Consider structurally impossible Sized predicates in MIR by qaijuang · Pull Request #156609 · rust-lang/rust

qaijuang · 2026-05-15T13:55:11Z

ImpossiblePredicates filters predicates mentioning local generics before invoking the trait solver, which avoids normalization and solver work in generic contexts, thereby skipping predicates such as dyn Trait<T>: Sized, even though they are impossible for every substitution.

This PR adds a narrow structural fast path for positive Sized predicates whose self type is known-unsized without normalization, letting the pass clear unreachable bodies before later MIR opts can evaluate SizedTypeProperties constants such as size_of::<Self>().

rustbot · 2026-05-15T16:05:02Z

Some changes occurred to MIR optimizations

cc @rust-lang/wg-mir-opt

rustbot · 2026-05-15T16:05:05Z

r? @TaKO8Ki

rustbot has assigned @TaKO8Ki.
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, mir, mir-opt
compiler, mir, mir-opt expanded to 73 candidates
Random selection from 17 candidates

qaijuang · 2026-05-15T16:05:45Z

cc @matthiaskrgr

rustbot · 2026-06-13T01:01:27Z

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.

qaijuang · 2026-07-10T07:13:23Z

Hi @TaKO8Ki are there any blockers on this PR that i should know?

qaijuang · 2026-07-13T12:32:52Z

Safe to @rustbot reroll

petrochenkov · 2026-07-13T12:53:15Z

@rustbot reroll

TaKO8Ki · 2026-07-13T13:02:48Z

Sorry for the delay, I completely overlooked this PR and your ping because I got a lot of notifications.

@bors r+ rollup

r? TaKO8Ki

rust-bors · 2026-07-13T13:02:51Z

📌 Commit f86ad47 has been approved by TaKO8Ki

It is now in the queue for this repository.

…predicates, r=TaKO8Ki Consider structurally impossible Sized predicates in MIR Fixesrust-lang#156051 `ImpossiblePredicates` filters predicates mentioning local generics before invoking the trait solver, which avoids normalization and solver work in generic contexts, thereby skipping predicates such as `dyn Trait<T>: Sized`, even though they are impossible for every substitution. This PR adds a narrow structural fast path for positive `Sized` predicates whose self type is known-unsized without normalization, letting the pass clear unreachable bodies before later MIR opts can evaluate `SizedTypeProperties` constants such as `size_of::<Self>()`.

…uwer Rollup of 13 pull requests Successful merges: - #159039 (resolve: fix effective visibilities for items in ambiguous glob sets) - #157706 (Deny `todo!()` in tidy) - #156712 (Pointer authentication config and user facing options) - #158535 (Support `#[track_caller]` on EII declarations) - #158632 (First steps of late-bound turbofishing (place FnDef behind a dummy binder)) - #159002 (Small refactorings in `need_type_info` module) - #159202 (Bump rustc-demangle to 0.1.28) - #159216 (Avoid using probe self_ty for delegation arguments) - #159222 (semicolon_in_expressions_from_macros: Lint on non-local macros too) - #156609 (Consider structurally impossible Sized predicates in MIR) - #158854 (Add `#[rustc_test_entrypoint_marker]`) - #158998 (Some minor ast validation and visiting cleanups) - #159123 (doc: clarify attr parser APIs)

…predicates, r=TaKO8Ki Consider structurally impossible Sized predicates in MIR Fixesrust-lang#156051 `ImpossiblePredicates` filters predicates mentioning local generics before invoking the trait solver, which avoids normalization and solver work in generic contexts, thereby skipping predicates such as `dyn Trait<T>: Sized`, even though they are impossible for every substitution. This PR adds a narrow structural fast path for positive `Sized` predicates whose self type is known-unsized without normalization, letting the pass clear unreachable bodies before later MIR opts can evaluate `SizedTypeProperties` constants such as `size_of::<Self>()`.

…uwer Rollup of 14 pull requests Successful merges: - #159039 (resolve: fix effective visibilities for items in ambiguous glob sets) - #157706 (Deny `todo!()` in tidy) - #158535 (Support `#[track_caller]` on EII declarations) - #158632 (First steps of late-bound turbofishing (place FnDef behind a dummy binder)) - #158846 (Fix unused variable warnings for diverging expressions) - #159002 (Small refactorings in `need_type_info` module) - #159202 (Bump rustc-demangle to 0.1.28) - #159216 (Avoid using probe self_ty for delegation arguments) - #159222 (semicolon_in_expressions_from_macros: Lint on non-local macros too) - #156609 (Consider structurally impossible Sized predicates in MIR) - #157993 (Expose more info about ADTs and functions in rustc_public) - #158854 (Add `#[rustc_test_entrypoint_marker]`) - #158998 (Some minor ast validation and visiting cleanups) - #159123 (doc: clarify attr parser APIs)

…predicates, r=TaKO8Ki Consider structurally impossible Sized predicates in MIR Fixesrust-lang#156051 `ImpossiblePredicates` filters predicates mentioning local generics before invoking the trait solver, which avoids normalization and solver work in generic contexts, thereby skipping predicates such as `dyn Trait<T>: Sized`, even though they are impossible for every substitution. This PR adds a narrow structural fast path for positive `Sized` predicates whose self type is known-unsized without normalization, letting the pass clear unreachable bodies before later MIR opts can evaluate `SizedTypeProperties` constants such as `size_of::<Self>()`.

