[IMAGE] A13(-WIFI) Debian (now Ubuntu) + xfce flashable (Updated: 08 Mar: R18)

Started by jwischka, December 17, 2012, 06:36:15 AM

Previous topic - Next topic

ontsurt

Hi John,

Sure I can do that and I will if there is no other way to resolve this.
However there are two thoughts I have around that.

  • If there are no serious ill effects of enabling the dynamic CPU frequency then enabling it will at least make all usages a little more efficient. Nothing wrong with that as far as I can see
  • having to turn the device on and off reduces the amount of locations one can usefully deploy it. I was thinking of mounting one in the garden so we can listen to music there. Having to turn it on or off means that it's more difficult to do that than when I can just lock it in a waterproofed box and only control it through Wifi

I realize these are all not significant items in the grand vast emptiness of space but if there are no ill effects why not?

jwischka

Quote from: ontsurt on March 06, 2013, 12:10:45 AM
The way I look at it: at 50 watt continuously the board uses about 95 Euro's worth of energy per year. That just feels so wasteful for a radio/squeezebox that I only listen too 2 or 3 hours per day when having breakfast or dinner.

Then again I don't know it there are any drawbacks from using cpu freq and allowing the board to go into near complete idle mode. If that makes it completely unsuitable for other purposes for instance then obviously it makes sense not to implement it.

I've tried to do some reading on it but this and came across some references to using cpufreq in combination with sched_mc_power_savings which apparently gives the best results or something but honestly this is beyond my current understanding of Linux.

If you do it you make me a happy man is about all I can 'meaningfully' say.

Kind regards,

Leon

I'm sorry, but there is absolutely no way the board is using 50 watts of energy. If it were, you wouldn't be able to hold it. Where are you getting that number? I suspect the board is using less than 5 watts at full power, if that. Do you actually have something that is measuring the power, or are you just looking at the rating of your supply? Put in perspective, many modern day laptops don't consume more than 40 watts - and even some desktop are down in the 45-50 watt range.

I'm in the process of doing some kernel updating, and I'll compile cpufreq support in, but I don't think you have to worry about the board using that much energy.

Kernel updates later today, perhaps.

ontsurt

Thanks for the pointer. I have to admit that I found it odd myself but it was technically possible as the PSU I am using delivers 12 volt at a maximum of 4 Amps. So I was just assuming it was maxing out the PSU. The PSU is regulated so it shouldn't consume this much power if it's not used.

However if you suspect it's only drawing about 5 watts of power it could well be that the sensor I am using is simply not sensitive enough and somehow bottoming out at 50 Watt. I'll borrow a more accurate sensor and see what that reads.

Cheers

jwischka

Quote from: ontsurt on March 08, 2013, 05:23:56 PM
Thanks for the pointer. I have to admit that I found it odd myself but it was technically possible as the PSU I am using delivers 12 volt at a maximum of 4 Amps. So I was just assuming it was maxing out the PSU. The PSU is regulated so it shouldn't consume this much power if it's not used.

However if you suspect it's only drawing about 5 watts of power it could well be that the sensor I am using is simply not sensitive enough and somehow bottoming out at 50 Watt. I'll borrow a more accurate sensor and see what that reads.

Cheers

Remember that in electronic terms, power always goes to heat. If the board is not hot, then you're not drawing that much power. If your PSU is hot, then the PSU may be consuming that much power, but it's not the board's fault.

R18 is posted.

Added FTDI support, and fixed some wireless driver issues in the video kernel. Preliminary support for cpufreqd, but it's not working yet. And I'm not going to spend a ton of time working on it at the moment. Maybe in future releases.

Kernels:
R15 Video Kernel
R15 Headless Kernel



Image:
Get current version (R18) here.

alp

Hi!

These Ubuntu images seem very nice. I would like to use them but I want to be
able to compile the kernel myself. To get the exactly same kernel it is however not
enough to use the kernel sources that can be downloaded according to the
general instructions about building the kernel. The vital part is the .config file.
Would it be possible to obtain the .config file for the kernels in these
images. I would like to have it at least for R17 and R18.

Thank you in advance!


jwischka

Quote from: alp on March 11, 2013, 10:50:27 AM
Hi!

