cubieboard2

[FUKAUMI Naoki]

hi

I tried hno's u-boot (https://github.com/hno/u-boot/tree/wip/a20) on Cubieboard2.

I used kernel from here: https://github.com/cubieboard2/linux-sunxi/tree/sunxi-3.3-cb2

I just tried to boot, and it can be booted.

output from serial console is here: http://androtab.info/cubieboard2/

thank you.

[Dmitriy B.]

Interestingly, I have same results, but I cant use anything Serial/UART related on my A20-olinuxino, while both boards should be pretty much same (except NAND/DDR used etc..). I tried both cubieboard2 kernel and "vanilla" git dump from allwinner, same results - if I have UART enabled - system hangs and panics on switch from bootconsole (earlyprintk) to the actual console, same goes for both 8250 standard driver and "winners serial driver".

I got linaro running and even made some benchmarks here http://linux-sunxi.org/Benchmarks#CPU_2 if anyone interested in A20 performance, you can compare results to info from A10 on the same page. Just keep in mind that these are most likely not maximum you can get from A20, escpecially, openssl results look quite low.

Best Regards,

Dmitriy Beykun

2013/6/16 FUKAUMI Naoki <nao...@gmail.com>

--
You received this message because you are subscribed to the Google Groups "linux-sunxi" group.
To unsubscribe from this group and stop receiving emails from it, send an email to linux-sunxi...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
 
 

[FUKAUMI Naoki]

hi

2013年6月17日月曜日 13時50分47秒 UTC+9 Dmitriy B.:

Interestingly, I have same results, but I cant use anything Serial/UART related on my A20-olinuxino, while both boards should be pretty much same (except NAND/DDR used etc..). I tried both cubieboard2 kernel and "vanilla" git dump from allwinner, same results - if I have UART enabled - system hangs and panics on switch from bootconsole (earlyprintk) to the actual console, same goes for both 8250 standard driver and "winners serial driver".

hmm. I used serial console for a while, but I didn't get any problem...

which compiler are you using? I'm using gcc-linaro-arm-linux-gnueabihf-4.7-2013.04-20130415_linux.tar.xz

how about stock(allwinner) bootloader + compiled kernel and/or hno's u-boot + stock kernel?

[alex allss]

I had issue with other arm device, some toolchain builded kernel worked some not. oliver just try use other toolchain.

cheers.

17.06.2013 16:25 пользователь "FUKAUMI Naoki" <nao...@gmail.com> написал:

--

[Dmitriy B.]

2013/6/17 FUKAUMI Naoki <nao...@gmail.com>

I am using same: "gcc version 4.7.3 20130328 (prerelease) (crosstool-NG linaro-1.13.1-4.7-2013.04-20130415 - Linaro GCC 2013.04)"

My setup is hno's uboot wip/a20 branch + cubieboard2 kernel. I didnt try the standard one.

[FUKAUMI Naoki]

hi

On Tuesday, June 18, 2013 12:47:23 PM UTC+9, Dmitriy B. wrote:

I am using same: "gcc version 4.7.3 20130328 (prerelease) (crosstool-NG linaro-1.13.1-4.7-2013.04-20130415 - Linaro GCC 2013.04)"

My setup is hno's uboot wip/a20 branch + cubieboard2 kernel. I didnt try the standard one.

can you get any panic message when uart is enabled? or just hang?

[Patrick Wood]

I am using gcc version 4.6.3 (Ubuntu/Linaro 4.6.3-1ubuntu5) (4.6.3-1ubuntu5cross1.63) on Ubuntu 12.04LTS for cross compiling hno's u-boot and the 3.3.0 kernel.  Successfully boots on the Cubieboard2.

[Hans de Goede]

Hi,

For those of you who, like me, have been trouble booting their
cubieboard2, because it is hanging on serial port output. I've
good news. This is the infamous designware uart busy bit problem,
and I've fixed this (in an ugly manner by simply copying over
8250_sunxi.c from sunxi-3.4 since that already has a fix).

Here is my git tree, which contains the (ugly) fix:
https://github.com/jwrdegoede/linux-sunxi/commits/sunxi-3.3-cubieboard2

While talking about A20 support I'm currently working on
integrating sun7i support into sunxi-3.4. My first target is to
simply get the kernel booting with a serial console (and nothing
else, no mmc, etc). And I'm not even there yet, so nothing to see
for now, but I just wanted people to know to avoid double work.