…uwer Rollup of 15 pull requests Successful merges: - #159039 (resolve: fix effective visibilities for items in ambiguous glob sets) - #157706 (Deny `todo!()` in tidy) - #156047 (Fix trait method resolution on an adjusted never type) - #158535 (Support `#[track_caller]` on EII declarations) - #158632 (First steps of late-bound turbofishing (place FnDef behind a dummy binder)) - #158846 (Fix unused variable warnings for diverging expressions) - #159002 (Small refactorings in `need_type_info` module) - #159202 (Bump rustc-demangle to 0.1.28) - #159216 (Avoid using probe self_ty for delegation arguments) - #156609 (Consider structurally impossible Sized predicates in MIR) - #157993 (Expose more info about ADTs and functions in rustc_public) - #158804 (Clarify `as_uninit_mut` may point to uninitialized memory) - #158854 (Add `#[rustc_test_entrypoint_marker]`) - #158998 (Some minor ast validation and visiting cleanups) - #159123 (doc: clarify attr parser APIs)

…uwer Rollup of 13 pull requests Successful merges: - #157706 (Deny `todo!()` in tidy) - #158535 (Support `#[track_caller]` on EII declarations) - #158632 (First steps of late-bound turbofishing (place FnDef behind a dummy binder)) - #158846 (Fix unused variable warnings for diverging expressions) - #159002 (Small refactorings in `need_type_info` module) - #159202 (Bump rustc-demangle to 0.1.28) - #159216 (Avoid using probe self_ty for delegation arguments) - #156609 (Consider structurally impossible Sized predicates in MIR) - #157993 (Expose more info about ADTs and functions in rustc_public) - #158804 (Clarify `as_uninit_mut` may point to uninitialized memory) - #158854 (Add `#[rustc_test_entrypoint_marker]`) - #158998 (Some minor ast validation and visiting cleanups) - #159123 (doc: clarify attr parser APIs)

Rollup merge of #156609 - qaijuang:mir-opt-impossible-sized-predicates, r=TaKO8Ki Consider structurally impossible Sized predicates in MIR Fixes#156051 `ImpossiblePredicates` filters predicates mentioning local generics before invoking the trait solver, which avoids normalization and solver work in generic contexts, thereby skipping predicates such as `dyn Trait<T>: Sized`, even though they are impossible for every substitution. This PR adds a narrow structural fast path for positive `Sized` predicates whose self type is known-unsized without normalization, letting the pass clear unreachable bodies before later MIR opts can evaluate `SizedTypeProperties` constants such as `size_of::<Self>()`.

…uwer Rollup of 13 pull requests Successful merges: - rust-lang/rust#157706 (Deny `todo!()` in tidy) - rust-lang/rust#158535 (Support `#[track_caller]` on EII declarations) - rust-lang/rust#158632 (First steps of late-bound turbofishing (place FnDef behind a dummy binder)) - rust-lang/rust#158846 (Fix unused variable warnings for diverging expressions) - rust-lang/rust#159002 (Small refactorings in `need_type_info` module) - rust-lang/rust#159202 (Bump rustc-demangle to 0.1.28) - rust-lang/rust#159216 (Avoid using probe self_ty for delegation arguments) - rust-lang/rust#156609 (Consider structurally impossible Sized predicates in MIR) - rust-lang/rust#157993 (Expose more info about ADTs and functions in rustc_public) - rust-lang/rust#158804 (Clarify `as_uninit_mut` may point to uninitialized memory) - rust-lang/rust#158854 (Add `#[rustc_test_entrypoint_marker]`) - rust-lang/rust#158998 (Some minor ast validation and visiting cleanups) - rust-lang/rust#159123 (doc: clarify attr parser APIs)

Kobzol · 2026-07-21T07:48:00Z

@rust-timer build 25028da

For #159246.

rust-timer · 2026-07-21T10:08:06Z

Finished benchmarking commit (25028da): comparison URL.

Overall result: ✅ improvements - no action needed

Benchmarking means the PR may be perf-sensitive. It's automatically marked not fit for rolling up. Overriding is possible but disadvised: it risks changing compiler perf.

