**** BEGIN LOGGING AT Wed Oct 23 02:59:58 2019 Oct 23 04:43:08 Matti Laine was added by: Matti Laine Oct 23 05:59:11 Darn. So the best option would be to build postmarketOS with the GUI I'd want Oct 23 07:11:43 Do Replicant OS & Halium have something in common regarding what they aim/require to do n accomplish? … (sry, i dont understand quite comprehensively what part these two aim to do - HALium wanna do Open HALs and Replicant wanna strip proprietary software at the core... so are these 2 things same??) … Ref: https://en.wikipedia.org Oct 23 07:11:43 /wiki/Replicant_(operating_system) (via @wiki) Oct 23 07:13:26 [Edit] Do Replicant OS & Halium have something in common regarding what they aim/require to do n accomplish? … (sry, i dont understand quite comprehensively what part these two aim to do - HALium wanna do Open HALs and Replicant wanna strip proprietary software at the core... so aren't these 2 things same atleast for HALs??) … Re Oct 23 07:13:26 f: https://en.wikipedia.org/wiki/Replicant_(operating_system) (via @wiki) Oct 23 07:14:20 [Edit] Do Replicant OS & Halium have something in common regarding what they aim/require to do n accomplish? … (sry, i dont understand quite comprehensively what part these two aim to do - HALium wanna do Open HALs and Replicant wanna strip proprietary software at the core... so aren't these 2 things same, atleast for HALs??) … R Oct 23 07:14:20 ef: https://en.wikipedia.org/wiki/Replicant_(operating_system) (via @wiki) Oct 23 07:16:41 [Edit] Do Replicant OS & Halium have something in common regarding what they aim/require to do n accomplish? If yes, then will it be a common good for both if they work together? … (sry, i dont understand quite comprehensively what part these two aim to do - HALium wanna do Open HALs and Replicant wanna strip proprietary software Oct 23 07:16:41 at the core... so aren't these 2 things same, atleast for HALs??) … Ref: https://en.wikipedia.org/wiki/Replicant_(operating_system) (via @wiki) Oct 23 07:20:38 The difference is replicant uses Android as Os and replace some parts. Halium tries to minimise what's needed by android at all. So in halium is only a small lxc android container started to talk with the propietary hardware drives Oct 23 07:52:32 of course, when replicant manages to "free" a driver, a halium port is probably happy to use it - but in the end, if the drivers are freed, the OS' Oct 23 07:52:52 [Edit] of course, when replicant manages to "free" a driver, a halium port is probably happy to use it - but in the end, if the drivers are freed, the OS-es that use halium would rather use mainlin Oct 23 09:37:44 What are namespaces in LXC? If I disable them in defconfig will I still get some hardware to test? Oct 23 09:38:11 @rixchy [The culprit was CONFIG_NAMESPACES=y], This was bootlooping my device. (Turns on, reboots,) Oct 23 09:39:30 @rixchy [What are namespaces in LXC? If I disable them in defconfig will I still get some …], Are they on the required flags for Halium/systemd list? ^^ Oct 23 09:39:44 Not sure. Oct 23 09:40:14 which kernel config checker did you use? from halium docs or from ubports Oct 23 09:40:29 From halium docs. The mer-kernel check. Oct 23 09:42:01 @rixchy namespaces are kernel support for containers Oct 23 09:42:05 lxc won't work without namespaces Oct 23 09:42:23 But it was bootlooping my device, how can I debug the reason, then? Oct 23 09:42:26 Well lxc will work without user ones Oct 23 09:42:54 These which we disable to resolve kernel compile errors Oct 23 09:43:23 Config_users_ns or how it was Oct 23 09:44:35 Wait, so I am confused now. Oct 23 09:45:00 test_lights will won't work too without namespaces, right? Oct 23 09:45:44 "root@localhost:~# /system/bin/logcat … logcat read failure" Oct 23 09:46:30 @rixchy [test_lights will won't work too without namespaces, right?], Its not using the lxc container, yes Oct 23 09:46:58 Can you give me a tip/push on what to do at this moment? Oct 23 09:47:07 Well whats your kernel version? Oct 23 09:47:12 3.4. Oct 23 09:47:21 3.4.113 probably Oct 23 09:47:21 I think this was the culprit, but the goal is UT. Oct 23 09:47:48 Something is conflicting this setting. ALl Android kernels are broken in some way, unfortunately. You do the smallest possible change and it crashes. You could ask in the lxc IRC channel if they know how to debug. Is there any dmesg log for it? Oct 23 09:48:15 Wait, BRB. Just used dmesg in SSH. Oct 23 09:48:46 Well wasn't namespaces responsible for kernel fault? Oct 23 09:48:51 Without lxc running? Oct 23 09:49:15 We wanted to see if it's kernel fault or lxc Oct 23 09:49:43 @Akatsu . [Well wasn't namespaces responsible for kernel fault?], I don't really know, but when CONFIG_NAMESPACES=y is configured in defconfig, sometimes a get SSH for a second or two. That means that the device boots into rootfs. Oct 23 09:50:14 Here is the dmesg log from SSH if you wanna look at it: https://pastebin.com/JKRtUzWJ Oct 23 09:51:33 @Flohack any idea how can I debug phone reboot after few 30s? Oct 23 09:51:52 I bet it's android in lxc which fail on some hardware Init Oct 23 09:52:10 Because getting android image off made phone to boot