Regards,

Hans

[Dmitriy B.]

For those of you who, like me, have been trouble booting their
cubieboard2, because it is hanging on serial port output. I've
good news. This is the infamous designware uart busy bit problem,
and I've fixed this (in an ugly manner by simply copying over
8250_sunxi.c from sunxi-3.4 since that already has a fix)

a20-olinuxino kernel needs the same fix. Was sure I saw same problem somewhere in the past. :)

Good to know that you are working on 3.4 for A20, poke me if you need testers.

Best Regards,

Dmitriy Beykun

[FUKAUMI Naoki]

if you can't boot your kernel from hno's u-boot, please try to use script.bin from Android 4.2 image on https://github.com/cubieboard2/manifests/blob/master/README.md or script.bin on my site(both are same).

if you're using .fex from https://github.com/linux-sunxi/sunxi-boards/commit/894b9d7aefa9731a4d382ca334a2866f582265c4 (probably it's from factory ubuntu image), you can see only 1st "aw_clkevt_init: ..." line. it seems [usbc0] section has a problem. if you modify whole [usbc0] section as same as Android image, problem will be fixed.

[FUKAUMI Naoki]

hi

is it possible to make branches for A20 on https://github.com/linux-sunxi/{u-boot-sunxi,linux-sunxi}?

[Oliver Schinagl]

Not really.

Both me and hno have a20 branches on our own githubs, volatile branches
that are really WIP. So sometimes they work, sometimes they don't. Once
this branch works on all boards (sun[457]i for now) this will be merged
into sunxi-current.

As for linux-sunxi, I assume the same is to be said for hansg, where he
has his own WIP branch and once he gets something working reliably,
it'll probably be merged into stage/sunxi-3.4.

I do not think it's desirable nor a good idea to have seperate branches
for each arch, especially since we are aiming at a unified tree.

oliver

[Oliver Schinagl]

On 06/22/13 08:47, FUKAUMI Naoki wrote:
> On Monday, June 17, 2013 4:09:34 AM UTC+9, FUKAUMI Naoki wrote:
>
> I tried hno's u-boot (https://github.com/hno/u-boot/tree/wip/a20

> <https://github.com/hno/u-boot/tree/wip/a20>) on Cubieboard2.

> I used kernel from here:
> https://github.com/cubieboard2/linux-sunxi/tree/sunxi-3.3-cb2
> <https://github.com/cubieboard2/linux-sunxi/tree/sunxi-3.3-cb2>
>
> I just tried to boot, and it can be booted.
> output from serial console is here:

> http://androtab.info/cubieboard2/ <http://androtab.info/cubieboard2/>

>
>
> if you can't boot your kernel from hno's u-boot, please try to use
> script.bin from Android 4.2 image
> on https://github.com/cubieboard2/manifests/blob/master/README.md or
> script.bin on my site(both are same).
>
> if you're using .fex
> from https://github.com/linux-sunxi/sunxi-boards/commit/894b9d7aefa9731a4d382ca334a2866f582265c4
> (probably it's from factory ubuntu image), you can see only 1st
> "aw_clkevt_init: ..." line. it seems [usbc0] section has a problem. if
> you modify whole [usbc0] section as same as Android image, problem will
> be fixed.

This is interesting. Because my device DOES boot when booting with tons
of power available.

If i boot an A20 device and stock android nand (olimex now, will double
the cubie2 soon) via USB-otg; i get 1 line.

If I boot via a much more powerfull 1 Amp powerblock over OTG, I get
about 5 seconds in (with uart output) and it just reboots there.

[FUKAUMI Naoki]

On Saturday, June 22, 2013 5:40:13 PM UTC+9, Oliver Schinagl wrote:

> if you're using .fex
> from https://github.com/linux-sunxi/sunxi-boards/commit/894b9d7aefa9731a4d382ca334a2866f582265c4
> (probably it's from factory ubuntu image), you can see only 1st
> "aw_clkevt_init: ..." line. it seems [usbc0] section has a problem. if
> you modify whole [usbc0] section as same as Android image, problem will
> be fixed.
This is interesting. Because my device DOES boot when booting with tons
of power available.

If i boot an A20 device and stock android nand (olimex now, will double
the cubie2 soon) via USB-otg; i get 1 line.

If I boot via a much more powerfull 1 Amp powerblock over OTG, I get
about 5 seconds in (with uart output) and it just reboots there.

