**** BEGIN LOGGING AT Sat Nov 16 02:59:57 2019 Nov 16 03:23:41 <庄桦番> (Photo, 1255x679) https://irc.ubports.com/Vh0mRGjG.png "精准币圈菠菜棋牌粉" "非小号/mytoken惊爆价"1、币圈电报群/微信群炸群推广 2、twitter、facebook关注 3.国内外KYC身份材料出售 4、电报群/微信群/potato拉币圈活人 5、国内外交易所APP实名注册认证 6、电报群客服活聊天 7、批量注册国内外交易所实名账户,,出twitter、facebook、微信带粉账号 @Cindyilove2 Nov 16 03:46:45 @rudi9719 [@Fuseteam - idk his handle here but one of my coworkers and I are exploring AOSP …], Oh das me Nov 16 03:47:19 I’m messing around with learning aosp yes Nov 16 03:47:44 I have a server at home I use for compiling which is nice Nov 16 03:48:10 Unfortunately I’m dealing with currently having made everything angry and having little in the way of debug messages Nov 16 04:10:29 @Anthony Olmstead [I have a server at home I use for compiling which is nice], I do the same thing, I have prob 15 different copies of local repos, half of which are broken or not happy Nov 16 04:12:39 Nice :) I have two poweredge r710s. One has dual 6c/12t xeons, the other one will be upgraded as such within a week. Both have 72G RAM. They’re nice little things, I love em. Really good for this kind of work Nov 16 04:13:00 Lowers compilation time to around an hour and a half instead of like... 6-8 Nov 16 08:56:09 @Flohack [THats the whole log from last_kmsg?], Hey, just woke up. Yes, that's the whole log, indeed. Nov 16 09:25:14 ArunG was added by: ArunG Nov 16 10:05:49 @Akatsu . [Joginder Singh there is halium port for redmi2 already], Ok, thank you. Nov 16 10:48:11 @rixchy [Hey, just woke up. Yes, that's the whole log, indeed.], Thats really few lines. Smth is terribly wrong here ^ Nov 16 11:10:56 @Flohack [Thats really few lines. Smth is terribly wrong here ^], Yep, previous log (when rootfs wasn't installed) was longer. I can try to get a log again, since this happened a few times to me, and getting another log, revealed other lines. Nov 16 11:13:20 Hm, I retried, and now we have a longer log, yay! Nov 16 11:13:32 (Document) https://irc.ubports.com/8SLGyeC4.txt Nov 16 11:13:44 This line sounds relevant … `EXT2-fs (loop0): error: couldn't mount because of unsupported optional features (244)` Nov 16 11:14:03 no actually I'm wrong, sorry Nov 16 11:14:14 @JBBgameich [This line sounds relevant … EXT2-fs (loop0): error: couldn't mount because of unsu …], I had this before, and I didn't pay attention to it. This is what needs to be fixed: kernel BUG at /home/rixchy/halium/kernel/samsung/exynos5420/drivers/net/loopback.c:205! Nov 16 11:14:46 I did fix it before, but I just deleted the damn line in the file, now I gotta fix it, so I don't run in other complications in the future. Nov 16 11:15:48 @rixchy [I had this before, and I didn't pay attention to it. This is what needs to be fi …], Actually, now I noticed something! It is the internet driver, when I got SSH on my device like a month ago, I couldn't ping to a Wi-Fi. So maybe this could be a reason! Nov 16 11:16:42 It maybe isn't useful, but it's this line: "BUG_ON(dev->ifindex != LOOPBACK_IFINDEX);" Nov 16 11:17:37 Well it's the loopback driver, the `127.0.0.1` thing. It's probably needed for quite a lot of things, but it isn't directly relatd to WiFI I think. Nov 16 11:17:46 Oh, alright. Nov 16 11:17:51 But I gotta fix it, anyways. Nov 16 11:17:58 Not by deleting the line, I guess. Nov 16 11:20:34 Would it be a good idea to replace the loopback.c in my local with this