These Ubuntu images seem very nice. I would like to use them but I want to be
able to compile the kernel myself. To get the exactly same kernel it is however not
enough to use the kernel sources that can be downloaded according to the
general instructions about building the kernel. The vital part is the .config file.
Would it be possible to obtain the .config file for the kernels in these
images. I would like to have it at least for R17 and R18.

Thank you in advance!

As John said, they are included in the tars, and in the image itself.

alp

Thank you, now I have found the config files. But then I had problems using
them. I followed the instructions in

https://www.olimex.com/wiki/Build_Bootable_SD_Card_with_Debian

under "Building the Kernel". But apparently this caused the download of the version
3.0.62 of kernel. According to the jwischka changelog I see I should have
version 3.4.29. I am confused now... Where exactly should I look to get this version of the
kernel source tree which includes all of the necessary support for the
Olimex A13-WIFI board? Can I simply use the general version 3.4.29 on www.kernel.org?
Do I need to apply some patches to it for the Olimex A13-WIFI board support?

Once again... thanks for the help!

jwischka

Quote from: alp on March 11, 2013, 07:44:49 PM
Thank you, now I have found the config files. But then I had problems using
them. I followed the instructions in

https://www.olimex.com/wiki/Build_Bootable_SD_Card_with_Debian

under "Building the Kernel". But apparently this caused the download of the version
3.0.62 of kernel. According to the jwischka changelog I see I should have
version 3.4.29. I am confused now... Where exactly should I look to get this version of the
kernel source tree which includes all of the necessary support for the
Olimex A13-WIFI board? Can I simply use the general version 3.4.29 on www.kernel.org?
Do I need to apply some patches to it for the Olimex A13-WIFI board support?

Once again... thanks for the help!

git checkout sunxi-3.4

Are you wanting to compile this for an academic exercise, or are you wanting to actually add/remove something from the kernel?

alp

Thank you very much, I will try that.

At the moment I do not have exact plan of
what I would do with the kernel. I want to use the GPIOs
and it is always good to have possibility to tweak the kernel.
I also can not get the touchscreen to work correctly and maybe
I will put some logs into the kernel driver to see what's
happening.

buzibus


I'm running R18 now. It's good to see some of the problems from old versions are gone :) thanks for that


Still two grudges:

- cpufreqd is not working correctly -maybe it's time to upgrade to something else?


- when i check who's on my board I always get a root user that doesnt go away :)

root     tty1                      Wed21    2days  0.09s  0.03s -bash

even when I reboot my board, he stays there. in this case I have rebooted my card last time 2 days ago lol


thanks for all the work you've place into this jwischka

torqu3e

I think this is my first post here, and I love my olinuxino wifi which was running the standard image from Olimex.

Today I switched to the R18 version of this image and the audio was dead, much head breaking and comparison between the other image and this one and reading from this thread itself, here is what fixed the issue...

olinuxino@ubuntu:~/drive/Songs$ grep audio /etc/group
audio:x:29:linaro,olinuxino


If the audio line does not have olinuxino as the user defined there, aplay -l will say something like this "aplay: device_list:207: no soundcards found..."

Happy to have the audio going but for some reason mplayer is hitting 100% load at times, and the playback is choppy, which doesn't happen on the other image, not sure what could be causing this but am investigating. This is with an mp3 not video.

Also I have the webcam that is integrated in a DELL SP2309W throwing video out using motion, it was a cakewalk there. :)

Installing anything using apt-get always throws this error...

Errors were encountered while processing:
cpufreqd
E: Sub-process /usr/bin/dpkg returned an error code (1)


What could be causing that?

jwischka

cpufreqd support is just broken. I shouldn't have included it. I'd recommend uninstalling for the moment. I may try to get it working in the future, but honestly it's pretty low on the priority list.


buzibus

Quote from: jwischka on March 17, 2013, 01:55:27 PM
cpufreqd support is just broken. I shouldn't have included it. I'd recommend uninstalling for the moment. I may try to get it working in the future, but honestly it's pretty low on the priority list.

There's no point in trying to fix it, just don't include it in future images.

If you really want there are other more recent options.

nvd

Do you know which headers to use to compile the source against the kernel?

I am talking about "olinuxino_xfce-r15.zip".

"uname -a" is giving:
"3.4.29-jwischka-3.4-video-20130209-R15+ #20 PREEMPT"

Can I get the "github" link for the kernel source, you have used?

Thanks.