**** BEGIN LOGGING AT Sun Jan 28 03:00:03 2018 Jan 28 03:04:53 well, looks like my N5 is a big android brick. maybe i should put webos on it to see if i can make it come up :| Jan 28 03:24:32 can't even get it to boot with the factory images :| Jan 28 07:15:47 Morning! Jan 28 08:28:14 Morning! Jan 28 08:28:50 Herrie: gosh, it's far too early for a Sunday ! Jan 28 08:35:30 Herrie: on N5 the graphics work, but I don't see adb Jan 28 09:00:33 Tofe: Same here Jan 28 09:00:38 Must be the android-tools upgrade Jan 28 09:04:53 yup Jan 28 09:33:53 Tofe: It might be as easy as adding the 2-4 lines from initramfs to android-gadget-setup script Jan 28 09:34:05 I'll give it a try Jan 28 09:37:44 Morning Jan 28 09:38:29 Herrie: functionfs is only active on mido, right? Jan 28 09:39:20 https://github.com/Halium/android_kernel_lge_hammerhead/blob/533a628f03d4d0be2f1c4e969aa1c6d7a7e564bb/arch/arm/configs/cyanogenmod_hammerhead_defconfig#L2765 seems so Jan 28 09:40:49 Tofe: It should work on others Jan 28 09:41:04 It worked in initramfs on mako, hammerhead when I tested Jan 28 09:41:11 Without defconfig changes Jan 28 09:41:44 ah ok good Jan 28 10:03:49 Herrie: mmh didn't work Jan 28 10:20:06 Tofe: Hmmz I can have a look later today could be we got the .service replaced somehow and it doesn't get started at right time by systemd? Jan 28 10:20:59 Well, the only dependency should be /system, and that's already in our rootfs now Jan 28 10:21:57 The only difference I see is that in initramfs you mount the Android /system, you don't take the one from our rootfs (i.e. Halium's) Jan 28 10:22:29 But build.prop should be quite the same, and I don't see what else would be needed Jan 28 10:24:09 Tofe: Yeah I only moun the system for build.prop anyway to pull the required info Jan 28 10:24:36 There was nothing mounted before so it would never return any inffo Jan 28 10:25:15 yup; and I checked, the current setup on N5 does set the IDs correctly already Jan 28 10:25:26 so it's not an issue of accessing build.prop Jan 28 10:29:29 Tofe: You tried adding the mount command for functionfs to android-gadget-setup? Jan 28 10:29:37 yes Jan 28 10:29:45 Hmmz Jan 28 10:29:59 https://paste2.org/68ADmFC7 Jan 28 10:30:15 Only other thing I can quickly think of is the systemd .service that's started at wrong time Jan 28 10:30:20 Anything in log? Jan 28 10:30:27 unfortunately not Jan 28 10:30:48 but it's a lead Jan 28 10:31:26 for instance, when I try to manually do "echo ffs > /sys/class/android_usb/android0/functions" it says resource busy; maybe by adbd, but maybe by the android container Jan 28 10:31:30 Alternatively you could check AsteroidOS to see what they do Jan 28 10:31:40 yes Jan 28 10:32:11 Since they use the same android-tools Jan 28 10:33:57 https://github.com/AsteroidOS/meta-asteroid/blob/master/recipes-devtools/android-tools/android-tools/0001-Allow-adbd-to-be-ran-as-root.patch ? Jan 28 10:35:00 Yeah I saw that one before... Adbd in initramfs ran without so thought we didn't need it Jan 28 10:35:04 Might be worth a try Jan 28 10:50:39 yup, that helped ! Jan 28 10:51:09 I guess ro.secure was "0" in your android install Jan 28 10:51:27 Whereas it's set to "1" in Halium Jan 28 10:53:21 Tofe: ah ok Jan 28 11:00:22 I'll try to push that soon, so that you can have adb to debug mido Jan 28 11:31:31 https://github.com/webOS-ports/meta-webos-ports/pull/274 tested on N5 Jan 28 11:42:00 Tofe: Thnx will try that later today, some guests now :P Jan 28 11:46:45 Tofe: seems we need to adjust android-gadget-setup to recover mtp support as well Jan 28 11:49:27 I do get mtp on N5, at least Jan 28 11:49:43 if you want you can set "mtp,ffs" instead of just "ffs", I think Jan 28 11:53:35 ok Jan 28 15:03:46 Tofe: i just wanted to note that we had mtp enabled previously: https://github.com/shr-distribution/meta-smartphone/commit/e6f2ad9efe830f71be990feda9e0cfa559e6aff1#diff-416e00a9330f50fa8e6c8f467e41076d