https://github.com/ianmacd/winnerx/blob/master/drivers/net/loopback.c Nov 16 11:21:01 The Git one has more lines than the original has. That means that it is missing somethimg/ Nov 16 11:27:17 That didn't work. (OFC I did save the original loopback.) Hybris-boot build failed. Well, now I literally don't know what to do, but I'll note this error and look if I would be able to fix it, for now I will just remove the line 205... Nov 16 11:33:08 This means that SSH's up? "4 GNU/Linux Device on rndis0 10.15.19.82" Or should I try telnet? Nov 16 11:33:24 [Edit] This means that SSH's up? "4 GNU/Linux Device on rndis0 10.15.##.##" Or should I try telnet? Nov 16 11:37:30 Hell yeah! Got SSH on my device! Nov 16 11:54:10 @JBBgameich [Well it's the loopback driver, the 127.0.0.1 thing. It's probably needed for qui …], For some reason, the Wi-Fi is working! Nov 16 11:55:02 But the test_lights - aren't. "test_lights: test_lights.c:34: main: Assertion `notifications != NULL' failed." Nov 16 12:16:32 (Document) https://irc.ubports.com/U007IdIz.txt Nov 16 12:16:35 Ok, so. Nov 16 12:16:40 My kernel OOPed. Nov 16 12:17:03 After trying to fix this error: lxc-start: utils.c: mkdir_p: 254 Invalid argument - failed to create directory '/sys/fs/cgroup/net_cls//lxc/android' Nov 16 12:17:29 Docs recommended this: "CONFIG_NET_CLS_CGROUP=n … CONFIG_NETPRIO_CGROUP=n" I did it, but it broke something. Nov 16 12:47:54 It's not possible that "EXT4-fs (loop0): couldn't mount as ext3 due to feature incompatibilities" OOPs the whole kernel, right? Nov 16 13:04:42 @rixchy [Docs recommended this: "CONFIG_NET_CLS_CGROUP=n … CONFIG_NETPRIO_CGROUP=n" I did i …], Use lxc-checkconfig from ssh to see what configs are missing Nov 16 13:05:12 @Mattia990 [Use lxc-checkconfig from ssh to see what configs are missing], Hm, okay. Let me disable those options, rebuild and flash. Hang tight there. Nov 16 13:07:48 I recommend in such cases to set lxc container startup to manual Nov 16 13:08:25 First get the ut init stable. And thats also enough for wifi normally. Then try to get the container running Nov 16 13:09:04 It is the lxc-android-config job that needs to be modified. Or similar name Nov 16 13:09:13 Okay. Nov 16 13:09:32 It is compiling now, I'll check the config and all that stuff. Nov 16 13:09:33 You know how to deal with upstart jobs? Nov 16 13:09:36 Nope. Nov 16 13:09:52 @Mattia990 [Use lxc-checkconfig from ssh to see what configs are missing], I think I need to fix all of those missing things before the lxc container starts? Nov 16 13:10:05 Ok then, I am on the road now but ppl here can help you Nov 16 13:10:05 Because in my case - if LXC starts, kernel panics. Nov 16 13:10:10 @Flohack [Ok then, I am on the road now but ppl here can help you], Ok, thank you. Nov 16 13:10:32 @rixchy [Because in my case - if LXC starts, kernel panics.], Got the same on my Nexus 6P Nov 16 13:10:58 Null pointer crash Nov 16 13:17:20 @Mattia990 [Use lxc-checkconfig from ssh to see what configs are missing], Here you go, the ouput https://pastebin.com/bGdJGdgz Nov 16 13:23:45 Hm, is "Veth pair device" configured with CONFIG_VETH? Nov 16 13:28:07 Yep, it was! Now "Veth pair device" is "Veth pair device: enabled". One missing down. Nov 16 13:59:18 Hm, "CONFIG_NF_NAT_IPV4: missing … CONFIG_NF_NAT_IPV6: missing" Even after settings these CONFIGs to y in defconfig? Nov 16 13:59:25 [Edit] Hm, "CONFIG_NF_NAT_IPV4: missing … CONFIG_NF_NAT_IPV6: missing" Even after settings these CONFIGs to y in defconfig. Nov 16 13:59:34 [Edit] Hm, "CONFIG_NF_NAT_IPV4: missing … CONFIG_NF_NAT_IPV6: missing" Even after setting these CONFIGs to y in defconfig. Nov 16 14:19:06 @rixchy [Hm, "CONFIG_NF_NAT_IPV4: missing … CONFIG_NF_NAT_IPV6: missing" Even after setting …], check dependencies Nov 16 14:19:19 How? Nov 16 14:26:27 @rixchy [How?], open net/ipv6/netfilter/Kconfig file and see its defined or not, maybe the name is changed like this IP6_NF_NAT Nov 16 14:27:04 It's in the kernel folder? Nov 16 14:30:01 (Document) https://irc.ubports.com/IbFl5qTz.txt Nov 16 14:30:10 I didn't understand you. Nov 16 14:30:20 (Document) https://irc.ubports.com/dBVtbH68.txt Nov 16 14:30:36 (Document) https://irc.ubports.com/GVKf9NYk.txt Nov 16 14:30:42 (Document) https://irc.ubports.com/72Nk2YXO.txt Nov 16 14:30:47 (Document) https://irc.ubports.com/TwW3SJN0.txt Nov 16 14:45:00 this file net/ipv6/netfilter/Kconfig Nov 16 14:45:09 in your kernel folder Nov 16 14:45:29 @rixchy [], It's this one. Nov 16 14:45:46 Oh wait! Nov 16 14:45:51 Sorry, now I understood! Nov 16 14:46:27 (Document) https://irc.ubports.com/s0aQVS3y.txt Nov 16 14:47:45 Fighting next day? Nov 16 14:48:16 Mhm! LXC crashes the kernel if started on startup, so now i am tryin to fix the missing CONFIGs and see f it makes any effect. Nov 16 14:52:47 Walid How's it look like? Nov 16 15:00:39 I just noticed that you can lxc-checkconfig without even connecting to SSH. Like this: "CONFIG=~/halium/kernel/samsung/exynos5420/arch/arm/configs/lineageos_chagalllte_defconfig /usr/bin/lxc-checkconfig". Nov 16 15:01:10 But for some reason, everything's there enabled, but when I check it through SSH, then something's missing... But vlan's also missing. Nov 16 15:04:46 Oh my. Nov 16 15:04:50 I think LXC started. Nov 16 15:05:07 I fixed all the things, and my device isn't rebooting. Nov 16 15:05:22 And it usually reboots with LXC on.. I gotta check ASAP. Nov 16 15:05:39 OH MY. Nov 16 15:05:40 101: lxcbr0 Nov 16 15:05:43 IN IP LINK. Nov 16 15:05:51 I AM FREAKING OUT. Nov 16 15:09:13 False alarm. -_- Nov 16 15:14:22 "failed to create directory '/sys/fs/cgroup/net_cls//lxc/android-999'" Could be, because I have a 3.4 kernel. Nov 16 15:35:08 @rixchy give full defconfig Nov 16 15:35:22 @Akatsu . [@rixchy give full defconfig], Okay, will give in a min. Nov 16 15:37:44 (Document) https://irc.ubports.com/2N2F8h7C.txt Nov 16 15:37:45 Here. Nov 16 15:39:41 I actually wanted to try Plasma Mobile now, because usually drivers are broken on reference rootfs. Nov 16 15:40:19 and what is output of lxc checkconfig on device? Nov 16 15:40:57 Here: https://pastebin.com/e4y3yhJR Nov 16 15:41:40 its documented inside docs Nov 16 15:41:55 What is documented, where? Nov 16 15:42:18 (Photo, 778x394) https://irc.ubports.com/gpg8KbhM.png Nov 16 15:42:32 Oh. Nov 16 15:44:44 I think it will crash. Nov 16 15:44:53 (Sticker, 512x365) https://irc.ubports.com/ghKE0MWv.webp Nov 16 15:45:09 (Sticker, 512x512) https://irc.ubports.com/h9PUhygX.webp Nov 16 15:46:39 no pain no gain Nov 16 15:46:59 That makes me cry. Nov 16 15:47:07 (Sticker, 512x443) https://irc.ubports.com/0i8sO615.webp Nov 16 15:54:31 @rixchy [I think it will crash.], It did. 