| rinigus | while looking into pulseaudio config, I stumbled on https://github.com/sailfishos-sony-tama/droid-hal-configs/blob/master/sparse/etc/pulse/xpolicy.conf#L519 installed on aarch64 device. should probably point to lib64 version, as that's the one we have on the device | 10:43 |
|---|---|---|
| rinigus | pulseaudio tries to load few 32-bit libs in vendor, fails and then doesn't try to get the available 64 bit version. this is according to journal message (below) and strace showing no try to load 64 bit one | 10:48 |
| rinigus | pulseaudio[4502]: "/vendor/lib/hw/sound_trigger.primary.sdm845.so" is 32-bit instead of 64-bit | 10:48 |
| rinigus | mal || jusa: any pointers? ^ | 10:48 |
| T42 | <eugenio_g7> rinigus: try setting HYBRIS_LD_LIBRARY_PATH like this: https://github.com/hybris-mobian/adaptation-fxtec-pro1/blob/bullseye/etc/profile.d/zz-fxtec-pro1.sh | 10:59 |
| T42 | <eugenio_g7> In your case both on /vendor I guess | 10:59 |
| T42 | <eugenio_g7> But lib64 before lib | 11:00 |
| rinigus | @eugenio_g7: thanks, will try | 11:04 |
| T42 | <eugenio_g7> if it works I guess a systemd drop-in on the pulseaudio service should be enough | 11:05 |
| rinigus | You mean adding that far into systemd unit file? | 11:11 |
| mal | rinigus: should not matter probably | 11:15 |
| T42 | <eugenio_g7> rinigus: yup, as a drop-in (/etc/systemd/user/pulseaudio.service.d/filename.conf) | 11:16 |
| mal | it seems to be quite common that android doesn't have all audio libs for 64-bit version | 11:25 |
| rinigus | mal: in my case, all the libs that failed to load this way are available in 64 bit as well. judging by names. | 11:26 |
| mal | hmm, that's odd | 11:30 |
| mal | rinigus: did you check strace? | 11:30 |
| rinigus | mal: yes, I did check strace. after writing message "... 32-bit instead of 64-bit" no attempt was made to load from 64 bit location. no idea where it also gets the list of libs and tries to load them. (haven't tested with hybris var, as suggested by eugenio_g7, will try soon). | 11:32 |
| rinigus | @eugenio_g7: running HYBRIS_LD_LIBRARY_PATH=/vendor/lib64:/vendor/lib strace -f pulseaudio -n --file=/etc/pulse/arm_droid_d | 11:37 |
| rinigus | efault.pa > logs/pa.strace 2>&1 did not help. still loading /vendor/lib/hw/sound_trigger.primary.sdm845.so . which I guess is expected assuming that PA is loading by itself.. | 11:37 |
| T42 | <eugenio_g7> 😐 try also putting /vendor/lib64/hw in the search path | 11:41 |
| rinigus | @eugenio_g7: did already - no difference. | 11:46 |
| T42 | <eugenio_g7> oh ok, well it was worth a try | 11:47 |
| rinigus | sure it was! thanks! | 11:48 |
| mal | rinigus: that needs some checking | 12:17 |
| rinigus | mal: thanks! I am trying to upload strace somewhere, but it is rather large. any idea where the search path for PA libs is set? | 12:18 |
| mal | rinigus: can you show the whole strace | 12:18 |
| mal | rinigus: you can email it to me if you want | 12:18 |
| rinigus | mal: just a sec. will see where to add it | 12:18 |
| rinigus | mal: it is attached at https://github.com/sailfishos-sony-tama/main/issues/107#issuecomment-785037503 | 12:20 |
| rinigus | mal: just to be sure that you see it. in PA config, we have /usr/lib/telepathy/telepathy-stream-engine and not lib64 version. | 12:32 |
| mal | rinigus: hmm | 12:34 |
| mal | rinigus: which config? | 12:34 |
| rinigus | mal: https://github.com/sailfishos-sony-tama/droid-hal-configs/blob/master/sparse/etc/pulse/xpolicy.conf#L519 | 12:35 |
| rinigus | looks to be merged with sparse-10. | 12:35 |
| rinigus | at least I found that on device | 12:35 |