@bors rollup=never rustc-perf
@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
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-1.8%	[-2.5%, -0.5%]	14
All ❌✅ (primary)	-	-	0

Max RSS (memory usage)

Results (primary 2.4%, secondary -20.0%)

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

	mean	range	count
Regressions ❌ (primary)	2.4%	[2.4%, 2.4%]	1
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-20.0%	[-23.3%, -11.7%]	13
All ❌✅ (primary)	2.4%	[2.4%, 2.4%]	1

Cycles

Results (secondary -14.1%)

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)	-14.1%	[-20.0%, -7.0%]	13
All ❌✅ (primary)	-	-	0

Binary size

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

Bootstrap: 488.117s -> 486.539s (-0.32%)
Artifact size: 389.87 MiB -> 389.83 MiB (-0.01%)

…uwer Rollup of 13 pull requests Successful merges: - rust-lang/rust#157706 (Deny `todo!()` in tidy) - rust-lang/rust#158535 (Support `#[track_caller]` on EII declarations) - rust-lang/rust#158632 (First steps of late-bound turbofishing (place FnDef behind a dummy binder)) - rust-lang/rust#158846 (Fix unused variable warnings for diverging expressions) - rust-lang/rust#159002 (Small refactorings in `need_type_info` module) - rust-lang/rust#159202 (Bump rustc-demangle to 0.1.28) - rust-lang/rust#159216 (Avoid using probe self_ty for delegation arguments) - rust-lang/rust#156609 (Consider structurally impossible Sized predicates in MIR) - rust-lang/rust#157993 (Expose more info about ADTs and functions in rustc_public) - rust-lang/rust#158804 (Clarify `as_uninit_mut` may point to uninitialized memory) - rust-lang/rust#158854 (Add `#[rustc_test_entrypoint_marker]`) - rust-lang/rust#158998 (Some minor ast validation and visiting cleanups) - rust-lang/rust#159123 (doc: clarify attr parser APIs)

rustbot added S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. T-compiler Relevant to the compiler team, which will review and decide on the PR/issue. labels May 15, 2026

qaijuang marked this pull request as ready for review May 15, 2026 16:04

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 May 15, 2026

rustbot assigned TaKO8Ki May 15, 2026

TaKO8Ki reviewed Jun 9, 2026
View reviewed changes

Comment threadtests/ui/const_prop/issue-102553.rs Outdated

This comment has been minimized.
Sign in to view

Add regression test for impossible dyn Sized predicate
b9a8c80

qaijuang force-pushed the mir-opt-impossible-sized-predicates branch from 5672c89 to eeacc59Compare June 13, 2026 01:01

This comment has been minimized.
Sign in to view

qaijuang added 2 commits June 12, 2026 21:35

Consider structurally impossible Sized predicates
2df8f1f

isolate regression test
f86ad47

qaijuang force-pushed the mir-opt-impossible-sized-predicates branch from eeacc59 to f86ad47Compare June 13, 2026 01:36

rustbot assigned petrochenkov and unassigned TaKO8Ki Jul 13, 2026

rustbot assigned oli-obk and unassigned petrochenkov Jul 13, 2026

rustbot assigned TaKO8Ki Jul 13, 2026

rust-borsBot added the S-waiting-on-bors Status: Waiting on bors to run and complete tests. Bors will change the label on completion. label Jul 13, 2026

rustbot unassigned oli-obk Jul 13, 2026

rust-borsBot removed the S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. label Jul 13, 2026

TaKO8Ki approved these changes Jul 13, 2026
View reviewed changes

JonathanBrouwer mentioned this pull request Jul 13, 2026
Rollup of 13 pull requests #159231
Closed

JonathanBrouwer mentioned this pull request Jul 13, 2026
Rollup of 14 pull requests #159236
Closed

JonathanBrouwer mentioned this pull request Jul 13, 2026
Rollup of 15 pull requests #159240
Closed

JonathanBrouwer mentioned this pull request Jul 13, 2026
Rollup of 13 pull requests #159246
Merged

rust-borsBot merged commit 3198c7a into rust-lang:mainJul 13, 2026
13 checks passed

rustbot added this to the 1.99.0 milestone Jul 13, 2026

qaijuang deleted the mir-opt-impossible-sized-predicates branch July 13, 2026 23:40

This comment has been minimized.
Sign in to view

Uh oh!

Uh oh!

Conversation

qaijuang commented May 15, 2026• edited by rustbot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

rustbot commented May 15, 2026

Uh oh!

rustbot commented May 15, 2026

Uh oh!

qaijuang commented May 15, 2026

Uh oh!

Uh oh!

This comment has been minimized.

rustbot commented Jun 13, 2026

Uh oh!

This comment has been minimized.

qaijuang commented Jul 10, 2026

Uh oh!

qaijuang commented Jul 13, 2026

Uh oh!

petrochenkov commented Jul 13, 2026

Uh oh!

TaKO8Ki commented Jul 13, 2026

Uh oh!

rust-borsBot commented Jul 13, 2026

Uh oh!

Uh oh!

Kobzol commented Jul 21, 2026

Uh oh!

This comment has been minimized.

rust-timer commented Jul 21, 2026

Overall result: ✅ improvements - no action needed

Instruction count

Max RSS (memory usage)

Cycles

Binary size

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

8 participants

qaijuang commented May 15, 2026•
edited by rustbot
Loading