**** BEGIN LOGGING AT Sat Feb 06 02:59:57 2021 **** BEGIN LOGGING AT Sat Feb 06 07:47:10 2021 Feb 06 08:20:11 Morning! Feb 06 08:20:39 Herrie: ok, glad you could fix it :) Feb 06 08:24:14 Tofe: Well it builds some dtb now, not sure it's 100% correct yet but well Feb 06 08:28:43 Currently I'm still investigating the wifi issue on tissot Feb 06 08:29:05 I think I'll have to reflash lineageos and compare logs Feb 06 08:29:29 it's probably some firmware not being loaded, but I don't see any obvious error Feb 06 08:54:34 4morning Feb 06 09:34:32 JaMa: o/ Feb 06 09:35:20 Herrie: ok I think I understand: the kernel we use today doesn't match the device tree; so I'll try to go back to the 4.x kernel of LineageOS Feb 06 09:48:45 Tofe: ah OK Feb 06 09:48:56 Not sure why adb doesn't work there though Feb 06 09:49:20 I have changes to init script for the configfs/functionfs I can PR Feb 06 09:49:32 So when it panics you can at least do something Feb 06 09:49:36 I'll try to fix that on the way. But it could just be related to the configfs stuff we discussed recently Feb 06 09:50:03 Could be we need newer android-tools but you were stuck there right? Feb 06 09:50:49 I had some difficulties with adb yes, with recent tissot's twrp Feb 06 09:51:02 though my android-tools on archlinux should be recent Feb 06 09:51:20 (and on my ubuntu chroot a bit older) Feb 06 09:51:36 "Android Debug Bridge version 1.0.40" Feb 06 09:52:35 ah, looks like I can upgrade it a bit on arch Feb 06 09:53:18 now "Android Debug Bridge version 1.0.41" we'll see Feb 06 09:56:26 Tofe: Laptop will be switched shortly, so I cleaned up some tools incl. git, but this is the init.sh for Halium I have now: https://paste.ubuntu.com/p/gFZGk9RkrT/ Feb 06 09:56:34 Should be pretty OK already Feb 06 10:45:33 ah, it doesn't build as-is with luneos Feb 06 11:31:08 Tofe: No we need kernel fixes/changes Feb 06 11:32:11 You might want to have a look at this 4.4 kernel: https://github.com/Herrie82/android_kernel_xiaomi_msm8998/commits/halium-9.0-24.1.21-LuneOS Feb 06 11:32:40 This is what I had for Volla: https://github.com/Herrie82/android_kernel_volla_mt6763/commits/halium-9.0-LuneOS Feb 06 12:26:24 Tofe: But I thought we had a 4.9 kernel building already? Feb 06 12:26:29 At least I recall I did somehow Feb 06 12:26:42 But you'd need to Switch Halium to 4.9 as well Feb 06 12:26:47 I.e. device tree Feb 06 12:32:01 on Halium it builds, that's not the problem Feb 06 12:32:18 but here it's some additional warnings introduced by newer gcc Feb 06 12:37:50 I'm pretty sure I built it already Feb 06 12:38:01 Gimme a few minutes to get my manifest Feb 06 12:38:25 I kept 4.9 and 3.18 for both mido & tissot I guess Feb 06 12:45:29 Tofe: https://github.com/Herrie82/android_kernel_xiaomi_msm8953-1/tree/halium-9.0-4.9 Feb 06 12:46:03 ah, I took "halium-9.0" Feb 06 12:46:04 That's one I had Feb 06 12:46:28 as I didn't really know what were the other commits about Feb 06 12:46:34 And there's also a LuciferKernel repo which has 2 branches Feb 06 12:47:04 A halium-9.0 and a halium-9.0-new Feb 06 12:47:19 Checking now here which one builds Feb 06 12:47:28 I want to stick to the same kernel as the device tree is using Feb 06 12:47:31 I kinda remember all 3 did Feb 06 12:47:51 Ok then no Lucifer Feb 06 12:49:15 Anyway, I'm getting there, slowly but surely Feb 06 12:50:16 some of the warnings pointed out a real mistake in the msm code, but for the rest I just disable the errors Feb 06 12:51:44 Msm code is garbage often Feb 06 12:51:51 Especially wlan etc Feb 06 12:52:14 I agree Feb 06 12:53:59 Needed to disable so many errors for Prima WLAN Feb 06 12:54:44 ok, my kernel code compiles now Feb 06 12:59:34 no adb, no UI, but no bootloop, as you had Feb 06 12:59:47 Tofe: OK Feb 06 13:00:04 You saw my updated init right? Feb 06 13:00:19 You also need to update cmdline I think Feb 06 13:01:26 To have androidboot.usbconfigfs=true in there Feb 06 13:01:40 As per BoardConfigCommon.mk Feb 06 13:02:00 let me check Feb 06 13:03:03 right; added that Feb 06 13:03:28 And I noticed that configfs and functionfs seem to bother each other as well sometimes Feb 06 13:04:00 So necessary to disable functionfs in defconfiig, at least I had this with a target recently Feb 06 13:04:12 But it would fail on compilation already Feb 06 13:05:58 See https://github.com/Herrie82/android_kernel_xiaomi_msm8998/commit/fd833e3807140661c9bc998583764b33a19a2bd0 