2016-04-03 12:36:00	@kido	damn... I tried your img_info and it produces the right offsets
2016-04-03 12:57:07	@kido	locusf: maybe you should contact that guy to ask him how he repacked his boot.img http://forum.xda-developers.com/zenwatch/development/kernel-lightly-modified-zenwatch-kernel-t3347924 there shouldn't be anything magic in there... from what I see mkboot with your config produces a correct layout
2016-04-03 12:58:24	@kido	it turned out that the "tinykernel" guy made no change to the kernel... http://forum.xda-developers.com/zenwatch/development/tinykernel-t3169754 "As I don't have a bootable modified kernel, I just made some changes to the ramdisk."
2016-04-03 16:03:00	locusf	hah
2016-04-03 16:26:30	locusf	kido: I wonder whats wrong then
2016-04-03 16:28:14	@kido	locusf: do you have a little bit of time ahead of you right now? if you do we can try a bunch of settings until we finaly get it working
2016-04-03 16:28:37	@kido	I'm sure it's not much than that, we have to find a way
2016-04-03 16:28:45	@kido	much more*
2016-04-03 16:28:47	locusf	okay I'll just finish my coffee :)
2016-04-03 16:29:06	@kido	great, I prepare my tea :)
2016-04-03 16:33:49	locusf	ok lets go
2016-04-03 16:35:10	@kido	ok so basicaly I'm pretty confident to say that the way we currently pack your boot.img is wrong so there are several possible point of failures and we should proceed incrementaly to isolate the "root of all evil"
2016-04-03 16:36:13	@kido	start by grabbing a working default android boot.img
2016-04-03 16:37:18	@kido	make sure you can boot it with "fastboot boot boot.img" if that works we will just unpack it and repack it, there are two ways because there are two tools, if you use "unmkbootimg" you get a bunch of warning and info but I wasn't able to repack it correctly with mkbootimg afterward
2016-04-03 16:37:37	@kido	if you use "mkboot" which has an unpack and repack option you might have more chance
2016-04-03 16:38:22	locusf	okay
2016-04-03 16:38:50	@kido	it is available in your asteroid/build/tmp-glibc/sysroots/x86_64-linux/usr/bin/mkboot
2016-04-03 16:39:59	@kido	if you use "mkboot default-boot.img extract-directory/" it will provide you with a img_info file in the extract-directory
2016-04-03 16:40:19	@kido	this file should be similar to the one you currently use in OE but well...
2016-04-03 16:40:59	@kido	to make sure we can repack the same boot.img with the same tools and same config just swap the two arguments of mkboot like that: "mkboot extract-directory new-boot.img"
2016-04-03 16:41:07	@kido	and try to fastboot boot this new-boot.img
2016-04-03 16:41:56	locusf	I pushed the current work to my repository
2016-04-03 16:42:11	@kido	theoreticaly it should boot, I tried to do that myself and it seemed that the file layout was the same, a hexdiff showed me a few different bytes but let's hope they are not the cause of an issue
2016-04-03 16:42:16	@kido	ok
2016-04-03 16:57:46	locusf	could my initial naming of athias -> anthias be an issue?
2016-04-03 16:59:19	@kido	we will soon see
2016-04-03 16:59:33	@kido	does the repacked kernel with the previous instructions work?
2016-04-03 17:02:51	locusf	booting, will see
2016-04-03 17:03:16	locusf	ejected partitions
2016-04-03 17:04:18	@kido	so the default boot.img works but once the exact same files are repacked with mkboot it enters this strange qualcom mode which expose partitions to the host via usb?
2016-04-03 17:05:23	@kido	so the problem is mkboot
2016-04-03 17:10:24	locusf	yes
2016-04-03 17:10:36	@kido	let's try with another tool then
2016-04-03 17:10:43	@kido	bitbake abootimg-native
2016-04-03 17:11:03	@kido	and then use asteroid/build/tmp-glibc/sysroots/x86_64-linux/usr/bin/abootimg -x default-boot.img
2016-04-03 17:12:37	@kido	and fastboot boot new.img
2016-04-03 17:13:04	@kido	on G Watch it didn't work because abootimg wasn't supporting the devicetree, you might have the same problem but anyway we can try
2016-04-03 17:15:32	locusf	okay trying that after hanging the laundry
2016-04-03 17:15:39	@kido	ok
2016-04-03 17:37:38	locusf	do I use the asteroid ramdisk?
2016-04-03 17:41:31	@kido	no
2016-04-03 17:41:45	@kido	we are just trying to find a working tool
2016-04-03 17:43:21	locusf	ok
2016-04-03 17:45:44	locusf	that worked
2016-04-03 17:48:57	@kido	with abootimg? :D
2016-04-03 17:49:21	locusf	ASTERROID BOOT LOG o/
2016-04-03 17:49:30	@kido	hell yeah! :D
2016-04-03 17:49:55	locusf	but nothing after :p
2016-04-03 17:50:04	@kido	what kernel do you use? where do you get those logs? what happens?
