Skip to content

[tester] IEEE 802.15.4 SubMAC on top of IEEE 802.15.4 Radio HAL#14787

Closed
jia200x wants to merge 31 commits intoRIOT-OS:masterfrom
jia200x:pr/ieee802154_submac
Closed

[tester] IEEE 802.15.4 SubMAC on top of IEEE 802.15.4 Radio HAL#14787
jia200x wants to merge 31 commits intoRIOT-OS:masterfrom
jia200x:pr/ieee802154_submac

Conversation

@jia200x
Copy link
Copy Markdown
Member

@jia200x jia200x commented Aug 19, 2020 •
edited
Loading

Contribution description

This draft provides a very rough implementation of #13376. In a nutshell, this module provides a uniform layer to send and receive IEEE 802.15.4 MAC frames with retransmissions and CSMA-CA. The SubMAC emulates in software hardware features (CSMA-CA, Retransmissions, etc) missing on a given radio.

Besides that, I added a a netdev_ieee802154_submac as a transition layer. With this, default and gnrc_networking are indirectly working on top on the Radio HAL (#13943 ).

I hope this server as a proof for the Radio HAL API :)

I had to include the work by @miri64 from #11263. Since the radios are asynchronous now, there are some lost fragments when the radio returns -EBUSY during the TX process.

Testing procedure

Check tests/ieee802154_submac for the "raw SubMAC API" calls. This example supports sending and receiving data with retransmissions and CSMA-CA.
A more interesting test is running default or gnrc_networking with netdev_ieee802154_submac. With that, all radios from #14655 (CC2538, nrf802154, at86rf2xx) are working with full CSMA-CA/retransmissions support!! :)
Try adding some noise and pinging between nodes.

Issues/PRs references

#13943
#13376
#14655 (base)
#11263

This was referenced Aug 19, 2020
Closed
Closed
@jia200x jia200x force-pushed the pr/ieee802154_submac branch from 2e83920 to 2359756 Compare August 19, 2020 09:43
@jia200x
Copy link
Copy Markdown
Member Author

jia200x commented Aug 19, 2020 •
edited
Loading

Here is an example of ping (300 ms interval, 1024 kb payload) between the nrf802154 and cc2538, without retransmissions:

2020-08-19 11:46:29,504 # --- fe80::2068:3123:fe42:2e25 PING statistics ---
2020-08-19 11:46:29,509 # 100 packets transmitted, 60 packets received, 40% packet loss
2020-08-19 11:46:29,514 # round-trip min/avg/max = 117.895/123.418/129.290 ms

Same example with retransmissions:

2020-08-19 11:47:12,495 # --- fe80::2068:3123:fe42:2e25 PING statistics ---
2020-08-19 11:47:12,501 # 100 packets transmitted, 100 packets received, 0% packet loss
2020-08-19 11:47:12,506 # round-trip min/avg/max = 123.342/131.136/141.767 ms

Both motes are next to me in my office. But I consistently get ~40% packet losses when I disable retransmissions (this is what we have in current master using these radios).
When I activate the SubMAC retransmissions, I consistently get ~1% packet losses.

Note the nrf52840 doesn't support Auto ACK (it's provided by the driver implementation) nor CSMA-CA/Retransmissions. The cc2538 supports Auto ACK and CSMA-CA, but no retransmissions. The SubMAC is handling the retransmission procedure for both radios and CSMA-CA for the nrf52840

@jia200x jia200x mentioned this pull request Aug 19, 2020
Merged
@jia200x jia200x mentioned this pull request Aug 19, 2020
Merged
@jia200x jia200x force-pushed the pr/ieee802154_submac branch from 2359756 to 8246844 Compare August 19, 2020 13:03
@jia200x
Copy link
Copy Markdown
Member Author

jia200x commented Sep 1, 2020

note: ACK request is disabled by default here. To enable it, use the ifconfig <pid> ack_req command

@jia200x jia200x force-pushed the pr/ieee802154_submac branch 2 times, most recently from ec6e7f4 to 82e0d5a Compare September 2, 2020 14:13
jia200x and others added 16 commits September 4, 2020 11:33
@jia200x
cc2538: adapt pan_id to pointer
6993733
@jia200x
nrf802154: adapt pan_id to pointer
c9cc2eb
@jia200x
at86rf2xx: adapt pan_id to pointer
433a05a
@jia200x
net/ieee802154: add default CSMA-CA params
20a8140
@jia200x
ieee802154_submac: add initial implementation
ac64a3a
@jia200x
tests/ieee802154_submac: add test application
b6bbea2
@jia200x
netdev/ieee802154_submac: add compatibility layer
a55c665
@jia200x
auto_init/nrf52840: add support for SubMAC
473a29d
@jia200x
auto_init/at86rf2xx: add support for SubMAC
705ed92
@jia200x
auto_init/cc2538_rf: add support for SubMAC
356f8dd
@miri64 @jia200x
gnrc_netif: add a send queue
e6293b0
@miri64 @jia200x
gnrc_netif: add packet to queue when device is busy
9c09d5e 
... and also send on send error (i.e. when *medium* was busy)
@miri64 @jia200x
gnrc_netif_pktq: add dequeue timer
22bc81a
@miri64 @jia200x
tests: provide unittests for gnrc_netif_pktq
8ba9eb5
@jia200x
gnrc_networking: use pktq by default
12f94f7
@jia200x
gnrc_networking: use netdev_ieee802154_submac
3631972
@jia200x jia200x force-pushed the pr/ieee802154_submac branch from 82e0d5a to 3631972 Compare September 4, 2020 09:36
@jia200x jia200x mentioned this pull request Sep 4, 2020
Merged
@jia200x
Copy link
Copy Markdown
Member Author

jia200x commented Oct 1, 2020

the SubMAC got merged. We can close this PoC now!

@jia200x jia200x closed this Oct 1, 2020
@kaspar030 kaspar030 mentioned this pull request Nov 20, 2020
Merged
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@benpicco benpicco Awaiting requested review from benpicco

@maribu maribu Awaiting requested review from maribu

@bergzand bergzand Awaiting requested review from bergzand bergzand is a code owner

@PeterKietzmann PeterKietzmann Awaiting requested review from PeterKietzmann

@miri64 miri64 Awaiting requested review from miri64 miri64 is a code owner

@leandrolanzieri leandrolanzieri Awaiting requested review from leandrolanzieri leandrolanzieri will be requested when the pull request is marked ready for review leandrolanzieri is a code owner

@MrKevinWeiss MrKevinWeiss Awaiting requested review from MrKevinWeiss MrKevinWeiss will be requested when the pull request is marked ready for review MrKevinWeiss is a code owner

@Teufelchen1 Teufelchen1 Awaiting requested review from Teufelchen1 Teufelchen1 will be requested when the pull request is marked ready for review Teufelchen1 is a code owner

@aabadie aabadie Awaiting requested review from aabadie aabadie will be requested when the pull request is marked ready for review aabadie is a code owner

@MichelRottleuthner MichelRottleuthner Awaiting requested review from MichelRottleuthner MichelRottleuthner will be requested when the pull request is marked ready for review MichelRottleuthner is a code owner

Assignees

No one assigned

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

2 participants

@jia200x @miri64