Fixed packets not being sent up by CSMA ALOHA on listen state

[dcostan]

While using an Evologics physical layer with CSMA ALOHA on the top, the later blocked incoming packets from a bidirectional communication from being processed. The problem is that the state CSMA_STATE_LISTEN was not handled and so all the incoming packets were discarded.

[fedefava86]

This should not be the case. The case you are mentioning, should be handled by the state "RX_LISTEN". In fact, during Phy2MacStartRx, if state is LISTEN, state is moved to RX_LISTEN, which is then handled in stateRxData method. Could you please verify this?

[dcostan]

Probably my solution is not the right one, but for sure a problem exists.

As it can be seen in the screenshot it happens that all the times in a bidirectional communication the Phy2MacStartRx function detects a Transmit data state so it never changes the state to CSMA_STATE_RX_LISTEN and all the incoming packets are so discarded.

[fedefava86]

No one said the problem does not exists, but we need to carefully debug it and my feeling is that your solution is a work-around. If you are receiving, this means that the previous transmission ended right? So someone should have called endTx and state should have gone to idle or wait_ack if you enabled acknowledgement.
My suggestion is to do the same test and enable the CSMA ALOHA debugging, checking carefully the state change to spot where the FSM is actually breaking.

[dcostan]

At this node there is no reception but the transmission works fine, below I am attaching a screenshot with CSMA ALOHA debug enabled. In any case I agree with you about the fact that my correction is a workaround, I have not analyzed the entire module yet.

I think there is a problem when the node is sending a packet but at the same time another one is arriving, in fact if the Phy2MacStartRx would work correctly the issue could be solved.

[fedefava86]

At this node there is no reception but the transmission works fine, below I am attaching a screenshot with CSMA ALOHA debug enabled. In any case I agree with you about the fact that my correction is a workaround, I have not analyzed the entire module yet. I think there is a problem when the node is sending a packet but at the same time another one is arriving, in fact if the Phy2MacStartRx would work correctly the issue could be solved.

got it, you are 100% right there's a problem, that probably is strictly related with the interaction with EvoLogics itself. I'll open an internal ticket for this. Your solution might be ok, but we'll need to validate it. If this is the case, we'll merge this first. Thanks for pointing out.