[CuBox-I] make kernel 3.14.14 available

[CurlyMoo]

SolidRun has announced that kernel 3.14.y will be an LTS version. They say it should have working GPIO support on the HB. It is possible to provide it (for me to test)?

Compiling it myself failed due to BTRFS unsupported features.

http://www.solid-run.com/community/topic1525.html

[CurlyMoo]

I installed the image from 20 aug. and upgraded it to use my kernel. xbian-copy sadly still doesn't work for the same reasons as earlier.

Few bugs i encountered:
@u-boot:
spl: error reading image u-boot.img, err - -1

@kernel boot:
/init: line 86: /bin/busybox: not found

@login:

Last login: Wed Aug 20 08:44:39 2014 from x.x.x.x
-bash: [: too many arguments

I also noticed that i can't change the FS of /dev/mmcblk0p1. Of course with that partition unmounted. It keeps telling me:
mkfs.msdos: unable to open /dev/mmcblk0p1: Device or resource busy

Also, can you make the serial setting default in the boot.src. Also making the boot partition fat32 would make my testing a lot easier :)

[mk01]

th middle bug is not related to imx6 nor 3.14 kernel. it is the one causing all the no booting also on RPI.

if that happened (no boot, that ERROR). it is needed to have initramfs-tools package at least -17a.
then also reinstall xbian-package-config-shell.

apt-get install --reinstall xbian-package-config-shell

(because on disk busybox is rewritten with some other file).
(((busybox has symlinks - standard cp command will no replace link. will follow it and replace busybox binary)))

the login is already fixed, otherwise it was no harm. only if check for correct locale was ok, "" was returned and badly tested.

the first one NOIDEA. but if you failed on initramfs that means that uboot was working ok. We can check with rabeeh but I would guess it is the try of SPL load uboot.img from partition.
(if then is read from sectors, the error should not be shown because is not errror)

[mk01]

I had the kernel running (as we agreed also for the cloning test, but after some longer hours I'm getting distrorsions on VPU decoding stuff. like bad reception on DVBT or when broken media file when pieces are missing. at first I was also thinking it is that, but is not.
so it is available but will not go out definitely for some time still.

btw: on reverting to 3.10.51 did btrfs send. all 70 GB ok.

[mk01]

what is reboot loop fix ?

[CurlyMoo]

Can you create a new 3.14.14 compilation that i can experiment with and upload it to the xbian apt? The current one is buggy.

[CurlyMoo]

I've currently got a nicely running 3.14.14+ with a XBMC nightly 14.0-ALPHA5 Git:2014-10-04-77f0939-dirty running on the Hummingboard. Only issue left is sound. Haven't bugfixed that very long though. BTRFS send/receive is working nicely as well as XBMC playback for longer periods of time.

I have created a XBMC package for it, will do the same for the kernel. Let me know if you're interested and want to test yourself.

[CurlyMoo]

Working well now.