Matti Virkkunen

Hi, thanks for the PR and sorry for the late reply! I wonder if this couldn't be done a bit more simply with just a few extra methods in `DescriptorWriter`,...

Do you think it'd make sense to just go all the way and remove the read/write and such methods and just expose the entire endpoint objects via accessors or even...

Doesn't `portable-atomic` already do feature detection and do essentially the same thing automatically? Edit: Okay, I guess it can be a bit overzealous sometimes if the target doesn't support _all_...

If you want to send actual binary data, I think it's better to write directly to the channel. The printing facility is really only meant for printing text. If it's...

This one is a bit annoying to fix due to the fact we can't rely on the availability of atomics on all platforms. I suppose I could live with atomics...

I think the `ptr::copy_nonoverlapping` was meant to try to persuade the compiler to never optimize the writes in any way, but I guess the real way to do that would...

I suppose an acceptable course of action would be to get rid of the volatiles and then just pepper `fence(SeqCst)` in some critical spots (e.g. writing the header, and also...

Does [this commit](e7defb0958fb58a9946c0bac68376b422d16ff4b) look any better? I used atomics for everything that changes on the fly, and volatile writes for all the initialization stuff.

Very mysterious. Which program on your computer are you using to look at the output?

What do you have configured as the runner in `.cargo/config.toml`? Is it probe-rs run or? ``` runner = 'probe-rs run --chip ' ```