| T42 | <TheVancedGamer> @elros34 I gave up on 9C, reported Redmi Note 9 with hybris-17.1 and now I have telnet on port 23, I had diagnosis.log previously saying AUTOFS4_FS is missing, after enabling that now I have no diagnosis.log and init.log only mentions "Failed to boot init in real rootfs" | 06:46 |
|---|---|---|
| T42 | <TheVancedGamer> I should mention this device is also dynamic partitions based, should I try to add parse-android-dynparts now or later? | 06:47 |
| T42 | <elros34> gave up without even trying? For sure there is more than that error and it's related to your data partition. You will need to handle dynamic partitions anyway but this is not related | 07:03 |
| T42 | <TheVancedGamer> that device was just pain | 07:03 |
| T42 | <TheVancedGamer> aarch64 kernel, armhf trees | 07:03 |
| T42 | <TheVancedGamer> anyway, for this device I have data mounted on /data, /target/data | 07:04 |
| T42 | <TheVancedGamer> what am I looking for in init.log? | 07:06 |
| T42 | <elros34> errors as usual | 07:06 |
| T42 | <TheVancedGamer> wait, I don't have sbin/ in /target | 07:07 |
| T42 | <TheVancedGamer> is that normal? | 07:07 |
| T42 | <elros34> did you install sailfish without error? | 07:08 |
| T42 | <TheVancedGamer> yes | 07:08 |
| T42 | <elros34> how did you managed that? | 07:10 |
| T42 | <TheVancedGamer> installed it from twrp after building mic? | 07:12 |
| T42 | <TheVancedGamer> i'm building for a device with much better trees, 9C was just a mess | 07:12 |
| T42 | <elros34> and it works just like that, impossible:) | 07:13 |
| T42 | <TheVancedGamer> / # ls /target/ | 07:13 |
| T42 | <TheVancedGamer> afs/ data/ etc/ home/ init.rc proc/ root/ usr/ | 07:13 |
| T42 | <TheVancedGamer> | 07:13 |
| T42 | <TheVancedGamer> i believe this is concerning | 07:13 |
| T42 | <elros34> show that log | 07:13 |
| T42 | <TheVancedGamer> how do I pull files from telnet? :P | 07:14 |
| T42 | <elros34> what do you want to pull? Format /data in twrp than then boot again and install sfos manually because twrp installation almost never works | 07:16 |
| T42 | <TheVancedGamer> ok | 07:16 |
| T42 | <elros34> boot again hybris-boot not twrp | 07:17 |
| T42 | <TheVancedGamer> so after formatting data and rebooting, I get telnet | 07:19 |
| T42 | <TheVancedGamer> I unpack rootfs in telnet? | 07:19 |
| T42 | <iDrinkCoffee> hello world, it is I, the troll | 07:20 |
| T42 | <TheVancedGamer> ok @elros34 I now have init-debug in real rootfs | 07:35 |
| T42 | <TheVancedGamer> thanks for the pointers | 07:35 |
| T42 | <TheVancedGamer> so I add | 08:03 |
| T42 | <TheVancedGamer> ``` | 08:03 |
| T42 | <TheVancedGamer> # For mount partitions | 08:03 |
| T42 | <TheVancedGamer> Requires: parse-android-dynparts``` | 08:03 |
| T42 | <TheVancedGamer> | 08:03 |
| T42 | <TheVancedGamer> to my patterns | 08:03 |
| T42 | <TheVancedGamer> but how will it get built? | 08:03 |
| T42 | <TheVancedGamer> where do I define the repo so that issuing just —mw would build it? | 08:04 |
| T42 | <elros34> you need to build it yourself, search recent channel log, mal gives more up-to-date instruction with fp4 example how to handle all of this | 08:05 |
| T42 | <TheVancedGamer> i've built it already, I want to autobuild it | 08:06 |
| T42 | <elros34> so read the script | 08:07 |
| T42 | <TheVancedGamer> found it, yes | 08:07 |
| T42 | <TheVancedGamer> but it's built by dhd, and that's pulled as a module from sfos repos | 08:07 |
| T42 | <TheVancedGamer> forking would become a bit unnecessary, no? | 08:09 |
| T42 | <TheVancedGamer> or should I fork it and add the buildmw line for parse-android-dynparts | 08:09 |
| T42 | <elros34> It's up to you unless you want to upstream that build but not sure this will be included | 08:10 |
| T42 | <TheVancedGamer> also, why isn't cpio just added into allowed paths? | 08:10 |
| T42 | <TheVancedGamer> I could make a PR for it if needed | 08:10 |