🙄 Nov 16 15:56:41 Kernel panic - not syncing: Attempted to kill init! exitcode=0x00000100 Nov 16 15:56:56 full log Nov 16 15:57:06 Alright, please wait. Nov 16 15:57:44 (Document) https://irc.ubports.com/UBjuuQ2i.txt Nov 16 16:06:54 Okay, so how's it look like? Nov 16 16:08:29 im idiot Nov 16 16:08:35 Hm, why? Nov 16 16:08:47 because i don;t really get what can be wrong there Nov 16 16:08:52 it looks like init crashes Nov 16 16:08:55 but why Nov 16 16:09:13 Well, that is true. It is just "calling for device" and then it is just "kernel crash" Nov 16 16:09:30 @Akatsu . [im idiot], But you are not an idiot, since I cannot spot an error there too.' Nov 16 16:10:02 @Akatsu . [], I just know that it happened after enabling these options. So maybe something with CGROUP or with LXC starting? Nov 16 16:11:28 For now, I am trying to install Plasma instead of Reference rootfs do test if it's okay. Do you think it would be a good idea, Akatsu ? Nov 16 16:12:02 [Edit] For now, I am trying to install Plasma instead of Reference rootfs to test if maybe it is just the reference rootfs. Do you think it would be a good idea, Akatsu ? Nov 16 16:12:12 i think its more related to kernel itself Nov 16 16:12:25 Oh, okay. Nov 16 16:12:26 i feel like it crashes on initramfs bundled with hybris-boot Nov 16 16:12:30 but you can try Nov 16 16:12:38 Okay. Nov 16 16:18:07 @rixchy ["failed to create directory '/sys/fs/cgroup/net_cls//lxc/android-999'" Could be, …], disable CONFIG_NET_CLS Nov 16 16:57:28 I got LXC working! Nov 16 16:58:30 Congratulations :) Nov 16 16:58:42 Mhm! But the graphics don't work. Nov 16 16:58:53 "library "libgui.so" wasn't loaded and RTLD_NOLOAD prevented it" Any idea? Nov 16 17:00:28 Hm, and test_recorder gave this: "cannot locate symbol "android_create_namespace" referenced by "/system/lib/libgui.so"..." Nov 16 17:00:41 Is libgui.so messed up or some? Nov 16 17:02:26 Edit: It shutdown Nov 16 17:02:32 [Edit] Edit: It shutdown'ED. Nov 16 17:02:42 [Edit] Edit: It shutdown'ED. (The LXC.) Nov 16 17:03:46 I have come to a conclusion: CONFIG_NETPRIO_CGROUP this CONFIG is crashing my system. Nov 16 17:12:18 i suggest you to save all changes Nov 16 17:12:59 I am. Nov 16 18:49:03 "/home/rixchy/halium/kernel/samsung/exynos5420/net/bluetooth/bnep/core.c: In function 'bnep_add_connection': … /home/rixchy/halium/kernel/samsung/exynos5420/net/bluetooth/bnep/core.c:559:2: error: implicit declaration of function 'l2cap_is_socket' [-Werror=implicit-function-declaration]" Nov 16 18:49:45 disabling bnep didnt work like last time? Nov 16 18:50:03 But wouldn't that affect Bluetooth connectivity? Nov 16 18:50:40 BNEP is about doing Ethernet emulation over bluetooth Nov 16 18:50:47 Ah, Nov 16 18:50:55 so like tethering over BT Nov 16 18:51:10 Hm, okay. Nov 16 18:51:19 (Photo, 425x858) https://irc.ubports.com/tJWAxaKl.png Nov 16 18:51:37 Hm, okay. Nov 16 18:56:55 (Document) https://irc.ubports.com/kLubAtX8.txt Nov 16 18:56:57 Okay, flashed it. Device's obviously rebooting. Here's last_kmsg. Nov 16 19:22:26 This means 110% that LXC is the culprit of kernel panic? "[ 17.412160] [B0: lxc-start: 959] Unable to handle kernel NULL pointer dereference at virtual address 00000000" Nov 16 19:23:51 @rixchy [This means 110% that LXC is the culprit of kernel panic? "[ 17.412160] [B0: …], Try to disable services in the init.devicename.rc Nov 16 19:24:06 Or make it empty. Then you see if lxc is still crashing Nov 16 19:24:14 @Flohack [Try to disable services in the init.devicename.rc], Where is that init.devicename.rc? Nov 16 