2016-04-03 17:50:13	locusf	boot logo
2016-04-03 17:50:17	@kido	ah logo!
2016-04-03 17:50:19	locusf	the selfbuilt kernel
2016-04-03 17:50:21	@kido	then you are on the rootfs
2016-04-03 17:50:23	locusf	and no logs
2016-04-03 17:50:24	locusf	ok
2016-04-03 17:50:31	@kido	do you have adb shell?
2016-04-03 17:50:41	locusf	nope
2016-04-03 17:50:47	locusf	null device
2016-04-03 17:50:57	@kido	ok, so we are getting further :)
2016-04-03 17:51:28	@kido	first thing first, we have to change your linux-anthias recipe to use abootimg instead of mkboot
2016-04-03 17:51:51	@kido	the meta-smartphone layer from SHR provided the tools to use abootimg but I had not to use it in dory
2016-04-03 17:52:11	locusf	okay
2016-04-03 17:53:35	@kido	you can get inspired by that https://github.com/shr-distribution/meta-smartphone/blob/shr/meta-samsung/recipes-kernel/linux/linux-samsung-tuna_git.bb#L11 
2016-04-03 17:54:03	@kido	inherit kernel_android instead of boot-img
2016-04-03 17:54:14	locusf	it didn't have enable_adb :p
2016-04-03 17:54:23	@kido	enable_adb is only used in the ramdisk
2016-04-03 17:54:26	@kido	to stop the boot
2016-04-03 17:54:44	@kido	adb is enabled in any way in the rootfs
2016-04-03 17:57:54	locusf	selinux=0 causes the partition throwup
2016-04-03 17:58:16	@kido	mhh
2016-04-03 17:58:28	@kido	ok
2016-04-03 18:04:14	locusf	if I'd just get developer mode :p
2016-04-03 18:05:19	@kido	this will come... :)
2016-04-03 18:05:49	locusf	does the first boot take long?
2016-04-03 18:06:01	locusf	also there is something odd in init.rc
2016-04-03 18:06:46	@kido	nop it doesn't take long
2016-04-03 18:06:51	@kido	and what is odd in init.rc ?
2016-04-03 18:07:04	locusf	it downclocks the cores
2016-04-03 18:07:52	@kido	yep those settings come from the default android wear init.rc from dory
2016-04-03 18:08:36	@kido	you might want to remove them though, I saw on xda that someone said that if he downclocked the CPU of zenwatch then he wasn't able to boot (on android)
2016-04-03 18:08:56	@kido	I'm currently reading the android wear default init.rc on anthias
2016-04-03 18:10:59	locusf	ok
2016-04-03 18:14:17	locusf	hmm
2016-04-03 18:14:20	locusf	resolution is wrong
2016-04-03 18:14:48	@kido	in psplash?
2016-04-03 18:14:52	locusf	no
2016-04-03 18:14:55	locusf	in machine.conf
2016-04-03 18:15:06	@kido	in fact the resolution here is not used
2016-04-03 18:15:13	@kido	you can remove the values
2016-04-03 18:15:19	locusf	oh
2016-04-03 18:15:28	@kido	there seems to be something to do with "/sys/devices/platform/msm_hsusb/gadget/wakeup"
2016-04-03 18:20:31	locusf	so whats next
2016-04-03 18:22:06	@kido	we gotta find why usb is not working, I believe that our best friend here will be adb shell inside android and tweaking the usb related values inside /sys/
2016-04-03 18:22:57	@kido	the /init.asus.usb.rc and /init.asus.usb.sh will also be useful to read
2016-04-03 18:24:03	@kido	then you will have to modify your src/meta-anthias-hybris/recipes/core/initrdscripts/initramfs-boot-android/init.sh
2016-04-03 18:24:43	@kido	to make usb work correctly
2016-04-03 18:25:28	@kido	so the current goal is to get a shell
2016-04-03 18:26:03	@kido	once you will have a shell you will be able to get back to the standard things you know from sfos porting... :)
2016-04-03 18:26:14	locusf	yes
2016-04-03 18:28:20	locusf	I can't get root anymore :/
2016-04-03 18:28:22	locusf	on android
2016-04-03 18:28:51	@kido	su is not working?