| rinigus | mal: ^ | 12:36 |
| Mister_Magister | mal: found this while running test_gps https://paste.opensuse.org/23d482ea as root | 12:37 |
| T42 | <edp_17> Hi all. Can somebody help me with sensorfwd, please? When the screen is on, sensorfwd uses more than 60% cpu and the log is flodded with messages "failed to write payload to the socket": https://paste.ubuntu.com/p/j9y4mvpfwf/ | 13:49 |
| T42 | <edp_17> Galaxy Note (n7000) CM14.1 base, sfos 4.0.1.48 | 13:49 |
| rinigus | mal: looks like few PA modules also fail to load: https://dpaste.org/jKUO | 14:14 |
| rinigus | looks like the missing libs are in /usr/lib64/pulse-12.2/modules which are not used during resolution | 14:24 |
| mal | rinigus: I'll have a look | 14:26 |
| rinigus | mal: thanks! | 14:29 |
| T42 | <adampigg> mal: remember the UCM lib issue too | 14:37 |
| mal | @adampigg that will be fixed | 14:39 |
| mal | rinigus: those don't seem to be critical, happens on other devices also, the sound volume issue is known and could be some sony issue | 14:52 |
| mal | there is some idea how to fix it | 14:52 |
| rinigus | mal: I presume that is related to aarch64 / aosp10. as on aosp9 32 bit I had no such issues | 14:53 |
| rinigus | looking forward to fix | 14:53 |
| rinigus | btw, it is not detecting headphones either | 14:54 |
| mal | hmm | 14:54 |
| rinigus | mal: those non-critical errors - this is regarding pulseaudio --dump-modules, right? | 14:55 |
| mal | yes | 14:57 |
| mal | I get those also for example on xa2 | 14:57 |
| rinigus | mal: that's right, dump-modules are making the same errors on aosp9/32bit xz2 as well | 15:00 |
| rinigus | should have checked, sorry. | 15:00 |
| rinigus | mal: but missing libs as only 32-bit versions are tried, that's strange. I guess something in PA sets the path and it is wrong for 64 bit version | 15:01 |
| mal | rinigus: more like android side does it | 15:02 |
| mal | rinigus: as I mentioned before android often has 32-bit audio server | 15:02 |
| rinigus | mal: maybe. is it also 32 bit server on seine? don't know what it is in AOSP for tama, I must say | 15:04 |
| mal | rinigus: looks like there might be another 64-bit bug related to audio routing, also speaker phone mode is probably not working | 15:05 |
| rinigus | mal: haven't got that far, no sim in that phone | 15:06 |
| rinigus | mal: do you have that volume bug in seine as well? | 15:07 |
| mal | rinigus: yes | 15:11 |
| T42 | <adampigg> whats the volume bug? | 15:13 |
| rinigus | @adampigg: https://github.com/sailfishos-sony-tama/main/issues/107 | 15:14 |
| T42 | <adampigg> ok, much different to the pinephone volume bug | 15:15 |
| mal | rinigus: try restarting ohmd and see if headphone detection works | 15:16 |
| rinigus | will do | 15:17 |
| mal | that helped for me | 15:17 |
| rinigus | mal: restarting ohmd while having headphones in helped. | 15:25 |
| mal | rinigus: should work fine after that | 15:28 |
| rinigus | mal: yes, inserting and removing reroutes sound accordingly. | 15:31 |
| rinigus | so, sounds like ohmd needs to be primed with headphones | 15:31 |
| mal | rinigus: no need to have headphone in when restarting ohm | 15:32 |
| mal | there is some timing issue at boot which causes initial detection to fail | 15:32 |
| rinigus | mal: ahh, that's better and should be easier to fix :) | 15:33 |
| T42 | <Verevka86> https://paste.ubuntu.com/p/3gYjxwTxJN/ | 17:56 |
| T42 | <Verevka86> Please help | 17:56 |
| T42 | <Verevka86> Not build aarch64😔 | 17:57 |
| mal | remove the repo from .ks file manually | 18:01 |
