Check signature WF when lowering MIR body

[compiler-errors]

Alternative to #137233.

#137233 (comment)

Fixes #137186

We do this check in mir_drops_elaborated_and_const_checked and not during mir_promoted because that may result in borrowck cycles if WF requires looking into an opaque hidden type. This causes some TAIT tests to fail unnecessarily.

r? lcnr

try-job: test-various

[rustbot]

Some changes occurred to MIR optimizations

cc @rust-lang/wg-mir-opt

[lcnr]

@bors r+ rollup

[bors]

📌 Commit bb7b912 has been approved by lcnr

It is now in the queue for this repository.

[workingjubilee]

#137339 (comment)
@bors r- rollup=maybe
@bors try

[bors]

⌛ Trying commit bb7b912 with merge 7527550...

[bors]

💔 Test failed - checks-actions

[jieyouxu]

Hm, the wasm tests may have symbol redefinition problems? 🤔

[lcnr]

I think that's a proper error, we somehow depend on the value of a static when checking that it's well-formed 🤔 this is surprising to me (and even more surprising that we don't have a separate UI test which triggers this)

[compiler-errors]

@bors try

[bors]

⌛ Trying commit eb94a63 with merge e7267fa...

[compiler-errors]

It's

rust/compiler/rustc_hir_analysis/src/check/mod.rs

Lines 148 to 195 in 71e06b9

	fn maybe_check_static_with_link_section(tcx: TyCtxt<'_>, id: LocalDefId) {
	// Only restricted on wasm target for now
	if !tcx.sess.target.is_like_wasm {
	return;
	}
	// If `#[link_section]` is missing, then nothing to verify
	let attrs = tcx.codegen_fn_attrs(id);
	if attrs.link_section.is_none() {
	return;
	}
	// For the wasm32 target statics with `#[link_section]` other than `.init_array`
	// are placed into custom sections of the final output file, but this isn't like
	// custom sections of other executable formats. Namely we can only embed a list
	// of bytes, nothing with provenance (pointers to anything else). If any
	// provenance show up, reject it here.
	// `#[link_section]` may contain arbitrary, or even undefined bytes, but it is
	// the consumer's responsibility to ensure all bytes that have been read
	// have defined values.
	//
	// The `.init_array` section is left to go through the normal custom section code path.
	// When dealing with `.init_array` wasm-ld currently has several limitations. This manifests
	// in workarounds in user-code.
	//
	// * The linker fails to merge multiple items in a crate into the .init_array section.
	// To work around this, a single array can be used placing multiple items in the array.
	// #[link_section = ".init_array"]
	// static FOO: [unsafe extern "C" fn(); 2] = [ctor, ctor];
	// * Even symbols marked used get gc'd from dependant crates unless at least one symbol
	// in the crate is marked with an `#[export_name]`
	//
	// Once `.init_array` support in wasm-ld is complete, the user code workarounds should
	// continue to work, but would no longer be necessary.
	if let Ok(alloc) = tcx.eval_static_initializer(id.to_def_id())
	&& alloc.inner().provenance().ptrs().len() != 0
	&& attrs
	.link_section
	.map(|link_section| !link_section.as_str().starts_with(".init_array"))
	.unwrap()
	{
	let msg = "statics with a custom `#[link_section]` must be a \
	simple list of bytes on the wasm target with no \
	extra levels of indirection such as references";
	tcx.dcx().span_err(tcx.def_span(id), msg);
	}
	}

I'll move that out of WF and into the regular "checks".

[compiler-errors]

@bors try

[compiler-errors]

@bors r-

[alexcrichton]

There was a bit of discussion on Zulip on what's going on here wasm-wise, and while it looks like you've diagnosed and solved it I wanted to lend a helping hand in that respect where if anything else wasm-related comes up here feel free to ping me and I can try to help discover what's going on if any extra help is needed

[compiler-errors]

A side-effect of this is that consts are only required to typeck if they're being documented.

[lcnr]

in rustdoc

[compiler-errors]

oh 💀 yes in roblox

[lcnr]

why this pub?

[compiler-errors]

so we can keep the error

[lcnr]

@bors rü

[lcnr]

@bors r+

[bors]

📌 Commit 0baee24 has been approved by lcnr

It is now in the queue for this repository.

[apiraino]

Since nobody excluded a beta backport, I'll nominate it myself (hope it's fine), this patch is fixing an ICE, I think it could be "a nice to have" but no particular urgency otherwise, afaics

@rustbot label +beta-nominated