Remove InstanceKind::generates_cgu_internal_copy by saethlin · Pull Request #136410 · rust-lang/rust

saethlin · 2025-02-01T23:03:32Z

This PR should not contain any behavior changes. Before this PR, the logic for selecting instantiation mode is spread across all of

instantiation_mode
cross_crate_inlinable
generates_cgu_internal_copy
requires_inline

The last two of those functions are not well-designed. The function that actually decides if we generate a CGU-internal copy is instantiation_mode, not generates_cgu_internal_copy. The function requires_inline documents that it is about the LLVM inline attribute and that it is a hint. The LLVM attribute is called inlinehint, this function is also used by other codegen backends, and since it is part of instantiation mode selection it is not a hint.

The goal of this PR is to start cleaning up the logic into a sequence of checks that have a more logical flow and are easier to customize in the future (to do things like improve incrementality or improve optimizations without causing obscure linker errors because you forgot to update another part of the compiler).

rustbot · 2025-02-01T23:03:40Z

r? @wesleywiser

rustbot has assigned @wesleywiser.
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

saethlin · 2025-02-01T23:03:48Z

@bors try @rust-timer queue

… r=<try> Remove InstanceKind::generates_cgu_internal_copy This is on top of rust-lang#136394 for now. `InstanceKind::generates_cgu_internal_copy` was called in two places: 1. By `reachable_non_generics`, but in that case the caller was guaranteed to not be generic and therefore we always fall through to just calling `cross_crate_inlinable`. 2. By `MonoItem::instantiation_mode`, but in that case the *only* thing that it does is some very complicated logic for selecting the instantiation mode for drop glue, and only very recently do we have a single codegen test for this. So I've touched up the logic in `cross_crate_inlinable` so that we don't try to claim that `drop_in_place` isn't cross-crate-inlinable, because it clearly is. Now off to perf, and if we get regressions I'll start reintroducing the logic about drop glue but, but this time into `MonoItem::instantiation_mode` where it always should have been.

bors · 2025-02-01T23:09:14Z

⌛ Trying commit 6e9db01 with merge 69c35dc...

bors · 2025-02-02T00:58:25Z

☀️ Try build successful - checks-actions
Build commit: 69c35dc (69c35dcc9e16d4d54ae9eab591266a070edf7235)

rust-timer · 2025-02-02T03:10:14Z

Finished benchmarking commit (69c35dc): comparison URL.

Overall result: ❌✅ regressions and improvements - please read the text below

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 the most reliable metric that we have; it was used to determine the overall result at the top of this comment. However, even this metric can sometimes exhibit noise.

	mean	range	count
Regressions ❌ (primary)	11.9%	[0.2%, 79.0%]	40
Regressions ❌ (secondary)	8.9%	[0.4%, 37.8%]	17
Improvements ✅ (primary)	-4.9%	[-9.4%, -0.7%]	4
Improvements ✅ (secondary)	-0.4%	[-0.4%, -0.4%]	1
All ❌✅ (primary)	10.4%	[-9.4%, 79.0%]	44

Max RSS (memory usage)

Results (primary 1.7%, secondary 1.5%)

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)	3.5%	[2.3%, 5.6%]	6
Regressions ❌ (secondary)	2.4%	[1.2%, 3.8%]	4
Improvements ✅ (primary)	-3.7%	[-3.8%, -3.6%]	2
Improvements ✅ (secondary)	-2.1%	[-2.1%, -2.1%]	1
All ❌✅ (primary)	1.7%	[-3.8%, 5.6%]	8

Cycles

Results (primary 15.7%, secondary 13.9%)

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)	17.3%	[1.0%, 87.6%]	29
Regressions ❌ (secondary)	13.9%	[2.5%, 34.3%]	11
Improvements ✅ (primary)	-8.2%	[-8.5%, -8.0%]	2
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	15.7%	[-8.5%, 87.6%]	31

Binary size

Results (primary 4.0%, secondary -0.2%)

