receive_any: Use trailing_zeros to improve performance

[showier-drastic]

In current code, receive_any needs to do loop operation to iterate over memory buffer to check whether there's new data which can consume a lot of CPU time.

This code changes this operation to use trailing_zeros to get the first available message directly from NDAT register without any loops.

Thank you!

Thank you for your contribution.
Please make sure that your submission includes the following:

Must

• The code compiles without errors or warnings.
• All tests pass and in the best case you also added new tests.
• cargo +stable fmt was run.
• (No, but it does not relate to my code) cargo +stable clippy yields no warnings.
• Your changes were added to the CHANGELOG.md in the proper section.
• You add a description of your work to this PR.
• You added proper docs (in code, rustdoc and README.md) for your
  newly added features and code.

[Ironedde]

Hello @showier-drastic!

Nice to see you showing interest in this crate.
And sorry for the late response.

I am not the original implementer of this code, but I'll try the best I can to be useful.
I know that this looping might seem a bit excessive, but it was (probably) made this way to enable highest priority (lowest ID) messages to be handled first, regardless of which buffer the message is placed in.

Since I haven't had any practical experience with the dedicated buffers, but instead have only used the FIFOs, do you mind providing some insight into your use-case and how you would like to use these buffers?

The reason I'm asking is because I do see the validity of having a configurable priority order, by which I mean the lowest buffer (in contrast to the lowest ID) gets handled first, but I am not sure if we want to eliminate the possibility for native ID ordering (like now) as well.

Maybe we could introduce some config option to specify the way we want the Iterator for RxDedicatedBuffer to function when reading out messages.