| T42 | <HengYeDev> @elros34 Boot header version is same, but lineage has BOARD_DTB_SIZE and BOARD_DTB_OFFSET that hybris-boot doesn't | 01:02 |
|---|---|---|
| T42 | <HengYeDev> also, lineage flashes a dtbo, does that change | 01:04 |
| T42 | <elros34> what if you will recreate boot.img, will it boot? If yes then replace only initramfs and see if that fails | 01:12 |
| T42 | <HengYeDev> it works if I reflash boot.img, hopefully thats what you mean by "recreate" (re @elros34: what if you will rec...) | 01:17 |
| T42 | <elros34> I just wondering whether there is something wrong in abootimg because of: "dt table header check fail". If that would be the case then lineage's boot.img with hybris initramfs should work. By recreate I mean create new boot.img with mkbotimg | 01:20 |
| T42 | <dcmmc> from dmesg of host, i got usb 3-3: Manufacturer: Oneplus and SerialNumber: af315f78 | 05:19 |
| T42 | <dcmmc> any tips for solving this? plz | 05:19 |
| T42 | <dcmmc> usb_moded[1822]: /config/usb_gadget/g1/functions/rndis.usb0/ethaddr: can't open for writing: No such file or directory | 05:24 |
| T42 | <dcmmc> I manually linked /usr/lib/systemd/system/usb-moded.service to /dev/null. | 05:40 |
| T42 | <dcmmc> But telnet still not work 🙁 | 05:40 |
| deathmist | @dcmmc set CONFIG_USB_CONFIGFS_RNDIS=y and CONFIG_USB_CONFIGFS_F_QCGSI=n iirc in kernel defconfig | 05:51 |
| deathmist | assuming if | 05:51 |
| deathmist | it's not an ancient device* | 05:51 |
| T42 | <dcmmc> it’s oneplus 6 with sdm845 | 05:52 |
| T42 | <dcmmc> i have CONFIG_USB_CONFIGFS_RNDIS=y in defconfig | 05:53 |
| T42 | <dcmmc> but missing CONFIG_USB_CONFIGFS_F_QCGSI=n | 05:53 |
| T42 | <dcmmc> i will try to add this one | 05:53 |
| T42 | <dcmmc> btw, what’s the mean of iirc? | 05:53 |
| Thaodan | @dcmmc: https://www.urbandictionary.com/define.php?term=iirc | 06:33 |
| T42 | <dcmmc> got it now | 06:37 |
| T42 | <elros34> @dcmmc you got telnet? | 10:49 |
| T42 | <elros34> @dcmmc your usb connection is working otherwise you wouldnt boot to sailfish so defconfig changes are not needed. Try "touch /data/.stowaways/sailfishos/init_enter_debug2" if that will help then probably some /init*rc script mess up with usb otherwise issue is with your pc/usb | 10:54 |
| T42 | <dcmmc> Elros: thank you for this tip | 10:55 |
| T42 | <dcmmc> will try it now | 10:55 |
| T42 | <dcmmc> Elros: u r a genius! | 11:29 |
| T42 | <dcmmc> after touch /data/.stowaways/sailfishos/init_enter_debug2, I finally got ``Serial Number: Mer Debug telnet on port 2323 on rndis0 192.168.2.15 - also running udhcpd” | 11:30 |
| T42 | <dcmmc> but i still cannot connect it, maybe issue with my pc | 11:39 |
| T42 | <dcmmc> cannot found 192.168.2.15 on host | 11:40 |
| T42 | <elros34> firewall, different usb port? | 12:01 |
| T42 | <dcmmc> ifconfig shows the usb network is on enp0s4u3 | 12:07 |
| T42 | <dcmmc> but not assign ipv4 address | 12:08 |
| T42 | <dcmmc> I manually assign an ip via ifconfig enp0s4u3 192.168.2.15 up | 12:09 |
| T42 | <dcmmc> ping 192.168.2.15 works but telnet 192.168.2.15 2323 not work | 12:09 |
| T42 | <elros34> oh so mac have all these tools, how about using instruction from halium-boot: ip address/route add | 12:11 |
| T42 | <dcmmc> will try it. ps, i’m now using debian via qemu | 12:13 |
| T42 | <elros34> nice instruction: https://docs.halium.org/en/latest/porting/debug-build/early-init.html | 12:15 |
| T42 | <dcmmc> it’s a very helpful instruction! | 12:21 |
| T42 | <dcmmc> now i can finally telnet to sfos | 12:21 |
| T42 | <dcmmc> Elros: export EGL_PLATFORM=hwcomposer && stace test_hwcomposer | 12:27 |
| T42 | <dcmmc> hwcomposer.default.so no such file | 12:27 |
| T42 | <elros34> just get logcat from normal boot when lipstick fails | 12:28 |
| T42 | <dcmmc> Elros: before normal boot should I change some /init*rc script to fix the usb telnet issue? | 12:29 |
