Implement EII for statics#154193

JonathanBrouwer · 2026-03-22T00:01:27Z

This PR implements EII for statics. I've tried to mirror the implementation for functions in a few places, this causes some duplicate code but I'm also not really sure whether there's a clean way to merge the implementations.

This does not implement defaults for static EIIs yet, I will do that in a followup PR

JonathanBrouwer · 2026-03-26T15:49:50Z

compiler/rustc_hir_analysis/src/check/compare_eii.rs

+
+    // Finally, resolve all regions. This catches wily misuses of
+    // lifetime parameters.
+    let errors = infcx.resolve_regions(external_impl, param_env, []);


I pass an empty array for the assumed_wf_tys here, I can't really find what this argument is for, is this correct?

Passing an empty array in the resolve_regions call inside compare_eii_function_types causes no tests to fail, so would also like to add a test for that once I understand what it does

seems to be related to the "implied bounds hack" which seems to be part of how we resolve GATs in traits. Since this code is mostly adapted from method comparison in traits that makes sense to me, but I don't think we need it for EIIs. @lcnr could you confirm that?

see https://rustc-dev-guide.rust-lang.org/return-position-impl-trait-in-trait.html#implied-bounds-in-rpitit-hidden-type-inference

relevant for static FOO<'a, 'b>: &'a &'b () = &&(); to add an assumption that 'b: 'a holds.

Given that statics don't have any where-bounds/generics anyways, this shouldn't matter

JonathanBrouwer

r? @jdonszelmann

This code touches quite a few areas of the compiler I'm not super familiar yet with, so please review carefully whether I'm doing any stupid things :3

View changes since this review

rustbot · 2026-03-26T15:51:18Z

jdonszelmann is currently at their maximum review capacity.
They may take a while to respond.

rustbot · 2026-03-26T15:51:33Z

Some changes occurred in compiler/rustc_passes/src/check_attr.rs

cc @jdonszelmann

Some changes occurred in compiler/rustc_attr_parsing

cc @jdonszelmann

The parser was modified, potentially altering the grammar of (stable) Rust
which would be a breaking change.

cc @fmease

Some changes occurred in src/tools/clippy

cc @rust-lang/clippy

compiler/rustc_hir_analysis/src/check/compare_eii.rs

jdonszelmann · 2026-03-30T15:29:15Z

compiler/rustc_codegen_llvm/src/mono_item.rs

+            assert!(
+                self.instances.borrow_mut().insert(instance, lldecl).is_none(),
+                "An instance was already present for {instance:?}"
+            );


I'm not sure I understand when this can fail

or why this sanity check is necessary. I believe it may be, but the comment doesn't explain what scenario this matters in or what bug we previously had with this.

I've added explanation to the comment, the assert is indeed simply a sanity check

compiler/rustc_hir_analysis/src/check/compare_eii.rs

bjorn3 · 2026-03-30T15:41:30Z

The fact that eii_impls is a list seems to imply that a single item can define multiple EIIs at the same time. Wouldn't that result in multiple statics aliasing each other with this PR? Especially with static mut that would be rather dangerous. But even without, I'm not sure that is not UB.

JonathanBrouwer · 2026-04-04T14:50:33Z

The fact that eii_impls is a list seems to imply that a single static can define multiple EIIs at the same time. Wouldn't that result in multiple statics aliasing each other with this PR? Especially with static mut that would be rather dangerous. But even without, I'm not sure that is not UB.

It is the other way around, a single item can implement multiple EIIs.

The definition of the EII is an extern static that "aliases" the static that implements the EII, I am far from an expert of rusts memory model so please correct me if I'm wrong, but I believe that should be fine?
Also if both the definition and implementation are static mut this should also be fine? Again far from an expert, please correct me

TLDR: It is exactly as sound as external statics are, since it's basically just an external static

rustbot · 2026-04-04T15:31:36Z

This PR was rebased onto a different main commit. Here's a range-diff highlighting what actually changed.

