**** BEGIN LOGGING AT Wed Jun 13 03:00:02 2018
Jun 13 10:11:01 <ant_work> VartiWork, greguu: the images I uploded are ok on qemu but give many alignment faults on real Zaurus
Jun 13 10:18:24 <VartiWork> :(
Jun 13 10:21:47 <ant_work> gcc7.2 seems the culprit
Jun 13 10:21:52 <ant_work> err  7.3
Jun 13 10:22:24 <ant_work> 7.2 was the one fixing las year's similar issue
Jun 13 20:50:28 <greguu[w]> Omegamoon: how are things with the Gemini ? Keen on getting the UART going ?
Jun 13 20:51:51 <Omegamoon> greguu[w], I'm a little reluctant on soldering a serial console on it myself... I know my weaknesses ;)
Jun 13 20:52:16 <Omegamoon> so I'll need to ask someone with the proper skills to get that done
Jun 13 20:52:58 <Omegamoon> but in the meantime I'm trying to get some development going
Jun 13 20:55:08 <Omegamoon> working on an ubuntu image, just to see how I can tweak into something that works
Jun 13 20:57:22 <Omegamoon> I really hate the MediaTek platform really... it is 150% crap... it feels like going back years in time, in the time I started with Rockchip devices
Jun 13 20:59:04 <greguu[w]> hm, yes I am also reluctant to solder atm. I use it as a main driver as my Nexus 5 broke just recently
Jun 13 20:59:25 <greguu[w]> I would however get some effort into a kexecboot kernel, mainline or stock kernel going.
Jun 13 20:59:46 <greguu[w]> once we have that, troubleshooting distros and kernels is much more easy than flashing all the time
Jun 13 21:00:44 <greguu[w]> but for that I would need serial console/uart
Jun 13 21:00:46 <greguu[w]> :/
Jun 13 21:01:10 <Omegamoon> without a serial console proper kernel development is a real pain... it is staring at black screen, not knowing what the issues are
Jun 13 21:01:14 <greguu[w]> maybe better to wait and see if Planet comes up with a 4.4 kernel as promised
Jun 13 21:01:51 <greguu[w]> one of the USB ports should provide also UART but that is not properly documented and tested afaik
Jun 13 21:02:01 <Omegamoon> I've found quite a few 4.4 kernels for a series of MTK SOCs... but not the X25/X27
Jun 13 21:02:31 <greguu[w]> yes, there are none for x2X at this stage.
Jun 13 21:02:52 <greguu[w]> I wonder if the kexec call for aarch64 could be backported to the 3.18 kernel
Jun 13 21:02:56 <Omegamoon> I got the full datasheets, so if USB acts as UART also, that should be documented
Jun 13 21:03:20 <greguu[w]> I rather buy an expensive adapter than solder on the board
Jun 13 21:03:25 <Omegamoon> backporting anything into the 3.18 kernel is a total waste of time to be honest
Jun 13 21:03:28 <greguu[w]> as I need it to be functional
Jun 13 21:03:52 <greguu[w]> if it would give us a kexecboot kernel, it would help in development
Jun 13 21:04:01 <Omegamoon> the arm64 kexec parts are huge... not something easy to port
Jun 13 21:04:48 <greguu[w]> and for this alone you would need console / uart to test.. so there is no way around getting uart going
Jun 13 21:05:45 <Omegamoon> I was thinking about either ask Planet for a plain gemini board... but I doubt if they have those available
Jun 13 21:06:23 <Omegamoon> other option would be to buy some of the other X27 devices around, strip it, and build it into a dev-device
Jun 13 21:07:33 <greguu[w]> true, a dev board would be usefull until the kernel issues are solved. Planet should have some returns, even a x25 board would do
Jun 13 21:08:13 <Omegamoon> I'd rather go that route than mess up my Gemini to be honest
Jun 13 21:08:46 <greguu[w]> I do have an aarch64 device I could use to build, but not much use for testing as it is S905x and not Mediatek
Jun 13 21:10:09 <Omegamoon> yeah, I've got loads of arm64 devices... that's not the problem :)
Jun 13 21:10:55 <Omegamoon> it needs to have the X25/X27 to make it useful for the Gemini
Jun 13 21:13:23 <greguu[w]> I will investigate the USB UART and check with TheKit about it. He seems to have soldered.