rootfs just fine Oct 23 09:52:41 @Akatsu . [@Flohack any idea how can I debug phone reboot after few 30s?], 2 options: YOur kernel gets stuck and the watchdog kills it, or the kernel encounters a kernel panic and reboots after 15s or so. 30s seems a bit long. Maybe its watchdog idk. In any case, you need dmesg logs or the newer ones from the other location which I always Oct 23 09:52:41 forget ^1 Oct 23 09:53:01 @Akatsu . [@Flohack any idea how can I debug phone reboot after few 30s?], connect.py -p ssh -u root ? Oct 23 09:53:07 And read dmesg maybe? Oct 23 09:53:08 Dmesg is pretty dead Oct 23 09:53:17 I mean it ends on audio Init logs Oct 23 09:53:22 When alsa is initialized Oct 23 09:53:33 @rixchy [Here is the dmesg log from SSH if you wanna look at it: https://pastebin.com/JKR …], This was running for over 200 seconds? Oct 23 09:53:44 No. More like 2 seconds. Oct 23 09:54:05 If its watchdog, maybe disabling it in defconfig? Oct 23 09:54:47 Well I tried yesterday switching values to n on defconfig and I narrowed down that the CONFIG_NAMESPACES was the culprit. Oct 23 09:54:50 CONFIG_WATCHDOG_NOWAYOUT=y Oct 23 09:54:56 This one? Oct 23 09:55:28 Yeah, try setting that to n Oct 23 09:55:45 I have introduced it myself Oct 23 09:55:52 I mean this option Oct 23 09:56:06 Should I log lxc-checkconfig? Oct 23 09:56:15 @Akatsu . [I have introduced it myself], Meaning you put it in? Oct 23 09:56:25 Yup Oct 23 09:56:57 Ok, try setting it to n, see if you get same behaviour Oct 23 09:57:18 Set the CONFIG_NAMESPACES to y or leave it on n? Oct 23 10:05:50 @Shakendo [Ok, try setting it to n, see if you get same behaviour], What is this option about? If I read it names it gives me impression that it makes kernel to panic Oct 23 10:06:05 If error appear Oct 23 10:07:37 Well its supposed to kill whatever crashes (like alsamixer using 100% kernel for too long) setting to n shouldnt hurt anything Oct 23 10:08:50 https://cateee.net/lkddb/web-lkddb/WATCHDOG.html Oct 23 10:10:44 Hmm Oct 23 10:10:48 It sounds right Oct 23 10:11:15 I remember from tizen docs there is issue with alsa and cpu optimization Oct 23 10:11:24 On 410 Oct 23 10:11:30 One of these days Im gonna go through and document every flag in my defconfig, but thats a side project Oct 23 10:12:49 @Shakendo [One of these days Im gonna go through and document every flag in my defconfig, b …], Woooow😂😂😂 Oct 23 10:13:21 lul. Oct 23 10:13:36 For me its a curiosity exercise Oct 23 10:16:46 CONFIG_WATCHDOG_NOWAYOUT=n and CONFIG_MANESPACES=y. Device still reboots in a loop. Oct 23 10:16:58 @rixchy which rootfs are you running? Oct 23 10:17:12 probably LXC startup causes a kernel panic Oct 23 10:17:37 The halium reference ones. Oct 23 10:18:05 try UBPorts one + halium-boot instead Oct 23 10:18:13 reference is too unmaintained by now Oct 23 10:18:35 But my device is ARM 32 bit, they published 64 bit one. Oct 23 10:18:55 armhf is 32-biy one Oct 23 10:18:59 [Edit] armhf is 32-bit one Oct 23 10:19:05 ...let me look. Oct 23 10:19:54 "Next we’ll need to download the rootfs (root filesystem) that’s appropriate for your device. Right now, we only have one available. Simply download ubports-touch.rootfs-xenial-armhf.tar.gz from our CI server. If you have a 64-bit ARM (aarch64) device, this same rootfs should work for you. If you have an x86 device, let us know. We do not Oct 23 10:19:54 have a rootfs available for these yet." Oct 23 10:20:00 I have a 32 bit device. Oct 23 10:20:16 Oh. Oct 23 10:20:25 I just understood, I apologize. Oct 23 10:20:29 Let me try to install. Oct 23 10:23:07 Halium boot requires more parameters Oct 23 10:23:25 I men kernel config Oct 23 10:28:58 @NotKit [try UBPorts one + halium-boot instead], "Uploading BOOT … 100% … ERROR: Failed to confirm end of file transfer sequence! … ERROR: BOOT upload failed!" Oct 23 10:30:14 boot image is too large? Oct 23 10:30:15 Redo Oct 23 10:30:18 Have you ran check kernel config from halium? Oct 23 10:30:41 Isn't boot.img like 12MB? Oct 23 10:30:50 Wait let me check. BRB. Oct 23 10:36:30 @rixchy ["Uploading BOOT … 100% … ERROR: Failed to confirm end of file transfer sequence! … ERR …], No, you are wrong Oct 23 10:36:39 @Mattia990 [No, you are wrong], Hm? Oct 23 10:36:52 @NotKit [try UBPorts one + halium-boot instead], He suggested halium-boot.. Oct 23 10:37:03 So I executed mka halium-boot and tried to flash it. Oct 23 10:37:06 Where was I wrong? Oct 23 10:37:22 @rixchy [Wait let me check. BRB.], You need to use another script to modify the defconfig, build halium-boot and redo the halium install with ut armhf rootfs Oct 23 12:04:54 Well lxc changes will persist Oct 23 12:04:54 That's good Oct 23 12:04:54 Less stuff to care anout Oct 23 12:04:54 So, what is that script? Oct 23 12:04:54 I do not know such script. Oct 23 12:04:54 Kernel check config inside halium-boot Oct 23 12:04:54 if he can't flash boot, kernel config does not matter at all Oct 23 12:04:54 It has - w parametr to fix config Oct 23 12:04:54 He flashed mer one already Oct 23 12:04:54 what are you suggesting will cause more issues at this moment, better start with known working config Oct 23 12:04:54 Will it boot with mer settings? Oct 