Rebasing is a normal part of keeping PRs up to date, so no action is needed—this note is just to help reviewers.

JonathanBrouwer · 2026-04-04T15:31:49Z

@rustbot ready

jdonszelmann · 2026-04-05T08:23:50Z

it's correct that with EIIs you can get multiple statics aliasing eachother. We might want to disallow static mut EIIs completely?

jdonszelmann · 2026-04-05T08:25:17Z

even with a single EII,

#[eii(a)]
static FOO: usize;

#[a]
static BAR: usize = 3;

fn main() {
    assert_eq!(
        &BAR as *const _ as usize, 
        &FOO as *const _ as usize
    );
}

JonathanBrouwer · 2026-04-05T08:54:07Z

The thing is that FOO is not a static, it's desugared to an extern static, is it still a problem then?

unsafe extern "Rust" {
    safe static FOO: usize;
}

mod imp {
    #[unsafe(no_mangle)]
    pub static FOO: usize = 3;
}

fn main() {
    assert_eq!(
        &FOO as *const _ as usize,
        &imp::FOO as *const _ as usize
    );
}

i.e. Is this code unsound?

JonathanBrouwer · 2026-04-05T09:02:17Z

Disallowing static mut EIIs makes sense to me, but more because they're hard to use correctly than that they're inherently unsound

bjorn3 · 2026-04-05T09:23:55Z

Currently if you have two extern statics with different symbol names, they would never alias. I can very much imagine that there will be crates which break if two EIIs defined by said crate alias. Breaking both as in causing UB and as in misbehaving without UB. Also it requires having support for aliasing at the object file level, while if each static can define only a single EII, then we can get away without any aliasing support at the object file level. The definition can just have the same symbol name as the EII. And if the EII has a default, this default can either be put in a separate object file only linked when the default is used or be defined as weak symbol.

JonathanBrouwer · 2026-04-05T09:39:12Z

Right, I agree this is also undesirable
So what I will do is:

Disallow multiple eii impls on 1 static
Disallow eii on mut statics

JonathanBrouwer · 2026-04-05T09:51:58Z

@rustbot author

rustbot · 2026-04-05T09:52:11Z

Reminder, once the PR becomes ready for a review, use @rustbot ready.

jdonszelmann · 2026-04-05T10:00:53Z

That doesn't help though, EIIs inherently generate two symbol names for one static. Even if one is extern. And the definition can't get the same name. It should be available under the name given to the implementation as well imo. The same is true for functions atm.

jdonszelmann · 2026-04-05T10:01:17Z

multiple impls does make that worse, but it's already a problem to begin with

jdonszelmann · 2026-04-05T10:01:29Z

also, could you add a test that both have the same address?

bjorn3 · 2026-04-05T11:02:16Z

That doesn't help though, EIIs inherently generate two symbol names for one static. Even if one is extern.

There is no inherent reason why the EII declaration and the impl have to use different symbol names afaik. If you only defining a single EII per static, then tcx.symbol_name() on the impl can return the symbol name of the declaration.

JonathanBrouwer added the F-extern_item_impls `#![feature(extern_item_impls)]` label Mar 22, 2026

JonathanBrouwer force-pushed the external-static branch from e3f6d77 to 30c073d Compare March 22, 2026 00:02

This comment has been minimized.

Sign in to view

JonathanBrouwer force-pushed the external-static branch from 30c073d to dd4e148 Compare March 22, 2026 00:12

rustbot added the T-clippy Relevant to the Clippy team. label Mar 22, 2026

JonathanBrouwer force-pushed the external-static branch 5 times, most recently from c7aa3cf to d9d3b13 Compare March 26, 2026 15:48

JonathanBrouwer commented Mar 26, 2026

View reviewed changes

rustbot assigned jdonszelmann Mar 26, 2026

JonathanBrouwer changed the title ~~[VERY WIP] Implement EII for statics~~ Implement EII for statics Mar 26, 2026

JonathanBrouwer marked this pull request as ready for review March 26, 2026 15:51

rustbot added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. and removed S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. labels Mar 26, 2026

This comment has been minimized.