Feb 06 13:06:19 And we might need this too: https://github.com/Herrie82/android_kernel_xiaomi_msm8998/commit/917d5d898766594acef1bffc2f661db06032863b Feb 06 13:07:53 Herrie: in your new init, you don't activate adb the old way anymore ? Feb 06 13:08:33 Tofe: Well this was only for panic mode I think? Feb 06 13:09:35 right Feb 06 13:10:13 I just incorporated some of the upstream Halium bits for this Feb 06 13:10:25 It's nothing too shocking by the looks of it Feb 06 13:10:45 but I don't see any adb activation with the configfs approach Feb 06 13:10:51 just rndis Feb 06 13:11:18 anyway, let's just try Feb 06 13:12:45 Tofe: ah your right Feb 06 13:12:58 That's because Halium doesn't do adb Feb 06 13:13:07 Should have adjusted slightly Feb 06 13:17:25 Herrie: but does it work ? we are starting adbd at the end of "panic" Feb 06 13:18:08 Tofe: Well not sure Feb 06 13:18:15 Adb didn't it seems Feb 06 13:18:33 were you able to connect to the device somehow? Feb 06 13:18:39 But I had someone reporting they got an additional modem :p Feb 06 13:18:49 So I guess it works but not for adb yet Feb 06 13:18:57 Need to do some different values Feb 06 13:22:19 Tofe: https://wiki.tizen.org/USB/Linux_USB_Layers/Configfs_Composite_Gadget/Usage_eq._to_g_ffs.ko Feb 06 13:22:35 That seems to give ideas as for required values for adb Feb 06 13:23:38 And this one: https://github.com/tonyho/adbd-linux/blob/master/adb.sh Feb 06 13:26:12 yes, that's a good lead. But I'm not even sure we're in panic mode here Feb 06 13:26:55 Well you need the skip initramfs patch I pointed out above it seems Feb 06 13:27:21 Thiis one: https://github.com/Herrie82/android_kernel_xiaomi_msm8998/commit/917d5d898766594acef1bffc2f661db06032863b Feb 06 13:29:23 Otherwise it would never even get to our init from what I understood Feb 06 13:30:32 that's waird Feb 06 13:30:34 weird Feb 06 13:31:43 if there is no "skip_initramfs" in cmdline, it shouldn't do that Feb 06 13:33:18 anyway, I'll do some tests Feb 06 13:33:34 it would be nice to have something solid here Feb 06 13:35:02 Both for Volla and Sagit we needed that patch somehow Feb 06 13:35:31 And they are both on 4.X Feb 06 13:36:17 mmmh in the recipe, the cmdline ended with " --" ? Feb 06 13:36:47 Not sure what that does and where it originates from Feb 06 13:37:05 Has something to do with selinux I think but not sure Feb 06 13:37:43 wow, the volla cmdline is quite minimal ! is it a mainllined device, or what Feb 06 13:37:53 https://www.kernel.org/doc/html/v4.14/admin-guide/kernel-parameters.html Feb 06 13:38:03 Tofe: MTK Feb 06 13:38:05 aaah, ok :) Feb 06 13:38:36 well, adding " -- " at the end of cmdline might solve some issues, as the bootloader can add its own, and could create unwanted behavior for us Feb 06 13:38:49 like, skipping initramfs ;) Feb 06 13:39:44 Well everything after -- it passes to init I think based on above Feb 06 13:40:04 but the kernel itself will ignore it Feb 06 13:43:27 Tofe: My understanding that we really need this patch for 4.X but will check cmdline for Volla and Sagit soon Feb 06 13:49:46 Ah ok I understand why it doesn't work on sagit :p Feb 06 13:50:37 I put datapart after -- while it shouldn't I guess although initramfs is where it's needed, so it might actually be correct Feb 06 13:50:56 Sagit has very limited console-ramoops with lots of stuff missing Feb 06 13:51:13 So should change rndis into adb so there are more useful logs Feb 06 14:12:25 Tofe: This might work better with ADB: https://paste.ubuntu.com/p/Dw3WBzrPNj/ Feb 06 14:29:57 Herrie: that's what I tried earlier, but no, I don't get anything Feb 06 14:30:41 I'll try some more basic debug Feb 06 14:31:03 Did you try this patch for initramfs already? Feb 06 14:31:20 not yet Feb 06 14:31:47 I really think you should try that one ;) Feb 06 14:31:55 Since it helped with the 2 other targets on 4.x kernel Feb 06 14:35:31 no need, I can confirm I do run the initrd Feb 06 14:36:56 Ah OK Feb 06 14:36:58 Cool Feb 06 14:37:13 I'm in panic (I forced it with cmdline) and just before the adbd line Feb 06 14:38:32 Will gives options to debug a bit at least Feb 06 14:39:27 "configfs-gadget 7000000.dwc3: failed to start g1: -19" Feb 06 14:41:25 https://patchwork.kernel.org/project/linux-usb/patch/20201103085456.15591-1-jackp@codeaurora.org/ maybe related Feb 06 14:47:31 oh, guests **** ENDING LOGGING AT Sun Feb 07 02:59:58 2021