| T42_ | <AntonlX> How do I can deal with this error on hybris-18.1? | 07:20 |
|---|---|---|
| T42_ | <AntonlX> library "libandroidicu.so" needed or dlopened by "/usr/libexec/droid-hybris/system/lib64/libmedia.so" is not accessible for the namespace "(default)" | 07:20 |
| *** rainemak_ is now known as rainemak | 07:27 | |
| T42_ | <AntonlX> default is not registered | 10:44 |
| T42_ | <AntonlX> Has anyone encountered with | 10:47 |
| T42_ | <AntonlX> Since android.hardware.graphics.composer@2.1::IComposer/default is not registered, trying to start it as a lazy HAL. | 10:47 |
| T42_ | <AntonlX> Unable to set property "ctl.interface_start" to "android.hardware.graphics.composer@2.1::IComposer/default": error code: 0x20 | 10:47 |
| T42_ | <AntonlX> What does 0x20 means? | 10:47 |
| T42_ | <Mister_Magister> either space or 32 | 11:13 |
| T42_ | <TheVancedGamer> it means HAL isn't running | 11:13 |
| T42_ | <b100dian> May be broken pipe | 11:28 |
| T42_ | <AntonlX> What pipe?) (re @b100dian: May be broken pipe) | 11:29 |
| T42_ | <b100dian> I think it is a generic comnunication error | 11:30 |
| T42_ | <b100dian> But you have that service or another version of it? | 11:30 |
| T42_ | <AntonlX> Founds out that it is namespace restrictions. https://paste.opensuse.org/pastes/398c8217bdd8 (re @b100dian: May be broken pipe) | 11:46 |
| T42_ | <AntonlX> Mounting is a solution, but I could not mount apex files before droid-hal-init starts | 11:47 |
| T42_ | <AntonlX> Mounting is the solution, but I could not mount apex files before droid-hal-init starts (edited) | 11:47 |
| T42_ | <AntonlX> Is there another way to deal with such errors on hybris-18? | 11:48 |
| T42_ | <elros34> the easiest is /opt symlink otherwise you would have to fight with /linkerconfig | 12:14 |
| T42_ | <AntonlX> I am actually fighting with linkerconfig) (re @elros34: the easiest is /opt ...) | 12:26 |
| T42_ | <AntonlX> Could not load /vendor/lib64/hw/hwcomposer.atoll.so from sphal namespace: dlopen failed | 12:26 |
| T42_ | <AntonlX> write(2, "library \"libbinder.so\" (\"/system/apex/com.android.vndk.current/lib64/libbinder.so\") needed or dlopened by \"/vendor/lib64/hw/hwcomposer.atoll.so\" is not accessible for the namespace: [name=\"sphal\", ld_library_paths=\"\", default_library_paths=\"/odm/lib64:/ven"..., 367library "libbinder.so" ("/system/apex/com.android.vndk.current/lib64/libbinder.so") needed or dlopened by "/vendor/lib64/hw/hwcomposer.atoll.so" is not accessi | 12:27 |
| T42_ | <AntonlX> But libbinder.so is already symlinked to /odm/lib64 | 12:27 |
| T42_ | <elros34> I don't remember exactly so what is mounted in /linkerconfig in your case? findmnt -R /linkerconfig | 12:29 |
| T42_ | <elros34> so strace shows that it found libbinder symlink? | 12:30 |
| T42_ | <AntonlX> Yes, it's test_hwcomposer strace (re @elros34: so strace shows that...) | 12:31 |
| T42_ | <AntonlX> https://paste.opensuse.org/pastes/b1be4173f3de (re @elros34: I don't remember exa...) | 12:31 |
| T42_ | <elros34> can you show full strace? | 12:35 |
| T42_ | <adampigg> ofono 2.5 is somewhat working now :) | 12:36 |
| T42_ | <AntonlX> https://paste.opensuse.org/pastes/b41c8cc40602 (re @elros34: can you show full st...) | 12:38 |
| T42_ | <AntonlX> Seems like I found problem /data/.stowaways/sailfishos/linkerconfig/bootstrap/ is empty | 12:49 |
| T42_ | <elros34> bootstrap is mounted during droid-hal-init (the patch you have added) then /linkerconfig/default -> /linkerconfig is used (or should) | 12:51 |
| T42_ | <elros34> this is tricky problem, you can't use symlink because it's resolved | 12:53 |
