Expand async drops during drop elaboration by cjgillot · Pull Request #156649 · rust-lang/rust

cjgillot · 2026-05-16T18:50:34Z

Async drops implemented by #123948 are expanded in await loops inside StateTransform. I find this odd, as StateTransform is tasked with removing yields from MIR.

This PR moves this async drop -> loop transformation to drop elaboration.

The first few commits create many mir-opt tests to gauge the effect of following commits. I can easily trim those.
The next few commits are from #156422.
Then come a few commits performing elementary simplifications to drop elaboration code.

The most important commit is Expand async drops during drop elaboration. It performs the main change and attempts to document what I had to understand doing that.

The last commit is a cleanup.

This is a large PR, but most of the diff is in generated files. I can easily split them, each commit compiles and passes tests on its own.

cc @zetanumbers

rustbot · 2026-05-16T18:50:37Z

This PR changes MIR

cc @oli-obk, @RalfJung, @JakobDegen, @vakaras

Some changes occurred to MIR optimizations

cc @rust-lang/wg-mir-opt

Some changes occurred to the CTFE / Miri interpreter

cc @rust-lang/miri

This PR changes rustc_public

cc @oli-obk, @celinval, @ouz-a, @makai410

The Cranelift subtree was changed

cc @bjorn3

Some changes occurred to the CTFE machinery

cc @RalfJung, @oli-obk, @lcnr

rustbot · 2026-05-16T18:50:40Z

r? @JohnTitor

rustbot has assigned @JohnTitor.
They will have a look at your PR within the next two weeks and either review your PR or reassign to another reviewer.

Use r? to explicitly pick a reviewer

Why was this reviewer chosen?

The reviewer was selected based on:

Owners of files modified in this PR: compiler
compiler expanded to 73 candidates
Random selection from 16 candidates

bjorn3 · 2026-05-16T19:02:42Z

Is there some relation to #156495 or would that PR be orthogonal to this one?

cjgillot · 2026-05-16T20:06:14Z

Is there some relation to #156495 or would that PR be orthogonal to this one?

It should be orthogonal.

zetanumbers · 2026-05-18T09:21:26Z

uuh
cc @azhogin

JohnTitor · 2026-05-27T10:09:09Z

@rustbot reroll
I'm not familiar with this 🙏

nnethercote · 2026-05-30T19:39:08Z

@cjgillot: looks like this is still evolving. Can you mark it as a draft? Or otherwise make it clear when it is ready for review? Thanks.

cjgillot · 2026-05-30T19:44:14Z

This is ready for review. Just rebasing regularly.

nnethercote · 2026-05-30T20:10:42Z

Ok! Thanks.

elaborate_drop: Avoid as_mut.unwrap dance, empty vectors are cheap. Nit I found while working on rust-lang#156649

nnethercote · 2026-05-31T10:53:10Z

+        patch.apply(&mut body);
+    }
+
+    // We did not bother respectig deref separation, do it here.


*respecting

View changes since the review

nnethercote · 2026-05-31T10:54:04Z

+        ),
+    );
+
+    // We did not bother respectig deref separation, do it here.


*respecting

View changes since the review

We generate a `*(_1.0)` projection in the `needs_sync_drop` case. Ensure that `deref_finder` is called on all paths.

We have `deref_finder` that introduces locals as needed. This allows to avoid bothering when writing the shim.

This makes code generation simpler, as we do not need to verify that successor blocks are only used once. This makes code generation simpler. The resulting MIR increase can easily be simplified by `SimplifyCfg`.

Same rationale as previous commit.

Drop elaboration runs early in the MIR pipeline. That pipeline is designed to work with unresolved function calls. Instance resolution during will call `codegen_select_candidate` when appropriate.

And also reduce the amount of intermediary locals. This shifts code to be mostly in reverse execution order.

This lets `StateTransform` focus on removing `Yield` terminators without adding new ones.

This is now unused.

rustbot · 2026-05-31T17:26:08Z

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.

cjgillot · 2026-05-31T17:54:29Z

@bors try @rust-timer queue

Expand async drops during drop elaboration

rust-bors · 2026-05-31T20:04:04Z

☀️ Try build successful (CI)
Build commit: c26a28f (c26a28fd306aa5dbe8978277159b97eeef3293cd, parent: 1ce45a011a3fe99ab08643d8eb7229bfc3b40bda)

rust-timer · 2026-05-31T20:45:24Z

Finished benchmarking commit (c26a28f): comparison URL.

