Allow passing a channel end for UIFM Set Address switch message

[larry-xmos]

Saves one channel end in endpoint 0 thread

[xross]

Odd version no?

[xross]

Hi Larry, why does this save a chan-end? Because the compiler only does one getr for the used end? Do you not get a build warning about only passing the channel to one core?

[xross]

Resolved conflicts

[larry-xmos]

Odd version no?

Yes, it's based on version 0.1.0 that comes with VocalFusion 2.2.2. The PR was meant as a placeholder, and it might make most sense to redo it (which I would hope to be able to do soon).

Furthermore, main XUD thread uses an extra channel end the same way, but in multiple places, making it difficult to hoist out.

[xross]

Okay, thanks

[larry-xmos]

@ross, is there any chance you will be using this soon? If yes, I would consider bringing it up to date. Or is it best closed?

[xross]

Hi Larry, this looks like useful work, I'd like to keep it open until I (or someone else) has time to review.

[larry-xmos]

I'd like to keep it open until I (or someone else) has time to review

OK, I'd always wanted to come back and look for a way to reduce the number of functions that have to change without using a global variable

[mbanth]

Given @larry-xmos schedule, it would be good to have a decision by 1 July, both for this PR and for lib_xua PR-77. That way he has time to bring them up-to-date if we decide to merge them in.

[xross]

Note, this is probably not relevant for XS3 since we no longer store a device address in HW. I believe all accesses to HW are now from the XUD core.

[larry-xmos]

On the basis that Phoenix no longer has to do the Set Address calls, I close this item