23 12:04:54 @Akatsu . [He flashed mer one already], but halium-boot is few mbs larger than hybris-boot due to ramdisk, so if it is larger than his boot partiton, it is a problem Oct 23 12:04:54 Welp, now I am confsued. Oct 23 12:04:54 @Akatsu . [Will it boot with mer settings?], it should Oct 23 12:04:54 Resize boot partition😂 **** BEGIN LOGGING AT Wed Oct 23 12:06:06 2019 Oct 23 12:09:30 [Edit] I indeed have kernel 3.4, but I do have Exynos processor. Oct 23 12:11:58 @rixchy: For 3.4 kernel you would need to apply some patches probably. We needed for QualComm ones. I suspect for Exynos as well. Oct 23 12:13:24 Ok, can you guide me please? Oct 23 12:13:28 @rixchy: See https://github.com/Halium/android_kernel_lge_hammerhead/commits/halium-5.1 and specifically: https://github.com/Halium/android_kernel_lge_hammerhead/commit/25437b2a54dd619a96e268ecaf303b089aa785e4, https://github.com/Halium/android_kernel_lge_hammerhead/commit/5754614eb43dea44a99e54898e3b83d4d96d8b83, https://github.com/Halium/android_kernel_lge_hammerhead/commit/bd221854de33b75 Oct 23 12:13:28 db7a7fa01cb34274b62a7cbf8 and https://github.com/Halium/android_kernel_lge_hammerhead/commit/17a8d0630b424d66981900e701ae2a452d4b8726 Oct 23 12:14:02 I think I did those patches already..... Oct 23 12:14:05 Most other commits are "LuneOS" specific since we use newer GCC toolkit (7 or 8) to build the kernel v.s. GCC 4.8/4.9 used by Halium and Android. Oct 23 12:14:44 Should I make a GitHub repo with all the build files I have? Oct 23 12:14:44 There seem to be various variants of https://github.com/Halium/android_kernel_lge_hammerhead/commit/17a8d0630b424d66981900e701ae2a452d4b8726 floating around for various 3.4 kernel versions. Rest should be pretty generic Oct 23 12:15:04 @rixchy: Yes that's a good approach. For the original kernel you use and apply the patches there Oct 23 12:15:18 Okay.. Oct 23 12:18:50 Where is this file located? https://github.com/Halium/android_kernel_lge_hammerhead/commit/5754614eb43dea44a99e54898e3b83d4d96d8b83 Oct 23 12:20:38 This isn't going to fix anything, since those patches are already applied on my build. Oct 23 12:23:20 Patches are already made, yup. Oct 23 12:24:55 Edit. This patch wasn't applied. -> https://github.com/Halium/android_kernel_lge_hammerhead/commit/17a8d0630b424d66981900e701ae2a452d4b8726 Oct 23 12:25:13 Going to rebuild hybris-boot again with CONFIG_NAMESPACES set to y. Oct 23 12:25:21 That one is more for systemd > 233, but doesn't hurt to have Oct 23 12:25:32 Alright. Oct 23 12:25:40 Where was that fix, in the docs? Oct 23 12:31:23 rixchy: the fix was backported from newer kernel; we found it with some tedious try & debug sessions Oct 23 12:31:55 @rixchy [/home/rixchy_/halium/kernel/samsung/exynos5420/fs/stat.c:65:17: note: each undec …], Do you know how to fix this one? Oct 23 12:32:49 please use a pastebin, I don't see the message you're talking about Oct 23 12:33:00 Oh, sorry! Oct 23 12:33:46 https://pastebin.com/3dDYTZfk Oct 23 12:36:12 mmh this might be a mismatch between more recent code and an old patch; for hammerhead we just use fput: https://github.com/Halium/android_kernel_lge_hammerhead/blob/17a8d0630b424d66981900e701ae2a452d4b8726/fs/stat.c#L65 Oct 23 12:36:30 Ok, I will check the patch in a sec. Oct 23 12:38:18 on tenderloin (HP Touchpad, old kernel too) we had to revert a problematic commit: https://github.com/shr-distribution/linux/commit/01141a353241d545a2aaf9addaa32afe39a98562 Oct 23 12:45:21 @Tofe [mmh this might be a mismatch between more recent code and an old patch; for hamm …], Did the trick. 👍 Oct 23 12:49:54 @rixchy: 3.4 is a bit tricky at times. It's really at it's last legs to be honest. There will be a point where it will be not possible to patch it anymore. But for now with minimal patches it's workable. Oct 23 12:50:30 @Herrie [@rixchy: 3.4 is a bit tricky at times. It's really at it's last legs to be hones …], Wonderful, now I am getting a "heimdall detect … ERROR: Failed to detect compatible download-mode device." error. Oct 23 12:50:35 Great! Oct 23 12:51:52 Heimdall, I'll pass, I don't know anything about that bootloader Oct 23 12:51:59 Same here ;) Oct 23 12:52:07 It's okay. Oct 23 12:52:22 I have a Samsung device, so I have to keep up with that. xD Oct 23 12:53:55 :) Oct 23 13:00:09 Hm. The device isn't rebootin- It rebooted. Oct 23 13:00:18 It is rebooting again. Oct 23 13:00:35 Why is the namespaces interfering with this?! Oct 23 13:00:54 If I set CONFIG_NAMESPACE to n in defconfig, it works! If it's y, then it is rebooting. Oct 23 13:02:48 Can you get a last_kmsg via a recovery? Oct 23 13:02:59 Sure, please wait. BRB. Oct 23 13:03:57 (Document) https://irc.ubports.com/8kYk5BCk.txt Oct 23 13:05:35 looks like the real error are the mount issues just before the crash Oct 23 13:06:00 The mountpoint 21? Oct 23 13:06:18 and loop0 Oct 23 13:06:35 Alright. Mountpoint 21 is userdata. But why. Oct 23 13:06:46 I reformatted it, because I had the same issue yesterday. Oct 23 13:07:04 Tho I have no idea what loop0 is. Oct 23 13:07:05 ACL ? journal ? Oct 23 13:07:13 ? Oct 23 13:07:28 loop0 is when halium mounts a .img, ramdisk for instance Oct 