Big query system cleanups by nnethercote · Pull Request #152636 · rust-lang/rust

nnethercote · 2026-02-15T02:33:15Z

Recent PRs have moved a lot of code from rustc_query_system to rustc_middle and rustc_query_impl, where this code now has access to TyCtxt, e.g. #152419, #152516. As a result, a lot of abstraction and indirection that existed to work around this limitation is no longer necessary. This PR removes a lot of it.

r? @Zalathar

`rustc_query_system` has been reduced so much that it's no longer needed. This avoids a lot of indirection and abstraction.

By removing the generic `D` parameter from `DepGraph`, `DepGraphData`, `CurrentDepGraph`, `SerializedDepGraph`, `SerializedNodeHeader`, and `EncoderState`.

It's no longer needed.

It's no longer needed now that we can access `TyCtxt` directly.

By moving most of it into `DepKind`, and changing two methods into free functions.

rustbot · 2026-02-15T02:33:20Z

Zalathar is not on the review rotation at the moment.
They may take a while to respond.

nnethercote · 2026-02-15T02:33:37Z

cc @cjgillot @oli-obk @Zoxc

Zalathar · 2026-02-15T02:55:18Z

Let's queue up the precautionary perf run while I'm reviewing the changes.

@bors try @rust-timer queue

Big query system cleanups

Zalathar · 2026-02-15T02:58:57Z

compiler/rustc_query_impl/src/plumbing.rs

-/// Implements [`QueryContext`] for use by [`rustc_query_system`], since that
-/// crate does not have direct access to [`TyCtxt`].


Suggestion: If we're removing this comment, maybe leave behind a FIXME for getting rid of QueryCtxt entirely.

(Though removing QueryCtxt turns out to not be particularly hard, so if we do it next then there's not much value in a FIXME.)

Suggestion: If we're removing this comment, maybe leave behind a FIXME for getting rid of QueryCtxt entirely.

+1, when reviewing I was also wondering why struct QueryCtxt itself wasn't removed.

Zalathar · 2026-02-15T03:12:26Z

Very minor commit style nit: When removing traits, I like to write “Remove trait Foo” to make it immediately obvious that the thing being removed is a trait and not a type.

Zalathar · 2026-02-15T03:21:13Z

compiler/rustc_middle/src/dep_graph/mod.rs

-    }
-
-    fn with_reduced_queries<T>(self, _: impl FnOnce() -> T) -> T;
+pub trait HasDepContext<'tcx>: Copy {


Suggestion: Maybe leave behind a FIXME to look into renaming or removing HasDepContext.

(Though this turns out to also be a fairly easy follow-up.)

Zalathar · 2026-02-15T03:32:47Z

Looks good, all very mechanical and straightforward.

r=me after considering the suggestions, assuming perf is good.

rust-bors · 2026-02-15T05:06:51Z

☀️ Try build successful (CI)
Build commit: 0ede0e7 (0ede0e7634f5edd2cdae2ab91b838980036b30ae, parent: a33907a7a5381473eec8bcfa0c56e05a856a911c)

rust-timer · 2026-02-15T05:47:14Z

Finished benchmarking commit (0ede0e7): comparison URL.

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

Benchmarking this pull request means it may be perf-sensitive – we'll automatically label it not fit for rolling up. You can override this, but we strongly advise not to, due to possible changes in compiler perf.

Next Steps: If you can justify the regressions found in this try perf run, please do so in sufficient writing along with @rustbot label: +perf-regression-triaged. If not, please fix the regressions and do another perf run. If its results are neutral or positive, the label will be automatically removed.

@bors rollup=never
@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)	0.3%	[0.2%, 0.4%]	12
Regressions ❌ (secondary)	0.3%	[0.2%, 0.3%]	8
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-0.3%	[-0.5%, -0.1%]	2
All ❌✅ (primary)	0.3%	[0.2%, 0.4%]	12

Max RSS (memory usage)

Results (primary 0.2%)

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

	mean	range	count
Regressions ❌ (primary)	2.6%	[2.6%, 2.6%]	1
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-2.1%	[-2.1%, -2.1%]	1
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	0.2%	[-2.1%, 2.6%]	2

Cycles

Results (secondary 1.0%)

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)	6.2%	[4.3%, 8.1%]	2
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-2.5%	[-2.8%, -2.3%]	3
All ❌✅ (primary)	-	-	0

Binary size

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

Bootstrap: 483.617s -> 481.529s (-0.43%)
Artifact size: 397.94 MiB -> 395.74 MiB (-0.55%)

nnethercote · 2026-02-15T20:25:32Z

To answer all the above questions: I have a todo list I am working through. I have various follow-up commits already written to remove more stuff. I haven't published those yet because I don't want individual PRs to get too big. I'm not bothering to put "FIXME: remove this" comments on things that I know I will be removing soon. QueryCtxt is definitely on the todo list and won't be difficult. HasDepContext is also on the list but is a little weirder; if it can't be removed it should at least be renamed HasTyCtxt.

I will look at the small perf regressions today. Hopefully they can be fixed with a small tweak to inlining or something like that.

nnethercote · 2026-02-15T22:01:48Z

I investigated the perf results.