Sign in to view

JonathanBrouwer force-pushed the external-static branch from d9d3b13 to 7b3a361 Compare March 26, 2026 16:01

JonathanBrouwer commented Mar 27, 2026

View reviewed changes

compiler/rustc_hir_analysis/src/check/compare_eii.rs Outdated Show resolved Hide resolved

JonathanBrouwer force-pushed the external-static branch from 7b3a361 to 2cb2f25 Compare March 27, 2026 21:21

jdonszelmann reviewed Mar 30, 2026

View reviewed changes

compiler/rustc_hir_analysis/src/check/compare_eii.rs Outdated Show resolved Hide resolved

jdonszelmann mentioned this pull request Apr 2, 2026

Tracking Issue for externally implementable items #125418

Open

48 tasks

Add eii_impls argument to StaticItem

66ca41c

JonathanBrouwer added 3 commits April 4, 2026 17:31

Analysis for externally implementable statics

fc134a4

Fix codegen for add_static_aliases

ed6fbb6

Uitests for external statics

d2c3767

JonathanBrouwer force-pushed the external-static branch from 2cb2f25 to d2c3767 Compare April 4, 2026 15:31

rustbot removed the S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. label Apr 5, 2026

rustbot added the S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. label Apr 5, 2026

Uh oh!

Conversation

JonathanBrouwer commented Mar 22, 2026 • edited by rustbot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

This comment has been minimized.

JonathanBrouwer Mar 26, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

JonathanBrouwer Mar 27, 2026

Choose a reason for hiding this comment

Uh oh!

jdonszelmann Apr 5, 2026

Choose a reason for hiding this comment

Uh oh!

jdonszelmann Apr 5, 2026

Choose a reason for hiding this comment

Uh oh!

lcnr Apr 6, 2026

Choose a reason for hiding this comment

Uh oh!

JonathanBrouwer left a comment • edited by rustbot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

rustbot commented Mar 26, 2026

Uh oh!

rustbot commented Mar 26, 2026

Uh oh!

This comment has been minimized.

Uh oh!

jdonszelmann Mar 30, 2026

Choose a reason for hiding this comment

Uh oh!

jdonszelmann Mar 30, 2026

Choose a reason for hiding this comment

Uh oh!

JonathanBrouwer Apr 4, 2026

Choose a reason for hiding this comment

Uh oh!

Uh oh!

bjorn3 commented Mar 30, 2026

Uh oh!

JonathanBrouwer commented Apr 4, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

rustbot commented Apr 4, 2026

Uh oh!

JonathanBrouwer commented Apr 4, 2026

Uh oh!

jdonszelmann commented Apr 5, 2026

Uh oh!

jdonszelmann commented Apr 5, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

JonathanBrouwer commented Apr 5, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

JonathanBrouwer commented Apr 5, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

bjorn3 commented Apr 5, 2026

Uh oh!

JonathanBrouwer commented Apr 5, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

JonathanBrouwer commented Apr 5, 2026

Uh oh!

rustbot commented Apr 5, 2026

Uh oh!

jdonszelmann commented Apr 5, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

jdonszelmann commented Apr 5, 2026

Uh oh!

jdonszelmann commented Apr 5, 2026

Uh oh!

bjorn3 commented Apr 5, 2026

Uh oh!

JonathanBrouwer commented Mar 22, 2026 •

edited by rustbot

Loading

JonathanBrouwer Mar 26, 2026 •

edited

Loading

JonathanBrouwer left a comment •

edited by rustbot

Loading

JonathanBrouwer commented Apr 4, 2026 •

edited

Loading

jdonszelmann commented Apr 5, 2026 •

edited

Loading

JonathanBrouwer commented Apr 5, 2026 •

edited

Loading

JonathanBrouwer commented Apr 5, 2026 •

edited

Loading

JonathanBrouwer commented Apr 5, 2026 •

edited

Loading

jdonszelmann commented Apr 5, 2026 •

edited

Loading