| T42 | <elros34> IIRC cpio is addedin hybris18 | 08:11 |
| T42 | <TheVancedGamer> not in 17.1, and it doesn't show it either | 08:12 |
| T42 | <TheVancedGamer> it's also possible to add the executable into soong config paths which eliminates the need for the temporary path restrictions var | 08:12 |
| T42 | <elros34> so this patch need to be added fro 17 base: https://github.com/mer-hybris/hybris-patches/commit/8dbd868f98d73781d2747764188d47f4db288cc6 | 08:15 |
| T42 | <TheVancedGamer> yes, basically | 08:15 |
| T42 | <TheVancedGamer> this should apply as-is | 08:15 |
| T42 | <k1gen> mal: hey, anything else to try with sound? | 09:06 |
| T42 | <TheVancedGamer> @elros34 I got a bit farther this time, product is mounted by dynparts service but for system I get | 12:53 |
| T42 | <TheVancedGamer> `Mar 29 15:35:32 RedmiNote9 bash[1572]: device-mapper: reload ioctl on dynpart-system failed: Invalid argument` | 12:53 |
| T42 | <TheVancedGamer> I get the same error if I run dmsetup manually as well | 12:55 |
| mal | how do you run dmsetup? using the command in my script? | 12:58 |
| T42 | <TheVancedGamer> yes | 12:58 |
| mal | @k1gen no ideas right now, need to check logs a bit more | 12:58 |
| T42 | <TheVancedGamer> more confusingly product is fine, I can access all files as well | 12:58 |
| T42 | <k1gen> mal: thank you! feel free to ping me if you need anything that can help you | 12:59 |
| T42 | <TheVancedGamer> ``` | 13:00 |
| T42 | <TheVancedGamer> / # dmsetup create --concise "$(parse-android-dynparts /dev/mmcblk0p43)" | 13:00 |
| T42 | <TheVancedGamer> device-mapper: reload ioctl on dynpart-system failed: Invalid argument | 13:00 |
| T42 | <TheVancedGamer> Command failed``` | 13:00 |
| T42 | <TheVancedGamer> but I only get /dev/mapper/dynpart-product | 13:00 |
| T42 | <TheVancedGamer> https://github.com/Redmi-MT6768/android_device_xiaomi_merlin/blob/lineage-17.1/BoardConfig.mk#L104 | 13:03 |
| T42 | <TheVancedGamer> it is ext4 so that shouldn't be an issue | 13:03 |
| T42 | <TheVancedGamer> https://paste.myself5.de/lukodezabi.pl | 13:06 |
| T42 | <TheVancedGamer> failing bit taken from strace | 13:06 |
| mal | @k1gen if you want to try to debug other issues in the meantime maybe you could get some information about sensors, check what sensor related processes are running, try to see if restarting sensorfwd systemd service helps | 13:06 |
| T42 | <TheVancedGamer> mal: I crosschecked with FP4 repo, everything is the same | 13:10 |
| mal | and partition if correct? | 13:29 |
| mal | can you show the list of partitions in for example /dev/block/bootdevice/by-name/ | 13:30 |
| T42 | <k1gen> mal: sensorfwd starts fine by systemd, here's log when I start it manually and try some sensor tests in csd: https://vomitb.in/raw/j5yWqCfnZ1. tell me if you need anything else | 13:31 |
| mal | what sensor related processes are running check "ps aux | grep -i sensor" | 13:34 |
| T42 | <k1gen> mal: https://vomitb.in/raw/ZEkB7UHEVs. sensorfwd without full path to executable is mine, not systemd's | 13:36 |
| T42 | <TheVancedGamer> mal: | 13:44 |
| T42 | <TheVancedGamer> / # ls /dev/block/platform/bootdevice/by-name/ | 13:44 |
| T42 | <TheVancedGamer> boot expdb lk misc persistbak sec1 tee2 | 13:44 |
| T42 | <TheVancedGamer> boot_para ffu lk2 nvcfg proinfo seccfg userdata | 13:44 |
| T42 | <TheVancedGamer> cache flashinfo logo nvdata protect1 spmfw vbmeta | 13:44 |
| T42 | <TheVancedGamer> cust frp md1img nvram protect2 sspm_1 vbmeta_system | 13:44 |
| T42 | <TheVancedGamer> devinfo gsort md_udc otp recovery sspm_2 vbmeta_vendor | 13:44 |
| T42 | <TheVancedGamer> dtbo gz1 metadata para scp1 super | 13:44 |
| T42 | <TheVancedGamer> exaid gz2 minidump persist scp2 tee1 | 13:44 |
| T42 | <TheVancedGamer> was I supposed to add super to fixup-mountpoints? | 13:45 |
