remoteproc: Call mbox_client_txdone() after mbox_send_message()

[ukleinek]

It's necessary to call one of the txdone functions to get the mailbox state machine moving.

Currently it still works by some luck: The message being sent is NULL. The framework code thus sets chan->active_req = NULL to signal that this message is currently sending out. On the next call to mbox_send_message() chan->active_req == NULL is interpreted differently though, that is as not busy and the next message is processed.

The inner workings of the mailbox subsystem will change in commit c58e945 ("mailbox: Fix NULL message support in mbox_send_message()") that is part of v7.1-rc2 where this inconsistency is somewhat fixed. (Then -1 is the value that cannot be properly sent instead of (the more usual) NULL. The upside is that this -1 is properly catched as bogous.)

To prepare updating the code to after v7.1-rc2 add a call to mbox_client_txdone() which is already now required conceptually, but deviating doesn't hurt while only sending NULL.

PR Type

• Bug fix (a change that fixes an issue, well somewhat)
• New feature (a change that adds new functionality)
• Breaking change (a change that affects other repos or cause CIs to fail)

PR Checklist

• I have conducted a self-review of my own code changes
• I have compiled my changes, including the documentation
• I have tested the changes on the relevant hardware
• I have updated the documentation outside this repo accordingly
• I have provided links for the relevant upstream lore

[ukleinek]

I failed to mention: I'm not entirely sure this is correct. If it is the mailbox API has a strange consumer contract. But it makes the echo example work on v7.2.