Shallowly bail from `coerce_unsized` more #142941

compiler-errors · 2025-06-24T02:07:52Z

We do a lot of coercion in HIR typeck. Most of the time we're just coercing a type to itself, but we always try coerce_unsized even if it's not necessary.

Let's avoid doing that by adding a fast path to coerce_unsized; see the comment in that function.

compiler-errors · 2025-06-24T02:07:59Z

@bors try @rust-timer queue

bors · 2025-06-24T02:09:10Z

⌛ Trying commit e0b68fe with merge 6a35eeb...

[perf] Shallowly bail from `coerce_unsized` more

compiler-errors · 2025-06-24T02:34:40Z

TODO: test with a fast path for &str: CoerceUnsized<&str>.

bors · 2025-06-24T04:37:02Z

☀️ Try build successful - checks-actions
Build commit: 6a35eeb (6a35eeb37ae3dfbce8f5159463405fb81cddd28a)

rust-timer · 2025-06-24T16:24:17Z

Finished benchmarking commit (6a35eeb): comparison URL.

Overall result: ✅ improvements - no action needed

Benchmarking this pull request means it may be perf-sensitive – we'll automatically label it not fit for rolling up. You can override this, but we strongly advise not to, due to possible changes in compiler perf.

@bors rollup=never
@rustbot label: -S-waiting-on-perf -perf-regression

Instruction count

Our most reliable metric. Used to determine the overall result above. However, even this metric can be noisy.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	0.0%	[0.0%, 0.0%]	1
Improvements ✅ (primary)	-0.7%	[-4.5%, -0.1%]	60
Improvements ✅ (secondary)	-6.7%	[-21.6%, -0.1%]	58
All ❌✅ (primary)	-0.7%	[-4.5%, -0.1%]	60

Max RSS (memory usage)

Results (primary -1.2%, secondary -5.0%)

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)	2.3%	[2.3%, 2.3%]	1
Improvements ✅ (primary)	-1.2%	[-1.2%, -1.2%]	2
Improvements ✅ (secondary)	-5.8%	[-10.6%, -1.6%]	9
All ❌✅ (primary)	-1.2%	[-1.2%, -1.2%]	2

Cycles

Results (primary -4.5%, secondary -15.4%)

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)	-4.5%	[-5.5%, -3.2%]	6
Improvements ✅ (secondary)	-15.4%	[-25.4%, -6.1%]	27
All ❌✅ (primary)	-4.5%	[-5.5%, -3.2%]	6

Binary size

Results (secondary -0.0%)

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
Improvements ✅ (secondary)	-0.0%	[-0.0%, -0.0%]	1
All ❌✅ (primary)	-	-	0

Bootstrap: 688.989s -> 689.016s (0.00%)
Artifact size: 371.96 MiB -> 372.08 MiB (0.03%)

compiler-errors · 2025-06-24T16:36:36Z

holy fuck lol

compiler-errors · 2025-06-24T16:53:04Z

@bors2 try @rust-timer queue

rust-bors · 2025-06-24T16:53:08Z

⌛ Trying commit 3d115ac with merge 1eb1a24…

To cancel the try build, run the command @bors2 try cancel.

[perf] Shallowly bail from `coerce_unsized` more

compiler-errors · 2025-06-24T17:01:49Z

Anyways this is ready to review.

r? @lcnr

rustbot · 2025-06-24T17:01:49Z

r? @petrochenkov

rustbot has assigned @petrochenkov.
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

rust-bors · 2025-06-24T19:22:11Z

☀️ Try build successful (CI)
Build commit: 1eb1a24 (1eb1a24f58bb3d49e9be66d408afa9adf2b30a78, parent: 36b21637e93b038453924d3c66821089e71d8baa)

compiler-errors · 2025-06-24T22:52:07Z

Crashes should be fixed by #142976; they're only coincidentally fixed here by the fast path bc we don't end up trying coerce_unsized.

rust-timer · 2025-06-25T00:47:12Z

Finished benchmarking commit (1eb1a24): comparison URL.

Overall result: ✅ improvements - no action needed

Benchmarking this pull request means it may be perf-sensitive – we'll automatically label it not fit for rolling up. You can override this, but we strongly advise not to, due to possible changes in compiler perf.

@bors rollup=never
@rustbot label: -S-waiting-on-perf -perf-regression

Instruction count

Our most reliable metric. Used to determine the overall result above. However, even this metric can be noisy.

	mean	range	count
Regressions ❌ (primary)	0.2%	[0.2%, 0.2%]	1
Regressions ❌ (secondary)	0.4%	[0.3%, 0.4%]	2
Improvements ✅ (primary)	-0.8%	[-4.5%, -0.1%]	80
Improvements ✅ (secondary)	-5.5%	[-21.6%, -0.0%]	73
All ❌✅ (primary)	-0.8%	[-4.5%, 0.2%]	81

Max RSS (memory usage)

Results (primary -0.9%, secondary -3.8%)

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

	mean	range	count
Regressions ❌ (primary)	1.3%	[1.3%, 1.3%]	1
Regressions ❌ (secondary)	3.9%	[3.9%, 4.0%]	2
Improvements ✅ (primary)	-3.1%	[-3.1%, -3.1%]	1
Improvements ✅ (secondary)	-5.2%	[-10.6%, -1.5%]	11
All ❌✅ (primary)	-0.9%	[-3.1%, 1.3%]	2

Cycles

Results (primary -2.0%, secondary -12.2%)

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)	-2.0%	[-4.1%, -0.8%]	26
Improvements ✅ (secondary)	-12.2%	[-24.0%, -0.9%]	31
All ❌✅ (primary)	-2.0%	[-4.1%, -0.8%]	26

Binary size

This benchmark run did not return any relevant results for this metric.

