| rinigus | @b100dian: good question. No idea what should we target | 02:16 |
|---|---|---|
| rinigus | Should write it down to remember to check both options | 02:17 |
| T42 | <elros34> @edp_17 are you sure that you remove encryption by formating /data before flashing sailfish? | 05:39 |
| T42 | <edp_17> Yes, I did format /data before that. | 07:35 |
| T42 | <edp_17> Isn't that mounted on /data in Sailfish? | 07:36 |
| T42 | <edp_17> Hmm, The ls /data/media/0/ shows me garbage. | 07:37 |
| T42 | <edp_17> Looks like encrypted: https://paste.ubuntu.com/p/8gD4YCCqHZ/ | 07:38 |
| T42 | <elros34> I don't know but "start time_daemon" based on encryption status in your time daemon.rc and error "required key ot available" in logcat suggest you did boot to android | 07:40 |
| T42 | <elros34> which enabled I guess FBE encryption | 07:41 |
| T42 | <edp_17> Hmm, could be. I remember after flashing Sailfish somehow Slot B was active and I had to reboot into TWRP to switch to Slot A. Is it possible to remove that encryption without breaking my Sailfish installation? Or easier to install it from scratch again? | 07:43 |
| T42 | <elros34> not sure, check /data/tombstone*, this throws mentioned error. Booting TWRP shouldn't encrypt enything | 07:46 |
| T42 | <edp_17> Probably Lineage started to boot when I didn't realize which slot was active. | 07:49 |
| T42 | <edp_17> There is /data/tombstone filder but it's empty | 07:49 |
| T42 | <elros34> what about getprop, anything encryption related like ro.crypt.state? | 07:51 |
| T42 | <edp_17> getprop: https://paste.ubuntu.com/p/gh5nHBmjch/ | 07:56 |
| T42 | <edp_17> There are a few *crypt* lines, but I am not sure. | 07:58 |
| T42 | <edp_17> Let me build a new zip, install it fresh and see what happens. At some points I will need to unit test the build anyway. | 07:59 |
| T42 | <elros34> not sure how to test this encryption, maybe play (recreate) with /data/tombstones directory, create some files there to see if you have access and related error is still in logcat | 08:01 |
| T42 | <edp_17> I've deleted and created that tombstone folder and touched a file into it (as root). | 09:45 |
| T42 | <edp_17> Now the "sudo logcat | grep required" gives me only one line: "10-09 15:18:10.245 4353 4353 I imsrcsd : CM service not required" | 09:47 |
| T42 | <edp_17> However, sailfish-fpd-community service is still time out. | 09:50 |
| T42 | <edp_17> I've built a new image and will try it on device with ensuring /data is not encrypted this time. | 09:51 |
| T42 | <elros34> there were other failing services: keystore fpc_fingerprint_hal an dos one with some file errors but anyway | 10:01 |
| T42 | <elros34> if you insist on reflashing | 10:02 |
| T42 | <edp_17> I'd like to try the new image, just in case somebody would like to install it. So yes, I'll do it. (Yeah, the big hassle is configuring the device and install all apps from scratch again.) | 10:11 |
| T42 | <edp_17> Hmm, After installing sailfish fresh, there is no /data/media/0 folder. | 11:06 |
| T42 | <edp_17> @b100dian : I've reinstalled sailfish with ensuring /data is not encrypted. The fingerprint service is still time out but now I do have /data/vendor/.fps and /data/misc/keystore and /data/misc/gatekeeper folders : https://paste.ubuntu.com/p/3Qb7np6mGC/ | 11:35 |
| T42 | <edp_17> I'll get logcat. | 11:36 |
| T42 | <edp_17> Logcat: https://paste.ubuntu.com/p/w29gG5jb9z/ | 11:37 |
| rinigus | @b100dian: any luck last night? no replies yet to your issues at sony :( | 12:01 |
| rinigus | and Thaodan hasn't given any info regarding state of nagara either. | 12:03 |
| T42 | <b100dian> rinigus: last night I picked some commits (and lowered the local_manifests version to 94006c86463125add9879eadc9d58b34c9f2dd61, basically without the last commit) | 12:35 |
| T42 | <b100dian> And got "vendor/qcom/opensource/audio-hal/st-hal/sound_trigger_platform.c:75:10: fatal error: 'linux/msm_audio_calibration.h' file not found" | 12:36 |
| T42 | <b100dian> Today I backed off to 5.10 and the build just passed | 12:36 |
| T42 | <b100dian> However I will not be able to test right now | 12:36 |
| mal | @b100dian about this https://github.com/libhybris/libhybris/pull/569 we were thinking that the type should be int instead of char, probably should not matter much but afaik enums are int | 12:37 |
| mal | I was planning on merging that when noticed that | 12:38 |