Jun 13 21:13:48 <greguu[w]> a dev board is worth looking into too..
Jun 13 21:14:19 <greguu[w]> but I gave TheKit a link from a guy who did it with x20 via USB, still a custom adaptor
Jun 13 21:21:50 <Omegamoon> so you're saying some sort of usb adaptor should work also?
Jun 13 21:22:22 <Omegamoon> let me see if I can find anything on that in the datasheets
Jun 13 21:34:20 <greguu[w]> old but may apply. http://www.stevenhoneyman.co.uk/2014/11/mtk-mediatek-debug-cable.html
Jun 13 21:35:20 <greguu[w]> it only works if the cable is present before boot.. otherwise the kernel inits it just as a normal usb port
Jun 13 21:36:29 <Omegamoon> just read that one :)
Jun 13 21:41:53 <Omegamoon> my guess is that you would need the left side usb port to make this work
Jun 13 21:43:04 <Omegamoon> that one is special anyway, since it is recognized during early boot by the flashing tool as well
Jun 13 21:44:35 <greguu[w]> yes, the little kernel does detect it too. I wonder if the 3.18 source contains some traces of this or not
Jun 13 21:51:35 <Omegamoon> this is a nice experiment... I wonder if I have all the stuff needed to make one myself
Jun 13 21:52:27 <greguu[w]> https://github.com/NotKit/kernel-3.18-geminipda/blob/2866c7ed283d1b7064299b9170dd79b221346c9a/drivers/misc/mediatek/uart/mt6797/platform_uart.c
Jun 13 21:52:32 <greguu[w]> it does contain it
Jun 13 21:52:51 <greguu[w]> so worth looking into
Jun 13 21:54:12 <Omegamoon> I have a usb-c to usb adaptor... then I have several FTDI boards... now "only" the connection part :P
Jun 13 21:55:16 <greguu[w]> you may not need to implement the "charging switch" as Steve did in his post.. not sure about this USB-C adaptor  - if it breaks anything
Jun 13 21:55:57 <Omegamoon> no, not sure myself...
Jun 13 21:57:34 <greguu[w]> important is that the port is 1.8v and not 3.3v
Jun 13 21:58:03 <greguu[w]> all a bit risky for my taste. :)
Jun 13 22:01:46 <greguu[w]> I have been using my gemini recently a lot for typing and find that using the keyboard with thumbs works just ok while holding it.
Jun 13 22:02:30 <greguu[w]> however I really find the missing back-light annoying in not bright environments
Jun 13 22:03:16 <greguu[w]> Omegamoon, does yours handle rotation when answering a call under Android ? (eg can hold upside down)
Jun 13 22:03:45 <Omegamoon> missing backlight is annoying to say the least
Jun 13 22:04:26 <Omegamoon> answering a call... there's no SIM in the device yet, so I wouldn't know :P
Jun 13 22:05:22 <Omegamoon> it is far from the point that I dare to switch phones... so currently it's only an experimental device
Jun 13 22:05:30 <Omegamoon> I hope that changes though
Jun 13 22:06:00 <greguu[w]> I wanted it for experiments too, just the nexus5 with sailfish was dropped on the day the gemini arrived
Jun 13 22:06:21 <greguu[w]> I have hopes for Sailfish for Gemini as an Android alternative. A dual boot, Sailfish / Linux would be my fav
Jun 13 22:07:09 <greguu[w]> on linux I would not need full phone features, nice would be a log file what records incoming / missed calls and txt while in pure Linux
Jun 13 23:14:22 <greguu[w]> ant_home: how is 4.14 going ? I noticed that after 4.14.30 the spi issue was upstream and did not require manual patching anymore
Jun 13 23:14:40 <greguu[w]> not sure what version it was introduced as i jumped to 4.14.47 or so
Jun 13 23:16:45 <ant_home> https://pastebin.com/ABrJie74
Jun 13 23:16:56 <ant_home> kernel is ok, libc spits align traps
Jun 13 23:17:35 <ant_home> kernel 4.14.48 https://pastebin.com/2X0Sjt9n
Jun 13 23:17:56 <ant_home> (I'll rebuild with gcc6 probably...)
