| T42 | <edp_17> @b100dian : There is nothing disabled in disabled_services.rc: https://paste.ubuntu.com/p/GkjHfRrJ42/. | 08:09 |
|---|---|---|
| T42 | <edp_17> How do you mean any version of services? | 08:10 |
| T42 | <edp_17> This is the result for searching for *iometri* and *ingerp* : https://paste.ubuntu.com/p/RrqpX9ymqw/ | 08:13 |
| T42 | <edp_17> I think I get it. I have to copy or symlink /system/lib64/android.hardware.biometrics.fingerprint@2.1.so to /usr/libexec/droid-hybris/system/lib64/android.hardware.biometrics.fingerprint@2.1.so | 08:29 |
| T42 | <edp_17> Let me try | 08:29 |
| T42 | <b100dian> I was gonna say, you need to figure out why it isn't running | 09:04 |
| T42 | <edp_17> Well, I symlinked and then copied that library but when restarted Fingerprint daemon: "Start operation timed out. Terminating." | 09:13 |
| T42 | <edp_17> So, the "openat(AT_FDCWD, "/usr/libexec/droid-hybris/system/lib64/android.hardware.biometrics.fingerprint@2.1.so", O_RDONLY|O_CLOEXEC) = -1 ENOENT (No such file or directory)" error is gone now: https://paste.ubuntu.com/p/3B2vhZ8GJC/ | 09:17 |
| T42 | <edp_17> Hah, there is another one missing: /usr/libexec/droid-hybris/system/lib64/android.hardware.gatekeeper@1.0.so", 0x7fdfff7508, AT_SYMLINK_NOFOLLOW) = -1 ENOENT (No such file or directory) | 09:17 |
| T42 | <edp_17> I've found four .so libraries that I copied from /system/lib64/ to /usr/libexec/droid-hybris/system/lib64/. The daemon still times out. | 09:30 |
| T42 | <edp_17> Now, the strace shows some other missing files like "usr/share/qt5/qtlogging.ini" and some others. Are these really missing or just red herring and I should ignore these? Strace: https://paste.ubuntu.com/p/ZFpzHdz2fV/ | 09:32 |
| T42 | <b100dian> Instead of copying you could link them in /odm/lib64 if you don't have that mount | 09:54 |
| T42 | <b100dian> What are you stracing? Which service? | 09:54 |
| T42 | <edp_17> I've deleted the copied libraries and instead I symlinked them into /usr/libexec/droid-hybris/system/lib64/ | 10:17 |
| T42 | <edp_17> The strace command is at the top, it is : strace /usr/bin/sailfish-fpd-community | 10:18 |
| T42 | <edp_17> Why should I link those libraries into /odm/lib64 ? | 10:18 |
| T42 | <edp_17> Isn't symlinking them into /usr/libexec/droid-hybris/system/lib64/ not enough? | 10:19 |
| T42 | <edp_17> The strace says the process is trying to access them in /usr/libexec/droid-hybris/system/lib64 rather than /odm/lib64 | 10:20 |
| T42 | <edp_17> It seems there is an issue with gatekeeper as /data/misc/gatekeeper is missing. It's in logcat at line 2000 : https://paste.ubuntu.com/p/SPFZgpbQb6/ | 10:21 |
| T42 | <edp_17> Also this looks like a problem: E fpc_tac : Unable to open '/sys/devices/platform/soc/619d000.etm/modalias' | 10:22 |
| T42 | <edp_17> and there are more similar | 10:22 |
| T42 | <edp_17> "E fpc_tac : Unable to open '/sys/devices/platform/soc/6019000.cti/modalias'" | 10:24 |
| T42 | <edp_17> On my device there is no '/sys/devices/platform/soc/6019000.cti' but I do have '/sys/devices/platform/soc/6019000.jtagmm' where there is 'modalias' | 10:24 |
| T42 | <b100dian> Hmm hard to believe that that is burned unto an executable or library, but you voukd try strings command on them | 10:39 |
| T42 | <b100dian> Maybe there's a config file that doesn't match the device your kernel is creating | 10:39 |
| T42 | <b100dian> Or a softlink maybe `find / -lname` would find it | 10:40 |
| T42 | <b100dian> Not sure what is jtagmm or cti, maybe your kernel sources or dts fikes have something in them | 10:41 |
