Remove `backend_optimization_level`. by veluca93 · Pull Request #165586 · rust-lang/rust

veluca93 · 2026-11-07T09:53:44Z

Turns out that all uses were effectively equivalent to using sess.opts.optimize.

In particular, the output of that function was only ever passed to target_machine_factory, which ignores it on anything but llvm, and on llvm it only passed it to to_llvm_opt_settings, where the difference between Size/SizeMin/More is only the second return value, which is immediately dropped. So replacing all uses of backend_optimization_level with sess.opts.optimize is effectively equivalent.

LLM usage: I asked $LLM to review the commit, and it didn't complain about anything. Also $LLM initially figured out (and I confirmed) that backend_optimization_level was effectively a no-op.

rustbot · 2026-11-07T09:53:47Z

rustc_codegen_gcc is developed in its own repository. If possible, consider making this change to rust-lang/rustc_codegen_gcc instead.

cc @antoyo, @GuillaumeGomez

rustc_codegen_cranelift is developed in its own repository. If possible, consider making this change to rust-lang/rustc_codegen_cranelift instead.

cc @bjorn3

rustbot · 2026-11-07T09:53:50Z

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: codegen, compiler
codegen, compiler expanded to 79 candidates
Random selection from 23 candidates

RalfJung · 2026-11-07T11:13:31Z

Cc @nikic

bjorn3 · 2026-11-07T11:23:56Z

-    ) -> TargetMachineFactoryFn<Self> {
-        back::write::target_machine_factory(sess, optlvl)
+    fn target_machine_factory(&self, sess: &Session) -> TargetMachineFactoryFn<Self> {
+        back::write::target_machine_factory(sess, sess.opts.optimize)


Can this be changed to only accept sess?

View changes since the review

I was considering that, but there is a "create_informational_machine" (or thereabouts) function that passes OptLevel::No, and I was unsure whether passing the opt level there would cause perf regressions.

That function is only used for fetching information from LLVM, not for any actual compilation, so I think it would't affect performance at all, but I don't know if LLVM does some work on creating the optimization pipeline in advance when creating a target machine.

Yep, that's exactly my doubt too :-)

nikic

The original intent was presumably to hook this up to the PassBuilder opt level, but given that this didn't actually happen and would not do anything on current LLVM anymore, dropping it entirely makes sense.

View changes since this review

JohnTitor · 2026-11-07T22:31:10Z

r? nikic

rustbot · 2026-11-10T00:13:05Z

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.

Turns out that all uses were effectively equivalent to using sess.opts.optimize.

nikic · 2026-11-10T08:46:41Z

@bors r+

rust-bors · 2026-11-10T08:46:45Z

📌 Commit b52077c has been approved by nikic

It is now in the queue for this repository.

Remove `backend_optimization_level`. Turns out that all uses were effectively equivalent to using sess.opts.optimize. In particular, the output of that function was only ever passed to target_machine_factory, which ignores it on anything but llvm, and on llvm it only passed it to to_llvm_opt_settings, where the difference between Size/SizeMin/More is only the second return value, which is immediately dropped. So replacing all uses of backend_optimization_level with sess.opts.optimize is effectively equivalent. LLM usage: I asked $LLM to review the commit, and it didn't complain about anything. Also $LLM initially figured out (and I confirmed) that backend_optimization_level was effectively a no-op.

…uwer Rollup of 16 pull requests Successful merges: - #165575 (Upgrade bundled musl to 1.2.6) - #163992 (add `core::arch::{x86, x86_64}::x87_f80`) - #165117 (std: use skip_unsupported_miri_tests to control when tests should be skipped in Miri due to missing shims) - #165458 (fix incorrect byte count suffix on MSVC) - #165487 (Migrate built-in macros `pattern_type` & `view_type` to macros 2.0 that expand to internal syntax) - #165490 (Mark the compile unit for `-Zdebuginfo-for-profiling`) - #165586 (Remove `backend_optimization_level`.) - #165646 (fix NumScalableVectors::from_field_count) - #165655 (Unignore passing GCC backend tests) - #165740 (miri subtree update) - #164559 (Windows: fix Dir::remove and Dir::rename following symlinks) - #165682 (interpret: ensure newly created unions have uninit in their padding) - #165693 (Move `check_attr` to attribute parsing) - #165715 (windows: check for null before wide encoding) - #165717 (mpsc_sync::stream_send_recv_stress: wait for all threads to finish) - #165733 (docs(num): remove needless whitespaces) Failed merges: - #165727 (Normalize "beta release channel" in `const-supertrait-unstable` test) - #153261 (Partially stabilize `ptr_alignment_type` as `alignment_type`) - #164218 (Tweak private re-export errors to point at correct span and have more context) - #165709 (remove 'x miri' smoke test from PR CI)

rustbot assigned JohnTitor Nov 7, 2026

bjorn3 reviewed Nov 7, 2026

View reviewed changes

nikic approved these changes Nov 7, 2026

View reviewed changes

Comment thread compiler/rustc_codegen_llvm/src/back/write.rs Resolved

veluca93 force-pushed the cleanup-opt-speed branch from 04c5c5d to 43aa445 Compare November 7, 2026 13:25

This comment has been minimized.

Sign in to view

JohnTitor added the llm-assisted An LLM-assisted PR as defined by the LLM policy. Requires ahead-of-time consent by assignee. label Nov 7, 2026

rustbot assigned nikic and unassigned JohnTitor Nov 7, 2026

veluca93 force-pushed the cleanup-opt-speed branch from 43aa445 to 17a8df9 Compare November 10, 2026 00:13

This comment has been minimized.

Sign in to view

Remove backend_optimization_level.

b52077c

Turns out that all uses were effectively equivalent to using sess.opts.optimize.

veluca93 force-pushed the cleanup-opt-speed branch from 17a8df9 to b52077c Compare November 10, 2026 00:34

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 Nov 10, 2026

JonathanBrouwer mentioned this pull request Nov 10, 2026

Rollup of 14 pull requests #165742

Closed

JonathanBrouwer mentioned this pull request Nov 10, 2026

Rollup of 16 pull requests #165747

Open

Conversation

veluca93 commented Nov 7, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

rustbot commented Nov 7, 2026

Uh oh!

rustbot commented Nov 7, 2026

Uh oh!

RalfJung commented Nov 7, 2026

Uh oh!

bjorn3 Nov 7, 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!

veluca93 Nov 7, 2026

Choose a reason for hiding this comment

Uh oh!

bjorn3 Nov 7, 2026

Choose a reason for hiding this comment

Uh oh!

veluca93 Nov 7, 2026

Choose a reason for hiding this comment

Uh oh!

nikic 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!

Uh oh!

This comment has been minimized.

JohnTitor commented Nov 7, 2026

Uh oh!

rustbot commented Nov 10, 2026

Uh oh!

This comment has been minimized.

nikic commented Nov 10, 2026

Uh oh!

rust-bors Bot commented Nov 10, 2026

Uh oh!

This branch has not been deployed

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

7 participants

veluca93 commented Nov 7, 2026 •

edited

Loading

bjorn3 Nov 7, 2026 •

edited by rustbot

Loading

nikic left a comment •

edited by rustbot

Loading