| T42 | <Verevka86> Boot up Xiaomi MI6 hybris17. 1 aarch64 😍 | 19:28 |
| rinigus | :) | 19:40 |
| rinigus | looks like I am missing miniafservice (miniaf not starting on boot). how do you package that? | 20:22 |
| *** daniel is now known as Guest18282 | 20:37 | |
| mal | rinigus: if you mean 17.1 base, do not use miniaf | 20:43 |
| rinigus | mal: yes, I mean aosp10. find it rather confusing with these mini services. right now I have minimediaservice and minisfservice | 20:44 |
| mal | rinigus: miniaf is not used anymore, it was replaced by https://github.com/mer-hybris/audiosystem-passthrough | 20:44 |
| rinigus | mal: that seems to be running (audiosystem-passthrough) | 20:45 |
| mal | yeah | 20:46 |
| rinigus | mal: I have also camera freezing if I switch from video to photo mode. https://github.com/sailfishos-sony-tama/main/issues/115 | 20:47 |
| rinigus | any tips for that? | 20:47 |
| rinigus | logs are in the issue | 20:48 |
| mal | rinigus: the reason is right there in the logs :) | 20:51 |
| rinigus | mal: media.audio_policy? | 20:52 |
| mal | yes | 20:52 |
| rinigus | mal: any clues on why wouldn't it start? or what should provide it? | 20:53 |
| mal | rinigus: try adding MINIMEDIA_AUDIOPOLICYSERVICE_ENABLE := 1 to env.mk in droidmedia sources | 20:53 |
| rinigus | mal: will do, thanks! | 20:54 |
| rinigus | mal: getting error on droidmedia compilation | 20:58 |
| rinigus | frameworks/av/services/audiopolicy/service/AudioPolicyService.h:34:10: fatal error: 'mediautils/ServiceUtilities.h' file not found | 20:58 |
| rinigus | I have made the change in env.mk and just fired make hybris-hal droidmedia | 20:59 |
| rinigus | such file does exist, but for some reason it is not included in the search path. | 21:02 |
| rinigus | btw, minisfservice and minimediaservice are 32 bit on tama aarch64 build | 21:06 |
| mal | rinigus: yes, that's how it's supposed to be afaik | 21:32 |
| mal | rinigus: I think I never built audio policy for android 10 | 21:32 |
| mal | rinigus: found the issue | 21:34 |
| T42 | Braillynn was added by: Braillynn | 22:18 |
| T42 | <Braillynn> Hello, idk if this is the place to reach out but I'm having some difficulty installing Sailfish OS, I'm at the final step. | 22:19 |
| mal | which device? | 22:21 |
| T42 | <Braillynn> @mal [which device?], Xperia 10 Plus | 22:21 |
| T42 | <Braillynn> (Photo, 1280x720) https://irc.thaodan.de/.imgstore/t8PpprzYFx.png It never stops searching device to flash. | 22:22 |
| mal | does the device led show blue color? | 22:22 |
| T42 | <Braillynn> @SailfishFreenodeIRCBridgeBot Yes | 22:23 |
| mal | what do you see if you run "fastboot devices" in the terminal? | 22:24 |
| T42 | <Braillynn> @SailfishFreenodeIRCBridgeBot Well I typed it in, nothing happened. The terminal didn't respond to it. | 22:26 |
| mal | do you have other usb ports on the computer, if yes try another port | 22:26 |
| T42 | <Braillynn> Okay, let me try another one. | 22:27 |
| mal | usb2 port if you have one | 22:27 |
| T42 | <Braillynn> I'm on an Intel Nuc 8th gen, I think all 4 ports are USB 3 | 22:27 |
| T42 | <Braillynn> @SailfishFreenodeIRCBridgeBot I moved to another usb port and it's telling me found I3223, baseband: FAILED, bootloader: waiting | 22:30 |
| T42 | <Braillynn> Found 1 devices: CQ3001SQ6Q | 22:30 |
| T42 | <Braillynn> Fastboot command: fastboot -s CQ3001SQ6Q | 22:30 |
| mal | hmm, interesting, does it still stop after that? | 22:32 |
| T42 | <Braillynn> @SailfishFreenodeIRCBridgeBot Not sure what you mean by stop, but phone still has blue led, and terminal looks like I can type more into it. | 22:32 |