2016-04-03 18:28:57	locusf	nope
2016-04-03 18:28:58	locusf	it hangs
2016-04-03 18:29:11	@kido	mhhh, then try to root once again, but asteroid should have nothing to do with that
2016-04-03 18:29:20	locusf	yeah ok
2016-04-03 18:29:24	@kido	strange
2016-04-03 18:30:12	@kido	ahah I'm sorry to have told you that porting asteroid to zenwatch would be easy peasy, since it is the same SoC as the currently working devices I thought you would have less issues
2016-04-03 18:32:02	locusf	heheh :)
2016-04-03 18:33:41	locusf	chown -h root.system /sys/devices/platform/msm_hsusb/gadget/wakeup
2016-04-03 18:33:41	locusf	chmod -h 220 /sys/devices/platform/msm_hsusb/gadget/wakeup
2016-04-03 18:33:50	locusf	that looks ... odd
2016-04-03 18:34:06	locusf	it is others write only by default
2016-04-03 18:34:49	@kido	yep chown and chmod shouldn't make usb work but maybe you would have to echo 1 > to the file or something in the directory
2016-04-03 18:36:48	locusf	ok
2016-04-03 18:36:51	@kido	locusf: ok, before evrything else try to add CONFIG_USB_ANDROID to your defconfig
2016-04-03 18:36:57	@kido	I'm not sure why it isn't in check-config
2016-04-03 18:37:13	@kido	I think this is mandatory
2016-04-03 18:37:51	locusf	okay
2016-04-03 18:38:01	locusf	kernel recompiling doesn't take too long
2016-04-03 18:38:35	@kido	mhh actualy it seems that it was CONFIG_USB_G_ANDROID which is already enabled
2016-04-03 18:41:19	@kido	there might also be a problem with OE trying to "simplify" the defconfig and sometime getting rid of important CONFIG_ values (yes, I know, this is really shitty)
2016-04-03 18:41:32	Nokius	locusf: yeah!
2016-04-03 18:41:38	@kido	so try to grep CONFIG_USB_G_ANDROID asteroid/build/tmp-glibc/work/athias-oe-linux-gnueabi/linux-athias/3.10+lollipop-r0/git/.config
2016-04-03 18:42:44	locusf	nothing found
2016-04-03 18:43:19	@kido	ok then this is the source of your problem
2016-04-03 18:43:22	@kido	I already met that
2016-04-03 18:43:41	@kido	I know this is hard to believe but your defconfig is actually missing... comments
2016-04-03 18:44:00	@kido	so make a backup of your current asteroid/build/tmp-glibc/work/athias-oe-linux-gnueabi/linux-athias/3.10
2016-04-03 18:44:07	@kido	+lollipop-r0/git/.config
2016-04-03 18:44:29	@kido	then add CONFIG_USB_G_ANDROID inside the .config
2016-04-03 18:45:06	@kido	run asteroid/build/tmp-glibc/work/athias-oe-linux-gnueabi/linux-athias/3.10+lollipop-r0/temp/run.do_compile and it should generate you a new .config that should differe from your backup from two or three comments (use diff here)
2016-04-03 18:45:42	@kido	those comments are important so add them inside the defconfig of your meta-athias-hybris
2016-04-03 18:47:34	locusf	gotta fee my cat :)
2016-04-03 18:47:35	locusf	brb
2016-04-03 18:48:22	@kido	:)
2016-04-03 18:54:31	locusf	nothing ..
2016-04-03 18:54:32	locusf	on boot
2016-04-03 18:56:54	@kido	?
2016-04-03 18:57:27	locusf	with new config
2016-04-03 18:57:29	locusf	and image
2016-04-03 18:57:35	locusf	I see the asteroid bootlogo
2016-04-03 18:57:39	locusf	but nothing appears in dmesg
2016-04-03 18:59:06	@kido	did you use bitbake or run.compile to rebuild the kernel?
2016-04-03 18:59:33	locusf	run.compile
2016-04-03 19:00:30	@kido	and where did you get your boot.img from?
2016-04-03 19:00:34	@kido	or kernel
2016-04-03 19:01:24	@kido	run.compile doesn't install the kernel in build/deploy, it only brings you a new arch/arm/boot/zImage-dtb
2016-04-03 19:01:53	locusf	ah
2016-04-03 19:01:57	locusf	that explains it
2016-04-03 19:11:46	locusf	partition throwup wtf ...
2016-04-03 19:12:22	@kido	:|
2016-04-03 19:34:15	locusf	W T F
2016-04-03 19:36:24	@kido	what's going on?
2016-04-03 19:37:17	locusf	still barfs partitions
2016-04-03 19:37:18	locusf	wtf
2016-04-03 19:38:40	locusf	even though I rolled back the config
2016-04-03 19:40:52	@kido	this is strange
2016-04-03 19:41:08	@kido	at least we now know that it *can* work
2016-04-03 19:42:24	locusf	yes
2016-04-03 19:48:30	locusf	it was the .sh commands in usb initialization
2016-04-03 19:48:39	locusf	which did the barfing
2016-04-03 19:49:12	@kido	you mean the chmod and chown on /sys/devices/platform/msm_hsusb/gadget/wakeup ?
2016-04-03 19:49:17	locusf	yeah
2016-04-03 19:49:22	locusf	also the echo 1
2016-04-03 19:49:32	@kido	ok
2016-04-03 19:49:50	@kido	so do you have a kernel with CONFIG_USB_G_ANDROID running?
2016-04-03 19:50:49	locusf	nope
2016-04-03 20:08:35	locusf	my arse and back started to hurt, will continue sometime later
2016-04-03 20:08:53	locusf	thanks for the awesome help so far kido :)
2016-04-03 20:09:15	@kido	alright, you're welcome :)