| T42_ | <elros34> according to strace maybe you could use /system/etc/ld.config.arm64.txt but this might break some services using bootstrap linker config during early boot | 12:57 |
| T42_ | <AntonlX> I have problems partition mounting, may be this is source. There is many such errors in journalctl: | 13:03 |
| T42_ | <AntonlX> Process '/bin/ln -s /dev/block/sda1 /dev/block/platform/soc/1d84000.ufshc/by-num/p1' failed with exit code 1. | 13:03 |
| T42_ | <AntonlX> And while he said: Mounted Droid mount for /vendor/bt_firmware. | 13:03 |
| T42_ | <AntonlX> /vendor/bt_firmware is empty | 13:03 |
| T42_ | <AntonlX> /vendor/bt_firmware is empty, the same for other mountpoints (edited) | 13:03 |
| T42_ | <elros34> do you use dynamic partition? | 13:04 |
| T42_ | <AntonlX> Yes, but super partitions does not mounting. They are packed in droid-system (re @elros34: do you use dynamic p...) | 13:05 |
| T42_ | <AntonlX> I don't have /dev/block/platform/soc/1d84000.ufshc/by-num | 13:05 |
| T42_ | <elros34> I wouldn't worry much about symlinks | 13:05 |
| T42_ | <elros34> so do you have dynamic vendor partition? Do you also package it? | 13:08 |
| T42_ | <AntonlX> Yes all files from super is here, but seems like script deletes mountpoints as empty folders (re @elros34: so do you have dynam...) | 13:09 |
| T42_ | <AntonlX> Yes all files from super.img is here, but seems like script deletes mountpoints as empty folders (edited) (re @elros34: so do you have dynam...) | 13:09 |
| T42_ | <AntonlX> Yes all files from super.img is here, but seems like script deletes mountpoints as empty folders(so I created them manually) (edited) (re @elros34: so do you have dynam...) | 13:10 |
| T42_ | <elros34> so I do not understand the problem, if 'findmnt /vendor/bt_firmware' partition is empty then you can't do much otherwise I do not get ho it could be | 13:12 |
| T42_ | <elros34> about initial issue: did you check whether copy of this libbinder.so in /odm even works correctly? | 13:13 |
| T42_ | <elros34> I wonder why do you even have this issue. Do you have some incorrect /odm/lib64 > /usr/libexec/droid-hybris/system/lib64? | 13:27 |
| T42_ | <AntonlX> /vendor mounting "overlayed" /vendor/dsp of my wrong script. Deal with mounting (re @elros34: so I do not understa...) | 13:35 |
| T42_ | <AntonlX> This is fresh link and its should work (re @elros34: about initial issue:...) | 13:36 |
| T42_ | <AntonlX> /odm/lib64/libbinder.so -> /system/apex/com.android.vndk.current/lib64/libbinder.so | 13:36 |
| T42_ | <AntonlX> /system/apex/com.android.vndk.current/lib64/libbinder.so | 13:36 |
| T42_ | <AntonlX> /system/apex/com.android.vndk.current/lib64/libbinder.so found by ls (edited) | 13:36 |
| T42_ | <AntonlX> https://paste.opensuse.org/pastes/aa47546abc0d Only reqired links (re @elros34: I wonder why do you ...) | 13:37 |
| T42_ | <elros34> so have you tried removing libbinder symlink and makeing modified ldconfig file in /system? | 13:39 |
| T42_ | <AntonlX> Could I modify it? My port is hybris-18.1. AOSP docs says it's runtime generated. (re @elros34: so have you tried re...) | 13:41 |
| T42_ | <elros34> /linkerconfig is runtime generated but I am talkig about different file | 13:42 |
| T42_ | <elros34> btw seems I mixup something about /linkerconfig/{bootstrap,default} mounting order so ignore my previous bollocks about it | 13:43 |
| T42_ | <AntonlX> The only ld.config.txt files I have located in apex partition. If I add ld.config.txt to /usr/libexec/droid-hybris/system/etc/init/ will it work? (re @elros34: /linkerconfig is run...) | 13:59 |
