gnrc/tcp : Expose configurations to Kconfig

[akshaim]

Contribution description

This PR exposes compile configurations in GNRC: TCP to Kconfig.

Testing procedure

1. New documentation was built using Doxygen

The build works fine.

2. Header and macro was introduced to tests/gnrc_tcp/

#include "net/gnrc/tcp/config.h"
#define STR(x)   #x
#define SHOW_DEFINE(x) printf("%s=%s\n", #x, STR(x))

The test file can be found here

Compiled binaries (BOARD=samr21-xpro) were uploaded to FIT IoT Test bed.

Default State:

Firmware Output

~~}!main(): This is RIOT! (Version: 2020.07-devel-737-g0b75d-Kconfig_tcp_tests)
~~}!CONFIG_GNRC_TCP_CONNECTION_TIMEOUT_DURATION=3000000
~~}!CONFIG_GNRC_TCP_MSL=1000000
~~}!CONFIG_GNRC_TCP_MSS_MULTIPLICATOR=(1U)
~~}!CONFIG_GNRC_TCP_RCV_BUFFERS=(1U)
~~}!CONFIG_GNRC_TCP_RTO_LOWER_BOUND=(1U * (1000000LU))
~~}!CONFIG_GNRC_TCP_RTO_UPPER_BOUND=(60U * (1000000LU))
~~}!CONFIG_GNRC_TCP_RTO_GRANULARITY=(10U * (1000LU))
~~}!CONFIG_GNRC_TCP_RTO_A_DIV=(8U)
~~}!CONFIG_GNRC_TCP_RTO_B_DIV=(4U)
~~}!CONFIG_GNRC_TCP_RTO_K=(4U)
~~}!CONFIG_GNRC_TCP_PROBE_LOWER_BOUND=(1U * (1000000LU))
~~}!CONFIG_GNRC_TCP_PROBE_UPPER_BOUND=(60U * (1000000LU))
~~}!RIOT GNRC_TCP test application

Usage with menuconfig [default values]

make menuconfig

Firmware Output

!main(): This is RIOT! (Version: 2020.07-devel-737-g82bda-Kconfig_tcp_tests)
~~}!CONFIG_GNRC_TCP_CONNECTION_TIMEOUT_DURATION=120000000
~~}!CONFIG_GNRC_TCP_MSL=30000000
~~}!CONFIG_GNRC_TCP_MSS_MULTIPLICATOR=1
~~}!CONFIG_GNRC_TCP_RCV_BUFFERS=1
~~}!CONFIG_GNRC_TCP_RTO_LOWER_BOUND=1000000
~~}!CONFIG_GNRC_TCP_RTO_UPPER_BOUND=60000000
~~}!CONFIG_GNRC_TCP_RTO_GRANULARITY=10000
~~}!CONFIG_GNRC_TCP_RTO_A_DIV=8
~~}!CONFIG_GNRC_TCP_RTO_B_DIV=4
~~}!CONFIG_GNRC_TCP_RTO_K=4
~~}!CONFIG_GNRC_TCP_PROBE_LOWER_BOUND=1000000
~~}!CONFIG_GNRC_TCP_PROBE_UPPER_BOUND=60000000
~~}!RIOT GNRC_TCP test application

Usage with menuconfig

make menuconfig

Firmware Output

!main(): This is RIOT! (Version: 2020.07-devel-737-g82bda-Kconfig_tcp_tests)
~~}!CONFIG_GNRC_TCP_CONNECTION_TIMEOUT_DURATION=110000000
~~}!CONFIG_GNRC_TCP_MSL=32000000
~~}!CONFIG_GNRC_TCP_MSS_MULTIPLICATOR=2
~~}!CONFIG_GNRC_TCP_RCV_BUFFERS=2
~~}!CONFIG_GNRC_TCP_RTO_LOWER_BOUND=1200000
~~}!CONFIG_GNRC_TCP_RTO_UPPER_BOUND=62000000
~~}!CONFIG_GNRC_TCP_RTO_GRANULARITY=20000
~~}!CONFIG_GNRC_TCP_RTO_A_DIV=4
~~}!CONFIG_GNRC_TCP_RTO_B_DIV=2
~~}!CONFIG_GNRC_TCP_RTO_K=2
~~}!CONFIG_GNRC_TCP_PROBE_LOWER_BOUND=1200000
~~}!CONFIG_GNRC_TCP_PROBE_UPPER_BOUND=50000000
~~}!RIOT GNRC_TCP test application

MACROS were successfully configured.

Issues/PRs references

#12888

[brummer-simon]

The changes did not break any existing tests (tested on "native" and "nucleo-f401re") and the custom configurations are propagated into the binary. From a technical point of view it works good, however there are some descriptions value descriptions that are a bit misleading.

If you fix those the PR is fine by me.

[brummer-simon]

I have some small change requests aside from that good work.

[akshaim]

I have some small change requests aside from that good work.

Thank you for the comments. I have updated the file. Introduced a new symbol to make sure that the default configuration remains intact. Directly configuring the CONFIG_GNRC_TCP_DEFAULT_WINDOW via menuconfig does not have a dependency on GNRC_TCP_MSS_MULTIPLICATOR.

[brummer-simon]

I have tested this PR on "native" and on "nucleo-f401re". The test suite works as expected an all configured values are present in the resulting binary. From my point of view, this PR can be merged, although I don't have the rights to do that.

@akshaim - Thanks for adding the Kconfig support.

[akshaim]

I have tested this PR on "native" and on "nucleo-f401re". The test suite works as expected an all configured values are present in the resulting binary. From my point of view, this PR can be merged, although I don't have the rights to do that.

@akshaim - Thanks for adding the Kconfig support.

Thanks @brummer-simon for the review. @leandrolanzieri Should I squash now ?

[leandrolanzieri]

I have tested this PR on "native" and on "nucleo-f401re". The test suite works as expected an all configured values are present in the resulting binary. From my point of view, this PR can be merged, although I don't have the rights to do that.
@akshaim - Thanks for adding the Kconfig support.

Thanks @brummer-simon for the review. @leandrolanzieri Should I squash now ?

Yes please

[akshaim]

I have tested this PR on "native" and on "nucleo-f401re". The test suite works as expected an all configured values are present in the resulting binary. From my point of view, this PR can be merged, although I don't have the rights to do that.
@akshaim - Thanks for adding the Kconfig support.

Thanks @brummer-simon for the review. @leandrolanzieri Should I squash now ?

Yes please

Done. I have reordered the commit history, to make better sense of, but github displays commits by order of author date.

[leandrolanzieri]

Changes look good and @brummer-simon confirmed that this works properly. ACK.

[akshaim]

Thanks @leandrolanzieri