| mal | but the flashing doesn't progress if you use the flash.sh script? | 22:33 |
| T42 | <Braillynn> @SailfishFreenodeIRCBridgeBot right, it doesn't appear anything is happening. | 22:34 |
| mal | try again | 22:35 |
| mal | usb3 and fastboot sometimes has issue | 22:35 |
| T42 | <Braillynn> @SailfishFreenodeIRCBridgeBot okay, just want to make sure I do this correct, I'm going into the Sailfish OS folder, opening the location in terminal, plugging in the Xperia 10+ while holding down the up volume button and waiting for the blue led, doing sudo bash ./flash.sh and it should take care of the rest? | 22:37 |
| mal | yes, it's known that usb3 sometimes misbehaves, we don't really know why | 22:39 |
| mal | do you happen to have any usb2 hub? | 22:39 |
| T42 | <Braillynn> @SailfishFreenodeIRCBridgeBot I do but on another PC. | 22:40 |
| T42 | <edp_17> Hi guys. Can I get help with my sensorfwd and/or rild issues, please? | 22:40 |
| mal | @edp_17 what is the issue? | 22:41 |
| T42 | <edp_17> When the screen is on, sensorfwd uses more than 60% cpu and the log is flodded with messages "failed to write payload to the socket": https://paste.ubuntu.com/p/j9y4mvpfwf/ | 22:41 |
| mal | @edp_17 do you have /run/sensord.sock ? | 22:43 |
| T42 | <edp_17> mal: yes. | 22:44 |
| T42 | <Braillynn> @SailfishFreenodeIRCBridgeBot I'm going to try on a Raspberry Pi 4, would that work? | 22:45 |
| mal | worth a try | 22:45 |
| mal | @edp_17 when did those messages start? | 22:45 |
| T42 | <Braillynn> @SailfishFreenodeIRCBridgeBot Wish me luck lol | 22:45 |
| T42 | <edp_17> mal: When I wake up the screen. | 22:45 |
| mal | @edp_17 I mean has that issue always been on that port or did it start after update or something? | 22:46 |
| T42 | <edp_17> mal: This is a fresh port, my first image for this device. | 22:47 |
| T42 | <edp_17> mal: What is also weird, right after I installed sfos, none of the sensors worked. After that I put hw-settings.ini into /usr/share/csd/settings.d/ and some of them started to work but got this behaviour. | 22:49 |
| T42 | <edp_17> mal: The usb charging icon and message was incorrect, so I added the /etc/mce/60-battery-udev-devices.ini file. | 22:50 |
| T42 | <edp_17> mal: vibra didn't work, so I changed /init.rc to set up the correct user and group for /sys/class/timed_output/vibrator/enable | 22:52 |
| T42 | <edp_17> This is all, I changed on the device. | 22:53 |
| mal | can you enable sensors one by one and see if it's some specific sensor causing the issue? | 22:55 |
| T42 | <edp_17> mal: How can I do that? (Enable one by one.) | 22:55 |
| T42 | <edp_17> How can I disable all of them? | 22:55 |
| mal | that csd config file for example | 22:59 |
| T42 | <edp_17> mal: Okay, I'll try. Should I restart something when I changed something in the file? | 23:00 |
| T42 | <edp_17> I set everything 0 in hw-settings.ini and restarted sensorfwd service. The device is a LOT snappier, like before I created thar hw-settings.ini. | 23:05 |
| T42 | <edp_17> mal: I think got it. When I enabled the LightSensor, I got that error in sensorfwd service and the device became sluggish again. | 23:11 |
| T42 | <edp_17> Does it mean that light sensor hardware faulty? | 23:14 |
| mal | @edp_17 not sure | 23:43 |
| T42 | <edp_17> I'll install an android and will test there to ensure it works. | 23:45 |
| T42 | <edp_17> I have another question. Leds do not work at all and I do not have /sys/class/leds. Do you knwo where else I can look for it? | 23:46 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!