| T42_ | <elros34> no the path I gave previously was /system/etc/ld.config.arm64.txt, you have packaged /systemd so my idea is to copy and modify /linkerconfig/ld.config.txt to that /system path | 14:01 |
| T42_ | <AntonlX> Your solution is working! Now test_hwcomposer and ever minimer is works, but still no UI and I can't get why. I am stuck again. (re @elros34: no the path I gave p...) | 19:12 |
| T42_ | <AntonlX> Logcat: https://paste.opensuse.org/pastes/6eec49d935a0 | 19:12 |
| T42_ | <AntonlX> journalctl: https://paste.opensuse.org/pastes/c6a47ef2a812 | 19:12 |
| T42_ | <AntonlX> I am intentionally moved sscrpcd because it spammed a lot into logcat | 19:15 |
| mal | no sign of lipstick even trying to start? are you sure user session is starting? | 19:19 |
| T42_ | <AntonlX> How to check that? (re @SailfishFreenodeIRCBridgeBot: <mal>no sign of lips...) | 19:22 |
| mal | have you masked it? what does output of "systemctl" tell | 19:26 |
| mal | I wonder if you have the usual issue that oneshot-root is stuck | 19:30 |
| T42_ | <AntonlX> Yes, but it unmasked already (re @SailfishFreenodeIRCBridgeBot: <mal>have you masked...) | 19:30 |
| T42_ | <AntonlX> https://paste.opensuse.org/pastes/bea697907692 (re @SailfishFreenodeIRCBridgeBot: <mal>have you masked...) | 19:30 |
| mal | it happens if minimedia is not working | 19:30 |
| T42_ | <AntonlX> Yes, no minimedia process (re @SailfishFreenodeIRCBridgeBot: <mal>it happens if m...) | 19:31 |
| mal | either fix minimedia or uninstall gmp-droid package | 19:31 |
| T42_ | <AntonlX> I forgot to mount 32bit libandroidicu.so. libandroidicu.so - is library of pain) (re @SailfishFreenodeIRCBridgeBot: <mal>either fix mini...) | 19:33 |
| T42_ | <AntonlX> Seems like I fixed minimedia starting, but they do not starting automatically (re @SailfishFreenodeIRCBridgeBot: <mal>either fix mini...) | 19:41 |
| T42_ | <AntonlX> I also have to start hwcomposer manually to get minimer and test_hwcomposer working | 19:42 |
| T42_ | <AntonlX> Removing gmp-droid helps (re @SailfishFreenodeIRCBridgeBot: <mal>either fix mini...) | 19:48 |
| T42_ | <pachof> @b100dian help, my kernel no boots | 21:01 |
| T42_ | <b100dian> Did you install modules too? | 21:27 |
| T42_ | <pachof> i dont know how install | 21:29 |
| T42_ | <b100dian> you can use the working kernel or adb shell to just copy them over in the correct subfolder | 21:29 |
| T42_ | <b100dian> without affecting the existing modules, because presumably this other kernel has another version | 21:30 |
| T42_ | <b100dian> so you just booted the hybris-boot.img I guess? | 21:30 |
| T42_ | <pachof> yes | 21:30 |
| T42_ | <b100dian> Ok, did you ever fixed telnet access? Might need it now or just install the older hybris-boot that works for this transfer | 21:31 |
| T42_ | <pachof> how fix the telnet access | 21:31 |
| T42_ | <b100dian> assuming it is not disabled it should have at least worked in the previous kernel. You just need to configure the host (PC) interface if it is not autoconfigured to something like 192.168.2.14/24 (notice 14, not 15 which is you phone) | 21:34 |
| T42_ | <b100dian> but better try that with the previous kernel | 21:34 |
| T42_ | <pachof> but my IP ending is 24, it should work, and the birdzhang kernel cannot be patched for the camera | 21:39 |
| T42_ | <b100dian> I mean the usb interface IP. Does ifconfig show a such interface, for starters? | 21:47 |
| T42_ | <b100dian> I know your camera issues, we're talking about telnet here. Get telnet to work at least with the working kernel | 21:48 |