19:33:39 inside device tree Nov 16 19:33:44 rootdir catalog Nov 16 19:34:02 @Akatsu . [rootdir catalog], I found it in ramdisk folder. Nov 16 19:34:18 its built as part of system.img Nov 16 19:34:28 so you need to rebuild systemimage Nov 16 19:34:46 Okay, but I found the .rc file in the ramdisk folder. Nov 16 19:37:39 ah its different device tree Nov 16 19:38:01 on qualcomms its inside rootdir, for you its indeed ramdisk one Nov 16 19:38:12 Okay, but I have .rc files. Nov 16 19:38:15 Shall I post them? Nov 16 19:38:27 I am not sure which one to edit + and what to edit. Nov 16 19:41:47 Here are the files. Nov 16 19:41:48 (Document) https://irc.ubports.com/cES4rm7e.txt Nov 16 19:41:48 (Document) https://irc.ubports.com/MCTa3Dch.txt Nov 16 19:55:27 @rixchy [Okay, but I found the .rc file in the ramdisk folder.], Check if you have something like this … https://github.com/halium-gemini/android_device_xiaomi_msm8996-common/blob/f6a36123a6924bd78d2b2a08fc959bd733b7afdd/rootdir/etc/init.qcom.rc#L681 … service "blablabla" etc... Nov 16 19:55:48 @peat_psuwit [Have anyone ever seen an Unable to handle kernel NULL pointer dereference at vir …], Was this issue fixed? Because I am having the same one. Nov 16 20:38:16 Okay, so update - I installed PM rootfs, and the device isn't rebooting anymore. Nov 16 20:38:29 NVM. Nov 16 20:38:37 Forget it. Nov 16 20:39:06 I got rekt. Nov 16 20:40:05 Same error. "[ lxc-start: ] Unable to handle kernel NULL pointer dereference at virtual address 00000000 Nov 16 20:45:10 @rixchy [Was this issue fixed? Because I am having the same one.], The issue is due to incomplete implementation of pidns. The is per-kernel-tree. Nov 16 20:45:21 Take a look at FP2's tree to see how I fixed it: https://github.com/Halium/android_kernel_fairphone_msm8974/commits/halium-7.1 Nov 16 20:45:49 The patch is applied, I made sure. At this point, I am lost. But I know that one specific CONFIG causes this. Nov 16 20:46:16 Starts from 'Revert "Finish implementation of PID namespace"' to 'userns: Require CAP_SYS_ADMIN for most uses of setns.' Nov 16 20:46:16 But I will take look at your commits. Nov 16 20:46:28 @peat_psuwit [Starts from 'Revert "Finish implementation of PID namespace"' to 'userns: Requir …], Okay, thanks. I will check it out now. Nov 16 20:46:53 @rixchy [I have come to a conclusion: CONFIG_NETPRIO_CGROUP this CONFIG is crashing my sy …], Here's the config that is crashing the kernel if set to n. Nov 16 20:49:02 I'm trying to update my kernel from 3.11. I've rebased it onto 3.16 and everything is compiling well so far. However, when it comes time to merge the dtbs, it says it has "found 0 unique dtb(s)" … At this point it fails. I've tried changing the dtb tool to the qualcomm one, but I get the same issue. Any ideas? It discovers the dtbs fine Nov 16 20:49:02 on 3.11 Nov 16 21:19:44 @Qiangong2 what do you mean under merging dtbs? Nov 16 21:19:58 it is not common process for upstream kernel Nov 16 21:21:08 When it creates the dzimage, it requires merging 3 dtbs. It always fails at finding the unique dtbs required Nov 16 21:21:20 Even though they are there Nov 16 21:22:22 seems to be Tizen-specific, unfortunately, so no idea, MediaTek and recent Qualcomms on Android just make Image.gz-dtb with dtbs literally appended to the end of gzipped kernel image Nov 16 21:25:33 Great. Okay, thanks Nov 16 21:39:34 Hm, I got phablet using PM rootfs, but