| mal | @k1gen install libgbinder-tools if it's not installed and then show output of "binder-list -d /dev/hwbinder" | 13:56 |
| mal | or maybe just the things related to sensors from that output | 13:56 |
| T42 | <k1gen> mal: https://vomitb.in/raw/HIHsdc1mZt | 13:57 |
| T42 | <k1gen> I don't want to accidentally grep something you might find useful | 13:57 |
| mal | @k1gen to get debug output from sensorfw you need to start it with: sensorfwd -c=/etc/sensorfw/primaryuse.conf --systemd --log-level=debug --no-magnetometer-bg-calibration | 14:00 |
| T42 | <k1gen> got it, let me send the new logs | 14:00 |
| T42 | <TheVancedGamer> mal: not sure you saw it, but https://paste.myself5.de/wuzimepuni.properties | 14:01 |
| mal | do you need anything in dmesg when you run that dmsetup | 14:03 |
| T42 | <k1gen> mal: https://vomitb.in/raw/LmxsCRCpbO. I tried to access sensors while sensorfwd was launched in the following order: light, proximity, gyro & accelerometer, compass | 14:03 |
| T42 | <TheVancedGamer> hold on, I didn't check (re @SailfishFreenodeIRCBridgeBot: <mal>do you need any...) | 14:04 |
| T42 | <TheVancedGamer> [ 5437.724160] (7)[1783:dmsetup]device-mapper: table: 253:1: linear: Invalid argument count | 14:05 |
| T42 | <TheVancedGamer> [ 5437.724178] (7)[1783:dmsetup]device-mapper: ioctl: error adding target to table | 14:05 |
| mal | strange error | 14:13 |
| mal | @k1gen just curious, why did you never fix the mount service I mentioned some days ago i.e. metadata.mount and mnt-vendor-persist.mount ? | 14:16 |
| mal | those could in theory have some effect on things | 14:16 |
| mal | just replace the device node with the actual device in those | 14:17 |
| mal | you can see if real device in for example output of ls -l /dev/block/bootdevice/by-name/ | 14:17 |
| mal | *the real device | 14:17 |
| mal | so some like /dev/sda1 or whatever instead of the by-name path in the mount service | 14:18 |
| T42 | <TheVancedGamer> so what do I do about this? | 14:18 |
| T42 | <k1gen> mal: thanks, let me try fixing those | 14:19 |
| T42 | <k1gen> mal: fixed them: https://vomitb.in/raw/m1VugItFov | 14:26 |
| T42 | <k1gen> thanks for instructions | 14:27 |
| T42 | <k1gen> mal: it fixed all the sensors! | 14:28 |
| T42 | <k1gen> only sound, camera & flashlight and vibro remain | 14:33 |
| mal | yeah, I thought it should fix sensors, sometimes sensors have configuration files in persist partition | 14:34 |
| mal | was there any difference to sound issue after fixing the mounts? | 14:34 |
| T42 | <k1gen> nope, same segfault with libraries not found. do you need logs? | 14:38 |
| mal | gdb of pulseaudio with debug log output in it | 14:40 |
| mal | maybe you could install back the pulseaudio-modules-droid and remove the jb2q one since that didn't help | 14:41 |
| T42 | <k1gen> mal: do I have to remake that custom .pa file you suggested earlier? | 14:45 |
| T42 | <k1gen> oh and do I have to mask both the user pulseaudio.service and the system one or just user's? | 14:49 |
| mal | you need to change custom .pa only if you have quirks in it | 14:50 |
| T42 | <k1gen> I deleted it after installing jb2q stuff | 14:51 |
| T42 | <k1gen> it contained load-module module-droid-card hw_volume=false | 14:51 |
| mal | no need to add it back now, we might need to do that later when testing | 14:53 |
| T42 | <k1gen> got it. what about masking question? | 14:54 |
| mal | @TheVancedGamer what does that parse-android-dynparts /dev/mmcblk0p43 return on your device? | 14:54 |
| mal | you can leave it masked | 14:55 |
| T42 | <TheVancedGamer> mal: it returns all 3 dynamic partitions | 14:56 |
| T42 | <TheVancedGamer> or if you want specific output: | 14:56 |
| T42 | <TheVancedGamer> dynpart-product,,,ro,0 1386960 linear /dev/mmcblk0p43 2048;dynpart-system,,,ro,0 5453792 linear /dev/mmcblk0p43 1904640,5453792 515632 linear /dev/mmcblk0p43 1389008,5969424 32 linear /dev/mmcblk0p43 7358432,5969456 2601888 linear /dev/mmcblk0p43 8676520;dynpart-vendor,,,ro,0 1318056 linear /dev/mmcblk0p43 7358464 | 14:56 |