| T42 | <elros34> if you are sure that caused issues because I am not | 12:30 |
| T42 | <elros34> just continue booting now according to instruction you see in terminal and you will see | 12:30 |
| T42 | <dcmmc> okay | 12:31 |
| T42 | <dcmmc> when normal boot, i cannot see the network device from `ip a` | 12:34 |
| T42 | <dcmmc> and in init_enter_debug2 mode, grep -r ‘import’ /*.rc shows that no import usb stuff | 12:36 |
| T42 | <elros34> so you lost connection after running "echo continue > /init-ctl/stdin"? | 12:36 |
| T42 | <dcmmc> oops, i misunderstood your advice (re @elros34: so you lost connecti...) | 12:37 |
| T42 | <dcmmc> i just removed /data/.stowaways/sailfishos/init_enter_debug2 and reboot | 12:38 |
| T42 | <dcmmc> Elros: you are right. after running "echo continue > /init-ctl/stdin" i lost the connection | 12:42 |
| T42 | <dcmmc> and the device in `ip a` disappeared | 12:43 |
| T42 | <elros34> so if you msaked usb-moded then find anything usb releated in all init*rc/sh files: / and in /system or /vendor | 12:43 |
| T42 | <dcmmc> Elros: can you tell me how to mask usb-moded | 12:47 |
| T42 | <dcmmc> echo ‘systemctl mask use-moded’ > /init-ctl/sbin | 12:47 |
| T42 | <dcmmc> ? | 12:47 |
| T42 | <elros34> just "systemctl mask usb-moded" | 12:49 |
| T42 | <dcmmc> okay | 12:49 |
| T42 | <dcmmc> i have init.usb.configfs.rc in / | 12:50 |
| T42 | <dcmmc> /system and /vendor are empty | 12:51 |
| T42 | <elros34> they are empty because you stop booting at /init-debug script, you must mount them manually | 12:52 |
| T42 | <dcmmc> okay | 12:52 |
| T42 | <dcmmc> after mount them, what should I found? import usb sutff? | 12:55 |
| T42 | <dcmmc> in /system/init.rc, i found `import /init.usb.rc` and `import /init.usb.configfs.rc` | 12:56 |
| T42 | <dcmmc> shoud i comment then? | 12:57 |
| T42 | <dcmmc> *them | 12:57 |
| T42 | <elros34> I am not sure /system/init.rc is used, don't you have /init.rc. Import is most common way but anything can call function even in .sh script | 12:58 |
| T42 | <dcmmc> yes i have /init.rc, and i found that import /init.usb.rc is already commented | 12:59 |
| T42 | <dcmmc> oh i found /init.rc import /vendor/etc/init/hw/init.qcom.rc and init.qcom.rc import /vendor/etc/init/hw/init.*.usb.rc | 13:02 |
| T42 | <dcmmc> does this lead problem? | 13:03 |
| T42 | <elros34> depends on what it contains but it might | 13:03 |
| T42 | <dcmmc> shoud i comment the usb import things in /vendor/etc/init/hw/init.qcom.rc? | 13:04 |
| T42 | <dcmmc> i think these will conflict with sf’s usb config | 13:04 |
| T42 | <dcmmc> after comment these, the telnet connection now is not lost after continue | 13:09 |
| T42 | <elros34> so you need to create bind mount unit in systemd to overwrite that *.rc file | 13:12 |
| T42 | <dcmmc> okay | 13:17 |
| T42 | <dcmmc> logcat: https://pastebin.com/UwUtiB0A | 13:27 |
| T42 | <dcmmc> Elros: could you plz help me | 13:27 |
| T42 | <dcmmc> test_hwcomposer running without segfault but the screen is black | 13:30 |
| T42 | <elros34> so which line in this enormous log is when lipstick starts? | 13:35 |
| T42 | <dcmmc> according to the journalctl, it started at 22:48:32 | 13:37 |
| T42 | <elros34> btw did you disable selinux? | 13:40 |
| T42 | <elros34> touch /mnt/vendor/mnt/vendor/persist/sensors/registry/temp.json to get rid of spam | 13:41 |
| T42 | <dcmmc> i set selinux=0 in cmdline | 13:41 |
| T42 | <elros34> but is it disabled according to early dmesg? | 13:41 |
| T42 | <elros34> same for "/data/.layout_version" | 13:42 |
| T42 | <dcmmc> from dmesg, i can found `selinux: loaded file_contexts` | 13:43 |
| T42 | <dcmmc> this is the only selinux related stuff in dmesg | 13:43 |
| T42 | <elros34> then its either too late or you did not disable it correctly | 13:44 |
| T42 | <dcmmc> so what’s the right way to disable selinux? | 13:45 |