Bootstrap: 689.383s -> 687.827s (-0.23%)
Artifact size: 371.94 MiB -> 372.03 MiB (0.02%)

lcnr · 2025-06-25T05:54:04Z

compiler/rustc_hir_typeck/src/coercion.rs

@@ -515,11 +515,7 @@ impl<'f, 'tcx> Coerce<'f, 'tcx> {
    ///
    /// [unsized coercion](https://doc.rust-lang.org/reference/type-coercions.html#unsized-coercions)
    #[instrument(skip(self), level = "debug")]
-    fn coerce_unsized(&self, mut source: Ty<'tcx>, mut target: Ty<'tcx>) -> CoerceResult<'tcx> {
-        source = self.shallow_resolve(source);
-        target = self.shallow_resolve(target);


why remove these, are we guaranteed that they are shallow resolved? if so, please convert them into debug asserts

yes, we shallowly resolve at the top of the coerce entrypoint

pls debug assert then

lcnr · 2025-06-25T05:56:45Z

compiler/rustc_hir_typeck/src/coercion.rs

+            | ty::Infer(ty::IntVar(_) | ty::FloatVar(_))
+            | ty::Str
+            | ty::Array(_, _)
+            | ty::Slice(_)


array to slice is always Deref an not CoerceUnsized? 🤔

Slices are the RHS for Unsize, not CoerceUnsized. The latter is for pointers not data.

ah yeah, remember how this works again/looked it up

lcnr · 2025-06-26T07:41:36Z

r=me after adding the debug asserts

compiler-errors · 2025-06-26T16:32:03Z

Added a bunch of debug assertions

@bors r=lcnr

bors · 2025-06-26T16:32:07Z

📌 Commit 0d37587 has been approved by lcnr

It is now in the queue for this repository.

rustbot added the T-compiler Relevant to the compiler team, which will review and decide on the PR/issue. label Jun 24, 2025

This comment has been minimized.

Sign in to view

rustbot added the S-waiting-on-perf Status: Waiting on a perf run to be completed. label Jun 24, 2025

bors added a commit that referenced this pull request Jun 24, 2025

Auto merge of #142941 - compiler-errors:shallow-bail, r=<try>

6a35eeb

[perf] Shallowly bail from `coerce_unsized` more

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 Jun 24, 2025

compiler-errors force-pushed the shallow-bail branch from e0b68fe to 3d115ac Compare June 24, 2025 16:52

This comment has been minimized.

Sign in to view

rust-bors bot added a commit that referenced this pull request Jun 24, 2025

Auto merge of #142941 - compiler-errors:shallow-bail, r=<try>

1eb1a24

[perf] Shallowly bail from `coerce_unsized` more

rustbot added the S-waiting-on-perf Status: Waiting on a perf run to be completed. label Jun 24, 2025

compiler-errors marked this pull request as ready for review June 24, 2025 17:01

rustbot assigned petrochenkov Jun 24, 2025

rustbot added the S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. label Jun 24, 2025

rustbot assigned lcnr and unassigned petrochenkov Jun 24, 2025

This comment has been minimized.

Sign in to view

compiler-errors changed the title ~~[perf] Shallowly bail from coerce_unsized more~~ Shallowly bail from coerce_unsized more Jun 24, 2025

rustbot removed the S-waiting-on-perf Status: Waiting on a perf run to be completed. label Jun 25, 2025

lcnr reviewed Jun 25, 2025

View reviewed changes

compiler-errors added 3 commits June 26, 2025 15:52

Assert shallow resolved args in coerce

8242b55

Shallowly bail from coerce_unsized more

904458f

Add a note and one more fast path

0d37587

compiler-errors force-pushed the shallow-bail branch from 3d115ac to 0d37587 Compare June 26, 2025 15:56

bors 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 Jun 26, 2025

Shallowly bail from coerce_unsized more #142941

Are you sure you want to change the base?

Shallowly bail from coerce_unsized more #142941

Conversation

compiler-errors commented Jun 24, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

compiler-errors commented Jun 24, 2025

Uh oh!

This comment has been minimized.

bors commented Jun 24, 2025

Uh oh!

compiler-errors commented Jun 24, 2025

Uh oh!

This comment has been minimized.

bors commented Jun 24, 2025

Uh oh!

This comment has been minimized.

rust-timer commented Jun 24, 2025

Overall result: ✅ improvements - no action needed

Uh oh!

compiler-errors commented Jun 24, 2025

Uh oh!

compiler-errors commented Jun 24, 2025

Uh oh!

This comment has been minimized.

rust-bors bot commented Jun 24, 2025

Uh oh!

compiler-errors commented Jun 24, 2025

Uh oh!

rustbot commented Jun 24, 2025

Uh oh!

This comment has been minimized.

rust-bors bot commented Jun 24, 2025

Uh oh!

This comment has been minimized.

compiler-errors commented Jun 24, 2025

Uh oh!

rust-timer commented Jun 25, 2025

Overall result: ✅ improvements - no action needed

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

compiler-errors Jun 25, 2025 • 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!

Choose a reason for hiding this comment

Uh oh!

lcnr commented Jun 26, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

compiler-errors commented Jun 26, 2025

Uh oh!

bors commented Jun 26, 2025

Uh oh!

Uh oh!

Shallowly bail from `coerce_unsized` more #142941

Shallowly bail from `coerce_unsized` more #142941

compiler-errors commented Jun 24, 2025 •

edited

Loading

compiler-errors Jun 25, 2025 •

edited

Loading

lcnr commented Jun 26, 2025 •

edited

Loading