Overall result: no relevant changes - no action needed

Benchmarking means the PR may be perf-sensitive. Consider adding rollup=never if this change is not fit for rolling up.

@rustbot label: -S-waiting-on-perf -perf-regression

Instruction count

This perf run didn't have relevant results for this metric.

Max RSS (memory usage)

Results (secondary 6.7%)

A less reliable metric. May be of interest, but not used to determine the overall result above.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	6.7%	[6.7%, 6.7%]	1
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-	-	0

Cycles

This perf run didn't have relevant results for this metric.

Binary size

Results (primary -0.2%, secondary -0.3%)

A less reliable metric. May be of interest, but not used to determine the overall result above.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-0.2%	[-0.4%, -0.1%]	28
Improvements ✅ (secondary)	-0.3%	[-0.4%, -0.1%]	25
All ❌✅ (primary)	-0.2%	[-0.4%, -0.1%]	28

Bootstrap: 513.507s -> 510.839s (-0.52%)
Artifact size: 400.69 MiB -> 400.62 MiB (-0.02%)

cjgillot · 2026-05-31T21:34:36Z

@nnethercote I added a few lines in each commit message. I'd like your review on those. And I added an extra commit with comment changes.

nnethercote · 2026-05-31T21:57:51Z

The expanded commit messages look great! Thanks.

@bors r+

rust-bors · 2026-05-31T21:57:53Z

📌 Commit d0e0cea has been approved by nnethercote

It is now in the queue for this repository.

@zetanumbers

Expand async drops during drop elaboration Async drops implemented by #123948 are expanded in await loops inside `StateTransform`. I find this odd, as `StateTransform` is tasked with removing yields from MIR. This PR moves this `async drop` -> loop transformation to drop elaboration. The first few commits create many mir-opt tests to gauge the effect of following commits. I can easily trim those. The next few commits are from #156422. Then come a few commits performing elementary simplifications to drop elaboration code. The most important commit is `Expand async drops during drop elaboration.` It performs the main change and attempts to document what I had to understand doing that. The last commit is a cleanup. This is a large PR, but most of the diff is in generated files. I can easily split them, each commit compiles and passes tests on its own. cc @zetanumbers

rust-bors · 2026-05-31T23:36:24Z

💔 Test for 6387066 failed: CI. Failed job:

auto - dist-loongarch64-musl (web logs, enhanced plaintext logs)

rust-log-analyzer · 2026-05-31T23:41:26Z

A job failed! Check out the build log: (web) (plain enhanced) (plain)

Click to see the possible cause of the failure (guessed by this bot)

rustbot added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. T-compiler Relevant to the compiler team, which will review and decide on the PR/issue. labels May 16, 2026

rustbot assigned JohnTitor May 16, 2026

cjgillot force-pushed the move-async-expand branch from 1ed6528 to 20114b0 Compare May 16, 2026 18:51

This comment has been minimized.

Sign in to view

cjgillot force-pushed the move-async-expand branch from 1a5cdbb to a4336bc Compare May 22, 2026 16:20

This comment has been minimized.

Sign in to view

cjgillot force-pushed the move-async-expand branch from a4336bc to 7fe62e4 Compare May 22, 2026 19:46

rustbot assigned nnethercote and unassigned JohnTitor May 27, 2026

cjgillot force-pushed the move-async-expand branch from 7fe62e4 to a3420ce Compare May 29, 2026 20:24

This comment has been minimized.

Sign in to view

cjgillot force-pushed the move-async-expand branch from a3420ce to 49f7a10 Compare May 30, 2026 15:07

This comment has been minimized.

Sign in to view

cjgillot mentioned this pull request May 30, 2026

elaborate_drop: Avoid as_mut.unwrap dance, empty vectors are cheap. #157162

Merged

This comment has been minimized.