| T42 | <elros34> its explained in collabedit.com/kr9xx you got: disabling/enabling selinux and checking whether is disabled | 14:00 |
| T42 | <dcmmc> I can canfirm that selinux is disable | 14:06 |
| T42 | <dcmmc> dmesg -w I got SELinux: disabled at boot | 14:06 |
| T42 | <dcmmc> my base is lineage-16 | 14:07 |
| T42 | <dcmmc> Elros: could you plz provide some tips for debuging lipstick crash? | 14:55 |
| T42 | <dcmmc> strace lipstick I saw some no such file of /data/vendor/gpu/esx_config_lipstick.txt | 14:56 |
| T42 | <dcmmc> is this important? | 14:56 |
| T42 | <elros34> do not strace lipstick, it's too noisy and requires more steps to do this correctly. If you want to strace then do it with minimer but I do not see anything hwcomposer related in logcat and that is important. Did you get rid of these spam messages I was talking about? Try to get all logs: dmesg, journalctl again and mark in logcat where lipstick is started/crashed | 15:06 |
| T42 | <dcmmc> okay, i will try | 15:08 |
| T42 | <dcmmc> dmesg: https://pastebin.com/YhqpYysG journalctl: https://pastebin.com/Ux176ex7 logcat: https://pastebin.com/YcF7XgBt | 15:33 |
| deathmist | mal: have you tried running build_packages.sh in a while? it appears pulseaudio-modules-droid-hidl builds fails 100% of the time the first time around due to seemingly not being able to find audiosystem-passthrough package which is built before it | 16:46 |
| deathmist | running build_packages.sh again lets it get past pulseaudio-modules-droid-hidl though :/ | 16:47 |
| mal | deathmist: it's been a while since I used it last time, need to test | 16:50 |
| T42 | <elros34> same strange dependencies issues started to happen for me together with snapshots introduction then I disable snapshots in scripts and didn't notice it anymore but could be just an accident. Could be related: https://github.com/mer-hybris/droid-hal-device/pull/320 | 16:53 |
| deathmist | mal: fwiw apparently just nuking droid-local-repo from my repo wasn't enough to trigger it again though | 17:02 |
| T42 | <elros34> @dcmmc journal log is partial only but anyway at least now you can see hwcomposer logs in logcat. There is nothing obvious. Do you see animation when you run /system/bin/surfaceflinger? If yes then stop it and try "systemctl restart user@100000" | 18:18 |
| T42 | <HengYeDev> @elros34 : When I replace the initramfs, it boots, shows a dark screen for a second or so, then enters the strange fastboot mode present on samsung devices | 22:49 |
| T42 | <HengYeDev> nothing in dmesg about hybris initramfs booting | 22:49 |
| T42 | <HengYeDev> i think this suggests that the dtb is whats wrong with the hybris-boot.img, are params like https://github.com/LineageOS/android_device_samsung_gta4xl-common/blob/7256dfa8a7acfdfd7d906e3c978cceecd735023d/BoardConfigPlatform.mk#L81 not applied in hybris-boot? | 22:59 |
| T42 | <elros34> sounds good at least you do not have that vbmeta errors. You can recreate boot.img now but with sailfish kernel and you will see if that is the case | 23:00 |
| T42 | <elros34> I think you should see full command used to create hybris-boot.img in make-hybris-boot output | 23:01 |
| T42 | <elros34> hmm interesting commit, did you see it?https://github.com/LineageOS/android_device_samsung_gta4xl-common/commit/aa86dbf022596d141df0c0059cc5a28f16c1e2f7 | 23:03 |
| T42 | <HengYeDev> yeah, that flashes the vbmeta which allows lineage to boot, lineage wont boot without that | 23:03 |
| T42 | <elros34> btw did you apply also skip_initramfs required changes to kernel? | 23:07 |
| T42 | <HengYeDev> no, where is that (re @elros34: btw did you apply al...) | 23:08 |
| T42 | <elros34> grep for skip_initramfs here https://collabedit.com/kr9xx | 23:08 |
| T42 | <HengYeDev> i only have Image in out, no Image.gz.dtb for aik (re @elros34: sounds good at least...) | 23:14 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!