| T42 | <elros34> If thats the code it use: https://github.com/edp17/android_kernel_samsung_smdk4412/blob/cm-12.1-mtpfix/drivers/input/keyboard/cypress/cypress-touchkey.c#L2104 then yes 1/0 are only allowed values. As you can se there are many printk(KERN_DEBUG "[TouchKey-BLN] which should be printed in dmesg | 00:02 |
|---|---|---|
| T42 | <edp_17> Thanks @elros34. Yes, that is the latest kernel I use. I'll check the messages in dmesg. | 00:05 |
| T42 | <edp_17> The 'dmesg | grep BLN' is full of: https://pastebin.com/ffNzR9BA | 00:08 |
| T42 | <elros34> I would disable timeout completely so write 0 to notification_timeout | 00:10 |
| T42 | <edp_17> By default that was -1. I wrote 0 into it and rebooted. | 00:14 |
| T42 | <edp_17> I unplugged the cable and a message appeared at the top saying: "Charging..." | 00:15 |
| T42 | <edp_17> As the battery is full, the leds are on. | 00:15 |
| T42 | <edp_17> Now I got the 'Charging completed' message (cable is still unplugged) and the leds started to blink. | 00:16 |
| T42 | <elros34> sysfs is virtual file system so if you reboot values you wrote are lost. So you are getting charging state when usb is unplugged and it stays like that? | 00:19 |
| T42 | <edp_17> I have created a service that writes 1 into notification_enabled. I added into the script to write 0 into notification_timeout. | 00:25 |
| T42 | <edp_17> And yes, I got 'Charging' message when unplugged the cable and a little later 'Charging completed' | 00:26 |
| T42 | <edp_17> Sometimes when there is no cable plugged in to the device I get a 'Not enough power to charge' message too. | 00:27 |
| T42 | <edp_17> That's why I think the led configs are ok-ish, and there is something underlying problem with setting the charging status somewhere. | 00:28 |
| T42 | <elros34> you could run: mce -Tq -l "modules/battery-udev.c:*" to see if everything looks sane when you disconnect cable | 00:40 |
| *** Oksana_ is now known as Oksana | 02:39 | |
| spiiroin | @edp_17 In case of "odd" battery full signaling, the last resort is something like "if charging stops at something resembling battery full percentage, assume maintenance charging until charger disconnect / battery level drops < threshold" - I guess you are hitting that | 04:04 |
| spiiroin | @edp_17 that unknown charger type warning is due: if charger device has no type property -> device name is used instead. generally speaking: qc devices have "usb" charger that has types like "cdp"/"usb"/... for dedicated-charger/pc-connection/... and mtk devices have charger names such as "ac" / "usb" / ... | 04:08 |
| spiiroin | and what you have is probably not matching either of these expectations -> interpreted as "some sort of charger" + diagnostic warning | 04:09 |
| spiiroin | and the other symptoms (not enough power to charge / battery full while cable is disconnected) look like false positive "charger is connected" | 04:14 |
| spiiroin | which is likely to come from broblems with battery-full interpretation. if battery reports it is getting charged / is full, an assumption is made that "there just has to be a charger connected for this to make sense" | 04:22 |
| spiiroin | now if there are issues with battery state notifications, the sum of heuristics can become: once percentage is high enough -> battery is full and therefore there is a charger | 04:26 |
| spiiroin | ways to fix include: get battery state notifications flowing / poll props periodically (if it produces more correct info than notifications) / reverse the assumption (=battery can't be full if no chargers are online) | 04:29 |
| *** Oksana_ is now known as Oksana | 06:04 | |
| T42 | Hans_Jo was added by: Hans_Jo | 07:33 |
| T42 | <edp_17> @elros34 : Thank you, I'll try that. | 08:46 |
| T42 | <edp_17> @spiiroin: Thank you! | 08:46 |
| T42 | KasperRib was added by: KasperRib | 08:48 |
| piggz | mal: did you get the config repo i sent | 08:52 |
| mal | piggz: yes, didn't see anything obviously wrong | 09:45 |
| walidh | when open image catured with camera its black | 10:06 |
