup21:38
Mister_Magisterenable_waekup says invalid argument21:38
malis that really correct kernel driver?21:39
malah missed the other link21:40
maldid you see anything in dmesg when you wrote to those paths?21:41
Mister_MagisterInvalid sensor class enable_wakeup handle21:42
malthat goodix_ts_gesture.c has some enable stuff21:42
Mister_Magisterbtw the dt-gesture is here https://github.com/VerdandiTeam/android_kernel_motorola_sm6375/blob/lineage-22.2/drivers/input/touchscreen/touchscreen_mmi/touchscreen_mmi_gesture.c#L52021:43
Mister_Magisterand hold up21:43
Mister_Magisterhttps://paste.opensuse.org/pastes/ddfdf1d1e16921:44
malhmm21:44
Mister_Magisterye i'm getting taps21:44
Mister_Magisterbut no double tap messages only single tap21:44
Mister_Magisterit should have single tap to wake too21:44
malthere is the mode_type in the driver21:45
Mister_Magisterig i'm seeing this https://github.com/VerdandiTeam/android_kernel_motorola_sm6375/blob/lineage-22.2/drivers/input/touchscreen/touchscreen_mmi/touchscreen_mmi_gesture.c#L25721:46
Mister_Magisterbut even that doesn't report anything in evdev, even though it should21:46
Mister_Magisterfrom that pr_info it should report to evdev nothin in between yet it doesnt21:47
malbut if mode_type is not showinf double tap support then it would not reach the other parts21:48
malthe mode_type checks at the beginning of each case21:48
Mister_Magisterbut if it reached the single tap it ought to report that in evdev no?21:49
malare you monitoring all input devices in evdev_trace ?21:50
Mister_Magisteryeap21:50
Mister_Magisterhttps://github.com/VerdandiTeam/android_kernel_motorola_sm6375/blob/lineage-22.2/drivers/input/touchscreen/touchscreen_mmi/touchscreen_mmi_class.c#L41721:51
Mister_Magisterso if double tap is set to 1 it will enable single tap to wake?21:51
malwhere is that double_tap_enabled ?21:52
malhave you found it in sysfs?21:52
Mister_Magistertouchscreen/primary/21:52
maland what does it contain?21:52
Mister_Magisterhttps://paste.opensuse.org/pastes/df217bdebda121:52
malalso in the those double_tap*21:53
maland maybe gesture21:53
Mister_Magisterthe enable is 0 unless i write 1, the pressed i described previously, is 0, when you double tap it's set to 1 until you turn the screen on then it resets, and gesture is 0721:54
maldo you see any of the events if display is on?21:54
malif you write 1 to enable21:55
Mister_Magisteri see the single tap events sent above with display off21:55
malit's possible that mce in configured correctly and it just ignores the events if it thinks those are unwanted21:55
mal*is not configured correctly21:55
Mister_Magisterhttps://paste.opensuse.org/pastes/4308c2c27edb nothing with display on or off21:56
Mister_Magisterbut if i tap the screen with it off i see single tap in dmesg i sent above21:57
Mister_Magisterand like single tap also should send evdev code21:57
malI suspect mce is filtering them21:57
Mister_Magistercan we stop mce and check?21:57
Mister_Magisterhttps://github.com/sailfishos/mce/blob/91198feec76f1069ed7b38f519e9a0f7852ed358/evdev.inc#L511 that's literally the only mention tho21:59
Mister_Magister>.>22:05
malhave you tried what you see in mce debug log22:10
Mister_Magistermal: doing cat /dev/input/event2 on android also doesn't report anything22:10
maland the wakeup works in android?22:11
Mister_Magister ye22:11
Mister_Magisterandroid listens to the sysfs file though22:11
Mister_Magisternot to evdev22:11
malmaybe mce needs patching then22:13
malunless you want to to debug why events are not sent22:13
Mister_Magisterideally i'd like not to debug that22:15
Mister_Magisterokay i'll leave dt2w for now22:19
Mister_Magisterso any way i can try jolla's volte plugin?22:19
Mister_Magisteri'll take that as a no then22:22
Mister_Magisterahh it was the fpd jolla's plugin that was available in repos not the volte one22:25
Mister_Magistermy mistake22:25
Mister_Magisterit gets confusing :P22:25
Mister_Magisteroh oh22:25
Mister_Magistermal: the display is 120Hz but in sfos it shows 60FPS so it seemingly runs in 60Hz mode, do you know how can we change that? last time i remember it wasn't straightforward22:26
Mister_Magistermy 144Hz device runs at 144Hz no issues22:26
maldoes android run with 120?22:29
Mister_Magisteri'm gonna disable TOUCHSCREEN_MMI becasue then goodix sends the events on its own instead of through it22:29
Mister_Magistermaybe that will fix it22:30
Mister_Magisterit's stab in the dark but might just work22:30
maldoes android have that mmi?22:30
Mister_Magisteryes its from preexisting defconfig22:31
Mister_Magisterabout android "switches between 60/120 by default, but 48 and 90hz are also available"22:31
Mister_Magisterlooks like it's dynamic22:31
malsysfs or via hwc?22:31
Mister_MagisterI couldn't tell you, I have no idea22:32
Mister_Magistertho in that sysfs i sent earlier there is refresh_rate22:32
Mister_Magisteri tried writing 120 and restarting lipstick and it did nothing22:32
Mister_Magisterno there is'nt anything in dmesg when i write to that22:36
maldid you restart hwcomposer?22:37
Mister_Magisteroh… no just lipstick22:37
malbtw, do you have correct hwc restart config for mce22:39
Mister_Magisterno but the lipstick restarts no issue22:39
malhttps://github.com/mlehtima/droid-config-fp5/blob/master/sparse/etc/mce/60-compositor-fp5.ini22:39
Mister_Magisterso i didn't think i needd it22:39
Mister_Magisterye but that was needed when lipstick failed to restart22:40
Mister_Magisterthere's no issue with restarting lipstick22:40
Mister_Magisterdo i still need it?22:40
maland boot splash looks ok?22:40
Mister_Magisteryes22:40
Mister_Magisteri restarted hwc still 6022:41
Mister_Magisterthere's CONFIG_DRM_DYNAMIC_REFRESH_RATE but it's not set to y22:43
Mister_Magisterthere's /sys/firmware/devicetree/base/soc/qcom,mdss_mdp@5e00000/qcom,mdss_dsi_visionox_r66451_fhd_plus_video/qcom,mdss-dsi-min-refresh-rate which returns < (60 in ascii) and max-refresh-rate which returns x (120 in ascii)22:51
Mister_Magister /sys/firmware/devicetree/base/soc/qcom,mdss_mdp@5e00000/qcom,sde-qos-refresh-rates also returns <x22:51
Mister_Magisterwell i really should sink my teeth into the suspend but idk even where to start22:55
Mister_Magisterthat or usb which I also have no idea where to even begin22:56
T42<CRK_11755> Any luck with volte on fairphone 4 on T-Mobile23:00
T42<CRK_11755> In the USA23:00
T42<CRK_11755> I tried sfos on my fairphone4 in the US I had no VOLTE on T-Mobile. Has the community port been updated at all23:01
Mister_Magistermal: i switched the if's so that goodix is the one sending events aaaand nada, nothing still :(23:43

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