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)	4.5%	[0.0%, 18.4%]	44
Regressions ❌ (secondary)	2.3%	[0.1%, 4.9%]	14
Improvements ✅ (primary)	-0.6%	[-0.8%, -0.6%]	5
Improvements ✅ (secondary)	-11.7%	[-12.6%, -9.8%]	3
All ❌✅ (primary)	4.0%	[-0.8%, 18.4%]	49

Bootstrap: 777.275s -> 785.803s (1.10%)
Artifact size: 328.69 MiB -> 328.60 MiB (-0.03%)

saethlin · 2025-02-08T02:39:58Z

@bors try @rust-timer queue

bors · 2025-02-08T02:41:10Z

⌛ Trying commit 2e0908b with merge 53335c7...

… r=<try> Remove InstanceKind::generates_cgu_internal_copy This is on top of rust-lang#136394 for now. `InstanceKind::generates_cgu_internal_copy` was called in two places: 1. By `reachable_non_generics`, but in that case the caller was guaranteed to not be generic and therefore we always fall through to just calling `cross_crate_inlinable`. 2. By `MonoItem::instantiation_mode`, but in that case the *only* thing that it does is some very complicated logic for selecting the instantiation mode for drop glue, and only very recently do we have a single codegen test for this. So I've touched up the logic in `cross_crate_inlinable` so that we don't try to claim that `drop_in_place` isn't cross-crate-inlinable, because it clearly is. Now off to perf, and if we get regressions I'll start reintroducing the logic about drop glue but, but this time into `MonoItem::instantiation_mode` where it always should have been.

bors · 2025-02-08T04:31:56Z

☀️ Try build successful - checks-actions
Build commit: 53335c7 (53335c70a1c5cf7c1e824d8b4da0285df47f6196)

rust-timer · 2025-02-08T05:47:52Z

Finished benchmarking commit (53335c7): comparison URL.

Overall result: ❌✅ regressions and improvements - please read the text below

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 the most reliable metric that we have; it was used to determine the overall result at the top of this comment. However, even this metric can sometimes exhibit noise.

	mean	range	count
Regressions ❌ (primary)	10.7%	[0.3%, 95.3%]	24
Regressions ❌ (secondary)	2.9%	[0.5%, 12.9%]	18
Improvements ✅ (primary)	-3.4%	[-11.1%, -0.2%]	7
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	7.5%	[-11.1%, 95.3%]	31

Max RSS (memory usage)

Results (primary -2.0%)

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)	2.1%	[2.1%, 2.1%]	1
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-3.4%	[-3.9%, -2.8%]	3
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-2.0%	[-3.9%, 2.1%]	4

Cycles

Results (primary 14.1%, secondary 4.8%)

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)	18.3%	[0.8%, 105.6%]	15
Regressions ❌ (secondary)	5.9%	[2.4%, 11.7%]	8
Improvements ✅ (primary)	-6.8%	[-10.5%, -1.7%]	3
Improvements ✅ (secondary)	-4.2%	[-4.2%, -4.2%]	1
All ❌✅ (primary)	14.1%	[-10.5%, 105.6%]	18

Binary size

Results (primary -0.1%, secondary 2.0%)

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)	0.5%	[0.0%, 1.3%]	27
Regressions ❌ (secondary)	2.0%	[0.1%, 4.9%]	17
Improvements ✅ (primary)	-0.8%	[-1.0%, -0.0%]	21
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-0.1%	[-1.0%, 1.3%]	48

Bootstrap: 778.956s -> 782.985s (0.52%)
Artifact size: 329.06 MiB -> 329.02 MiB (-0.01%)

saethlin · 2025-02-08T14:44:31Z

codegen tests are expected to fail

@bors try @rust-timer queue

bors · 2025-02-08T23:35:35Z

☀️ Try build successful - checks-actions
Build commit: 6a752a8 (6a752a87a5c108106cfdd90e2ff6ee784585c55d)

rust-timer · 2025-02-09T00:53:20Z

Finished benchmarking commit (6a752a8): comparison URL.

