Don't walk the bodies of free constants for reachability.

[oli-obk]

follow-up to #122371

cc #119214

This avoids codegening items (e.g. functions) that are only used during const eval, but do not reach their final constant value (e.g. via function pointers).

r? @tmiasko

[rustbot]

Some changes occurred to the CTFE / Miri engine

cc @rust-lang/miri

These commits modify the Cargo.lock file. Unintentional changes to Cargo.lock can be introduced when switching branches and rebasing PRs.

If this was unintentional then you should revert the changes before this PR is merged.
Otherwise, you can ignore this comment.

[oli-obk]

@bors try @rust-timer queue

[bors]

⌛ Trying commit d46c98d with merge 5620969...

[fmease]

What about generic_const_items? :P Or is that not what you meant?

[oli-obk]

Yea, those xD. I keep forgetting they already exist. Will add a test that hits my new span_bug

[bors]

☀️ Try build successful - checks-actions
Build commit: 5620969 (5620969908da123892bbe342730e38a9a6141180)

[rust-timer]

Finished benchmarking commit (5620969): comparison URL.

Overall result: ❌✅ regressions and improvements - ACTION NEEDED

Benchmarking this pull request likely means that it is perf-sensitive, so we're automatically marking it as not fit for rolling up. While you can manually mark this PR as fit for rollup, we strongly recommend not doing so since this PR may lead to changes in compiler perf.

Next Steps: If you can justify the regressions found in this try perf run, please indicate this with @rustbot label: +perf-regression-triaged along with sufficient written justification. If you cannot justify the regressions please fix the regressions and do another perf run. If the next run shows neutral or positive results, the label will be automatically removed.

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

Instruction count

This is a highly reliable metric that was used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	2.3%	[2.0%, 2.7%]	3
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-1.2%	[-1.6%, -0.3%]	4
All ❌✅ (primary)	2.3%	[2.0%, 2.7%]	3

Max RSS (memory usage)

Results

This is a less reliable metric that may be of interest but was not used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-2.7%	[-4.3%, -0.9%]	6
Improvements ✅ (secondary)	-2.3%	[-3.8%, -0.8%]	26
All ❌✅ (primary)	-2.7%	[-4.3%, -0.9%]	6

Cycles

Results

This is a less reliable metric that may be of interest but was not used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	3.8%	[3.1%, 4.5%]	3
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	3.8%	[3.1%, 4.5%]	3

Binary size

Results

This is a less reliable metric that may be of interest but was not used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-0.0%	[-0.1%, -0.0%]	6
Improvements ✅ (secondary)	-0.7%	[-3.7%, -0.0%]	6
All ❌✅ (primary)	-0.0%	[-0.1%, -0.0%]	6

Bootstrap: 669.602s -> 670.27s (0.10%)
Artifact size: 310.86 MiB -> 310.86 MiB (0.00%)

[tmiasko]

Left a question in a review comment, but other than that r=me. Thanks.

[oli-obk]

@bors r=tmiasko

[bors]

📌 Commit 66cca33 has been approved by tmiasko

It is now in the queue for this repository.

[oli-obk]

huh... this still worked yesterday and is failing now...

[tmiasko]

Minimized reproducer (doesn't require changes from this pull request):

$ cat a.rs
pub static A: fn() -> fn() = {
    const {
        #[inline(never)]
        fn f() {}
        || -> fn() { f }
    }
};
$ cat b.rs
static A: fn() -> fn () = a::A;
fn main() { A()() }
$ rustc a.rs --crate-type=lib -Zprint-mono-items=lazy -O
MONO_ITEM fn <{closure@a.rs:5:9: 5:19} as std::ops::FnOnce<()>>::call_once - shim @@ a.a67665f9196cd4be-cgu.0[Internal]
MONO_ITEM fn A::f @@ a.a67665f9196cd4be-cgu.0[Internal]
MONO_ITEM fn A::{closure#0} @@ a.a67665f9196cd4be-cgu.0[Internal]
MONO_ITEM static A @@ a.a67665f9196cd4be-cgu.0[External]
$ rustc b.rs --crate-type=bin --extern a -L. 
error: missing optimized MIR for an item in the crate `a`
  |
note: missing optimized MIR for this item (was the crate `a` compiled with `--emit=metadata`?)
 --> /tmp/a.rs:4:9
  |
4 |         fn f() {}
  |         ^^^^^^

error: aborting due to 1 previous error
$ rustc --version
rustc 1.80.0-nightly (7c52d2db6 2024-06-03)

[oli-obk]

oof, I hope this isn't hitting stable next week. Tho it would be easy to fix

[oli-obk]

@bors r=tmiasko

[bors]

📌 Commit a0358f4 has been approved by tmiasko

It is now in the queue for this repository.

[bors]

⌛ Testing commit a0358f4 with merge d0ccb54...

[bors]

☀️ Test successful - checks-actions
Approved by: tmiasko
Pushing d0ccb54 to master...

[rust-timer]

Finished benchmarking commit (d0ccb54): comparison URL.

Overall result: ❌ regressions - ACTION NEEDED

Next Steps: If you can justify the regressions found in this perf run, please indicate this with @rustbot label: +perf-regression-triaged along with sufficient written justification. If you cannot justify the regressions please open an issue or create a new PR that fixes the regressions, add a comment linking to the newly created issue or PR, and then add the perf-regression-triaged label to this PR.

@rustbot label: +perf-regression
cc @rust-lang/wg-compiler-performance

Instruction count

This is a highly reliable metric that was used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	1.8%	[0.2%, 2.6%]	4
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	1.8%	[0.2%, 2.6%]	4

Max RSS (memory usage)

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

Cycles

Results (primary 2.8%, secondary -2.1%)

This is a less reliable metric that may be of interest but was not used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	2.8%	[2.3%, 3.2%]	3
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-2.1%	[-2.1%, -2.1%]	1
All ❌✅ (primary)	2.8%	[2.3%, 3.2%]	3

Binary size

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

Bootstrap: 667.497s -> 669.842s (0.35%)
Artifact size: 319.04 MiB -> 319.03 MiB (-0.01%)

[Kobzol]

This had a few instruction count wins, and also binary size wins before merge, but no binary size wins after merge. Was that expected? In any case, the regression is small and only on a single benchmark, so I don't need that we need to investigate further, just wanted to ask, especially about the missing binary size wins.

[RalfJung]

I don't think I understand why being a free constant vs an associated constant makes a difference here. Even for free constants we store their MIR, not their computed result, in the rmeta, right?

[RalfJung]

Ah I think I get it... I have clarified the comments in #126259.

[oli-obk]

Yea, we could evaluate associated constants, often even generic ones can be evaluated to a concrete value.

[RalfJung]

Sure, but that would be a pretty fundamental departure from our current evaluation strategy.

[Kobzol]

Marking as triaged after a discussion with Oli, the small perf. regression is expected.

@rustbot label: +perf-regression-triaged