I cannot login as root, is this normal? Nov 16 21:40:29 Good job, @peat_psuwit Thank you for the patches! The LXC container starts successfully, without failing and rebooting now! Nov 16 22:52:31 SO, I made some experiments/conclusions. Nov 16 22:53:36 I ran this: "EGL_PLATFORM=hwcompose strace test_hwcomposer" To see the "verbose" output, and it showed many missing files expected in `/vendor/lib/` Nov 16 22:53:47 (Document) https://irc.ubports.com/2usl1khS.txt Nov 16 22:55:47 Here's the list, but if I go to another vendor tree and into the lib folder, they don't have these files into there! So, I think that the test is expecting the .so files to be in the `/vendor/lib/` folder, while they clearly aren't. So what would I do, is try to find and copy those files to the `/vendor/lib/` folder, to see the graphics wo Nov 16 22:55:47 uld start working, (and also other tests.) Nov 16 23:04:22 [Edit] Here's the list, but if I go to another vendor tree and into the lib folder, they don't have these files into there! So, I think that the test_hwcomposer (or any other test,) is expecting the .so files to be in the `/vendor/lib/` folder, while they clearly aren't. So what would I do, is try to find and copy those files to the `/vend Nov 16 23:04:22 or/lib/` folder, to see the graphics would start working, (and also other tests.) Nov 16 23:47:00 [ 54.331954] 54.331954 / 01-01 07:29:15.389] type=1400 audit(26955.389:69): apparmor="DENIED" operation="open" parent=4249 profile="/usr/lib/telepathy/telepathy-ofono" name="/dev/tty" pid=4256 comm="telepathy-ofono" requested_mask="r" denied_mask="r" fsuid=32011 ouid=0 Nov 16 23:47:12 what can I do? Nov 16 23:48:28 Any idea how to fix: "cannot locate symbol "__popcountsi2" referenced by "/system/vendor/lib/egl/libGLES_mali.so"..."? Nov 16 23:48:40 [Edit] Any idea how to fix: "`cannot locate symbol "__popcountsi2" referenced by "/system/vendor/lib/egl/libGLES_mali.so"...`"? Nov 16 23:55:15 grep __popcountsi2 /system in original LineageOS partition Nov 16 23:57:01 @NotKit [grep __popcountsi2 /system in original LineageOS partition], ? Nov 16 23:58:36 I mean find which file it is defined in original LineageOS build Nov 16 23:58:45 you seem to be missing either stubs or patches Nov 16 23:59:28 @NotKit [grep __popcountsi2 /system in original LineageOS partition], The point which I didn't understand is, "/system in original LineageOS partition." Nov 17 00:00:17 I misphrased, I meant real system partition with LineageOS installed as opposed to Halium system.img Nov 17 00:00:55 So, you are saying I gotta sideload LOS now on the device? Will it delete the rootfs and system.img? Nov 17 00:01:13 you can do it on PC as well if you extract the image Nov 17 00:01:21 Ah! Nov 17 00:01:30 Okay, will be back in a bit! Nov 17 00:01:46 or link me to your device tree just in case I see something obvious Nov 17 00:02:11 @NotKit [or link me to your device tree just in case I see something obvious], https://github.com/LineageOS/android_device_samsung_chagalllte Nov 17 00:02:18 Here, check it out. Nov 17 00:02:23 Do your magic. :) Nov 17 00:03:08 https://github.com/LineageOS/android_device_samsung_exynos5420-common/tree/cm-14.1/libpopcountsi2 Nov 17 00:03:25 https://github.com/LineageOS/android_device_samsung_exynos5420-common/commit/ed6357185e65fe8280f6ae98d335e5a786e7b819 Nov 17 00:03:38 Okay, when I apply the changes, system should