**** BEGIN LOGGING AT Sat Nov 03 02:59:58 2018 Nov 03 05:43:23 @VeryOriginalUsername this? http://docs.halium.org/en/latest/porting/debug-build/dmesg.html?highlight=last_kmsg#linux-kernel-3-4 Nov 03 05:43:34 CONFIG_ANDROID_RAM_CONSOLE_ENABLE_VERBOSE=y Nov 03 05:56:07 p4p4r4zi was added by: p4p4r4zi Nov 03 06:09:03 @dohniks, What does this do? Nov 03 06:10:11 It mentioned in the docs (link above) that you need this to get last_kmsg (along with the other config) Nov 03 06:10:26 @VeryOriginalUsername only has one of the two configs Nov 03 08:36:35 Gonna try that after I get back home, thanks Nov 03 14:03:13 skye_fentras was added by: skye_fentras Nov 03 14:33:46 okay, nothing in /proc/last_kmsg Nov 03 14:33:54 perhaps the kernel either didn't boot at all or didn't crash Nov 03 14:34:35 you should have logs even if it doesn't crash Nov 03 14:34:49 I'll do make clean and rebuild hybris-boot again Nov 03 14:34:56 maybe it didn't quite rebuild Nov 03 14:43:13 rebuilt it, still nothing Nov 03 14:44:15 how exactly to you test? `fastboot boot hybris-boot.img` Nov 03 14:44:28 give it a few seconds Nov 03 14:44:32 then boot into recovery Nov 03 14:44:41 and check proc Nov 03 14:45:03 oh, I flashed hybris-boot and then flashed recovery Nov 03 14:45:06 maybe that's the issue Nov 03 14:45:38 flashing hybris-boot should be fine, but why do you need to flash recovery? Nov 03 14:45:55 to check /proc/last_kmesg Nov 03 14:46:07 ok, that doesn't work Nov 03 14:46:10 hybris-boot doesn't make device show up at all Nov 03 14:46:19 no lsusb no dmesg no nothing Nov 03 14:46:34 ah Nov 03 14:46:36 i'm pretty sure it doesn't survive a flash process Nov 03 14:47:04 do you need to re-flash recovery? Nov 03 14:47:13 wait Nov 03 14:47:17 I flashed recovery Nov 03 14:47:22 and fastboot boot hybris-recovery Nov 03 14:47:26 and something showed up in dmesg Nov 03 14:47:27 yikes Nov 03 14:47:57 "usb mass storage detected" Nov 03 14:49:44 I wonder if my defconfig isn't missing some usb ethernet options Nov 03 14:50:05 which dmesg are you looking at now? Nov 03 14:50:09 host Nov 03 14:50:30 ah Nov 03 14:51:33 try to get a kernel log from your self built device kernel Nov 03 14:51:47 so 1 make sure you have a recovery that you can adb into Nov 03 14:51:55 2 boot your kernel Nov 03 14:51:56 @dohniks, done Nov 03 14:52:06 3 reboot into your recovery without turning device off Nov 03 14:52:06 @dohniks, only recovery seems to have any effect Nov 03 14:52:34 4 adb in and check the last_kmesg Nov 03 14:53:04 lets judge whether it "has any effect" when we see the log Nov 03 14:54:55 for now I can get it to "boot" with hyb-recovery image Nov 03 14:55:16 let me upload my dmesg Nov 03 14:56:14 https://pastebin.com/F4CtMELh Nov 03 14:57:27 yeah I see. well if you can telnet into hybris-recovery then you could also get a kernel log Nov 03 14:57:55 and yeah the fact it talks about storage mode might be an indication about some config missing ... dunno Nov 03 14:58:14 nothing new in `ip a` Nov 03 14:58:20 let me change some configs Nov 03 14:58:30 why not try 1,2,3,4? Nov 03 14:58:30 postmarketos wiki mentioned some Nov 03 14:59:11 ```CONFIG_USB_G_ANDROID … CONFIG_USB_ETH … CONFIG_USB_ETH_RNDIS``` could these help Nov 03 14:59:18 @dohniks, what do you mean by 1234? Nov 03 14:59:54 that was 1 Nov 03 14:59:54 Fwd from dohniks: so 1 make sure you have a recovery that you can adb into Nov 03 14:59:59 ah Nov 03 15:00:02 it doesn't boot Nov 03 15:00:11 how can you tell? Nov 03 15:00:11 or I didn't understand you Nov 03 15:00:26 it doesn't boot to the state where I can connect to it at least Nov 03 15:00:31 it might very well boot, but you can't see any output from it Nov 03 15:00:55 1,2,3,4 doesnt require you to connect to your booted kernel Nov 03 15:01:06 but how does one reboot without powering offf Nov 03 15:01:07 [Edit] but how does one reboot without powering off Nov 03 15:01:41 right. THAT might be the problem, I think thats