Jun 13 23:18:36 <ant_home> greguu[w], only patch is th esharpslpart stuff
Jun 13 23:19:13 <ant_home> I'll post the OE recipe
Jun 13 23:19:58 <ant_home> greguu[w], anyway 4.14.4x builds with gcc8
Jun 13 23:20:08 <ant_home> 4.4.1xx still fails
Jun 13 23:20:30 <ant_home> they are backporting the fixes right these days
Jun 13 23:20:37 <ant_home> to stable
Jun 13 23:20:54 <ant_home> sigh..every summer a gcc breakage :/
Jun 13 23:23:37 <ant_home> I don't know whether disassembling libc right now or going straight to bed :)
Jun 13 23:23:43 <greguu[w]> hm.. strange
Jun 13 23:23:52 <greguu[w]> I got to check but I build void with 7.3 afaik
Jun 13 23:23:56 <greguu[w]> kernel and userland
Jun 13 23:23:58 <greguu[w]> musl
Jun 13 23:24:08 <ant_home> heh, OE gcc7.3 is flawed
Jun 13 23:24:15 <ant_home> so is 8.1
Jun 13 23:24:26 <greguu[w]> it may be 7.2 , once home I can confirm :)
Jun 13 23:24:29 <ant_home> fun is all runs fine in qemu
Jun 13 23:24:40 <ant_home> as it did last year with kernel
Jun 13 23:24:40 <greguu[w]> what cflags ? xscale ?
Jun 13 23:24:44 <ant_home> yes
Jun 13 23:25:02 <ant_home> TUNE_FEATURES        = "arm armv5 thumb dsp"
Jun 13 23:25:02 <ant_home> TARGET_FPU           = "soft"
Jun 13 23:25:12 <ant_home> seems we have some issues with thumb
Jun 13 23:25:33 <ant_home> at least I was told this coul be a culprit
Jun 13 23:25:48 <greguu[w]> thumb is enabled in kernel but I dont set it in cflags
Jun 13 23:25:58 <greguu[w]> as I doubt we need it anymore
Jun 13 23:26:28 <ant_home> NOTE: make -j 8 HOSTCC=gcc  -isystem/tmp/build/tmp-musl/work/c7x0-oe-linux-musleabi/linux-handheld/4.14-r0/recipe-sysroot-native/usr/include -O2 -pipe -L/tmp/build/tm
Jun 13 23:26:28 <ant_home> p-musl/work/c7x0-oe-linux-musleabi/linux-handheld/4.14-r0/recipe-sysroot-native/usr/lib -L/tmp/build/tmp-musl/work/c7x0-oe-linux-musleabi/linux-handheld/4.14-r0/recip
Jun 13 23:26:28 <ant_home> e-sysroot-native/lib -Wl,-rpath-link,/tmp/build/tmp-musl/work/c7x0-oe-linux-musleabi/linux-handheld/4.14-r0/recipe-sysroot-native/usr/lib -Wl,-rpath-link,/tmp/build/t
Jun 13 23:26:28 <ant_home> mp-musl/work/c7x0-oe-linux-musleabi/linux-handheld/4.14-r0/recipe-sysroot-native/lib -Wl,-rpath,/tmp/build/tmp-musl/work/c7x0-oe-linux-musleabi/linux-handheld/4.14-r0
Jun 13 23:26:34 <ant_home> age CC=arm-oe-linux-musleabi-gcc  -mno-thumb-interwork -marm -fuse-ld=bfd -fdebug-prefix-map=/tmp/build/tmp-musl/work/c7x0-oe-linux-musleabi/linux-handheld/4.14-r0=/u
Jun 13 23:26:37 <ant_home> sr/src/debug/linux-handheld/4.14-r0 -fdebug-prefix-map=/tmp/build/tmp-musl/work/c7x0-oe-linux-musleabi/linux-handheld/4.14-r0/recipe-sysroot= -fdebug-prefix-map=/tmp/
Jun 13 23:26:40 <ant_home> build/tmp-musl/work/c7x0-oe-linux-musleabi/linux-handheld/4.14-r0/recipe-sysroot-native=  -fdebug-prefix-map=/tmp/build/tmp-musl/work-shared/c7x0/kernel-source=/usr/s
Jun 13 23:26:43 <ant_home> rc/kernel   LD=arm-oe-linux-musleabi-ld.bfd
**** ENDING LOGGING AT Thu Jun 14 03:00:04 2018