be rebuilt? Nov 17 00:04:01 you have those changes already Nov 17 00:04:12 Wha- Nov 17 00:04:13 the problem is that our build of libhybris ignores shims... Nov 17 00:04:41 Okay. Nov 17 00:04:42 https://github.com/libhybris/libhybris/pull/377 Nov 17 00:04:46 this was never merged Nov 17 00:05:30 if you want a quick fix, I would probably copy __popcountsi2 somewhere into bionic and rebuild system Nov 17 00:05:51 @NotKit [if you want a quick fix, I would probably copy __popcountsi2 somewhere into bion …], What's bionic? Where is the bionic folder? Nov 17 00:06:45 bionic is the implementation of libc that ships on android based devices Nov 17 00:07:23 Okay, but how do I understand a phrase "copy somewhere into bionic." Nov 17 00:08:13 https://github.com/Halium/android_bionic/blob/cm-14.1/libc/bionic/ndk_cruft.cpp Nov 17 00:08:44 try to copy the whole int __popcountsi2(int a) { ... } implementation into that file before } // extern "C" Nov 17 00:09:02 then make libc and rebuild systemimage Nov 17 00:10:32 And how do you make libc? Nov 17 00:10:44 "make libc" is the command :) Nov 17 00:10:52 Okay, thanks. :) Nov 17 00:11:16 Another question, where is that file located? Nov 17 00:11:20 @NotKit [https://github.com/Halium/android_bionic/blob/cm-14.1/libc/bionic/ndk_cruft.cpp], This? Nov 17 00:11:44 in Halium root/bionic/... Nov 17 00:11:52 Okay, thanks. Nov 17 00:13:39 @NotKit ["make libc" is the command :)], Am I doing something wrong? Or I am just too dumb? "rixchy@toasterpc:~/halium/halium-scripts$ make libc … make: *** No rule to make target 'libc'. Stop. … #### make failed to build some targets ####" Nov 17 00:15:27 is halium-scripts Halium root? Nov 17 00:15:46 -_- As I said, I am dumb. Nov 17 00:16:18 you need the normal "source build/envsetup.sh" and "lunch" as when you make systemimage Nov 17 00:16:28 @NotKit [you need the normal "source build/envsetup.sh" and "lunch" as when you make syst …], Yeah, I ran it. Nov 17 00:22:55 "cannot locate symbol "__popcountsi2" referenced by "/system/vendor/lib/egl/libGLES_mali.so"..." Nov 17 00:22:59 :/ Nov 17 00:23:27 you rebuilt and replaced the system already? Nov 17 00:23:32 Yep. Nov 17 00:23:35 It was fast. Nov 17 00:24:04 "grep __popcountsi2" -r out/target/product on host in Halium root Nov 17 00:24:34 [Edit] "grep __popcountsi2 -r out/target/product" on host in Halium root Nov 17 00:24:43 "rixchy@toasterpc:~/halium$ grep __popcountsi2" -r out/target/product … > []" Nov 17 00:24:48 A blinking cursor. Nov 17 00:24:55 made a typo with " Nov 17 00:24:58 [Edit] "rixchy@toasterpc:~/halium$ grep __popcountsi2" -r out/target/product … > " Nov 17 00:25:04 so Ctrl+C Nov 17 00:26:10 Ok, it is done. Nov 17 00:26:42 (Document) https://irc.ubports.com/0NR4HnTC.txt Nov 17 00:28:26 so the symbol got compiled just fine Nov 17 00:28:38 Okay..? Nov 17 00:28:55 what about grep -r __popcountsi2 /system on device? Nov 17 00:29:14 Through SSH? Nov 17 00:29:17 yes Nov 17 00:30:32 @NotKit [what about grep -r __popcountsi2 /system on device?], "Binary file /system/lib/libc.so matches … Binary file /system/lib/libpopcountsi2.so matches … Binary file /system/vendor/lib/egl/libGLES_mali.so matches … Binary file /system/vendor/lib/liblvverx.so matches … Binary file /system/vendor/lib/liblvvetx.so matches" Nov 17 00:33:11 It's ok? @NotKit Nov 17 00:33:29 yes, but I'm trying to guess why it is still not found if libc.so exports