| T42 | <b100dian> It was not very clear to me that they are int, since 6.7.2.2 Enumeration specifiers says "Each enumerated type shall be compatible with char, a signed integer type, or an unsigned integer type" according to stackoverflow :)) | 12:39 |
| T42 | <b100dian> So char was chosen as the safest / narrowest access, and should work - provided big endianness | 12:40 |
| T42 | <b100dian> Link would be: https://stackoverflow.com/a/366033 | 12:40 |
| rinigus | @b100dian: interesting. let's see what it will do when you get there | 12:41 |
| rinigus | thanks! | 12:41 |
| mal | @b100dian please change to int, link says "The expression that defines the value of an enumeration constant shall be an integer constant expression that has a value representable as an int." | 12:44 |
| mal | afaik the type is in but can be converted always to the other types listed | 12:44 |
| T42 | <b100dian> Sure. For a normal function call I can imagine that conversion happening, for a bionic to glibc adapter I wasn't that sure. Maybe they're both in one register and it doesn't matter | 12:47 |
| T42 | <b100dian> mal: updated | 12:52 |
| mal | thanks, I'll merge it in a moment, along with the usage PR I made | 12:53 |
| T42 | <b100dian> @edp_17 maybe you neeed to strace `/vendor/bin/hw/android.hardware.biometrics.fingerprint@2.1-fpcservice` to see what "fpc_hal_open" and "sensor_command, Failed to send command" mean | 16:31 |
| T42 | <elros34> mind that there are more relevant or not output lines but they are suppresed: chatty | 16:43 |
| T42 | <b100dian> rinigus: not booting to GUI the build I finished today. Starting another one | 17:24 |
| rinigus | @b100dian: rather remarkable how much struggle is to bring aosp up on that device. | 17:26 |
| T42 | <b100dian> Well, if we'd knew the right combination of commits it probably wouldn't be | 17:27 |
| T42 | <b100dian> If they used tags :D... | 17:29 |
| rinigus | @b100dian: why they don't is really strange. | 17:30 |
| rinigus | @b100dian: but which blobs did you use? the ones for 5.15 or 5.10? | 17:35 |
| T42 | <b100dian> That was easy, the 5.10 ones. For the next build I am again trying 5.15 and will try a previous version of 5.15 blobs too | 17:36 |
| rinigus | OK, makes sense | 17:36 |
| T42 | <b100dian> Unfortunately I have the same error that made me think switching to 5.10 was a good idea | 18:18 |
| T42 | <b100dian> in vendor/qcom/opensource/audio-hal/st-hal/sound_trigger_platform.c, "fatal error: 'linux/msm_audio_calibration.h' file not found" | 18:19 |
| T42 | <b100dian> I am not sure how this has built before, the st-hal is not something new according to the manifest | 18:19 |
| rinigus | @b100dian: could that be some of the repos that stayed after new sync? | 18:29 |
| T42 | <b100dian> I was thinking st-hal was introduced by me dismissing this commit but it is not https://github.com/sonyxperiadev/local_manifests/commit/ee0f1b1f3694c5f781865bb72a5fbc31c73e02e9 | 18:31 |
| rinigus | @b100dian: looks like that was there for a while already | 18:53 |
| rinigus | st-hal-ar was added in aug - maybe that's why you had this impression | 18:54 |
| T42 | <b100dian> I thought soong would just build ./**/Android.mk files could it be it is pulled in by something? | 18:55 |
| T42 | <b100dian> That header file was definitely missing from 5.15 all along | 18:55 |
| rinigus | @b100dian: but you were able to work around it before, right? | 18:58 |
| T42 | <b100dian> rinigus: not quite. the initial repo sync "just built" (with an unrelated revert). | 19:04 |
| T42 | <b100dian> After trying other commits in the repos, this error message lead me to think that 5.10 was neede | 19:05 |
| T42 | <b100dian> d. Because i thought it was st-hal-ar's fault | 19:05 |
| T42 | <b100dian> but now I am not sure why it would've built | 19:06 |
| rinigus | @b100dian: quite blank over here and hard to say why it managed to do so before | 19:12 |
| T42 | <b100dian> yeah, don't worry, I'll probably copy some struct to get this going :D | 19:13 |
| rinigus | :) | 19:14 |
| T42 | <b100dian> enum actually. so much safer | 19:16 |
| T42 | <b100dian> Hopefully it is just not used as a build output | 19:19 |
| T42 | <rfasciani> Good day all. Anyone know how can I install waydroid on Sailfish OS running on a Xperia 5? Cheers. | 19:59 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!