I didn't investigate deeply. my information is just for workaround ;)

it may be related to power thing, usb_port_type or something may affect power...

[Oliver Schinagl]

I agree, it's a very good finding, because I can't imagine A20 requiring
10 watts to boot without any peripherals.

[Hans de Goede]

Hi,

On 06/22/2013 08:47 AM, FUKAUMI Naoki wrote:
> On Monday, June 17, 2013 4:09:34 AM UTC+9, FUKAUMI Naoki wrote:
>

> I tried hno's u-boot (https://github.com/hno/u-boot/tree/wip/a20 <https://github.com/hno/u-boot/tree/wip/a20>) on Cubieboard2.
> I used kernel from here: https://github.com/cubieboard2/linux-sunxi/tree/sunxi-3.3-cb2 <https://github.com/cubieboard2/linux-sunxi/tree/sunxi-3.3-cb2>

>
> I just tried to boot, and it can be booted.

> output from serial console is here: http://androtab.info/cubieboard2/ <http://androtab.info/cubieboard2/>

>
>
> if you can't boot your kernel from hno's u-boot, please try to use script.bin from Android 4.2 image on https://github.com/cubieboard2/manifests/blob/master/README.md or script.bin on my site(both are same).
>
> if you're using .fex from https://github.com/linux-sunxi/sunxi-boards/commit/894b9d7aefa9731a4d382ca334a2866f582265c4 (probably it's from factory ubuntu image), you can see only 1st "aw_clkevt_init: ..." line. it seems [usbc0] section has a problem. if you modify whole [usbc0] section as same as Android image, problem will be fixed.

I had already seen this and fixed this locally a few days ago, and it saved me
a ton of headache (some mysterious problems I were having went away).

So although a bit late I still want to say: many many thanks!

Regards,

Hans

[FUKAUMI Naoki]

minor updates,

https://groups.google.com/d/msg/linux-sunxi/DyIwRsEo0Bk/SoVEGy_gCqIJ

these kernel patches works fine with sunxi-3.3-cb2.

I'm using gcc-linaro-arm-linux-gnueabihf-4.8-2013.06_linux now.

http://forum.cubietech.com/forum.php?mod=viewthread&tid=653

script.bin in Android 4.2.2 firmware v1.02 has many difference from v1.01.

I think it's better(but I'm not sure everything is perfect).

[Hans de Goede]

Hi,

Can you perhaps put up the new script.bin somewhere for easy access ?

Thanks,

Hans

[FUKAUMI Naoki]

hi

On Monday, July 1, 2013 11:32:45 PM UTC+9, Hans de Goede wrote:

> http://forum.cubietech.com/forum.php?mod=viewthread&tid=653
> script.bin in Android 4.2.2 firmware v1.02 has many difference from v1.01.
> I think it's better(but I'm not sure everything is perfect).

Can you perhaps put up the new script.bin somewhere for easy access ?

script.bin/fex links in http://androtab.info/cubieboard2/ is updated now.

new one has machine = "cubieboard2"

[Hans de Goede]

Hi,

On 07/01/2013 04:44 PM, FUKAUMI Naoki wrote:
> hi
>
> On Monday, July 1, 2013 11:32:45 PM UTC+9, Hans de Goede wrote:
>

> > http://forum.cubietech.com/forum.php?mod=viewthread&tid=653 <http://forum.cubietech.com/forum.php?mod=viewthread&tid=653>

> > script.bin in Android 4.2.2 firmware v1.02 has many difference from v1.01.
> > I think it's better(but I'm not sure everything is perfect).
>
> Can you perhaps put up the new script.bin somewhere for easy access ?
>
>
> script.bin/fex links in http://androtab.info/cubieboard2/ is updated now.
>
> new one has machine = "cubieboard2"

Thanks!

Nothing really useful in there, seems they started with a different fex,
and modified that, most of the changes are in csi and lcd, and other sections
for unused peripherals on the cubie.

Regards,

Hans

[FUKAUMI Naoki]

On Monday, July 1, 2013 11:58:12 PM UTC+9, Hans de Goede wrote:

> script.bin/fex links in http://androtab.info/cubieboard2/ is updated now.
>
> new one has machine = "cubieboard2"

Thanks!

Nothing really useful in there, seems they started with a different fex,
and modified that, most of the changes are in csi and lcd, and other sections
for unused peripherals on the cubie.

thank you for your analysis!