23 13:07:49 Ok. Oct 23 13:08:15 (or system rootfs; it depends on how you've put it in the userdata partition) Oct 23 13:08:40 Alright. Oct 23 13:12:54 Do you have any tips for me to do? Oct 23 13:14:22 you could check your defconfig, see if ACL are enabled, for instance; also compare with a working defconfig like hammerheads Oct 23 13:14:52 Could you explain to me, what ACL is. Oct 23 13:18:48 Access control list Oct 23 13:20:24 CONFIG_NFS_ACL_SUPPORT? Oct 23 13:21:03 Or CONFIG_NFS_V3_ACL? Oct 23 13:29:53 yes, I was just more thinking about the ones for ext4: https://github.com/Halium/android_kernel_lge_hammerhead/blob/533a628f03d4d0be2f1c4e969aa1c6d7a7e564bb/arch/arm/configs/cyanogenmod_hammerhead_defconfig#L3060 (and next lineà Oct 23 13:29:54 ) Oct 23 13:30:35 but I said ACL at random; it just seems related to some FS capability Oct 23 13:33:17 but I might be wrong, as it crashes in the loopback network device init, not in the loop fs driver Oct 23 13:33:30 (I misread the crash stack :) ) Oct 23 13:37:47 Maybe boot to hybris recovery and then mount 21 and then rootfs.img Oct 23 13:38:06 Can you mount? ls inside? Oct 23 13:38:28 @dohniks [Maybe boot to hybris recovery and then mount 21 and then rootfs.img], Can I do this in TWRP? Oct 23 13:38:56 Better hybris. Oct 23 13:39:15 We want to see whether your kernel can handle them Oct 23 13:39:43 There was some way to force hybris-boot into recovery ... Oct 23 13:40:01 Otherwise flash hybris-recovery Oct 23 13:40:13 Alright.. Oct 23 13:42:55 (Photo, 1280x719) https://irc.ubports.com/4k6x1dCd.png hybris-recovery just spammed my screen. Oct 23 13:43:02 @rixchy [], Can someone explain this log to me? It has no "booting linux" and it races from 0s to 12s without doing much ... it looks very very "unlinuxy" to me Oct 23 13:44:24 @rixchy [], Its normal that it shows you all partitions it can find, can be a lot. Also maybe your cable is faulty? Oct 23 13:44:45 No, it always does that when it is in kernel/hybris-recovery. Oct 23 13:44:51 I don't know why. Oct 23 13:51:55 @dohniks [Maybe boot to hybris recovery and then mount 21 and then rootfs.img], Alright, I got in telnet now. Oct 23 13:52:23 How do you mount userdata and rootfs in telnet? Oct 23 13:53:12 Can i somehow find which dtb is my phone using? Oct 23 13:53:33 somehow cmdline isnt respected Oct 23 13:53:41 and i want to append it in dtsi Oct 23 13:54:48 ok will try it in msm8916.dtsi Oct 23 13:58:55 @rixchy [How do you mount userdata and rootfs in telnet?], http://docs.halium.org/en/latest/porting/debug-build/logging-in.html#logging-in Oct 23 13:59:26 `mount /data` Oct 23 14:04:36 Thank you. Oct 23 14:05:25 ~ # mount /userdata … mount: can't find /userdata in /etc/fstab … ~ # mount /data … mount: can't find /data in /etc/fstab Oct 23 14:05:29 Uhm. Oct 23 14:06:25 That was the issue all along? Oct 23 14:06:42 No way that is an issue in mountpoints. Oct 23 14:06:46 mount /dev/mmcbla/blub /data Oct 23 14:07:01 ~ # mount /dev/mmcbla/blub /data … mount: mounting /dev/mmcbla/blub on /data failed: No such file or directory Oct 23 14:07:08 Hehe Oct 23 14:07:11 Yes Oct 23 14:07:20 "Hehe"? Oct 23 14:07:26 bla and blub are placeholders Oct 23 14:07:35 Wot. Oct 23 14:07:42 Whatever the proper device name is for you Oct 23 14:07:51 xD Oct 23 14:08:13 How can I find what to put in the place? Oct 23 14:09:06 What did you put in your fixmountpoints Oct 23 14:14:00 from your last_kmsg, I'd say your userdata is mmcblk0p21 Oct 23 14:16:15 It is. Oct 23 14:17:20 And what should in place of blub? Oct 23 14:18:35 Nothing I guess Oct 23 14:18:40 You want Oct 23 14:18:41 nothing, just use "mkdir -p /data ; mount /dev/mmcblk0p21 /data" Oct 23 14:18:52 this :) Oct 23 14:19:09 "~ # mkdir -p /data ; mount /dev/mmcblk0p21 /data … mount: mounting /dev/mmcblk0p21 on /data failed: Device or resource busy" Oct 23 14:19:39 dmesg Oct 23 14:20:37 (Document) https://irc.ubports.com/O8n8zszY.txt Oct 23 14:33:21 So, how does it look like? :) Oct 23 14:33:31 I'm on my mobile phone, so I can't read this too well, but seems there is neither a "mount" nor a "mmc" to be found ... so dunno :( Oct 23 14:33:50 Oh, I didn't know. I apologize for bothering you. :) Oct 23 14:33:56 No worries Oct 23 14:34:29 Do you know what file system you have on p21? Oct 23 14:35:04 Maybe the kernel doesn't understand it Oct 23 14:35:18 Hm, let me take a look. Oct 23 14:35:39 "-e 's block/platform/dw_mmc.0/by-name/USERDATA mmcblk0p21 ' \" Oct 23 14:35:44 It is userdata, of course. Oct 23 14:35:57 I have been getting alot of issues between kernel and userdata. Oct 23 14:37:18 @rixchy ["-e 's block/platform/dw_mmc.0/by-name/USERDATA mmcblk0p21 ' \"], No that's not what I mean Oct 23 14:37:22 Try this: Oct 23 14:37:37 Boot back onto twrp Oct 23 14:37:47 Oh, wait. Then I have to flash it. Oct 23 14:38:01 Mount it there and then check the output of `mount` Oct 23 14:38:19 It should show you the file system type of userdata Oct 23 14:38:34 It is ext4. Oct 23 14:38:52 I already know that, because I reformatted it yesterday using TWRP. Oct 23 14:38:52 I guess ext2,3,4 are ok, 4 preferred, others might give trouble Oct 23 14:38:54 Oh Oct 23 14:39:02 