| T42 | <edp_17> @elros34 : I ran the suggested mce -Tq -l "modules/battery-udev.c:*" | 11:32 |
| T42 | <edp_17> When it ran in fingerterm on the device, I got: https://pastebin.com/wyAbqGEH | 11:32 |
| T42 | <edp_17> When I ran it on the PC I got: https://pastebin.com/Tyah4FTC | 11:33 |
| walidh | mal: when open image got black screen , i added #define WANT_ADRENO_QUIRKS 1 but nothing change | 12:54 |
| T42 | <elros34> walidh: this is fix for black images: https://github.com/mer-hybris/droid-config-sony-nile/blob/master/sparse/var/lib/environment/nemo/99-qtscenegraph.conf | 12:55 |
| T42 | <elros34> @edp_17 example from my device for comparison: https://pastebin.com/WHADETKF | 12:57 |
| T42 | <edp_17> @elros34 : Thanks. Have you | 13:41 |
| T42 | <edp_17> I think know the issue. My device is looking for different type of charger (max8997 and max8922) and the result is wrong "fuelgauge: battery @ cap" value. This would explain why I receive charging messages when the usb is not even plugged in. The relevant comparison against your log is here: https://pastebin.com/itu1Cgas | 13:58 |
| walidh | repo problem: nothing provides libbinder.so needed by droid-config-h918-1-1.armv7hl ??? libbinder.so is in out dir | 14:19 |
| T42 | P@t %lastname% was added by: P@t %lastname% | 14:38 |
| mal | walidh: what are you trying to do? | 14:55 |
| walidh | mal: i'm trying to change some vendor blobs files (exp :etc/firmware/a530_zap.b01, vendor/lib/libscalar.so...), because now i must push them under recovery, | 14:58 |
| mal | you should not do that | 14:58 |
| mal | why are those not on the device? | 14:59 |
| mal | I mean in /system or /vendor or some firmware partition | 14:59 |
| mal | that is the wrong of trying to fix some thing | 15:00 |
| walidh | because cm14.1 based on old bootloader and firmware | 15:00 |
| walidh | so i think i must create twrp file to change thoes files | 15:02 |
| T42 | <elros34> @edp_17 I connected it to wall charger and started mce via ssh | 15:13 |
| vknecht | walidh, maybe you'd be better off making your own lineageos build, fixing stuff in there directly | 15:17 |
| vknecht | I ended up doing that for fixing proximity sensor, dualsim and useless 3rd "eyeverify" camera which crashed cam apps on switching rear/front | 15:18 |
| vknecht | and lineageos user were happy too :-) | 15:19 |
| T42 | <elros34> @edp_17 when you disconnect usb (I assume it happens in line 41) you get led_pattern_activate_pipe PatternBatteryFull instead deactivate. Probably spiiroin could tell you what to do if you show him full log | 15:21 |
| T42 | <elros34> There is a ways to blacklist some devices: https://git.sailfishos.org/mer-core/mce/blob/master/inifiles/battery-udev-devices.ini | 15:23 |
| T42 | <edp_17> @elros34 : Good idea to connect via wifi because if I use the usb cable, it looses the connection when I unplug it. Thank you. I will ask spiiroin about that. He already told a lot about it but I need time to get my head around that. | 15:23 |
| walidh | vknecht yes i think this is the best solution | 15:44 |
| spiiroin | @elros34 @edp_17: look out for battery.POWER_SUPPLY_STATUS initial state (line 17) vs changes (only line 55) -> initially "charging", stays that way over disconnect (triggers "battery is charging -> there is a charger" heuristics), changes to "discharging" briefly after reconnect -> discharging while charger connected -> triggers "is full" heuristics | 15:45 |
| spiiroin | i.e. notifications about POWER_SUPPLY_STATUS = do not to reflect reality. and since that it is given more weight than charger.online=0 -> hiccups | 15:48 |
| T42 | <elros34> what if he blacklist: max8997-charger, max8922-charger, ac? If I get it right usb reports correct "present=-1 online=0 -> active=0 | 15:50 |
| walidh | mal: in camera app i can't switch to front camera there is no button | 16:40 |
| walidh | mal: app camera not show front camera and can't acces to it | 20:34 |
| vknecht | walidh, try with GST_DEBUG : https://sailfishos.org/wiki/Sailfish_OS_Cheat_Sheet#GStreamer | 20:56 |
| vknecht | (to get better logs) | 20:56 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!