| T42 | <edp_17> "Strings command" what's that? and how to do that? | 10:44 |
| T42 | <b100dian> Ok, here's what jtagmm https://android.googlesource.com/kernel/msm/+/android-7.1.0_r0.2/Documentation/devicetree/bindings/arm/msm/jtag-mm.txt | 10:45 |
| T42 | <b100dian> `strings libfile.so | less` would allow you to see if there's that file path burned into the library | 10:46 |
| T42 | <b100dian> Or | grep 6019000 | 10:46 |
| T42 | <b100dian> Or `find /path -type f -iname \*.so -exec grep -q 6019000 {} \; -print` | 10:48 |
| T42 | <b100dian> Path being vendor or system lib or wherever you found those libs | 10:48 |
| T42 | <edp_17> So the path should be /system/lib64/ where I found android.hardware.biometrics.fingerprint@2.1.so? Right? | 10:50 |
| T42 | <edp_17> I am sorry but really I have no idea what am doing here. | 10:51 |
| T42 | <edp_17> Output of 'strings /system/lib64/android.hardware.biometrics.fingerprint@2.1.so' : https://paste.ubuntu.com/p/Z4Y9Cw92wx/ | 10:54 |
| T42 | <edp_17> The 'find /system/lib64/ -type f -iname \*.so -exec grep -q 6019000 {} \; -print' didn't find nothing. | 10:55 |
| T42 | <edp_17> When I did a "strings /system/lib64/android.hardware.biometrics.fingerprint@2.1.so | less | grep IBiometricsFingerprint/default", it didn't find nothing. So the correspondent error in logcat is true: "droid-hal-init: Could not find 'android.hardware.biometrics.fingerprint@2.1::IBiometricsFingerprint/default' for ctl.interface_start" | 11:00 |
| T42 | <b100dian> So that path is not hatdcoded into an executable | 12:31 |
| T42 | <b100dian> *hardcoded | 12:31 |
| T42 | <edp_17> Which path? The /system/lib64/ ? | 12:56 |
| rinigus | @b100dian: could you please add "logcat -b radio" to the issue https://github.com/sonyxperiadev/bug_tracker/issues/847 | 13:30 |
| T42 | <b100dian> In about an hour yes | 14:07 |
| rinigus | @b100dian: thanks! | 14:07 |
| T42 | <b100dian> @edp_17 that soc path | 14:08 |
| T42 | <b100dian> @edp_17 I nitice you have camera recording issues on L17.1. Did you apply all recent patches, dhd submodule update and other submodule updates? | 14:10 |
| T42 | <b100dian> https://piggz.co.uk/sailfishos-porters-archive/index.php?log=2022-11-03.txt#line79 | 14:14 |
| T42 | <edp_17> @b100dian : No, I haven't applied them yet. I was planning to come back to the video-recording issue after the Fingerprint. However, I'll try this if it helps. | 14:19 |
| T42 | <edp_17> I need to remember what device that was back in 2022. :) | 14:21 |
| T42 | <edp_17> I've just checked and can confirm, I have applied all those patches. | 14:27 |
| T42 | <edp_17> Also, before started building 4.6, I've deleted/recreated sdk, target & tooling and updated all submodules. | 14:28 |
| T42 | <edp_17> The only issue I have with camera is when trying to record a video with the main camera, the app freezes and it only creates an empty *.mp4 file. | 14:29 |
| mal | anything about waiting in logcat? | 14:30 |
| T42 | <edp_17> mal, Hi there. Yes, Waiting for service 'media.audio_flinger' : logcat: https://paste.ubuntu.com/p/KyDX6XtJ8q/ | 14:32 |
| mal | are you sure dummy af is installed? | 14:35 |
| mal | and working | 14:35 |
| T42 | <edp_17> @b100dian : Ahhhhhh, now I understand why you asked that. After I cloned libhybris, I did not applied the patches. (However, at the end didn't need to build libhybris locally as I built in OBS and installed it from from there. | 14:40 |
