Pluto & M2k production scripts

This is version 2 of the Pluto production scripts, and it also supports M2k.

The main scripts are production_pluto.sh & production_m2k.sh.

Before running the production scripts, run setup_env.sh once.

These scripts must be run with sudo, so sudo ./production_pluto.sh.

What these scripts will do is loop indefinitely, during which:

Initialize FTDI channels to default state
If there was an error, or settings not loaded from EEPROM, it will load them from EEPROM
Turn on Ready LED, Turn off In-Progress LED, Wait for button "Go" to be pressed
When "Go" button pressed, it will

Turn off Done,Error,Ready LEDs
Turn on In-Progress LED

Run the pre-Flash tests (voltage measurements) and check valid values (script lib/preflash.sh)

If an error happens, or invalid measurements persist 4 retries, Error LED is turned on and will go to 1.

Run the Flash step (script lib/flash.sh)

Run OpenOCD to load bootloader via JTAG
Run dfu-util to flash uboot, uboot-env and firmware to proper partitions
power-cycle the board

Post flash step is specific to each board

for Pluto the cal_ad9361 script will be run, the linux env will be validated

Configuration structure

All config for all supported boards is in the config folder.

Board specific stuff is under config/<board>.

The digilent-hs2.cfg & ftdi4232.cfg files are used by OpenOCD to operate with these JTAG adapters. However, each board has it's own specific override for these, in order to include it's own ps7_init.tcl which differs per-board.

Under config/<board>/values.sh are stored the values for ADC for measurement.

For Pluto specifically (and maybe it will be similar for other boards) the config/pluto/postflash.sh is used during the post flash step. Since this was written in v1 of the script, it is largely unchanged. Similarly this is true for config/pluto/linux.exp.

This may change at some point in time to common-alize things.

Important utilities

setup_env.sh

Run this on a new system when cloning this repo with ./setup_env.sh <pluto|m2k>. Requires sudo access and the script will request it. This script will also install the current user.

This script sets up all necessary pieces for flashing to work:

downloads & builds libiio
downloads plutosdr_scripts and builds cal_ad9361
builds ft4232h_pin_ctrl utility that controls the FTDI channels/pins for SPI and Bitbang functions
downloads builds Scopy for M2k
disable some default XUbuntu auto-start services that may complicate things
installs running user in /etc/sudoers with NOPASSWD: directive
installs init.sh in the udev rules to be called whenever the JIG board is plugged in
installs production_<pluto|m2k>.sh to the .config/autostart folder to start production script on startup
installs autosave_logs.sh to the .config/autostart folder ; this will save logs on anything mounted under /media/ that has an autosave_logs folder
installs call_home to the .config/autostart folder ; this will periodically try to open a reverse SSH tunnel back to testjig.hopto.org on port 2222

Tested/working on XUbuntu 18.04. Currently XUbuntu is the preferred distro; it could work on other Ubuntu distros as well, as long as XFCE4 is available & installed.

autosave_logs.sh

It's running as a service installed by setup_env.sh.

This will save logs on anything mounted under /media/ that has an autosave_logs folder.

Once the logs are saved, it will wait for 1 minute and save them again if the folder is still available.

The script itself is pretty dumb. All logs under the logs subfolder will be tar + gzipped to the autosave_logs folder in a file with format "-.tar.gz".

init.sh

Initiates all FTDI pins + GPIO expander pins to default values. Only initiates channels A, B & D on FTDI, since C is reserved for UART talk.

This is is called by udev whenever a JIG board is plugged in.

The logic is also called when the production_<m2k|pluto>.sh scripts run.

This is important to run because it inits all the GPIOs to valid states.

m2k_power_calib_meas.sh

Contains the logic for calibrating the power-supplies for M2k. It checks that the values are withing valid ranges.

It's called by the config/m2k/scopy1.js script.

config/<m2k|pluto>/postflash.sh, config/m2k/scopy1.js, config/m2k/scopy2.js

Post-flashing scripts for Pluto & M2k have been added to the config/ folders.

These steps are very specific to each board.

For M2k, Scopy is being run with the scopy1.js & scopy2.js. The reason it was split into 2 parts, is because scopy1.js is run without showing the Scopy GUI; in that part it's more important to show the console. There is currently a problem when showing the Scopy GUI [when it was hidden initially], in that it doesn't render properly.

ref_measure_ctl.sh

Used for M2k in config/m2k/scopy1.js.

It toggles some GPIOs on the GPIO expander, and then will measure the values. The step for measuring these refs, requires enabling them via some GPIOs on the GPIO expander, and then measuring, so the step was hidden behind this script.

scp.sh

Used for M2k in config/m2k/scopy1.js. Writes the calibration file (generated by config/m2k/scopy1.js together with m2k_power_calib_meas.sh) via SCP to the M2k device.

SCP is more reliable than using the auto-mounted drive from M2k.

This also handles passwd login via the sshpass tool.

Outputs ok if all went well.

Called with scp.sh <src> <dst> <pass>

wait_pins.sh

Used in production_<pluto|m2k>.sh and config/m2k/scopy1.js to wait for a pin to be asserted.

Called with wait_pins.sh <chan> <pin>, where chan can be A,B,C,D or GPIO_EXP1 and pin can be pinX [ X = 0..7 ].

scopy

Wrapper for calling the Scopy binary that was built by setup_env.sh.

call_home

A simple loop that will open a reverse SSH tunnel to testjig.hopto.org on port 2222.

This tunnel can then be used to connect back to the test-jigs from the testjig.hopto.org host via ssh -p <port> localhost, where port is the REVERSE_SSH_PORT=2000 from the call_home script.

General/other utilities

Most scripts will load config.sh which overrides paths to make sure that the proper iio_attr, cal_ad9361 and ft4232h_pin_ctrl utilities are used. i.e. the utilities built by setup_env.sh.

eeprom_cfg.sh

This will write the board's settings to EEPROM.

Must be called with:

sudo ./eeprom_cfg.sh save VGAIN=<gain> VREF=<vref> VOFF=<voffset>

Order is not important, but all 3 values must be provided.

Similarly:

sudo ../eeprom_cfg.sh load

will load configuration from EEPROM and display it.

Loading it via this shell script, won't make it available in the shell's env.

selftest.sh

This will run the ADC self-test for the digital interface.

Run with:

sudo ./self_test.sh

Expected result is:

AA AA 55 55 

!!All good!!

toggle_pins.sh

Generic bitbang wrapper for all FTDI pins.

Can be called with:

sudo ./toggle_pins <chan> [pin names]

Where:

is A,B,C,D - FTDI channels
pin names are pin0,pin1,...,pin7 to make pins output high, pin0i,pin1i,...,pin7i to make them input, and if omitted, should be active low

measure.sh

Run a measurement with ADC on the board.

Run with:

sudo ./measure.sh <channel>

When channel is V0A,V1A,...,V7A,V0B,V1B,...V7B or all. Default is all.

flash_<pluto|m2k|sidekiqz2>.sh, postflash_<m2k|pluto>.sh, preflash_<m2k|pluto>.sh

These scripts handle flashing, preflash, postflash part of the boards.

Are added for convenience.

The flashing scripts require files to be presend in release/<target-name>

update_<m2k|pluto>_release.sh

These files are also called by the setup_env.sh script.

They will download the latest version of the files required for flashing for Pluto/M2k.

Files will be placed in release/pluto & release/m2k.

Note: they will delete existing files in those folders.

If needed, an older version can be specificed as an argument.

Example: update_m2k_release.sh v0.18

AlexandraTrifan/board-test