Kodi Community Forum

Full Version: [i.MX6] XBMC running on Freescale SoC's
You're currently viewing a stripped down version of our content. View the full version with proper formatting.
Pages: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39
@ccelebiler I have made my own kernel and put wolfgars rootfs onto it. NB!!!To run this you need to modify the cooling of the dongle as I guess it would fry dead running xbmc on it. I don't have wifi working, but use a usbdongle with ethernet. To set it up you need:
4.0.0 kernel for gk802 (https://github.com/mtx512/linux-imx/tree....0.0-gk802 is one (defconfig need CONFIG_DEVTMPFS=y CONFIG_DEVTMPFS_MOUNT=y ),
a uboot for gk802 like http://jas-hacks.blogspot.no/2013/05/imx...-boot.html and boot.scr
wolfgars rootfs (extract it from the image).
Hi Stephan,

Your iMX6 VPU HW decode implementation for XBMC should in the future also build for Android support too right?

As asked in this other thread http://forum.xbmc.org/showthread.php?tid=172974 ?

Thanks again for working on iMX, and hope to see a pull request that will be accepted into XBMC mainline soon!

(2013-09-05, 22:37)wolfgar Wrote: [ -> ]Hi Hedda,

You seem to be really disappointed by the original cubox. I cannot tell as I don't know about it nor about solidrun support...

On the paper at least the cubox-i seems to be an interesting product and solidrun seems to prepare a SDK and to open their box.
Of course it is not a substitute for a ready to use solid software offer but it is, at least, appealing for developers who like to hack with such products...
Yes I was seriously disappointed by SolidRun's support as a company for the original CuBox, however it could be mostly due to that SolidRun did not marketed it as purely a development kit, which they should have done instead with the poor support that they offered.

But I was initially not disappointed by the hardware of the original CuBox, as at the time it was announced the original CuBox looked really impressive on paper, just as the new CuBox-i looks great on paper today.

At the very least I hope that SolidRun as a company just does not make the same mistakes with the new CuBox-i as they did with the original CuBox, and that they realize that they have to step-up with Linux-kernel and Android-kernel development themselves to add full support for all its hardware features, and sooner rather than later.

They should really ONLY marked it as a development kit at first, and not solely rely on their community to add Linux-kernel and Android-kernel support for their hardware, because that simply takes to long.

And if and when they are ready to market the new CuBox-i to none low-level developers and non-developer consumers after it has mainline code support upstream, then they should get proxy warehouses within Europe and USA/America for local shipping so people do not have to pay hidden custom fees and high shipping charges.
Hi Hedda

I will try to mainline my code . It is my intend.
But my code is not ready for now.
When I will issue the pull request I will be confident that it is stable enough and I will have removed a few shortcomings Wink

I hope it will happen soon but you have to remember it is a project that I handle on my free time so it is obviously not as fast as everyone (including me) would expect...
(For your information, I have locally synced with the very latest dev branch but it is not stable enough... Should be solved in coming days)

Regarding cubox, I have no opinion : I was able to have constructive exchanges with compulab but I wonder whether it will be the same with solidrund. Way and see...
(But I perfectly understand your remarks and your disappointment as a end user...)

To come back to your question, my work targets mainly raw linux because android support is already available without additional code thanks to libstagefright support which has been recently merged in XBMC. I am unsure about the maturity of the libstagefright provided by freescale for use with XBMC but when it comes to android, it should be the most standard approach I guess...

Best regards
Stephan
xbmc for GK802:

I have now made a script that can be run to install this xbmc on a sdcard or usb for GK802 with the xbmc rootfs from Stephan (wolfgar).
WARNING!!!!!! You should have a proper cooling solution as it get hot when idling in GUI, and you run this at your own risk:
http://stende.no-ip.info/files/xbmc/mkxbmc.sh

The rootfs is the one in http://stephan-rafin.net/blog/2013/09/08...ge-part-2/

I have broken my wifi so I don’t know if wifi works. For network I use an usb ethernet-dongle. If you want to try wifi you have to set it up in /etc/network/interfaces
The rootfs has ssh and password is not set.

I use uboot that support booting from internal,external and usb, but if you choose internal in the script it should just work after putting the sdcard into the internal slot and boot.
You can read more on the uboot here:
http://jas-hacks.blogspot.no/2013/05/imx...-boot.html

UPDATE!!! 15.09.2013 Added option for chosing resolution between 1920x1080 and1280x720

hste
Hi hste,

Thanks a lot for sharing this build script with us.
I will release rootfs alone for any release to come so that gk802 users will be able to use your script to follow my new deliveries
(my sdcard image is for now targeted at wandboard quad)...

and again for users : Setup a PROPER COOLING solution before trying xbmc. It is not an option : I will flame you if you report black screens while using an improperly cooled GK802 and you may even destroy your device...

Best regards
Stephan
Quick question:
(for general knownledge, I don't have an i.MX6 device)

Is the current implemantation of i.MX6 support in XBMC allow the use of alsa devices ? (mainly I would like to know if it possible to pair a bluetooth device via bluez and use it for XBMC)
Hi dhead

Yes XBMC enumerates ALSA devices and is able to use any of them. So if you have a working ALSA BT device, it should be handled ...
Of course you may have to enable a kernel option or to install a package but no reason not to to get the whole thing up and running ....

Best regards
Stephan
Thanks wolfgar,
this is definitely an advantage over the current state of the Raspberry Pi and Amlogic devices.
To be honest, I don't know so well the current implementation on such devices...
Yet, at first sight, I don't know why they cannot behave the same as the xbmc code that is is charge of enumerating and handling alsa devices is not specific to the platform...
Generally speaking if you are able to play a wav with the alsa utility aplay then you should be able to use this device in XBMC (in gotham, you have to switch to expert mode to be able to select the alsa device of your choice while in audio settings)

Regards
Thanks wolfgar, it seems that I was wrong about the Amlogic devices, it looks this had been changed recently, current J1nx's builds for 8726-M3 devices doesn't support alsa sound cards but at least one build for 8726-M6/MX (by CronicCorey) does support.

Anyway thanks for your adding support for i.MX6 devices (one Cubox-i2 pre-ordered Smile )
Hi dhead,

Thanks for these precisions.
Hopefully you will be able a have decent experience with your cubox.
As far as I am concerned, I was contacted by solidrun (thats why I said I may get an early access developer) but after I answered their email, they did not even acknowledge this answer...
I hope for you that they will prove to be more serious when it comes to support their product ...

Regards
Stephan
@hste: thanks for your script, I will try it on my gk802
By reading it, I think there is a little bug: when "downloading and extract kernel again", script wgets rootfs instead of kernel (http://stende.no-ip.info/files/xbmc/mkxbmc.sh)
Thank you again
@olepage: thanks for reporting. I have now fixed the script.
Updated kernel for the gk802 installer script:
http://stende.no-ip.info/files/xbmc/mkxbmc.sh

Think this fix some soundproblems and wifi. Wifi must still be setup in /etc/network/interfaces and probably you have to use static ip and not dhcp


hste
And how to set the correct time and date on the xbmc. And today it has established, and the date of September 6, 2013 and the time incorrectly. And many additions write "script error"
Pages: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39