| T42 | <edp_17> mal: are you referring to audiosystem-passthrough-dummy-af ? No, that is not installed. | 14:41 |
| T42 | <edp_17> I only have audiosystem-passthrough installed. | 14:42 |
| mal | install dummy af | 14:42 |
| mal | it's usually needed for camera | 14:43 |
| T42 | <edp_17> I thought that was for audio. | 14:43 |
| mal | video recording uses audio :) | 14:43 |
| T42 | <edp_17> Yeah. :) | 14:43 |
| T42 | <edp_17> How should I know when it is working? | 14:43 |
| mal | if the waiting is gone, also checking binder-list for hwbinder or binder to see if audio flinger is there | 14:44 |
| T42 | <edp_17> Rebooting now. | 14:44 |
| T42 | <edp_17> The "logcat | grep media.audio_flinger" doesn't give me nothing back. Good sign. | 14:46 |
| T42 | <edp_17> How to check binder-list? | 14:46 |
| mal | it starts to wait only once you use camera | 14:46 |
| T42 | <edp_17> Oh, okay. | 14:46 |
| mal | I think only after trying to record a video | 14:48 |
| T42 | <edp_17> Yeah, that was it. Video-recording is working now. Many thanks! (again) | 14:49 |
| T42 | <b100dian> rinigus: https://github.com/sonyxperiadev/bug_tracker/issues/847#issuecomment-2402574579 | 14:56 |
| rinigus | @b100dian: thanks! let's see if sodp devs can help | 14:58 |
| T42 | <b100dian> looks like a missing so file | 15:00 |
| rinigus | @b100dian: maybe that the one needed in fresher blob. will try to look through latest changes, maybe there was some config that worked. | 15:03 |
| T42 | <b100dian> I'll look into a less recent odm sometime later today | 15:06 |
| rinigus | @b100dian: it is also possible that they made changes on open source side that have not been covered by older blobs. so, its possible that blobs follow some older version of aosp. in this respect, testing older sources against current blobs could make sense | 15:38 |
| rinigus | its "just" finding a right combination is very hard. pity that they don't post use this+that and it will wortk | 15:39 |
| rinigus | @b100dian: looks like the work on audio updates for nagara started at the end of august (https://github.com/sonyxperiadev/device-sony-pdx224/commits/t-mr1 https://github.com/sonyxperiadev/device-sony-nagara/commits/t-mr1/ and https://github.com/sonyxperiadev/local_manifests/commits/android-13.0.0_r75/ ). as local manifests are not tagged, tags would have to be added to get older configs. | 18:18 |
| rinigus | but it is some work to start changing the build to follow its state around april-may (or at least before end of aug). | 18:19 |
| rinigus | whether it will work with blobs from the end of april - hard to say | 18:20 |
| rinigus | problem of course would be to deviate from the current build as bug reports for such case should be probably ignored | 18:21 |
| T42 | <b100dian> the two repos seem to be tagged as https://github.com/sonyxperiadev/local_manifests/blob/android-13.0.0_r75/devices.xml#L43 | 18:38 |
| T42 | <b100dian> (nagara and pdx224 I mean use branches t-mr1) | 18:38 |
| T42 | <b100dian> rinigus: not sure I follow why are we looking for e.g. end of april. Since you mention audio updates from august. Do you think that is breaking changes? | 18:39 |
| rinigus | @b100dian: problem that these branches are changing. btw the same branches were used for _r30 which seemed to have 5.10 kernel. | 18:40 |
| rinigus | based on the commit comments, 5.15 is using new audio system. whether it used it earlier in april, I am not sure | 18:41 |
| rinigus | it is also possible that it did use it already and nagara + yodo devices had non-functioning audio with 5.15. but that would be rather strange to have devices on platform in such state for so long | 18:43 |