Overall result: ❌ regressions - please read the text below

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 the most reliable metric that we have; it was used to determine the overall result at the top of this comment. However, even this metric can sometimes exhibit noise.

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

Max RSS (memory usage)

Results (secondary -0.4%)

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)	1.7%	[1.7%, 1.7%]	1
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-2.4%	[-2.4%, -2.4%]	1
All ❌✅ (primary)	-	-	0

Cycles

Results (primary 3.6%, secondary -3.3%)

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)	3.6%	[3.6%, 3.6%]	1
Regressions ❌ (secondary)	2.1%	[2.1%, 2.1%]	1
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-6.0%	[-8.3%, -3.7%]	2
All ❌✅ (primary)	3.6%	[3.6%, 3.6%]	1

Binary size

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

Bootstrap: 779.565s -> 781.025s (0.19%)
Artifact size: 329.13 MiB -> 329.08 MiB (-0.01%)

saethlin · 2025-02-09T01:02:38Z

@rustbot label: +perf-regression-triaged
bimodal benchmark, not a regression.

saethlin · 2025-03-13T04:13:13Z

r? compiler

compiler-errors · 2025-03-24T23:38:46Z

                return None;
            }

-            // Functions marked with #[inline] are codegened with "internal"


Why'd you remove the comment? Seems valid still to mention the interaction w/ inline.

I removed it because the logic for the inline attribute and extern indicators is in cross_crate_inlinable. I can put back a comment if you want 🤷 but one of the purposes of tcx.cross_crate_inlinable is to be the single point in the compiler that juggles those interactions.

compiler-errors · 2025-03-24T23:40:04Z

+fn opt_incr_drop_glue_mode<'tcx>(tcx: TyCtxt<'tcx>, ty: Ty<'tcx>) -> InstantiationMode {
+    // Non-ADTs can't have a Drop impl. This case is mostly hit by closures whose captures require
+    // dropping.
+    let Some(adt_def) = ty.ty_adt_def() else {


nit: I prefer let ty::Adt(adt_def, _) = ty.kind() over ty_adt_def.

I agree that's better.

compiler-errors · 2025-03-24T23:48:11Z

+
+    // Types that don't have a direct Drop impl, but have fields that require dropping.
+    let Some(dtor) = adt_def.destructor(tcx) else {
+        if adt_def.is_enum() {


Do you know what motivates this choice of making enums act differently than structs? I mean, both of them have fields after all.

I know something. I've added a comment that explains what I know and references the PR that landed the logic that I'm rewriting.

compiler-errors · 2025-03-24T23:49:10Z

r=me after nits or not

rustbot · 2025-03-25T00:30:45Z

Some changes occurred in compiler/rustc_codegen_ssa

cc @WaffleLapkin

compiler-errors · 2025-03-25T02:12:47Z

@bors r+

bors · 2025-03-25T02:12:49Z

📌 Commit 817e2c5 has been approved by compiler-errors

It is now in the queue for this repository.

bors · 2025-03-25T06:36:45Z

⌛ Testing commit 817e2c5 with merge 7d49ae9...

bors · 2025-03-25T09:44:54Z

☀️ Test successful - checks-actions
Approved by: compiler-errors
Pushing 7d49ae9 to master...

github-actions · 2025-03-25T09:46:42Z

What is this?

This is an experimental post-merge analysis report that shows differences in test outcomes between the merged PR and its parent PR.

Comparing e61403a (parent) -> 7d49ae9 (this PR)

Test differences

Show 4 test diffs

Additionally, 4 doctest diffs were found. These are ignored, as they are noisy.

Job group index

rust-timer · 2025-03-25T11:01:50Z

Finished benchmarking commit (7d49ae9): comparison URL.