| *** Daaanct12 is now known as Danct12 | 14:58 | |
| mal | does running it with only the last part work (vendor) is system the only issue | 15:00 |
| mal | I mean giving vendor info to dmsetup | 15:00 |
| T42 | <TheVancedGamer> dynpart-vendor,,,ro,0 1318056 linear /dev/mmcblk0p43 7358464 ? | 15:01 |
| mal | yes | 15:01 |
| T42 | <TheVancedGamer> that works, yes | 15:01 |
| mal | now try manually only system | 15:01 |
| mal | that system looks strange, it has so many things in it | 15:02 |
| T42 | <TheVancedGamer> doesn't work | 15:02 |
| T42 | <iDrinkCoffee> this build of los is not using a opensource vendor (re @TheVancedGamer: doesn't work) | 15:02 |
| T42 | <TheVancedGamer> I just flashed lineage, booted it and installed sfos | 15:02 |
| T42 | <TheVancedGamer> doesn't matter | 15:02 |
| T42 | <iDrinkCoffee> so it a lot bigger than the actual size of the os so it fills the remaining space in super (re @iDrinkCoffee: this build of los is...) | 15:02 |
| T42 | <k1gen> mal: pa gdb: https://vomitb.in/raw/Y0pthJkQ8Y | 15:03 |
| mal | @TheVancedGamer I wonder if we have too old dmsetup | 15:05 |
| T42 | <TheVancedGamer> hmm, I don't know | 15:06 |
| T42 | <TheVancedGamer> this one works as-is on Ubuntu Touch | 15:06 |
| mal | which lvm2 does ubuntu touch have? | 15:06 |
| T42 | <TheVancedGamer> need to check, let me get my device | 15:07 |
| T42 | <TheVancedGamer> ah, parse-android-dynparts is triggered in initrd | 15:08 |
| T42 | <TheVancedGamer> mal: https://github.com/NotKit/initramfs-tools-halium/tree/dynparts | 15:09 |
| mal | @TheVancedGamer that doesn't really tell which dmsetup version they have | 15:23 |
| T42 | <TheVancedGamer> I can't seem to make it run on sailfish | 15:24 |
| T42 | <TheVancedGamer> from what I do gather dmsetup in sailfish is much older | 15:24 |
| T42 | <TheVancedGamer> library version 1.02.155 (2018-12-18) instead of 1.02.146 (2017-12-18) of sailfish | 15:28 |
| T42 | <TheVancedGamer> library version 1.02.155 (2018-12-18) for UT instead of 1.02.146 (2017-12-18) of sailfish (edited) | 15:28 |
| T42 | <TheVancedGamer> oookay, copying the library from UT ramdisk into sailfish rootfs did make dmsetup pick up the library, but still doesn't work | 15:30 |
| mal | @TheVancedGamer which file did you copy? | 15:50 |
| T42 | <TheVancedGamer> libdevmapper.so | 15:50 |
| T42 | <k1gen> mal: what do you think of gdb? is it different with metadata and persist mounted? | 16:19 |
| mal | looks the same | 16:21 |
| T42 | <k1gen> sad | 16:22 |
| T42 | <k1gen> well, ping me if you want me to try something or need more logs on pulse | 16:27 |
| T42 | <k1gen> mal: if you have time, we can debug camera until you have any ideas on sound | 16:28 |
| T42 | <k1gen> if you want to, of course | 16:28 |
| mal | how did camera fail? | 16:32 |
| T42 | <k1gen> I open the camera app after reboot and it works until: | 16:34 |
| T42 | <k1gen> 1) I make a photo and swipe left to check it | 16:34 |
| T42 | <k1gen> 2) I switch to video mode | 16:34 |
| T42 | <k1gen> 3) I switch to selfie mode | 16:34 |
| T42 | <k1gen> 4) I close the app and open it again | 16:34 |
| T42 | <k1gen> in all cases but the app freezes, in 4 it's a black screen and freeze | 16:34 |
| mal | do you see anything in logcat when that happens, about something waiting | 16:35 |
| T42 | <k1gen> let me reboot and check with working camera. I'll send logcat file | 16:36 |
| mal | does it successfully take that first picture? | 16:37 |
| T42 | <k1gen> yeah | 16:37 |
| T42 | <k1gen> by the way, there's an issue with reboots: sometimes phone goes in bootloader and stays there with "no valid slot to boot" message, it can be fixed in no time with fastboot --set-active=a, but it's still quite weird and uncomfortable. also 3 times out of 4 ui freezes on pin code entry, and ssh disconnects if already connected or doesn't let you in if you wasn't. sometimes hybris-boot.img reflash helps | 16:40 |
