| T42 | <Coltor %lastname%> @rinigus [Coltor, Herrie1982: https://bugreports.qt.io/b …], Does sfos's qtwebengine support hardware acceleration now ? | 02:35 |
|---|---|---|
| T42 | <Coltor %lastname%> it seems sfos the 3.3.4 version switch to Gecko Engine,not use qtwebengine ? | 03:13 |
| r0kk3rz | sailfish has never used qtwebengine for anything | 03:20 |
| r0kk3rz | sailfish browser was always gecko | 03:20 |
| T42 | <Coltor %lastname%> i got it~~ it seems 3.4 updates Gecko Engine https://github.com/sailfishos/sailfish-browser | 03:20 |
| r0kk3rz | email client used qtwebkit before, now also uses gecko | 03:21 |
| T42 | <Coltor %lastname%> However, the WebView widget used by other apps is still qtwebengine? | 03:23 |
| r0kk3rz | no | 03:25 |
| r0kk3rz | it was never qtwebengine | 03:25 |
| r0kk3rz | its probably still qtwebkit, with maybe an option to move to gecko | 03:27 |
| T42 | <Coltor %lastname%> @r0kk3rz [sailfish browser was always gecko], After upgrading gecko, I found that the fps was improved a lot. It seems that gecko already supports hardware acceleration | 03:30 |
| rinigus | Coltor: not sure which bug report do you refer to in qt. Assuming that it's one that I filed regarding webengine, it is when running from flatpak on hybris. It is partially hw accelerated, as described in the issue | 04:47 |
| T42 | <edp_17> Good morning. | 08:27 |
| T42 | <edp_17> I have one more outstanding and annoying problem on my treltexx. The USB is set to 'Always ask' but when I plug the cable in, the selection doesn't pop up. If I restart the usb-moded service, the question pops up. Do you have an idea how to fix it? Thanks. | 09:29 |
| T42 | <edp_17> Journal: https://paste.ubuntu.com/p/R37hJBBv94/ | 09:34 |
| T42 | <edp_17> Dmesg: https://paste.ubuntu.com/p/TCRzGB955Y/ | 09:34 |
| T42 | <edp_17> There is nothing in logcat. | 09:34 |
| T42 | <Darcy %lastname%> (Photo, 702x628) https://irc.thaodan.de/.imgstore/pOvRVgFchZ.png 推广，炸群，拉人，增粉，推特/脸书关注 ，联系上方图片 | 10:05 |
| T42 | <Anton %lastname%> How do we do audio nowadays? Do we still use audioflingerglue? I seem to have troubles building it? | 14:41 |
| mal | do not use that | 14:42 |
| mal | we use https://github.com/mer-hybris/audiosystem-passthrough now | 14:43 |
| mal | those are built automatically by build_packages.sh | 14:44 |
| T42 | <Anton %lastname%> So, it shouldn't be in patterns then? | 14:45 |
| mal | no, read the instructions in that the repo I linked | 15:11 |
| mal | it says "For new adaptations audioflingerglue (and pulseaudio-modules-droid-glue) should not be used" | 15:12 |
| T42 | <Anton %lastname%> Here's a marvelous revelation for you all. Apparently starting /vendor/bin/sensors.qti and restarting sensorfwd fixes sensors, camera and sound. Wonder why it didn't start automatically. Permissions to the file are 755 system system, should I throw a 777 on it? | 16:48 |
| T42 | <Anton %lastname%> My concern is that it's on vendor partition and I can't remember how to do it gracefully (I recall something about udev, but that was a whole another story) | 16:55 |
| mal | Anton which android base are you using? | 17:18 |
| T42 | <Anton %lastname%> Android 9, hybris-16, full treble support | 17:20 |
| mal | which .rc file has the sensors.qti ? | 17:22 |
| mal | most likely in /vendor/etc/init/ or something | 17:22 |
| T42 | <Anton %lastname%> Found it in /vendor/etc/init/hw/init.qcom.rc | 17:27 |
| mal | so why is that not starting it | 17:31 |
| mal | check dmesg or journal to see if that was loaded | 17:31 |
| mal | I assume you setup the needed things to /usr/libexec/droid-hybris/system/etc/init/ ? | 17:32 |
| T42 | <Anton %lastname%> It's trying to start it, but says something about unsufficient permissions | 17:36 |
| mal | hmm, can you pastebin the relevant part of the rc file, also check the permissions of the sensors.qti file | 17:39 |