| rinigus | so, if it worked in april, then those audio changes may have made some of it. | 18:44 |
| rinigus | @b100dian: but there are rather large number of guesses over here. and not sure how was it impacting gsm | 18:48 |
| T42 | <b100dian> There's also the fact that I had to revert a commit in device-sony-common | 18:51 |
| T42 | <b100dian> so basically I should try to navigate a bit back down the lane in those three repositories. Or are there more? | 18:51 |
| rinigus | @b100dian: maybe device-sony-sepolicy, but that can probably be ignored as it is mainly about camera | 18:58 |
| rinigus | and who cares about camera if you cannot use it as a phone? :) | 18:58 |
| rinigus | @b100dian: others, as far as I looked, seem to be changing less | 19:01 |
| T42 | <b100dian> Ok, taking a shot at picking some april-ish commits, mainly 5.10 is the next target | 19:24 |
| rinigus | @b100dian: thanks for trying. note that in april and already before that the blobs were for kernel 5.15. so, its bit strange that corresponding change for kernel was done on apr 4 (https://github.com/sonyxperiadev/device-sony-nagara/commit/f0f937134189a8d18b5f629e44eb7d333f58756b) | 19:38 |
| rinigus | thus, you could choose the kernel (and blobs) based on whether you use earlier or later state :) | 19:38 |
| rinigus | good luck! | 19:38 |
| T42 | <b100dian> Thaodan: you seem to have had a stab at nagara around 5 month ago, did you build AOSP for it and if so, did it worked and what is your `repo status` ? | 19:53 |
| T42 | <b100dian> rinigus: yes, thanks for the hint. It would make sense to try 5.15 first and switch only that (and odm) if failed | 19:54 |
| T42 | <b100dian> rinigus: what do you think it's this TARGET_VENDOR_VERSION https://github.com/sonyxperiadev/device-sony-common/commit/6b13b6a2cc33a96dd29a3f62d436c45a52354b36 ? | 20:37 |
| T42 | <b100dian> I wonder if I should go before that commit or after | 20:37 |
| T42 | <edp_17> Any advise on the fingerprint related error on my port? Can I get an example where it is implemented and I can check repos/commits/etc? | 20:41 |
| T42 | <b100dian> @edp_17 what is your kernel source? Maybe something is missing | 20:45 |
| T42 | <b100dian> I did not have to do any more than the packaging part for my xiaomi | 20:45 |
| T42 | <b100dian> (actually I had to change the sources to fit the optical in-screen sensor, but nothing that was not present out of the box in the droid layer) | 20:46 |
| T42 | <edp_17> @b100dian : Ocean kernel : https://github.com/edp17/android_kernel_motorola_sdm632 | 20:48 |
| T42 | <edp_17> Branch : lineage-17.1 | 20:48 |
| T42 | <edp_17> Thanks in advance. | 20:49 |
| T42 | <b100dian> do you have `/data/vendor/.fps` folder? | 21:09 |
| T42 | <b100dian> also `/data/misc/keystore` | 21:11 |
| T42 | <b100dian> These are from your previous paste at https://paste.ubuntu.com/p/SPFZgpbQb6/ | 21:11 |
| T42 | <b100dian> also gatekeeper seems to be used, and `/data/misc/gatekeeper` is also missing? | 21:30 |
| T42 | <b100dian> I have these folders on tucana, except .fps, I have fpc for example (empty) | 21:37 |
| T42 | <edp_17> None of these exist on my device. No /data/misc/keystore no /data/misc/gatekeeper and no /data/vendor/.fps | 22:28 |
| T42 | <edp_17> Should I just create those folders? | 22:30 |
| T42 | <edp_17> Or change something in kernel/device and that will automatically create them? | 22:30 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!