| T42_ | <pachof> usb interface ip is eno1 or lo? | 21:50 |
| T42_ | <b100dian> eno is probably ethernet, lo is local, so neither | 21:51 |
| T42_ | <b100dian> again, when you connect the phone it should pop up messages in your PC dmesg, do you see any | 21:51 |
| T42_ | <pachof> "[ 2355.682201] usb 1-1: new high-speed USB device number 3 using xhci_hcd | 21:52 |
| T42_ | <pachof> [ 2355.829501] usb 1-1: New USB device found, idVendor=18d1, idProduct=d00d, bcdDevice= 1.00 | 21:52 |
| T42_ | <pachof> [ 2355.829529] usb 1-1: New USB device strings: Mfr=1, Product=2, SerialNumber=3 | 21:53 |
| T42_ | <pachof> [ 2355.829539] usb 1-1: Product: Android | 21:53 |
| T42_ | <pachof> [ 2355.829547] usb 1-1: Manufacturer: Google | 21:53 |
| T42_ | <pachof> [ 2355.829555] usb 1-1: SerialNumber: 1db7c300 | 21:53 |
| T42_ | <pachof> [ 2514.398357] perf: interrupt took too long (3150 > 3137), lowering kernel.perf_event_max_sample_rate to 63300 | 21:53 |
| T42_ | <pachof> [ 2687.328158] usb 1-1: USB disconnect, device number 3 | 21:53 |
| T42_ | <pachof> [ 2700.522010] usb 1-1: new high-speed USB device number 4 using xhci_hcd | 21:53 |
| T42_ | <pachof> [ 2700.666578] usb 1-1: New USB device found, idVendor=05c6, idProduct=9039, bcdDevice= 4.14 | 21:53 |
| T42_ | <pachof> [ 2700.666605] usb 1-1: New USB device strings: Mfr=1, Product=2, SerialNumber=3 | 21:53 |
| T42_ | <pachof> [ 2700.666616] usb 1-1: Product: Xiaomi Redmi Note 8 | 21:53 |
| T42_ | <pachof> [ 2700.666624] usb 1-1: Manufacturer: Xiaomi | 21:53 |
| T42_ | <pachof> [ 2700.666632] usb 1-1: SerialNumber: 1db7c300 | 21:53 |
| T42_ | <pachof> [ 2893.732074] usb 1-1: USB disconnect, device number 4 | 21:53 |
| T42_ | <pachof> [ 3079.738605] usb 1-1: new high-speed USB device number 5 using xhci_hcd | 21:53 |
| T42_ | <pachof> [ 3079.886318] usb 1-1: New USB device found, idVendor=05c6, idProduct=9039, bcdDevice= 4.14 | 21:53 |
| T42_ | <pachof> [ 3079.886344] usb 1-1: New USB device strings: Mfr=1, Product=2, SerialNumber=3 | 21:53 |
| T42_ | <pachof> [ 3079.886353] usb 1-1: Product: Xiaomi Redmi Note 8 | 21:53 |
| T42_ | <pachof> [ 3079.886360] usb 1-1: Manufacturer: Xiaomi | 21:53 |
| T42_ | <pachof> [ 3079.886367] usb 1-1: SerialNumber: 1db7c300 | 21:53 |
| T42_ | <pachof> [ 3125.857715] usb 1-1: USB disconnect, device number 5 | 21:53 |
| T42_ | <pachof> [ 3130.718607] usb 1-1: new high-speed USB device number 6 using xhci_hcd | 21:53 |
| T42_ | <pachof> [ 3130.871974] usb 1-1: New USB device found, idVendor=18d1, idProduct=d00d, bcdDevice= 1.00 | 21:53 |
| T42_ | <pachof> [ 3130.871999] usb 1-1: New USB device strings: Mfr=1, Product=2, SerialNumber=3 | 21:53 |
| T42_ | <pachof> [ 3130.872007] usb 1-1: Product: Android | 21:53 |
| T42_ | <pachof> [ 3130.872014] usb 1-1: Manufacturer: Google | 21:53 |
| T42_ | <pachof> [ 3130.872020] usb 1-1: SerialNumber: 1db7c300 | 21:53 |
| T42_ | <pachof> [ 3828.727846] perf: interrupt took too long (3938 > 3937), lowering kernel.perf_event_max_sample_rate to 50700 | 21:53 |
| T42_ | <pachof> [ 4266.394895] usb 1-1: USB disconnect, device number 6 | 21:53 |
| T42_ | <pachof> [ 4266.821748] usb 1-1: new high-speed USB device number 7 using xhci_hcd | 21:53 |
| T42_ | <b100dian> you just wrote 50 lines of IRC messages :) | 21:55 |
| T42_ | <b100dian> It's in HADK 9.2 what you should expect | 21:55 |
| T42_ | <b100dian> Was the output from the working kernel? | 21:55 |
| T42_ | <pachof> how see the output? | 21:56 |
| T42_ | <b100dian> I mean, if you boot the kernel from the original porter, what do you see in dmesg on PC (without pasting here!)? | 21:57 |