| T42 | <k1gen> "freezes on pin code entry" | 16:41 |
| T42 | <k1gen> ^ only after reboots, and on forth-fifth digit of six-digit passcode | 16:41 |
| mal | you probably have bootctl service failing | 16:42 |
| mal | it informs some part of the system that the boot was successful if it's not done then every 7 boots or was it 5 it fails | 16:43 |
| T42 | <k1gen> now after metadata and persist mount fine, device freezes after 3 of 4 reboots not on passcode entry, but on spinning circle animation, then screen turns off and no ssh access | 16:44 |
| T42 | <k1gen> also if I try to reboot device by holding power button, ui turns on for half a second literally to say "goodbye" | 16:45 |
| T42 | <k1gen> 10 minutes of constant rebooting already, since I said I'd take camera logcat | 16:47 |
| mal | hmm | 16:47 |
| mal | that spinning circle could suggest ofono is not ready | 16:48 |
| T42 | <k1gen> let me reflash the kernel and see if spinning circle disappears | 16:48 |
| T42 | <k1gen> nope, stuck 10-15 frames past than last time | 16:50 |
| T42 | <k1gen> mal: when I boot successfully, what should I send besides the journal? | 16:50 |
| T42 | <k1gen> ok, I booted into system | 16:51 |
| T42 | <k1gen> mal: journal of clean boot, no camera attempts, just one lockscreen unlock: http://ix.io/4sad | 16:52 |
| T42 | <k1gen> yeah, Failed to start Droid bootctl. | 16:53 |
| T42 | <k1gen> mal: any idea how to fix these reboot issues, or do you need more info/logs? | 17:01 |
| mal | that bootctl issue should be fixed by this https://github.com/mer-hybris/droid-hal-configs/pull/260 | 17:38 |
| T42 | <k1gen> mal: how do I build specific pull request with build_packages.sh? | 17:43 |
| T42 | <edp_17> With what command can I check/set the volume from command line? | 18:14 |
| mal | @k1gen you should not build that, it's too new, you can just copy the file from your droid-configs-device/sparse/ to sparse/ and edit it | 18:45 |
| mal | @edp_17 I would assume pactl | 18:46 |
| T42 | <edp_17> Thanks. I'll try to play with it. | 18:50 |
| mal | why do you want to do that? | 18:50 |
| T42 | <k1gen> mal: sorry it took so long, here's camera logcat: https://transfer.sh/Y1vyf3/logcat.txt. while logcat was launched I opened the camera app first time a minute after a reboot, was in photo mode for 5 seconds, switched to video mode, chose to close the app after sailfish dialogue apperared, then opened the app again and switched to photo mode | 18:58 |
| T42 | <edp_17> mal : I thought there was no sound because it was muted or something. So, wanted to test whether that's the case. | 19:08 |
| T42 | <k1gen> mal: with this https://github.com/mer-hybris/droid-hal-configs/blob/1fac5d12b0618b36dbe143f827516eca0b2ab20f/sparse/lib/udev/rules.d/998-droid-system.rules in my /lib/udev/rules.d bootctl still fails: http://ix.io/4sb0 | 19:29 |
| mal | @k1gen try to check the command here manually https://github.com/mer-hybris/droid-hal-configs/blob/master/sparse-11/usr/bin/droid/droid-bootctl.sh | 19:37 |
| T42 | <k1gen> mal: Marking boot as successful | 19:40 |
| mal | what happens if you run it second time? | 19:41 |
| T42 | <k1gen> no output | 19:41 |
| mal | wondering if it really marked it successful | 19:41 |
| T42 | <k1gen> I also executed /usr/bin/droid/droid-bootctl.sh before last reboot, but rebooted straight to bootloader with "no valid slot to boot" | 19:42 |
| mal | looks like something odd happens with that bootctl | 19:46 |
| T42 | <k1gen> something odd happens a lot on my device :) | 19:47 |
| T42 | <k1gen> it'd be nice to fix this no valid slot to boot stuff before we go anywhere where I need to reboot a lot | 19:52 |
| mal | yeah, it needs some investigation | 21:24 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!