Sign in to view

JonathanBrouwer added a commit to JonathanBrouwer/rust that referenced this pull request May 30, 2026

Rollup merge of rust-lang#157162 - cjgillot:as_mut_unwrap, r=mejrs

d662f5a

elaborate_drop: Avoid as_mut.unwrap dance, empty vectors are cheap. Nit I found while working on rust-lang#156649

nnethercote reviewed May 31, 2026

View reviewed changes

cjgillot added 9 commits May 31, 2026 15:36

Fix deref separation for async drop shim.

25b459e

We generate a `*(_1.0)` projection in the `needs_sync_drop` case. Ensure that `deref_finder` is called on all paths.

Simplify build_adrop_for_coroutine_shim.

4b09215

We have `deref_finder` that introduces locals as needed. This allows to avoid bothering when writing the shim.

Avoid modifying succ block in build_async_drop.

210b8b4

This makes code generation simpler, as we do not need to verify that successor blocks are only used once. This makes code generation simpler. The resulting MIR increase can easily be simplified by `SimplifyCfg`.

Avoid modifying existing blocks.

a352c54

Same rationale as previous commit.

Do not resove trait impl during elaboration.

ed0471b

Drop elaboration runs early in the MIR pipeline. That pipeline is designed to work with unresolved function calls. Instance resolution during will call `codegen_select_candidate` when appropriate.

Simplify type computation.

358881e

And also reduce the amount of intermediary locals. This shifts code to be mostly in reverse execution order.

Expand async drops during drop elaboration.

30782fc

This lets `StateTransform` focus on removing `Yield` terminators without adding new ones.

Remove async_fut.

a27155e

This is now unused.

Move outdated comment.

d0e0cea

cjgillot force-pushed the move-async-expand branch from 49f7a10 to d0e0cea Compare May 31, 2026 17:26

This comment has been minimized.

Sign in to view

rust-bors Bot pushed a commit that referenced this pull request May 31, 2026

Auto merge of #156649 - cjgillot:move-async-expand, r=<try>

c26a28f

Expand async drops during drop elaboration

This comment has been minimized.

Sign in to view

rustbot removed the S-waiting-on-perf Status: Waiting on a perf run to be completed. label May 31, 2026

rust-bors Bot added S-waiting-on-bors Status: Waiting on bors to run and complete tests. Bors will change the label on completion. and removed S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. labels May 31, 2026

This comment has been minimized.

Sign in to view

rust-bors Bot added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. and removed S-waiting-on-bors Status: Waiting on bors to run and complete tests. Bors will change the label on completion. labels May 31, 2026

Uh oh!

Conversation

cjgillot commented May 16, 2026 • edited by rustbot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

rustbot commented May 16, 2026

Uh oh!

rustbot commented May 16, 2026

Uh oh!

bjorn3 commented May 16, 2026

Uh oh!

This comment has been minimized.

cjgillot commented May 16, 2026

Uh oh!

This comment has been minimized.

zetanumbers commented May 18, 2026

Uh oh!

This comment has been minimized.

This comment has been minimized.

JohnTitor commented May 27, 2026

Uh oh!

This comment has been minimized.

This comment has been minimized.

nnethercote commented May 30, 2026

Uh oh!

cjgillot commented May 30, 2026

Uh oh!

nnethercote commented May 30, 2026

Uh oh!

This comment has been minimized.

nnethercote May 31, 2026 • 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!

nnethercote May 31, 2026 • 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 May 31, 2026

Uh oh!

cjgillot commented May 31, 2026

Uh oh!

This comment has been minimized.

This comment has been minimized.

rust-bors Bot commented May 31, 2026

Uh oh!

This comment has been minimized.

rust-timer commented May 31, 2026

Overall result: no relevant changes - no action needed

Instruction count

Max RSS (memory usage)

Cycles

Binary size

Uh oh!

cjgillot commented May 31, 2026

Uh oh!

nnethercote commented May 31, 2026

Uh oh!

rust-bors Bot commented May 31, 2026

Uh oh!

This comment has been minimized.

rust-bors Bot commented May 31, 2026

Uh oh!

rust-log-analyzer commented May 31, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

8 participants

cjgillot commented May 16, 2026 •

edited by rustbot

Loading

nnethercote May 31, 2026 •

edited by rustbot

Loading

nnethercote May 31, 2026 •

edited by rustbot

Loading