Mhm! Oct 23 14:39:02 Good Oct 23 14:39:47 @rixchy ["-e 's block/platform/dw_mmc.0/by-name/USERDATA mmcblk0p21 ' \"], How did you get this solution for userdata? Oct 23 14:40:13 What do you mean? That is a line from fixup-mountpoints. Oct 23 14:40:48 I understand, but what command did you run to figure out that is userdata? Oct 23 14:41:18 Wait, BRB. (I set up mountpoints from docs, BTW.) Oct 23 14:42:15 @Shakendo [I understand, but what command did you run to figure out that is userdata?], I looked in fixup-mountpoints. And fixup-mountpoints was setup after fstab and readlink -f [src]. Oct 23 14:43:40 Ok, I did that step a little differently, but thats niether here nor there, one sec while I check my fixup_mountpoints, havent looked at mine in months Oct 23 14:45:20 Your output wasnt "/block/platform/..." Oct 23 14:45:35 .? Oct 23 14:46:12 When you did those commands it didnt output /block/platform/by-name/userdata? Oct 23 14:46:24 No? Oct 23 14:46:37 Interesting Oct 23 14:46:39 readlink gave me the mm..... output. Oct 23 14:47:11 I setup my mountpoints after these instructions: http://docs.halium.org/en/latest/porting/build-sources.html#include-your-device-in-fixup-mountpoints Oct 23 14:47:59 I have a feeling your troubles are related to mountpoints cause if I understand correctly, userdata is not mounting, correct? Oct 23 14:48:07 Correct! Oct 23 14:48:23 Ok, do you still have twrp on your device? Oct 23 14:48:40 @Shakendo [Ok, do you still have twrp on your device?], Nope, but I can flash it real fast. (I have it downloaded.) Oct 23 14:49:18 That would help, it would likely be easier to leave it on your device for now too Oct 23 14:49:25 Alright, BRB. Oct 23 14:50:02 Alright, flashedit. Oct 23 14:50:48 Ok, can you adb shell in and do "ls -ln /dev/block/" for me Oct 23 14:50:53 Sure. Oct 23 14:51:51 (Document) https://irc.ubports.com/Q0tutjAZ.txt Oct 23 14:52:40 I cant open that on my mobile :( Oct 23 14:53:00 Oh, sorry. BRB. Oct 23 14:53:18 Is there a platform? Oct 23 14:53:23 https://pastebin.com/9Biusj5i Oct 23 14:53:26 @Shakendo [Is there a platform?], ? Oct 23 14:54:28 Same thing but in /dev/block/platform/ Oct 23 14:55:36 So that's the issue why device reboots, and creates other issues, right? Oct 23 14:56:08 Your picking up my train of thought ... I hope, whatd you get for output? Oct 23 14:56:25 What output? Oct 23 14:56:28 Haha. Oct 23 14:56:49 "Ls -ln /dev/block/platform/" Oct 23 14:57:33 "~ # ls -ln /dev/block/platform/ … __bionic_open_tzdata: couldn't find any tzdata when looking for localtime! … __bionic_open_tzdata: couldn't find any tzdata when looking for GMT! … __bionic_open_tzdata: couldn't find any tzdata when looking for posixrules! … drwxr-xr-x 4 0 0 580 Oct 23 14:50 dw_mmc.0" Oct 23 14:58:15 Ok, and do that again but add dw_mmc.0 to the end, Im trying to just figure it out Oct 23 14:58:58 Here. https://pastebin.com/pseZ3hKN Oct 23 14:59:52 And lastly again with by-name on the end please Oct 23 15:01:02 Here, again. :) https://pastebin.com/dQ0wMPxd Oct 23 15:04:57 Hmm from those outputs it should be working, when you added them you followed the other devices as an example for spacing (using the space bar not tab) right? Oct 23 15:05:46 Hm, let me check, Oct 23 15:06:00 Believe it or not that makes a difference Oct 23 15:06:24 @Shakendo [Believe it or not that makes a difference], Yes, I know that. Oct 23 15:06:29 @Shakendo [Hmm from those outputs it should be working, when you added them you followed th …], Yes. Oct 23 15:07:00 Ok, out of curiosity compile halium-boot for me and see what happens Oct 23 15:07:21 I tried, it didn't flash. Oct 23 15:07:36 What was the error when you tried to flash it? Oct 23 15:07:54 @rixchy ["Uploading BOOT … 100% … ERROR: Failed to confirm end of file transfer sequence! … ERR …], Here. Oct 23 15:08:21 I get that same error using heimdall, try flashing it in twrp Oct 23 15:08:31 Alright. Oct 23 15:08:46 Adb push it to your device then flash it in twrp to boot Oct 23 15:09:42 Believe it or not but that makes a difference as well Oct 23 15:10:15 "Size of image is larger than target device" Oct 23 15:10:36 Thats from trying to flash it in twrp? Oct 23 15:10:40 Yup. Oct 23 15:10:54 I think the boot partition is too small. Oct 23 15:11:18 Halium boot is 9,3 MB for me. Oct 23 15:11:22 Ok, interesting, where did you push the halium-boot.img file too? Oct 23 15:11:33 I pushed it to / Oct 23 15:12:00 Try pushing it to /data/ and delete the one at / cause that could affect things Oct 23 15:12:15 Okay..? Oct 23 15:12:52 Nope. Oct 23 15:12:57 Still the same. Oct 23 15:12:57 Same error? Oct 23 15:13:01 Yup. Oct 23 15:14:10 Ok, thats something youll need to address at some point, try it with hybris-boot then Oct 23 15:15:15 hybris-boot is rebooting all the time in a loop. Oct 23 15:15:43 Ok, did you copy a rootfs and system.img? Oct 23 15:15:49 Yes. Oct 23 15:16:44 Hmm, what script did you use to install them? Oct 23 15:17:02 halium-install. Oct 23 15:17:06 From JBB. Oct 23 15:17:54 Ok, and you used "-p hybris" when you ran it? Oct 23 15:18:17 Nope. Just "halium-install rootfs.tar.gz system.img" Oct 23 15:18:43 Ok, try running that again