different for different devices Nov 03 15:01:47 I just hold some buttons Nov 03 15:01:58 power and vol down or something Nov 03 15:01:58 for now I can only store recovery on /boot Nov 03 15:02:05 cause it's an old implementation of sony bootloader Nov 03 15:02:39 on custom android kernels it's stored on /fota-kernel but that requires an actual android kernel to work Nov 03 15:02:59 mhm, yeah sonys have some funky recoveries not too familiar with this Nov 03 15:03:07 but you have a recovery that you can adb into? Nov 03 15:03:12 twrp or smt Nov 03 15:03:13 fortunately they fixed it in 2014 I think Nov 03 15:03:16 @dohniks, yes Nov 03 15:03:20 it's flashed as /boot Nov 03 15:03:30 can you `fastboot flash boot twrp.img` Nov 03 15:03:41 it is done like that xdf Nov 03 15:03:46 [Edit] it is done like that xd Nov 03 15:03:48 then `fastboot boot hybris-boot.img` Nov 03 15:03:57 then reboot with some button combo into recovery? Nov 03 15:04:02 yeah and that doesn't output anything to twrp Nov 03 15:04:06 I'll try that in a moment tho Nov 03 15:04:29 [Edit] yeah and that doesn't output anything to dmesg (duh_ Nov 03 15:04:32 [Edit] yeah and that doesn't output anything to dmesg ( Nov 03 15:04:39 [Edit] yeah and that doesn't output anything to host dmesg Nov 03 15:08:15 hey @dohniks https://github.com/Halium/docs/pull/116 can we merge it ? Nov 03 15:09:01 @VeryOriginalUsername, nope Nov 03 15:09:01 nothing Nov 03 15:09:07 no combo works Nov 03 15:09:29 I even pulled out battery for split second and instantly pressed power button but it didn't work Nov 03 15:10:06 well, @abhishek_0 I'm not convinced. right now the docs promote two scripts such that people can use a or b whatever works best, your change basically deletes a. I don't see the big gain Nov 03 15:10:33 @VeryOriginalUsername does the device rebooting with any button combo at all? Nov 03 15:10:41 I think battery out is too much Nov 03 15:10:47 no reboots Nov 03 15:10:58 sony only implemented button combo reboots on devices with non-removable battery Nov 03 15:11:15 I have a sony sitting on my desk as well and mocking me :) so I'm curious how to get past this Nov 03 15:11:29 here comes a far fetched idea Nov 03 15:11:47 maybe you can deliberately crash your kernel or force it to reboot somehow Nov 03 15:12:03 I thought about it too Nov 03 15:12:24 if you flash recovery and then just `fastboot boot` your kernel it should magically fall back to the flashed one Nov 03 15:12:39 @dohniks, both a and b solutions were supposed to be in the docs until Jbb's script could get stable, … well it is stable now, and i dont see any point in bloating the docs. Nov 03 15:13:43 rebuilding hyb-b and hyb-r with forementioned (is that a word?) usb_eth options Nov 03 15:13:46 maybe that will help Nov 03 15:13:52 @VeryOriginalUsername I cobbled this together once Nov 03 15:13:54 https://github.com/doniks/crashandburn Nov 03 15:14:02 also I saw android_usb_g_serial or something and enabled it Nov 03 15:14:41 the original idea (if I remember) was to help newer kernels crash because they don't ALWAYS write the kernel log to memory, but ONLY when they crash Nov 03 15:15:51 @VeryOriginalUsername, yeah lets see, if the last_kmesg wont work, I think you're pretty much down to guessing Nov 03 15:16:15 if my thing doesn't work I'll try and use your crasher Nov 03 15:16:40 you might only need a line out of it or so Nov 03 15:17:44 https://github.com/doniks/crashandburn/blob/master/crashandburn.sh#L37 Nov 03 15:17:56 those two lines I think Nov 03 15:18:30 another thing you might consider is the blinking I do in that same file Nov 03 15:19:01 It could provide you some VERY coarse grained debugging to tell if your kernel ever does anything Nov 03 15:19:16 maybe it doesnt drive the screen or the usb, but maybe you can make it blink :) Nov 03 15:19:31 will the recovery image boot to rootfs by itself tho? Nov 03 15:19:36 I think it forces telnet shell Nov 03 15:20:20 