it Nov 17 00:33:39 Okay. Nov 17 00:33:45 Sorry, to interrupt you. Nov 17 00:35:00 could you send the libc.so? Nov 17 00:35:07 Location? Nov 17 00:35:40 or out/target/$device/system/lib/libc.so Nov 17 00:36:15 (Document) https://irc.ubports.com/WQ2Mj5xX.null Nov 17 00:36:16 Here. Nov 17 00:43:13 (Document) https://irc.ubports.com/6gYwNJZk.null Nov 17 00:43:36 @NotKit [], So, what am I lookin' at? Nov 17 00:43:58 apply that change to your bionic/libc files Nov 17 00:44:36 [Edit] apply that change to your bionic/libc files, then compile libc and resend it Nov 17 00:46:23 Okay, do I resend the libc.so again? Nov 17 00:47:03 yes Nov 17 00:48:08 "prebuilts/gcc/linux-x86/arm/arm-linux-androideabi-4.9/arm-linux-androideabi/bin/ld: warning: using 'LIBC' as version for '__popcountsi2' which is also named in version 'LIBC_PRIVATE' in script … prebuilts/gcc/linux-x86/arm/arm-linux-androideabi-4.9/arm-linux-androideabi/bin/ld: error: version script assignment of LIBC_PRIVATE to symbol __ Nov 17 00:48:08 popcountsi2 failed: symbol not defined … clang++: error: linker command failed with exit code 1 (use -v to see invocation)" Nov 17 00:48:17 Sorry for this long, fat stack of text. Nov 17 00:48:33 @rixchy ["prebuilts/gcc/linux-x86/arm/arm-linux-androideabi-4.9/arm-linux-androideabi/bin …], "make libc" fail. Nov 17 00:49:41 oh, revert the changes you made to bionic, the function was there already, but made private by default Nov 17 00:49:49 Okay. Nov 17 00:50:49 Okay, it compiled (after reverting, OFC.) Nov 17 00:50:56 Anything else to do? Nov 17 00:51:15 (Document) https://irc.ubports.com/3Lz0hATm.null Nov 17 00:51:26 and remove your additions to ndk_cruft as well Nov 17 00:51:33 Okay. Nov 17 00:51:45 then recompile libc and send it Nov 17 00:55:17 Okay, it compiled. I applied the patches, here's the libc.so Nov 17 00:55:26 (Document) https://irc.ubports.com/ffxKY8u8.null Nov 17 00:56:07 arm-eabi-objdump -T libc.so | grep __popcountsi2 … 00072f30 g DF .text 00000044 LIBC __popcountsi2 Nov 17 00:56:10 looks fine now Nov 17 00:56:20 Okay, rebuild systemimage? Nov 17 00:56:26 I will go to sleep, but rebuild systemimage and it may either work or fail somewhere else Nov 17 00:56:39 test_egl_configs is a bit more useful test just to check if libs can be loaded properly Nov 17 00:56:40 @NotKit [I will go to sleep, but rebuild systemimage and it may either work or fail somew …], Okay, good night. Thank you for your help. :) Nov 17 00:56:48 @NotKit [test_egl_configs is a bit more useful test just to check if libs can be loaded p …], Okay. :) Nov 17 01:05:15 (Photo, 750x1000) https://irc.ubports.com/COeUV4pz.png Nov 17 01:05:28 YALL WE HAVE A SUCCESSFUL PORT BY A 12 YEAR OLD. Nov 17 01:06:28 [Edit] YALL WE HAVE A SUCCESSFUL PORT BY A 12 YEAR OLD. Well, almost. Nov 17 01:07:13 Should I try installing PM on it?? Omigosh. Nov 17 01:09:02 @NotKit That wad the final piece to get graphics working! Nov 17 01:09:16 [Edit] @NotKit That was the final piece to get graphics working! Nov 17 01:09:47 TBH, I don't really care that almost the other things don't work, I am just glad and freaked out because the graphics are working. LOL. Nov 17 01:10:20 getting the graphics is just the beginning of journey though :), but cool! Nov 17 01:10:45 @NotKit [getting the graphics is just the beginning of journey though :), but cool!], Thank you! **** ENDING LOGGING AT Sun Nov 17 02:59:57 2019