Jan 28 16:26:46 nizovn: fair point Jan 28 16:27:02 Does the idProduct matter for something? Jan 28 16:29:28 yes, that's how pc can identify that device supports mtp Jan 28 16:29:44 nizovn: ok Jan 28 16:30:37 in general it should be unique for device, i think Jan 28 16:30:55 so touchpad was claiming it's n4 Jan 28 16:38:02 Herrie: I've been trying to catch up where you are with the onyx image Jan 28 16:38:49 Herrie: I think you must have a local build for the onyx root file system and probably some meta-luneos additions Jan 28 16:41:06 My local luneos build is currently stuck trying to download a Halium root file system from build.webos-ports.org Jan 28 17:07:43 elvispre: Ah yes I have something local Jan 28 17:08:04 elvispre: If you have the android image you can simply have it point to it locally Jan 28 17:08:50 Herrie: The only android image I have is a LineageOS one. I guess that would be OK? Jan 28 17:09:25 As opposed to a Halium one. Jan 28 17:09:46 elvispre: No ;) Jan 28 17:09:54 But I guess I could put the onyx one I have online Jan 28 17:10:49 Ah I deleted my local onyx one Jan 28 17:10:53 I can re-do it I guess Jan 28 17:11:03 Is there a recipe? Jan 28 17:11:17 elvispre: No this is the halium bit Jan 28 17:11:40 Which generates a minimal android build which you then link into the build with https://github.com/shr-distribution/meta-smartphone/blob/pyro/meta-oneplus/recipes-core/android-system-image/android-system-image-onyx.bb#L22 Jan 28 17:12:14 I do have an android build environment (is what I meant). Jan 28 17:14:34 elvispre: No you need to run a .sh to generate the required file Jan 28 17:14:56 Herrie: Like https://bpaste.net/show/4d03c8973d52 ? Jan 28 17:15:03 Something like https://bpaste.net/show/a125ab55d6f8 Jan 28 17:15:12 Yes ;) Jan 28 17:15:17 :) Jan 28 17:16:07 That will generate something in the halium-luneos-7.1/results that you can use ;) Jan 28 17:16:33 Thanks. I'll see how much further I get with that. Jan 28 17:21:19 elvispre: I can upload the file once it's ready at my end Jan 28 17:23:53 Herrie: Yes, please. This is a steep learning curve. Jan 28 19:09:03 elvispre: http://build.webos-ports.org/halium-luneos-7.1/halium-luneos-7.1-20180128-1-onyx.tar.bz2 Jan 28 19:10:17 You'd need to update your recipe with this contents: https://bpaste.net/show/b85734e6ac93 Jan 28 19:10:23 And it should work more or less I guess Jan 28 19:17:16 nizovn: yes, we already had mtp, but the addition of functionfs eventually overrides it Jan 28 19:17:33 I still have mtp on my N5, that's fine Jan 28 19:29:55 ok; we maybe need to change code here then: https://github.com/webOS-ports/nyx-modules/blob/webOS-ports/master/src/rockhopper/mass_storage_mode_mtp/msm.c#L194 Jan 28 19:30:54 Herrie: Thanks! Jan 28 19:56:25 nizovn: this code might still be ok, it's just that instead of "mtp,adb" we'll have "mtp,ffs". I didn't test it, but I guess it should work like that Jan 28 19:57:49 ok Jan 28 19:58:02 Herrie: oops, my PR broke tendereloin's build... Jan 28 19:59:53 because, well, as nizovn pointed out, we have a specific android-gadget-setup script for TP, so my patch doesn't work... Jan 28 20:16:30 Ok, now patched for tenderloin, it builds. Jan 28 21:01:50 Tofe: OK great! Jan 28 21:02:06 Any clues as to display issues or still waiting for a working build to debug that? Jan 28 21:58:26 Herrie: memnotify module fails to compile for onyx. Seems to be because CONFIG_MODULES is (was) not defined for onyx in your suggested patch to the Halium kernel. Does that ring a bell? Jan 28 22:42:44 elvispre: Ehm that could be or you could simple remove it from the packagegroup-luneos-extended file Jan 28 22:44:16 Herrie|Pre3: OK. Thanks for the suggestion. Jan 28 23:09:30 Herrie: I'll debug that after re-flash **** ENDING LOGGING AT Mon Jan 29 03:00:01 2018