This repository contains a little bit of software targeted for the Olimex Olinuxino-A13 computer board. Whatever is included may work for other Allwinner-based Olimex boards, but I only have these boards here, so I cannot test on anything else.
The software is packaged as a Mruby gem. Mruby is a small-footprint implementation of the Ruby language written by the Ruby creator targeted to small harware and embedded application.
Currently, the code focuses on:
- controlling the onboard LED
- talking to the SPI serial interface
- talking to an Olimex MOD-MRF89-868 board via the UEXT interface of the A13 board.
- a beginning of a library for managing the framebuffer and touchscreen of an Olimex A13-LCD43TS 4.3inch touch-screen.
The A13 board has an on-board green LED that can be controlled via GPIO pin PG09. Once the pin is appropriately configured (see below), you gain control of it with:
l=Led::newand you switch it on and of as follows:
l.on
l.offThe MRF89 board is a radio receiver/transmitter board for the 868MHz (SRD) frequency band.
Around April 2013 I saw the possibility to develop a work that could make use of this hardware. I thus purchased three boards, and saw if I could make sense of the mechanism.
When I started working on this, I found out that there was practically no available information about how to use SPI with the Olinuxino boards. I also had very little experience on this.
One month, and much sweat later, I was able to have one board sending a data packet, and a second board reliably receiving it. Soon after, the project was canceled and I had to deal with other things.
Six months later, I found myself with a couple of weeks' time and with the desire to learn a bit about Mruby. I thus decided to port the old code (which I had written in Ruby/C) to Mruby, and to publish it on Github, in the hope that it may be useful to others.
My way is to run Debian on my boards. I use Debian since it was created (I am that old), and I am at home there. Of course, the A13 is not very quick in compiling, but I found it much preferrable to do my compilations onboard. You find here a PDF that describes how you can end up with a bootable SD-card for your board, that includes your self-compiled kernel.
I talk to my boards via Ethernet, using a USB-to-USB cable from my main PC. All remote operations are performed from a SSH-connected session. I make sure my main machine offers IP masquerading. This is not the place to describe how I do that. The important thing is that, once logged in to the A13, I have full internet connectivity.
First thing, you will need to include the appropriate support for SPI in your kernel. I have the following selected:
CONFIG_SPI_SUN5ICONFIG_SUN5I_SPI_NDMACONFIG_SPI_SPIDEV
You also have to have GPIO active:
GPIO_SUNXI
The second important thing to do is to open the A13 pins that are used
to talk to SPI#2 (the interface that is wired to the UEXT socket), and
to configure them to be used for SPI. For this, you must modify the
script.bin file that you will include in your SD-card.
The script.bin file is a 'compiled' version of a text file including
a series of parameters that are accessible by the sunxi modules
included in the kernel.
Here is what you should do.
-
You should already have downloaded the
A13_script_files.zipfile here. -
Inside that file, you will file another archive, called
fex2bin_bin_fex_tools.tar.gz. Take it out, extract it to some directory, and, if you are as paranoid as me, do amake clean && make. As a result, you will have an executable calledfexc, and two links to it, calledfex2binandbin2fex -
Pick your preferred
script.bin, copy it to that directory, call itoldscript.binand run:./bin2fex oldscript.bin > script.fex -
Edit the resulting
script.fexfile:-
Comment out this block:
[uart_para] uart_debug_port = 1 uart_debug_tx = port:PG03<4><1><default><default> uart_debug_rx = port:PG04<4><1><default><default>by prepending a
;to each line (pin PG04 is used by SPI). -
Change line
twi2_used = 1to
twi2_used = 0(pins PB17 and PB18 are used by SPI).
-
Search the block beginning with
[spi2_para]. Change it so that it reads:[spi2_para] spi_used = 1 spi_cs_bitmap = 1 spi_cs0 = port:PE00<4><default><default><default> spi_sclk = port:PE01<4><default><default><default> spi_mosi = port:PE02<4><default><default><default> spi_miso = port:PE03<4><default><default><default> -
Add these two blocks:
[spi_devices] spi_dev_num = 1 [spi_board0] modalias = "spidev" max_speed_hz = 1000000 bus_num = 2 chip_select = 0 mode = 0 full_duplex = 0 manual_cs = 0All this should make sure that the port connected to SPI bus#2 of your A13 (the one wired to the UEXT socket) is seen at boot, and managed, within the kernel, by
spidev(see here to learn more about spidev). -
Make sure you can access the GPIO pins you need to access, by adding this other block:
[gpio_para] gpio_used = 1 gpio_num = 5 gpio_pin_1 = port:PG03<1><default><default><default> gpio_pin_2 = port:PG04<0><default><default><default> gpio_pin_3 = port:PB17<0><default><default><default> gpio_pin_4 = port:PB18<1><default><default><default> gpio_pin_5 = port:PG09<1><default><default><default>(you can see all IO ports of the A13 here). Note that I am also adding pin PG09, which is used by the LED module (see above).
-
-
You should then create your own
script.binfile . run:./fex2bin script.fex > script.bin`You should make sure that you put this file into your SD card.
With the above kernel and script.bin, you should be able to see file
spidev2.0 in your /dev/ tree. This means that the SPI port is
recognized.
The library is in communications both with the framebuffer surface
(mmapped) and with the output of the touchscreen (/dev/input
file). As of now, it is capable of:
- returning the current touchscreen position, and whether the screen is currently being touched or not.
- filling the screen
- drawing lines
- drawing rectangles
- performing a simple calibration procedure. The calibration results are saved in a file in the user's home directory.
It goes without saying that I have no idea whether all of this works for any of the other screens that Olimex produces.
Log in to your board. Make sure you have the normal compilation environment. Make sure you install a version of Ruby! It is needed for compiling Mruby. Install Git. Then go to a suitable directory and type
git clone https://github.com/mruby/mruby.git
Before compiling, you must add a couple of lines to file
build_config.rb:
conf.gem(:github=>"asfluido/olinuxino",:branch=>"master")
conf.gem(:github=>"mattn/mruby-onig-regexp",:branch=>"master")This makes sure that my gem, as well as the regular expression gem, are included into your version of the mruby executable.
It is then sufficient to run make, and you will be able to run my
test programs. ./minirake -p connects to the remote gem directories
and syncronizes with them if needed.
The /test folder includes a few scripts I wrote for testing my code.
By running led.rb, you will invoke a very simple script that toggles
the onboard LED on/off every time you press return.
In order to test my code you need to have two boards. The first one will be transmitting, the second one will be receiving.
Say that you have your mruby on /usr/src/mruby. Start the
transmitter by running:
/usr/src/mruby/bin/mruby /usr/src/mruby/build/olinuxino/test/transmit.rb 1 2 3 4 5 6 7 8
The digits represent the channels that will be transmitted on. 32 channels are available. As soon as the program is started, it will start sending a test message to each of the channels that you specify on the command line, endlessly cycling.
On the second board, start the receiver by running:
/usr/src/mruby/bin/mruby /usr/src/mruby/build/olinuxino/test/receive2.rb 4
Whenever a message is received on channel 4, it will be printed on your screen.
Script fb.rb exercises the current capabilities of my library.
The first time you run it, you must calibrate the screen - just touch on the four crosses in sequence.
Then, multicoloured segments and rectangles are drawn very quickly on the screen. If you touch the screen, a coloured square will follow your finger around.