yeah I think hybris-recovery always provides telnet Nov 03 15:21:25 @abhishek_0 "a and b were supposed to be in the docs until b is stable" Nov 03 15:21:38 I'm not aware that there ever was any such conclusion Nov 03 15:22:00 not saying it couldn't be a good plan, but you make it sound like it's obvious and a done deal Nov 03 15:25:05 (Photo, 715x533) https://irc.ubports.com/8qY8HOL0.png Nov 03 15:25:06 :P Nov 03 15:25:30 but you also know that the conversation didnt end there Nov 03 15:26:04 happy to summarize my thoughts but telegram feels a bit awkward to do so. I'll comment on your pr Nov 03 15:27:47 @dohniks, umm , that is the last message there Nov 03 15:28:06 oh really, shoot Nov 03 15:28:08 lemme check Nov 03 15:32:31 well, ok, *I* see the fine print in the "After", but I get it how it's maybe not clear for third parties. as I said. I comment on your PR Nov 03 15:32:57 btw are you making that other one for the flex? Nov 03 15:33:19 as long as it's not copying binaries around I think that could be a simple quick to merge change Nov 03 15:33:49 @dohniks, Actually it is copying binaries around Nov 03 15:33:57 It's a hack only needed on arch Nov 03 15:34:10 I think I needed it as well on ubuntu Nov 03 15:34:16 well not the copying of binaries Nov 03 15:34:29 but just setting some lang var Nov 03 15:34:29 @dohniks, Maybe newer ones, idk Nov 03 15:34:34 Haven't used it in a while Nov 03 15:34:37 does the lang var help on arch? Nov 03 15:34:55 @dohniks, by copying the binary Nov 03 15:34:55 It's a different issue Nov 03 15:35:05 ah Nov 03 15:35:37 @dohniks, yes Nov 03 15:35:38 [Edit] @vanyasem ah Nov 03 15:36:57 `device descriptorr read/64, error -110` Nov 03 15:36:59 @VeryOriginalUsername how's it going over there with the sony? Nov 03 15:36:59 what did I f up xd Nov 03 15:37:06 [Edit] `device descriptor read/64, error -110` Nov 03 15:38:37 actually let me try another usb port Nov 03 15:42:24 okay got it to boot Nov 03 15:42:25 i mean Nov 03 15:42:29 network interface is up Nov 03 15:42:39 no freaking idea how, but it works Nov 03 15:42:52 telnet? ping? Nov 03 15:44:47 I tried assigning ssh address but not yet perhaps Nov 03 15:44:49 I'll try telnet Nov 03 15:51:21 okay it boots for sure since host dmesg mentions rndis internet gadget `Linux 3.4.0-CM+` Nov 03 15:51:32 and I can see it in `ip l` and assign ip to it Nov 03 15:51:34 but it won't go up Nov 03 15:51:46 even with `ip l set enp2s0f0u3 up` Nov 03 15:55:43 yeah no matter what I do it stays at DOWN state Nov 03 15:55:58 perhaps I should enable some more things in kernel? Nov 03 15:56:26 if you can think of any :) Nov 03 15:56:48 does it say anything else in your host dmsg? Nov 03 16:11:42 https://pastebin.com/VMxXbBan Nov 03 16:11:48 only this Nov 03 16:12:27 further up or down, no announcement of the ip and telnet? Nov 03 16:12:49 `enp2s0f0u4: mtu 1500 qdisc fq_codel state DOWN group default qlen 1000 … link/ether 0e:c1:6a:80:b9:fb brd ff:ff:ff:ff:ff:ff` Nov 03 16:12:54 [Edit] ```enp2s0f0u4: mtu 1500 qdisc fq_codel state DOWN group default qlen 1000 … link/ether 0e:c1:6a:80:b9:fb brd ff:ff:ff:ff:ff:ff``` Nov 03 16:12:58 output of ip a Nov 03 16:13:12 it won't go to UNKNOWN or UP no mater what I do Nov 03 16:13:35 it stays DOWN even after using dhclient/dhcpcd (I tried on xenial vm and arch host) Nov 03 16:14:50 @VeryOriginalUsername, Is that correct syntax? I'm not familiar with it Nov 03 16:14:56 l = link Nov 03 16:14:58 up was when we still used ifconfig Nov 03 16:15:19 http://docs.halium.org/en/latest/porting/debug-build/early-init.html Nov 03 16:15:20 no up here Nov 03 16:15:29 yes up in ssh part Nov 03 16:15:35 but it won't go up anyway Nov 03 16:15:59 @VeryOriginalUsername, what do you mean? Nov 03 16:16:05 you said no up i said yes up Nov 03 16:16:19 aka `ip l [..] up` is present in ssh chapter of docs Nov 03 16:19:33 I see. interesting. If the sections are different, I wonder whether one of them are wrong Nov 