| T42_ | <pachof> from the original kernel? | 22:00 |
| T42_ | <pachof> is same https://paste.opensuse.org/pastes/570c10e87e52 | 22:02 |
| T42_ | <b100dian> I really don't know how this was developed without telnet access | 22:22 |
| T42_ | <b100dian> maybe usb-moded breaks it, try masking that service | 22:22 |
| T42_ | <b100dian> or maybe there is another thing blocking it the /init_disable_telnet file | 22:23 |
| T42_ | <b100dian> *than the | 22:23 |
| T42_ | <pachof> how mask the service | 22:23 |
| T42_ | <b100dian> you could blindly copy the built modules in /lib/module/$(newkernelversion) and try again maybe | 22:24 |
| T42_ | <b100dian> systemctl mask usb-moded | 22:24 |
| T42_ | <b100dian> as root | 22:24 |
| T42_ | <pachof> but where do I install the modules? | 22:25 |
| T42_ | <pachof> /init_disable_telnet | 22:30 |
| T42_ | <pachof> /init_disable_telnet | 22:32 |
| T42_ | <b100dian> not sure, I would search for ko files in $ANDROID_ROOT/out/ Their location should reveal a version string for the kernel | 22:37 |
| T42_ | <b100dian> And copy them under /lib/modules/$KERNELVERSION on the device | 22:37 |
| T42_ | <b100dian> respecting the existing structure of the existing kernel version there | 22:38 |
| T42_ | <b100dian> that is because you've not build a rpm with them, if I understand correctly, only the Android bits | 22:38 |
| T42_ | <b100dian> maybe someone knows of more clear instructions here, it is not something I do | 22:39 |
| T42_ | <b100dian> also: did you made at least the config changes the original kernel had? | 22:44 |
| T42_ | <pachof> yes | 22:45 |
| T42_ | <pachof> using the mer kernel check | 22:45 |
| T42_ | <b100dian> And more or less the same changes, even outside those recommended by mer-kernel-check but maybe added by the original porter for some reason? | 22:48 |
| T42_ | <b100dian> Here's how the rpm with kernel modules would find them: https://github.com/mer-hybris/droid-hal-device/blob/master/droid-hal-device.inc#L926 | 22:48 |
| T42_ | <pachof> the kernel module only exists in vendor flder | 22:54 |
| T42_ | <pachof> the kernel module only exists in vendor folder (edited) | 22:54 |
| T42_ | <b100dian> ok | 22:58 |
| T42_ | <pachof> its no working | 23:31 |
| T42_ | <b100dian> there's nothing you can debug without telnet though. Where did you copy the modules in the end? Do they have the same permission as the other modules? | 23:33 |
| T42_ | <pachof> kernel module folder name is 4.14.98-perf | 23:38 |
| T42_ | <pachof> ? | 23:38 |
| T42_ | <b100dian> Also show you kernel changes | 23:39 |
| T42_ | <pachof> yes, this kernel version is 4.14.98, so the name is 4.14.98-perf | 23:40 |
| T42_ | <pachof> but it no works | 23:40 |
| T42_ | <b100dian> it doesnt boot to gui does it reboot? | 23:42 |
| T42_ | <pachof> It does not restart, it is static | 23:42 |
| T42_ | <elros34> first think to test is existance of /data/.stowaways/sailfishos/init.log. That way you will know whether you are even booting sailfis, which Idoubt | 23:44 |
| T42_ | <pachof> possible patch https://android.googlesource.com/device/lge/bullhead-kernel/+/b6708ca998288cb9bf46a77cff243c569cdbe37d or this https://android.googlesource.com/kernel/msm/+/c01894df6ba74d8a71cfb3bfad4036da5f05bb4a%5E%21/ | 23:47 |
| T42_ | <pachof> @elros34 the file init.log dont exist | 23:48 |
| T42_ | <elros34> then you don't boot sailfish. Show your kernel changes in github | 23:49 |
| T42_ | <pachof> this is a only change : https://irc.thaodan.de/.imgstore/2c29324e/ginkgo_perf_defconfig | 23:57 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!