[OT]: ATMEL AT91SAM9G25/ARM926EJ-S <Helpneededforinterpretationofadmesgmessage

Hi,

I am using a Arietta G25 (http://www.acmesystems.it/arietta) with
Robert Nelsons Kernel 3.17.3 and the kernel configuration provided
by Acmesystems for their Debian Linux for the Arietta G25 based on
Linux kernel 3.16.1.
As far as I know the only patches they used are agains dts/dtb files
(I use the same ones) and to configure the kernel without the need of
using makeconfig.

So far so nice...the system boots with a Gentoo stage3 rootfs.

There only two things left:
1.) shuthdown -h -H now (as said to be used by Acmesystems) reboots
    the system instead of powering it off (which works with their
    image).

2.) The at91 RTC is missing the "Real"...the difference are several
    minutes per hour.

I compared the dmesg output of the Arietta G25 booted with one and
the other system. With the Gentoo system there is one line which
irritates me and may indicate a way to a solution for the RTC problem:

[ 0.000000] Main crystal frequency not set, using approximate value
[ 0.000000] master clk is overclocked
[ 0.000000] sched_clock: 32 bits at 128 Hz, resolution 7812500ns, wraps every 16777216000000000ns
[ 0.000000] Console: colour dummy device 80x30
[ 0.054687] Calibrating delay loop... 220.07 BogoMIPS (lpj=862208)

The clock frequency of the CPU is unknown??? The Debian system does
not have that problem...

Any suggestion how to fix one or (better! :slight_smile: both problem is VERY 8)
appreciated! Thank you very much in advance for any help!

Best regards,
Meino

About the time clock , as I know , its about the crystal , called jitter ?
You can correct it with NTP client or GPS , ( if you have gps )

Hi liyaoshi,

thanks for your posting!

I am not searching for workaround -- the small Arietta will
become a mobile device which has no connection to the
internet.

I am searching for the reasons of the two problems and want
to fix the (currently unknown) bugs behind that.

Even without sync an RTC should not become THAT wrong as
the RTC of the Arietta currently.

What does this dump from dmesg mean?

Best regards,
Meino

liyaoshi <liyaoshi@gmail.com> [14-12-01 02:57]:

seems dts config file error , but from

http://www.acmesystems.it/compile_linux_3_16

maybe you should try this patch http://www.acmesystems.it/www/compile_linux_3_16/acme.patch

or can you show up your patch ?

That's for the Cortex-A5 from Atmel, the Arietta G25 should be fully
supported on mainline.

There's been lots of rtc changes, you should retry with 3.18-rc7, as
the free-electrons guys took over mach-at91..

Regards,

Hi liyaoshi,

in the meanwhile I digged deeper into the kernel
and found, that in Linux 3.17.3 there is a new
configuration item AT91_USE_OLD_CLK which is set
to <n> which in turn set USE_OLD_CLK to <n>.
But last get set when using Linux 3.16.1.
I compiled Linux 3.16.1 with the instructions
of Acmesystems (including the patch, which inserts
some dts files and the kernel configuration itsself)
and it seems (no long period done though) that the
RTC problem is gone.

As Robert mentioned in his mail to this thread,
there were a lot of RTC changes AND (TADA!)
the Arietta should be supported by the mainline kernel,
which in turn makes me completly independant from any
patches by acmesystems (*NO* inplicite or explicite cirtism
intended here!!!), which are GOOD news.

Gentlemen...start your compilers!
Best regards,
Meino

liyaoshi <liyaoshi@gmail.com> [14-12-01 04:04]:

Robert Nelson <robertcnelson@gmail.com> [14-12-01 04:04]:

> seems dts config file error , but from
>
> http://www.acmesystems.it/compile_linux_3_16
>
> maybe you should try this patch
> http://www.acmesystems.it/www/compile_linux_3_16/acme.patch

That's for the Cortex-A5 from Atmel, the Arietta G25 should be fully
supported on mainline.

There's been lots of rtc changes, you should retry with 3.18-rc7, as
the free-electrons guys took over mach-at91..

Regards,

--
Robert Nelson
http://www.rcn-ee.com/

--
For more options, visit http://beagleboard.org/discuss
---
You received this message because you are subscribed to the Google Groups "BeagleBoard" group.
To unsubscribe from this group and stop receiving emails from it, send an email to beagleboard+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Hi Robert,

thank you for your posting! :slight_smile:

THATS really good news: Independance from any specialized
patches! :slight_smile:

Only to ensure, that I am not confusing things again:
I interpret "Mainline" as: "Take a kernel from ftp.kernel.org"
Or does "the free-electrons guys took over mach-at91" mean,
that I should search their site for their kernel sources?

Best regards,
Meino

Hi Robert,

thank you for your posting! :slight_smile:

THATS really good news: Independance from any specialized
patches! :slight_smile:

Only to ensure, that I am not confusing things again:
I interpret "Mainline" as: "Take a kernel from ftp.kernel.org"
Or does "the free-electrons guys took over mach-at91" mean,
that I should search their site for their kernel sources?

The free-electron guys only work on mainline, so all their patches are
heading to mainline. (kernel.org) For the last 2 kernel releases's
they've cleaning up everything in mach-at91 and pushing everything to
device tree's.

Regards,

Robert Nelson <robertcnelson@gmail.com> [14-12-01 16:56]:

> Hi Robert,
>
> thank you for your posting! :slight_smile:
>
> THATS really good news: Independance from any specialized
> patches! :slight_smile:
>
> Only to ensure, that I am not confusing things again:
> I interpret "Mainline" as: "Take a kernel from ftp.kernel.org"
> Or does "the free-electrons guys took over mach-at91" mean,
> that I should search their site for their kernel sources?

The free-electron guys only work on mainline, so all their patches are
heading to mainline. (kernel.org) For the last 2 kernel releases's
they've cleaning up everything in mach-at91 and pushing everything to
device tree's.

Regards,

--
Robert Nelson
http://www.rcn-ee.com/

--
For more options, visit http://beagleboard.org/discuss
---
You received this message because you are subscribed to the Google Groups "BeagleBoard" group.
To unsubscribe from this group and stop receiving emails from it, send an email to beagleboard+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Hi Robert,

THANK YOU ! :))

Best regards,
Meino

Robert Nelson <robertcnelson@gmail.com> [14-12-01 16:56]:

> Hi Robert,
>
> thank you for your posting! :slight_smile:
>
> THATS really good news: Independance from any specialized
> patches! :slight_smile:
>
> Only to ensure, that I am not confusing things again:
> I interpret "Mainline" as: "Take a kernel from ftp.kernel.org"
> Or does "the free-electrons guys took over mach-at91" mean,
> that I should search their site for their kernel sources?

The free-electron guys only work on mainline, so all their patches are
heading to mainline. (kernel.org) For the last 2 kernel releases's
they've cleaning up everything in mach-at91 and pushing everything to
device tree's.

Regards,

--
Robert Nelson
http://www.rcn-ee.com/

--
For more options, visit http://beagleboard.org/discuss
---
You received this message because you are subscribed to the Google Groups "BeagleBoard" group.
To unsubscribe from this group and stop receiving emails from it, send an email to beagleboard+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Hi Robert,