03 16:19:49 what exact error do you get if you just follow the telnet docs? Nov 03 16:19:59 no errors Nov 03 16:20:04 oh telnet part Nov 03 16:20:24 I mean the whole `ip blabla` stuff Nov 03 16:20:41 and let me rephrase Nov 03 16:20:47 what exact *output* do you get Nov 03 16:21:02 pinging: destination host unreachabler Nov 03 16:21:02 [Edit] pinging: destination host unreachable Nov 03 16:21:56 aaand the state of connection is down Nov 03 16:22:17 mhm. dunno. Trying harder to get a kernel log is the best I think think of Nov 03 16:22:52 oh, maybe just do the `ip` stuff multiple times Nov 03 16:23:01 hosts network manager might interfere Nov 03 16:26:51 your device is nicki? whats your manifest, maybe it helps to start for that xperia I have here Nov 03 16:27:10 -7.1 manifest is uploaded to halium github Nov 03 16:27:40 -5.1 is on my local drive but I need to write a guide on how to compile 5.1 for nicki since it needs few tricks Nov 03 16:27:44 what's your device btw? Nov 03 16:28:06 amami Nov 03 16:28:10 z1c? Nov 03 16:28:13 yes Nov 03 16:28:39 I think it will be quite easy since it shares the soc with hammerhead as far as I'm aware Nov 03 16:29:23 ah. but it has no official lineage, just some stuff on xda Nov 03 16:29:37 it has mainline but no lineage Nov 03 16:29:39 what an irony Nov 03 16:29:59 it was supported by sony open devices project Nov 03 16:30:02 you can use the aosp tree Nov 03 16:30:24 where's that? Nov 03 16:30:57 https://github.com/sonyxperiadev/device-sony-amami Nov 03 16:31:03 supported up to 7.1 Nov 03 16:31:15 sweet. thanks@ Nov 03 16:31:19 [Edit] sweet. thanks! Nov 03 16:31:32 also this https://developer.sony.com/develop/open-devices/guides/aosp-build-instructions/build-aosp-nougat-7-1-2#tutorial-step-5 Nov 03 16:31:55 you'll need to find the correct branch in their kernel repo (there are around 100 iirc), this guide or some other on this site should cover it Nov 03 16:32:17 ok. I'll check Nov 03 16:42:28 can you modify your script to export the log to some file? Nov 03 16:42:41 I don't think I can access recovery without reflashing Nov 03 16:42:51 what do you mean? Nov 03 16:43:00 the crash thingy? Nov 03 16:43:06 yes Nov 03 16:43:08 you don't need all of that Nov 03 16:43:25 anyway I think it's only for a systemd rootfs Nov 03 16:43:36 your still working with hybris-boot right? Nov 03 16:43:44 yeah Nov 03 16:43:46 the initrd is just a shell script Nov 03 16:43:52 you can add the two lines there Nov 03 16:44:01 and you know what ... Nov 03 16:44:08 I think you might not even need that Nov 03 16:44:14 because you don't NEED to crash Nov 03 16:44:20 lol Nov 03 16:44:33 you could just insert a `reboot` somwhere Nov 03 16:44:47 the script is for later kernels that actually NEED to crash to produce a log Nov 03 17:04:03 actually I might just try connecting to it on my laptop Nov 03 17:04:17 it runs ubu bionic so it should have slightly other network setup than arch Nov 03 17:16:36 no luck whatshowever Nov 03 17:17:07 device boots for sure to some extend and shows up as a network interface but it stays on DOWN state Nov 03 17:17:17 also interestingly enough it has a NO-CARRIER flag Nov 03 17:17:33 quick google told me that on wired connection it indicates that cable is unplugged xd Nov 03 17:34:49 okay on raspberry pi it seems more interesting Nov 03 17:34:54 rpi is used as host ofc Nov 03 17:35:30 it shows the device, without the NO-CARRIER flag and assigns an address to it (169.254.136.211/16) Nov 03 17:35:48 but every time I run any `ip` command the network driver of rpi spams dmesg with exceptions lol Nov 03 18:49:06 https://github.com/konradybcio/android_kernel_sony_msm8x27/blob/halium-5.1/arch/arm/configs/cyanogenmod_nicki_defconfig could someone review if I'm missing something important again? Nov 03 19:56:39 alexeysergey was added by: alexeysergey Nov 03 22:18:24 Marin Dauriat was added by: Marin Dauriat **** ENDING LOGGING AT Sun Nov 04 02:59:59 2018