Remove redundant call to `check_codegen_attributes_extra` in Inliner by stellanomia · Pull Request #152984 · rust-lang/rust

stellanomia · 2026-02-22T18:25:44Z

Inside try_inlining, check_codegen_attributes is called first. This function internally invokes check_codegen_attributes_extra.
However, immediately after that returns, try_inlining calls check_codegen_attributes_extra again.

First call:

rust/compiler/rustc_mir_transform/src/inline.rs

Lines 800 to 814 in 7ec34de

    
           fn check_codegen_attributes<'tcx, I: Inliner<'tcx>>( 
        
               inliner: &I, 
        
               callsite: &CallSite<'tcx>, 
        
               callee_attrs: &CodegenFnAttrs, 
        
           ) -> Result<(), &'static str> { 
        
               let tcx = inliner.tcx(); 
        
               if let InlineAttr::Never = callee_attrs.inline { 
        
                   return Err("never inline attribute"); 
        
               } 
        
               if let OptimizeAttr::DoNotOptimize = callee_attrs.optimize { 
        
                   return Err("has DoNotOptimize attribute"); 
        
               } 
        
               inliner.check_codegen_attributes_extra(callee_attrs)?;

Second call:

rust/compiler/rustc_mir_transform/src/inline.rs

Lines 598 to 612 in 7ec34de

    
           fn try_inlining<'tcx, I: Inliner<'tcx>>( 
        
               inliner: &I, 
        
               caller_body: &mut Body<'tcx>, 
        
               callsite: &CallSite<'tcx>, 
        
           ) -> Result<std::ops::Range<BasicBlock>, &'static str> { 
        
               let tcx = inliner.tcx(); 
        
               check_mir_is_available(inliner, caller_body, callsite.callee)?; 
        
               let callee_attrs = tcx.codegen_instance_attrs(callsite.callee.def); 
        
               let callee_attrs = callee_attrs.as_ref(); 
        
               check_inline::is_inline_valid_on_fn(tcx, callsite.callee.def_id())?; 
        
               check_codegen_attributes(inliner, callsite, callee_attrs)?; 
        
               inliner.check_codegen_attributes_extra(callee_attrs)?; 
        
               let terminator = caller_body[callsite.block].terminator.as_ref().unwrap();

    // Inside try_inlining:
    check_codegen_attributes(inliner, callsite, callee_attrs)?; // Internally calls `check_codegen_attributes_extra`
    inliner.check_codegen_attributes_extra(callee_attrs)?;      // Called again here

In try_inlining, inliner is held as a shared reference (&I). Since check_codegen_attributes_extra takes &self and does not rely on interior mutability or external state, there does not seem to be any state change between the two calls.

Therefore, the second call appears to be redundant.

Currently, check_codegen_attributes is only called from try_inlining, and check_codegen_attributes_extra appears to be called only from these two locations. It seems reasonable for the check_codegen_attributes wrapper to handle the hook automatically.

…s_extra

rustbot · 2026-02-22T18:25:47Z

Some changes occurred to MIR optimizations

cc @rust-lang/wg-mir-opt

rustbot · 2026-02-22T18:25:49Z

r? @chenyukang

rustbot has assigned @chenyukang.
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 68 candidates
Random selection from 15 candidates

JonathanBrouwer · 2026-02-22T21:34:50Z

@bors r+ rollup

rust-bors · 2026-02-22T21:34:53Z

📌 Commit febd3b9 has been approved by JonathanBrouwer

It is now in the queue for this repository.

…e-check, r=JonathanBrouwer Remove redundant call to `check_codegen_attributes_extra` in Inliner Inside `try_inlining`, `check_codegen_attributes` is called first. This function internally invokes `check_codegen_attributes_extra`. However, immediately after that returns, `try_inlining` calls `check_codegen_attributes_extra` again. First call: https://github.com/rust-lang/rust/blob/7ec34defe9e62a1a6946d3e700b5903d8dc89ece/compiler/rustc_mir_transform/src/inline.rs#L800-L814 Second call: https://github.com/rust-lang/rust/blob/7ec34defe9e62a1a6946d3e700b5903d8dc89ece/compiler/rustc_mir_transform/src/inline.rs#L598-L612 ```rust // Inside try_inlining: check_codegen_attributes(inliner, callsite, callee_attrs)?; // Internally calls `check_codegen_attributes_extra` inliner.check_codegen_attributes_extra(callee_attrs)?; // Called again here ``` In `try_inlining`, inliner is held as a shared reference (`&I`). Since `check_codegen_attributes_extra` takes `&self` and does not rely on interior mutability or external state, there does not seem to be any state change between the two calls. Therefore, the second call appears to be redundant. Currently, `check_codegen_attributes` is only called from `try_inlining`, and `check_codegen_attributes_extra` appears to be called only from these two locations. It seems reasonable for the `check_codegen_attributes` wrapper to handle the hook automatically.

