esp-storage: Fix incorrect usage of MaybeUninit

[DBLouis]

Submission Checklist 📝

• I have updated existing examples or added new ones (if applicable).
• I have used cargo xtask fmt-packages command to ensure that all changed code is formatted correctly.
• My changes were added to the CHANGELOG.md in the proper section.
• My changes are in accordance to the esp-rs API guidelines

Extra:

• I have read the CONTRIBUTING.md guide and followed its instructions.

Pull Request Details 📖

Description

I found incorrect usage of MaybeUninit while browsing the source that might cause UB. Specifically, calling assume_init_mut on a MaybeUninit that has not been initialized ref.

Testing

WIP

[DBLouis]

I am not sure how to run the tests, can someone help please

[Dominaezzz]

You need to add an entry to the changelog for CI to continue.

[DBLouis]

You need to add an entry to the changelog for CI to continue.

Do I need to bump the version then?

[Dominaezzz]

No, just add an entry for your PR

[DBLouis]

I just renamed the commit

[jessebraham]

Sorry for the delay, thanks for the PR. This looks fine to me but I would like either @bjoernQ or @MabezDev to take a look as well.

[bjoernQ]

Maybe I'm missing something but is the words field ever really used? (Yes, by the accessors but those seem to be unused?)

[MabezDev]

If we do end up sticking with the union, we also need to make sure its #[repr(C, align(4))]. I checked on the rust playground, and it does currently have an align of 4 anyways, but I think that it's best to be explicit

[DBLouis]

I don't think specifying the alignment is needed according to this: https://doc.rust-lang.org/reference/type-layout.html#reprc-unions

[DBLouis]

Maybe I'm missing something but is the words field ever really used? (Yes, by the accessors but those seem to be unused?)

I have a following PR coming that uses those. I split in two to make more focused PR.

[ProfFan]

Wonder what's the status on this? Just checking in :)

[DBLouis]

Wonder what's the status on this? Just checking in :)

I've been busy, I finish this soon 🙂

[bjoernQ]

Converted to draft for now - set it to Read-for-Review when it's updated

[jessebraham]

Just checking in here, @DBLouis would you like one of us to take over here to wrap this up? Or do you expect to be able to get back to this in the next couple weeks? No worries either way, would just like to get this across the finish line :)

[DBLouis]

Just checking in here, @DBLouis would you like one of us to take over here to wrap this up? Or do you expect to be able to get back to this in the next couple weeks? No worries either way, would just like to get this across the finish line :)

Yes it is probably better that someone takes over, I will not have time soon, life got in the way :)

[MabezDev]

I think the easier solution to get the perf required and to avoid unsafe is to store the flash buffer inside FlashStorage itself? This way it's only initialized once, and this will probably reduce stack usage too.

@bjoernQ are you aware of any reason not to do it like this instead?