drivers/at86rf2xx: Switch SRT mode to get random values.

[CW-75]

Contribution description

This Pr is dedicated to resolve the random generation numbers from at86rf233, this should establish the device in a basic operation mode states. also needs to the preamble detector is disabled (RX_PDT_DIS=1). Something else that should be considered is the state of Smart Receiver Technology (SRT), this is an element used to manage the Reception Power Consumption (RPC) and it's affecting the random_number_generator() of the at86rf2xx. Well this Pull request add the functionality to disable temporarily the smart_idle_mode because the random number generator needs take values from RSSI register. When the SRT is active could won´t get the wished random values).

Also to integrate the number generator by the driver implementation, also let's to manipulate the functionality by network options (netopts), adding support in the NETOPT_RANDOM

Testing procedure

run and falsh the test/driver_at86rf_2xx.

make -C tests/driver_at86rf2xx flash term

When it has loaded the shell type random ( This is a command added only to the test ).

2022-07-01 16:20:56,261 # random
2022-07-01 16:20:56,267 # Random number for device 0 via native API: a2f979f9
2022-07-01 16:20:56,271 # Random number for device 0 via netopt: c200fa9d

Also try checking resetting the device to be sure that the self value is not repeating.

2022-07-01 16:21:03,653 # main(): This is RIOT! (Version: 2022.07-devel-900-g162cf-feature/at86rf2xx_random_numbers)
2022-07-01 16:21:03,655 # AT86RF2xx device driver test
2022-07-01 16:21:03,659 # Initializing AT86RF2xx radio at SPI_0
2022-07-01 16:21:03,665 # gnrc_netdev: possibly lost interrupt.
2022-07-01 16:21:03,670 # Initialization successful - starting the shell now
> random
2022-07-01 16:21:06,149 # random
2022-07-01 16:21:06,154 # Random number for device 0 via native API: a8f63524
2022-07-01 16:21:06,159 # Random number for device 0 via netopt: c0fe3169
> reboot
2022-07-01 16:21:17,349 # reboot
2022-07-01 16:21:17,381 # main(): This is RIOT! (Version: 2022.07-devel-900-g162cf-feature/at86rf2xx_random_numbers)
2022-07-01 16:21:17,383 # AT86RF2xx device driver test
2022-07-01 16:21:17,386 # Initializing AT86RF2xx radio at SPI_0
2022-07-01 16:21:17,392 # gnrc_netdev: possibly lost interrupt.
2022-07-01 16:21:17,397 # Initialization successful - starting the shell now
> random
2022-07-01 16:21:19,013 # random
2022-07-01 16:21:19,018 # Random number for device 0 via native API: eccf2f1c
2022-07-01 16:21:19,023 # Random number for device 0 via netopt: 7abc92e6

Issues/PRs references

Fix #18285
You can get more information about the SRT in the samr21-xpro-datasheet in the page 1034, section 40.10.

[maribu]

Thanks for PRing a fix. Some nitpicks inline.

[CW-75]

Thanks for your suggestions @maribu, I applied your recommendations, this could be reviewed again

[maribu]

ACK. I tested on the samr21-xpro and I can confirm that in master 0xff bytes are returned:

> random
2022-07-02 10:05:56,785 # Random number for device 0 via native API: ffffffff
2022-07-02 10:05:56,789 # Random number for device 0 via netopt: 0
> 2022-07-02 10:05:58,269 # random
2022-07-02 10:05:58,274 # Random number for device 0 via native API: ffffffff
2022-07-02 10:05:58,278 # Random number for device 0 via netopt: 0
> 2022-07-02 10:05:58,893 # random
2022-07-02 10:05:58,898 # Random number for device 0 via native API: ffffffff
2022-07-02 10:05:58,902 # Random number for device 0 via netopt: 0

(the netopt version doesn't change the zero-initialized value as it is not implemented in master)

but with this PR, the output appear to contain a high level of entropy:

> random
2022-07-02 10:04:52,783 # random
2022-07-02 10:04:52,788 # Random number for device 0 via native API: 2a613db0
2022-07-02 10:04:52,793 # Random number for device 0 via netopt: 38ec93e2
> random
2022-07-02 10:04:55,725 # random
2022-07-02 10:04:55,730 # Random number for device 0 via native API: dda44bc2
2022-07-02 10:04:55,735 # Random number for device 0 via netopt: 452e984d
> random
2022-07-02 10:04:57,700 # random
2022-07-02 10:04:57,705 # Random number for device 0 via native API: dff3f25d
2022-07-02 10:04:57,710 # Random number for device 0 via netopt: ca0be5c9

Thanks for addressing the issue!

[CW-75]

Hi, I've updated the PR with the new changes required, Thanks @maribu

[maribu]

The shell command in the test is making use of the random function even when the driver is not providing, resulting in the remaining compilation errors. But that is quick to fix.

Please squash afterwards :)

[maribu]

Thanks for fixing the issue!

[CW-75]

I'm happy to contribute with RIOT, thanks for the help @maribu