Rollup of 9 pull requests Successful merges: - #152229 (Remove deterministic picking from query cycle handling) - #152970 (extend unpin noalias tests to cover mutable references) - #149783 (stabilize `cfg_select!`) - #151744 (fix refining_impl_trait suggestion with return_type_notation) - #152366 (Add try_shrink_to and try_shrink_to_fit to Vec) - #152640 (Add direct link to rustc-dev-guide in README) - #152963 (Revert "Stabilize `str_as_str`") - #152984 (Remove redundant call to `check_codegen_attributes_extra` in Inliner) - #152987 (Use `HashStable` derive in more places)

Rollup merge of #152984 - stellanomia:remove-redundant-inline-check, r=JonathanBrouwer Remove redundant call to `check_codegen_attributes_extra` in Inliner Inside `try_inlining`, `check_codegen_attributes` is called first. This function internally invokes `check_codegen_attributes_extra`. However, immediately after that returns, `try_inlining` calls `check_codegen_attributes_extra` again. First call: https://github.com/rust-lang/rust/blob/7ec34defe9e62a1a6946d3e700b5903d8dc89ece/compiler/rustc_mir_transform/src/inline.rs#L800-L814 Second call: https://github.com/rust-lang/rust/blob/7ec34defe9e62a1a6946d3e700b5903d8dc89ece/compiler/rustc_mir_transform/src/inline.rs#L598-L612 ```rust // Inside try_inlining: check_codegen_attributes(inliner, callsite, callee_attrs)?; // Internally calls `check_codegen_attributes_extra` inliner.check_codegen_attributes_extra(callee_attrs)?; // Called again here ``` In `try_inlining`, inliner is held as a shared reference (`&I`). Since `check_codegen_attributes_extra` takes `&self` and does not rely on interior mutability or external state, there does not seem to be any state change between the two calls. Therefore, the second call appears to be redundant. Currently, `check_codegen_attributes` is only called from `try_inlining`, and `check_codegen_attributes_extra` appears to be called only from these two locations. It seems reasonable for the `check_codegen_attributes` wrapper to handle the hook automatically.

Rollup of 9 pull requests Successful merges: - rust-lang/rust#152229 (Remove deterministic picking from query cycle handling) - rust-lang/rust#152970 (extend unpin noalias tests to cover mutable references) - rust-lang/rust#149783 (stabilize `cfg_select!`) - rust-lang/rust#151744 (fix refining_impl_trait suggestion with return_type_notation) - rust-lang/rust#152366 (Add try_shrink_to and try_shrink_to_fit to Vec) - rust-lang/rust#152640 (Add direct link to rustc-dev-guide in README) - rust-lang/rust#152963 (Revert "Stabilize `str_as_str`") - rust-lang/rust#152984 (Remove redundant call to `check_codegen_attributes_extra` in Inliner) - rust-lang/rust#152987 (Use `HashStable` derive in more places)

rustc_mir_transform: Remove redundant call to check_codegen_attribute…

febd3b9

…s_extra

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 Feb 22, 2026

rustbot assigned chenyukang Feb 22, 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 Feb 22, 2026

JonathanBrouwer mentioned this pull request Feb 22, 2026

Rollup of 5 pull requests #152992

Closed

jhpratt mentioned this pull request Feb 22, 2026

Rollup of 5 pull requests #152993

Closed

Zalathar mentioned this pull request Feb 23, 2026

Rollup of 9 pull requests #153000

Merged

rust-bors bot merged commit 499fca9 into rust-lang:main Feb 23, 2026
11 checks passed

rustbot added this to the 1.95.0 milestone Feb 23, 2026

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Comments

Remove redundant call to `check_codegen_attributes_extra` in Inliner#152984

Remove redundant call to `check_codegen_attributes_extra` in Inliner#152984
rust-bors[bot] merged 1 commit intorust-lang:mainfrom
stellanomia:remove-redundant-inline-check

stellanomia commented Feb 22, 2026

Uh oh!

rustbot commented Feb 22, 2026

Uh oh!

rustbot commented Feb 22, 2026

Uh oh!

JonathanBrouwer commented Feb 22, 2026

Uh oh!

rust-bors bot commented Feb 22, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

	fn check_codegen_attributes<'tcx, I: Inliner<'tcx>>(
	inliner: &I,
	callsite: &CallSite<'tcx>,
	callee_attrs: &CodegenFnAttrs,
	) -> Result<(), &'static str> {
	let tcx = inliner.tcx();
	if let InlineAttr::Never = callee_attrs.inline {
	return Err("never inline attribute");
	}

	if let OptimizeAttr::DoNotOptimize = callee_attrs.optimize {
	return Err("has DoNotOptimize attribute");
	}

	inliner.check_codegen_attributes_extra(callee_attrs)?;

	fn try_inlining<'tcx, I: Inliner<'tcx>>(
	inliner: &I,
	caller_body: &mut Body<'tcx>,
	callsite: &CallSite<'tcx>,
	) -> Result<std::ops::Range<BasicBlock>, &'static str> {
	let tcx = inliner.tcx();
	check_mir_is_available(inliner, caller_body, callsite.callee)?;

	let callee_attrs = tcx.codegen_instance_attrs(callsite.callee.def);
	let callee_attrs = callee_attrs.as_ref();
	check_inline::is_inline_valid_on_fn(tcx, callsite.callee.def_id())?;
	check_codegen_attributes(inliner, callsite, callee_attrs)?;
	inliner.check_codegen_attributes_extra(callee_attrs)?;

	let terminator = caller_body[callsite.block].terminator.as_ref().unwrap();

Uh oh!

Comments

Conversation

stellanomia commented Feb 22, 2026

Uh oh!

rustbot commented Feb 22, 2026

Uh oh!

rustbot commented Feb 22, 2026

Uh oh!

JonathanBrouwer commented Feb 22, 2026

Uh oh!

rust-bors bot commented Feb 22, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants