spv-in parsing Op::AtomicIIncrement

[schell]

Connections
Partially adresses #4489.

This is mostly for @jimblandy and @cwfitzgerald to ensure that I'm not totally off course on my quest to add atomics.
I'm trying to make my PRs as small as possible to reduce the overhead of reviewing these as they come in :)

Description
Begins work adding support for atomics in naga's SPIR-V frontend.

• adds a temporary feature spv-in-atomics that enables the new feature
• adds support for parsing Op::AtomicIIncrement
  • parse instruction
  • insert result expression
  • insert atomic statement
• adds a lookup map to Frontend to track atomics use for a later pass

This does not generate a valid Module as it predictably fails to type check ;)

Testing
Added a test to parse and validate (with empty flags) a .spv file with an AtomicIIncrement instruction.

Checklist

• Run cargo fmt.
• Run cargo clippy. If applicable, add:
  • --target wasm32-unknown-unknown
  • --target wasm32-unknown-emscripten
• Run cargo xtask test to run tests.
• Add change to CHANGELOG.md. See simple instructions inside file.

[cwfitzgerald]

Great to see this!

Can't comment on the code itself (though it generally looks good), but I don't think we need a special feature for it, this can just be part of the main spirv backend.

[jimblandy]

This looks fine so far. Before we merge this I'd like the test to be turned into a snapshot test.

I'm kind of guessing at a bunch of stuff here so if I say something that doesn't sound right, definitely push back until it makes sense.

[cwfitzgerald]

LGTM

[jimblandy]

Two small issues, but otherwise looks good for now. Once you've addressed these, give me a ping and I'll merge it.

[jimblandy]

This is fine for now, but ideally commit comments should be a clearer explanation of the changes, of the sort one might include in CHANGELOG.txt or something, but with more detail. For example, see: 8d73e5a

[edit: well, maybe that message talks a bit too much about a boring change. But the hope is that at least occasionally the commit message could save someone the time of actually reading the diff.]

[jimblandy]

Thanks very much! Looking forward to the next episode!