Overall result: no relevant changes - no action needed

@rustbot label: -perf-regression

Instruction count

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

Max RSS (memory usage)

Results (primary 2.8%, secondary 0.4%)

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%	[2.8%, 2.8%]	1
Regressions ❌ (secondary)	2.7%	[2.6%, 2.8%]	3
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-2.9%	[-3.3%, -2.6%]	2
All ❌✅ (primary)	2.8%	[2.8%, 2.8%]	1

Cycles

Results (secondary 4.9%)

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)	4.9%	[2.8%, 6.6%]	3
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-	-	0

Binary size

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

Bootstrap: 776.063s -> 777.441s (0.18%)
Artifact size: 365.84 MiB -> 365.81 MiB (-0.01%)

rustbot assigned wesleywiser Feb 1, 2025

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 1, 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 Feb 1, 2025

This comment has been minimized.

Sign in to view

rustbot added perf-regression Performance regression. and removed S-waiting-on-perf Status: Waiting on a perf run to be completed. labels Feb 2, 2025

saethlin force-pushed the clean-up-cgu-internal-copy branch from 6e9db01 to d22354c Compare February 8, 2025 02:02

This comment has been minimized.

Sign in to view

saethlin force-pushed the clean-up-cgu-internal-copy branch 2 times, most recently from 8a8bb0c to 2e0908b Compare February 8, 2025 02:39

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 Feb 8, 2025

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 Feb 8, 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 Feb 8, 2025

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 Feb 9, 2025

saethlin marked this pull request as ready for review February 9, 2025 00:59

rustbot added the perf-regression-triaged The performance regression has been triaged. label Feb 9, 2025

rustbot assigned estebank and unassigned wesleywiser Mar 13, 2025

compiler-errors reviewed Mar 24, 2025

View reviewed changes

compiler-errors assigned compiler-errors and unassigned estebank Mar 24, 2025

Remove InstanceKind::generates_cgu_internal_copy

817e2c5

saethlin force-pushed the clean-up-cgu-internal-copy branch from 2c32926 to 817e2c5 Compare March 25, 2025 00:30

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 Mar 25, 2025

bors added the merged-by-bors This PR was explicitly merged by bors. label Mar 25, 2025

bors merged commit 7d49ae9 into rust-lang:master Mar 25, 2025

rustbot added this to the 1.87.0 milestone Mar 25, 2025

rustbot removed the perf-regression Performance regression. label Mar 25, 2025

saethlin deleted the clean-up-cgu-internal-copy branch May 4, 2025 13:12

Uh oh!

Conversation

saethlin commented Feb 1, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

rustbot commented Feb 1, 2025

Uh oh!

saethlin commented Feb 1, 2025

Uh oh!

This comment has been minimized.

bors commented Feb 1, 2025

Uh oh!

bors commented Feb 2, 2025

Uh oh!

This comment has been minimized.

rust-timer commented Feb 2, 2025

Overall result: ❌✅ regressions and improvements - please read the text below

Uh oh!

This comment has been minimized.

saethlin commented Feb 8, 2025

Uh oh!

This comment has been minimized.

bors commented Feb 8, 2025

Uh oh!

bors commented Feb 8, 2025

Uh oh!

This comment has been minimized.

rust-timer commented Feb 8, 2025

Overall result: ❌✅ regressions and improvements - please read the text below

Uh oh!

saethlin commented Feb 8, 2025

Uh oh!

This comment has been minimized.

bors commented Feb 8, 2025

Uh oh!

This comment has been minimized.

rust-timer commented Feb 9, 2025

Overall result: ❌ regressions - please read the text below

Uh oh!

saethlin commented Feb 9, 2025

Uh oh!

saethlin commented Mar 13, 2025

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!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

compiler-errors commented Mar 24, 2025

Uh oh!

rustbot commented Mar 25, 2025

Uh oh!

compiler-errors commented Mar 25, 2025

Uh oh!

bors commented Mar 25, 2025

Uh oh!

bors commented Mar 25, 2025

Uh oh!

bors commented Mar 25, 2025

Uh oh!

github-actions Bot commented Mar 25, 2025

Test differences

Uh oh!

rust-timer commented Mar 25, 2025

Overall result: no relevant changes - no action needed

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

saethlin commented Feb 1, 2025 •

edited

Loading