the -p hybris see if it makes a difference Oct 23 15:18:48 Alright. Oct 23 15:19:02 Where should I add it? At the end? Oct 23 15:19:25 Like this: "halium-install rootfs.tar.gz system.img -p hybris"? Oct 23 15:19:32 Before you specify rootfs Oct 23 15:19:42 Alright, let me try that. Oct 23 15:20:28 "rixchy_@ubuntu-asus:~/Desktop/halium-install-master$ ./halium-install -p hybris halium_rootfs.tar.gz system.img … Unsupported post install release! Exiting now" Oct 23 15:24:20 Latest version? Oct 23 15:24:26 Yup. Oct 23 15:24:47 Hmm, I wonder if it was changed, try changing hybris to halium Oct 23 15:25:16 @rixchy [I think the boot partition is too small.], is boot partition 8 mb? Oct 23 15:25:31 How do you check that? Oct 23 15:25:58 iirc my /boot/ is 10mb Oct 23 15:26:27 maybe TWRP shows it? otherwise through fdisk -l or similar Oct 23 15:27:25 Believe it, or not. I can't find a boot folder in TWRP. Oct 23 15:29:13 @NotKit [maybe TWRP shows it? otherwise through fdisk -l or similar], Doesnt fstab have that info? Oct 23 15:29:26 I don't think so, let me check. Oct 23 15:29:48 Nope. Oct 23 15:31:05 Hmm, I know I have something in my source that specifies partition sizes ... anyway fdisk -l should show it Oct 23 15:31:19 adb shell? Oct 23 15:31:41 Unless you wanna know partition sizes on your developer machine Oct 23 15:31:51 Haha. Oct 23 15:31:55 It returned nothing. Oct 23 15:31:58 Just emptiness. Oct 23 15:32:12 Fdisk -l has no output? Oct 23 15:32:21 Nope. Oct 23 15:32:34 Does fdisk by itself show anything? Oct 23 15:32:39 (Photo, 1280x719) https://irc.ubports.com/T0zu8QDy.png Oct 23 15:32:52 Yep. Oct 23 15:33:00 Try fdisk -l /boot/ Oct 23 15:33:21 fdisk: can't open '/boot/': No such file or directory Oct 23 15:33:38 That's true, because I couldn't find the boot folder in TWRP too. Oct 23 15:34:16 "fdisk -l /dev/block/mmcblk0p9 Oct 23 15:34:37 Ay. We got a return! Oct 23 15:34:39 "Disk /dev/block/mmcblk0p9: 8 MB, 8388608 bytes … 4 heads, 16 sectors/track, 256 cylinders … Units = cylinders of 64 * 512 = 32768 bytes … Disk /dev/block/mmcblk0p9 doesn't contain a valid partition table" Oct 23 15:34:51 This is where the ls commands I told you earlier pay off Oct 23 15:35:07 Okay. Oct 23 15:35:39 Thats interesting, but it answers @NotKit question Oct 23 15:37:27 So now you know, you have to get halium-boot to within that size, should be doable by changing some of the =y flags in your defconfig to =m Oct 23 15:38:08 Yeah, but the problem is. Which values should I change? Oct 23 15:39:23 I would go with ones pertaining to things like BT, as a start and go from there, try to keep it to things that arent crucial to system boot like wifi and radio services Oct 23 16:04:30 @Shakendo [So now you know, you have to get halium-boot to within that size, should be doab …], I don't think so honestly Oct 23 16:05:40 Ive been wrong before Oct 23 16:08:06 what is realistic is to have small ramdisk that can extract halium-boot ramdisk from partition Oct 23 16:08:40 but that requires at least basic bash knowledge and understanding how things work Oct 23 16:10:09 Well I explained things as I understand them Oct 23 17:06:06 do halium rootfs also use pulseaudio? Oct 23 17:07:22 I think it doesn't have anything for sound Oct 23 17:08:19 i wonder what is a reason behind this buggy audio Oct 23 17:08:32 i know sheduler based audio is choppy Oct 23 17:08:36 but why it happens Oct 23 17:20:59 Vamsi QT was added by: Vamsi QT Oct 23 17:58:54 @NotKit if i look into logcat and i see app wanting to call for file inside etc do it want to call it for etc located inside of lxc container or rootfs? Oct 23 17:59:22 likely lxc if it is logcat Oct 23 17:59:31 its says like No .acdb files found in /etc! Oct 23 18:01:22 it finds them in /android/system/etc/General_cal.acdb however there is no such thing in /var/lib/lxc/android/rootfs/etc Oct 23 18:02:15 uh these are in rootfs/system/etc, any idea how can i symlink it to be in /etc instead of /system/etc ? Oct 23 18:02:56 no question sorry Oct 23 18:03:11 Hi. Back again. Oct 23 18:04:18 I started wondering, boot partition shouldn't be set in fixup-mountpoints, right? Oct 23 18:04:33 @rixchy ["Disk /dev/block/mmcblk0p9: 8 MB, 8388608 bytes … 4 heads, 16 sectors/track, 256 c …], . Oct 23 18:04:47 That is boot parition right? Oct 23 18:05:01 yes, 8 mb Oct 23 18:05:16 Ok, but it says something about the format. What's that? Why? Oct 23 18:05:24 I mean the partition table. Oct 23 18:29:13 because a single partition does not have a partition table on it Oct 23 18:31:20 How can I fix that? Oct 23 18:59:53 Arun Nidhi was added by: Arun Nidhi Oct 23 19:01:48 (Photo, 606x1280) https://irc.ubports.com/AI2CPfz3.png Oct 23 19:02:00 This is my phone info Oct 23 19:02:23 Android 8.1, no Oct 23 19:02:25 Tell me how I install Ubuntu on my phone Oct 23 19:02:47 @Arun Nidhi [Tell me how I install Ubuntu on my phone], First, you port halium to your device. Oct 23 19:03:10 Or just get already ported build, and install the UT rootfs with halium-install. Oct 23 19:03:20 And the hybris-boot. Oct 23 19:04:06 @rixchy [First, you port halium to your device.], Could you please let me Oct 23 19:04:15 @Arun Nidhi [This is my phone info], seeing it is Vivo, they don't seem to be caring about GPL to release kernel source code. Is it available for you device? If answer is no, all I can do is to advice to get another device Oct 23 19:04:16 https://docs.ubuntu.com/phone/en/devices/porting-new-device Oct 23 19:05:37 @Arun Nidhi [], No LOS build for your device. Oct 23 19:05:54 @NotKit [seeing it is Vivo, they don't seem to be caring about GPL to release kernel sour …], Can't he like extract the kernel files somehow from his existing ROM? Oct 23 19:06:44 no, you can't trivialy decompile binary back into C code Oct 23 19:06:49 Tell me the supporting devices Oct 23 19:07:05 @Arun Nidhi [Tell me the supporting devices], They are on the UBports website, please wait. Lemme link to it. Oct 23 19:07:30 Here are the Ubuntu Touch supported devices: https://docs.ubuntu.com/phone/en/devices/devices Oct 23 19:07:40 @rixchy [They are on the UBports website, please wait. Lemme link to it.], Sure and thank you Oct 23 19:07:48 @rixchy [Here are the Ubuntu Touch supported devices: https://docs.ubuntu.com/phone/en/de …], Thank you Oct 23 19:08:07 These devices are already built, but you can always port your own device, if it's files are found in CyanogenMod or LineageOS repositories. Oct 23 19:08:38 Ok Oct 23 19:08:53 @rixchy [These devices are already built, but you can always port your own device, if it' …], beware that we don't have support for Android 8/9/10 devices yet Oct 23 19:09:10 Oh, yeah. I apologize for that. Oct 23 19:09:27 But, if the oldest version of Android can be installed, then it works. Oct 23 19:10:50 (Photo, 949x705) https://irc.ubports.com/7ugrKGfT.png I am so mad at myself that I didn't find this earlier. This just freed myself from the confusion. Oct 23 19:10:50 I mean the problem are devices that got released with Android 8 and newer Oct 23 19:11:03 I understood you. Oct 23 19:15:20 (Photo, 949x705) https://irc.ubports.com/0nBTl4Z4.png I am so mad at myself that I didn't find this earlier. This just freed myself from the confusion. Especially about LXC. Oct 23 19:39:54 @NotKit [but that requires at least basic bash knowledge and understanding how things wor …], Could he just change kernel compression? Oct 23 19:40:09 can be tried Oct 23 19:40:18 but 1.4 mb is a lot Oct 23 19:40:41 though ramdisk is also gzipped, maybe xz for both would give enough improvement Oct 23 19:45:40 I didn't understand anything totally. 👍😂 Oct 23 19:48:16 Im not in front of my laptop ... so a bit vague ... there is a config that determines the compression of initrd and kernel, so basically of halium-boot.img Oct 23 19:48:42 Look for lzma , gzip, stuff like that Oct 23 19:49:00 There should be a way to configure xc instead Oct 23 19:49:17 Should give higher compression, maybe it fits then Oct 23 19:49:22 Clear? Oct 23 19:53:38 Yes..? Oct 23 20:16:18 https://m.youtube.com/watch?v=2aegP8j5al0 Oct 23 20:24:14 anyone with a device supported by phh-treble 9.0 GSI who wants to experiment with Android 9 support in UBPorts? Oct 23 20:27:11 Won't it work on any gsi device with 9.0 vendor? Oct 23 20:27:35 @dohniks [Should give higher compression, maybe it fits then], Guide me please, because I generally do not understand what you are talking about. xD Oct 23 20:28:24 Something about compressing kernel and initramfs Oct 23 20:28:59 .. Oct 23 20:29:28 Kk. Lemme try again :( Oct 23 20:29:31 :) Oct 23 20:29:42 [Edit] Kk. Lemme try again :) Oct 23 20:29:57 @Akatsu . [Won't it work on any gsi device with 9.0 vendor?], phh-treble should, but no idea about rest Oct 23 20:30:05 still need custom kernel at least Oct 23 20:30:34 Well phusson is one of gsi devs Oct 23 20:30:58 Gsi is gsi, there is also sgsi which is different Oct 23 20:31:27 Sadly I have no device with gsi Oct 23 20:31:53 Maybe after nullbyte will release treble for s2 Oct 23 20:33:44 If I would flash UT rootfs on my device.. it wouldn't make any difference from halium refernce ones.. right? Oct 23 20:34:22 Maybe rootfs alone won't be broken Oct 23 20:34:38 Halium reference is from what I know bit outdated Oct 23 20:34:46 But would that change somethings and maybe something will work? Oct 23 20:34:50 Is that even worth a try? Oct 23 20:35:05 @rixchy [Guide me please, because I generally do not understand what you are talking abou …], add CONFIG_KERNEL_XZ=y and remove any other CONFIG_KERNEL_something , similarly add CONFIG_RD_XZ=y and remove any other CONFIG_RD_whatever Oct 23 20:35:22 (I think ....) Oct 23 20:36:24 But trying to install UT rootfs is worth a try? Akatsu Oct 23 20:36:32 Is it going to affect anything? Oct 23 20:37:23 Tbh I dont know Oct 23 20:37:28 I started with ut Oct 23 20:37:47 BTW lzma seems to be best performing one Oct 23 20:38:01 https://wiki.postmarketos.org/wiki/Kernel_configuration Oct 23 20:38:18 I will port plasma tomorrow on le2 Oct 23 20:38:52 @Akatsu . [I started with ut], Ok, I will try UT, but I won't get any graphics, right? Oct 23 20:39:11 No lxc no graphics Oct 23 20:39:14 No LXC = no graphics? Oct 23 20:39:17 OH. Oct 23 20:39:22 Haha. Oct 23 20:39:47 I also have issues with my builds Oct 23 20:39:48 So if I set