| T42 | <Anton %lastname%> There are files in that directory, but none related to that specific one | 17:39 |
| T42 | <Anton %lastname%> Is it that codeword about service being disabled? https://pastebin.com/Lp9YGGyR | 17:43 |
| T42 | <Anton %lastname%> Please don't tell me it is | 17:43 |
| mal | it's possible some other part in .rc files is starting that, grep for sensors.qti in all .rc files | 17:44 |
| mal | in both /system and /vendor | 17:44 |
| T42 | <Anton %lastname%> Well, it's the only one "find" could find | 17:46 |
| T42 | <Anton %lastname%> "cat $(find /system -name "*.rc") | grep sensors.qti" if you're curious | 17:46 |
| mal | btw, easier way is "grep -RI --include=*.rc sensors.qti /system" | 17:48 |
| T42 | <elros34> its probably in the same file init.qcom.rc in /vendor. Something like init.qcom.sensors.sh | 17:48 |
| mal | that is also possible | 17:48 |
| T42 | <Anton %lastname%> Only one line returns sensors.qti, so unlikely | 17:49 |
| T42 | <Anton %lastname%> But there is | 17:49 |
| mal | but you said there is some permission error? | 17:49 |
| T42 | <elros34> because you are searching in .rc files but this can be started in .sh script | 17:50 |
| T42 | <elros34> so just grep whole device repo | 17:50 |
| T42 | <Anton %lastname%> Yes, indeed. The service is started from /vendor/bin/init.qcom.sensors.sh, which is 755 root shell | 17:51 |
| T42 | <Anton %lastname%> https://pastebin.com/FahcA3qj here it is | 17:52 |
| mal | so where was the error you were talking about | 17:53 |
| T42 | <Anton %lastname%> https://pastebin.com/KdtiLDGf something along these lines | 17:57 |
| T42 | <elros34> Anton Klimanov I bet this is because nemo is no longer in system group. Add it back to confirm it | 18:17 |
| T42 | <Anton %lastname%> Isn't nemo being in system group a bad practice? | 18:18 |
| mal | yes, that's why it was removed, some device might need to small fixes to udev rules for example | 18:20 |
| T42 | <elros34> yes that is why it was removed recently but IMHO adding it back to confirm issue is just a matter of 1 minute test | 18:20 |
| T42 | <Anton %lastname%> Didn't make a difference | 18:22 |
| T42 | <elros34> did you reboot? | 18:22 |
| T42 | <Anton %lastname%> Yep | 18:22 |
| T42 | <elros34> ah ok | 18:22 |
| T42 | <Anton %lastname%> At least one thing's for sure. It's definitely because nemo starts it and not root | 18:35 |
| T42 | <Anton %lastname%> And if it is run by nemo, it crashes with "Diag_LSM_Init: Failed to open handle to diag driver, error = 13" | 18:37 |
| T42 | <elros34> based on some random kernel source this function check for net_raw group | 19:16 |
| T42 | <elros34> I mean msm_ipc_router_bind | 19:19 |
| T42 | <Anton %lastname%> Could I take a look at that source? | 19:29 |
| T42 | <Anton %lastname%> There is a possibility of it being unable to open the third and last MSM_IPC | 19:30 |
| T42 | <elros34> search for that function in your kernel, I don't know what device you use | 19:31 |
| T42 | <Anton %lastname%> I got it, but it's useless | 19:33 |
| T42 | <Anton %lastname%> That'd be so great to know which exact pipe or file is denying | 19:33 |
| T42 | <elros34> how about you add net_raw group to sensors.qti in init.qcom.rc? | 19:40 |
| T42 | <Anton %lastname%> Excuse me, but how do you get that? Surprisingly, that worked! | 19:46 |
| T42 | <elros34> like I said msm_ipc_router_bind check for that group and fail with "Do not have permissions" same as in your logs. Better question is why is that even needed if it works in android | 19:50 |
| mal | maybe something in /dev has wrong permissions? | 19:51 |
| mal | just a theory | 19:51 |
| mal | but that would have been solved by system group | 19:52 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!