CONFIG_NAMESPACES, then LXC is disabled.. right? Oct 23 20:40:43 @Akatsu . [BTW lzma seems to be best performing one], oh, ok, LZMA instead of XZ then ! Oct 23 20:41:17 @doniks do you know if we can use newer pulseaudio on Ubuntu? Oct 23 20:41:19 @rixchy [So if I set CONFIG_NAMESPACES, then LXC is disabled.. right?], ..? Oct 23 20:42:15 @Akatsu . [@doniks do you know if we can use newer pulseaudio on Ubuntu?], no clue 😱 Oct 23 20:42:39 I want to get working audio ehh Oct 23 20:43:05 Any idea why I get no sound in phone app? Oct 23 20:44:05 I mean dialer Oct 23 20:44:10 all I have ever picked up about that was whats in the ubports porting faq Oct 23 20:49:08 If CONFIG_NAMESPACES set to n, then LXC is not working or what.? Oct 23 20:49:26 Is CONFIG_NAMESPACES required to set to y to make something appear on the screen? Oct 23 20:54:01 @Akatsu . [I mean dialer], like during a phone call? hearing the voice from the other side? I think this was mentioned in the portcast ... but just that it's complicated and I think sony xperia x was mentioned as one of the few (?) where it works. maybe poke around in that repo... Oct 23 20:54:22 Dilaer where you press numbers Oct 23 20:54:52 I feel like mainline is the way to go for that shit Oct 23 20:55:09 (shall we step over into ubports porting chat?) Oct 23 20:55:18 I should look at other devices on postmarketos and follow Oct 23 20:55:30 Probably Oct 23 21:55:38 @NotKit [anyone with a device supported by phh-treble 9.0 GSI who wants to experiment wit …], yup? Oct 23 21:56:53 well, I have rootfs that can supposedly boot to SSH over USB, then with proper mounts Android init can be started and Mir works Oct 23 21:57:46 without kernel change? Oct 23 21:58:43 with custom kernel and ramdisk, trying without would be further step Oct 23 21:59:01 LXC is not used though Oct 23 21:59:13 ok. do you have a variant with ramdisk and rootfs merged together? Oct 23 21:59:45 it is problematic. rootfs is too large to fit onto /system partition, so something has to mount it from /data/rootfs.img and setup bind-mounts Oct 23 21:59:56 Is it possible to resize boot partition on Samsung devices? Oct 23 22:00:05 uh? how large is rootfs? Oct 23 22:00:21 we usually create rootfs.img with 3G Oct 23 22:00:30 it is not used fully, but I think it is more than 2G Oct 23 22:00:36 many devices have 3G system partition :P Oct 23 22:01:09 2GB is the bare minimum for 64bits android really Oct 23 22:01:13 it can be squashed then on smaller ones, but it will break remount rw Oct 23 22:01:23 anyway, I guess you could do an ext4 ramdisk that I can just flash as system partition? Oct 23 22:02:25 Is it possible to resize the boot parition? My halium-boot is 9 MB, and it won't fit on my boot partition. Oct 23 22:02:28 this could work, but system.img with AOSP build is needed Oct 23 22:02:44 @rixchy [Is it possible to resize the boot parition? My halium-boot is 9 MB, and it won't …], depends on device, no idea with your Samsung Oct 23 22:02:59 I guess we can pack AOSP's system.img inside ext4 ramdisk which would be real rootfs Oct 23 22:03:06 Hm, alright. Let me look in google. Oct 23 22:03:22 @rixchy keyword on samsung devices is "FIT" Oct 23 22:03:35 PIT? Oct 23 22:03:48 ah uh yeah Oct 23 22:05:18 @phhusson but as I see, Magisk patches boot.img and force-enables ramdisk? Oct 23 22:05:57 maybe similar is easier to start with Oct 23 22:14:31 well i really don't like changing my device's boot.img Oct 23 22:15:04 anyway if you give me a rootfs, a system.img and a ramdisk for arm32 device, I should be able to cook something Oct 23 22:16:03 arm32? which device do you have in mind, just curious? Oct 23 22:17:13 well, I have a tv box on which I fully control the boot chain and I have a serial port, so it'll be easier to prototype this on Oct 23 22:17:54 does it have 64-bit binder? Oct 23 22:18:11 good point. yes it does Oct 23 22:33:01 @rixchy [But trying to install UT rootfs is worth a try? Akatsu], For ut you need halium-boot, but you said it is too large for boot partition so nothing Oct 23 22:33:31 I don't remember if plasma uses halium or hybris, i think hybris Oct 23 22:33:53 @Mattia990 [For ut you need halium-boot, but you said it is too large for boot partition so …], I tried repartitioning the boot with heimdall, no dice. Oct 23 22:34:41 @Mattia990 [I don't remember if plasma uses halium or hybris, i think hybris], I started the port after UT afterall. Oct 23 22:34:58 @rixchy [I tried repartitioning the boot with heimdall, no dice.], I partitioned cache partition on an old phone and worked, idk if it works with boot too Oct 23 22:35:22 How did you do it? Oct 23 22:35:29 Worth a try. Oct 23 22:39:55 @rixchy [How did you do it?], Like this: https://www.google.com/amp/s/forum.xda-developers.com/xiaomi-mi-3/general/howto-resize-extend-partition-capacity-t3011918/amp/ … But if you want to try you need a loooot of attention on what you are doing, because you can break everything Oct 23 22:41:11 Alright, thank you. Have to go to sleep, maybe will work on the port, since I've got a road trip tomorrow. **** BEGIN LOGGING AT Thu Oct 24 01:31:54 2019 **** ENDING LOGGING AT Thu Oct 24 02:59:57 2019