Local measurements show similarly small regressions, but the set of affected benchmarks is entirely different. The worst-affected primary benchmark on CI is bitmaps-3.2.1 which doesn't show up in my local regressions list. The worst-affected primary benchmark locally is libc-0.2.172 which doesn't show up on the CI regressions list.
The cachegrind diffs show small perturbations over a range of functions, some a little better, some a little worse. Although the net result is worse, just scanning through the list it's not obvious which direction the net effect is, and there's no clear single cause. The differences are a few million instructions here and there, which are very small numbers. Hash table functions (e.g. insertion) do show up a few times.
Based on the diffs, I think slight changes in inlining is responsible for some or most of this. The code changes are purely "administrative", e.g. changes in types, and a bit of moving code around. There is no change to the amount of actual work (computation, data traversal, etc.) being done.
So I think this is all-but-meaningless perturbation. Only 22 runs out of 584 (3.7%) reach the significance threshold, and the worst regression is only 0.40%.

Based on all this, and the fact that these are good code simplification changes and there are lots more follow-up changes waiting on these, I will declare this perf loss as negligible and acceptable and not worth further investigation.

@rustbot label: +perf-regression-triaged
@bors r=Zalathar

rust-bors · 2026-02-15T22:01:51Z

📌 Commit fcea886 has been approved by Zalathar

It is now in the queue for this repository.

nnethercote · 2026-02-16T00:05:52Z

@bors p=1 (conflict-prone, and blocking follow-up work)

@Zalathar

Big query system cleanups Recent PRs have moved a lot of code from `rustc_query_system` to `rustc_middle` and `rustc_query_impl`, where this code now has access to `TyCtxt`, e.g. #152419, #152516. As a result, a lot of abstraction and indirection that existed to work around this limitation is no longer necessary. This PR removes a lot of it. r? @Zalathar

rust-bors · 2026-02-16T04:12:30Z

💔 Test for a61003d failed: CI. Failed job:

auto - dist-x86_64-msvc (web logs, enhanced plaintext logs)

Zalathar · 2026-02-16T04:13:06Z

A CI runner flaked out for no observable reason.

@bors retry

Zalathar · 2026-02-16T04:13:38Z

@bors p=5

rust-bors · 2026-02-16T08:37:59Z

☀️ Test successful - CI
Approved by: Zalathar
Duration: 4h 17m 23s
Pushing fef627b to main...

github-actions · 2026-02-16T08:40:53Z

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 1396514 (parent) -> fef627b (this PR)

Test differences

Show 6 test diffs

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

Test dashboard

Run

cargo run --manifest-path src/ci/citool/Cargo.toml -- \
    test-dashboard fef627b1ebdc7369ddf8a4031a5d25733ac1fb34 --output-dir test-dashboard

And then open test-dashboard/index.html in your browser to see an overview of all executed tests.

Job duration changes

dist-apple-various: 1h 42m -> 2h 2m (+19.3%)
x86_64-gnu-llvm-21-2: 1h 36m -> 1h 24m (-12.5%)
dist-x86_64-apple: 2h 10m -> 2h 24m (+10.9%)
dist-arm-linux-gnueabi: 1h 30m -> 1h 20m (-10.8%)
dist-aarch64-apple: 2h -> 2h 12m (+9.4%)
x86_64-gnu: 2h 16m -> 2h 4m (-8.6%)
aarch64-apple: 4h 34m -> 4h 10m (-8.5%)
dist-ohos-aarch64: 1h 13m -> 1h 19m (+8.2%)
dist-i686-mingw: 2h 48m -> 2h 35m (-7.9%)
x86_64-msvc-ext3: 1h 58m -> 1h 49m (-7.8%)

How to interpret the job duration changes?

Job durations can vary a lot, based on the actual runner instance
that executed the job, system noise, invalidated caches, etc. The table above is provided
mostly for t-infra members, for simpler debugging of potential CI slow-downs.

Zoxc · 2026-02-16T08:41:12Z

HasDepContext is also on the list but is a little weirder; if it can't be removed it should at least be renamed HasTyCtxt.

#151968 does show how to remove it, but it needs a rebase or two.

rust-timer · 2026-02-16T09:19:13Z

Finished benchmarking commit (fef627b): comparison URL.

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

Our benchmarks found a performance regression caused by this PR.
This might be an actual regression, but it can also be just noise.

Next Steps:

If the regression was expected or you think it can be justified,
please write a comment with sufficient written justification, and add
@rustbot label: +perf-regression-triaged to it, to mark the regression as triaged.
If you think that you know of a way to resolve the regression, try to create
a new PR with a fix for the regression.
If you do not understand the regression or you think that it is just noise,
you can ask the @rust-lang/wg-compiler-performance working group for help (members of this group
were already notified of this PR).

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

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)	0.3%	[0.2%, 0.4%]	10
Regressions ❌ (secondary)	0.2%	[0.1%, 0.3%]	9
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-0.4%	[-0.4%, -0.3%]	4
All ❌✅ (primary)	0.3%	[0.2%, 0.4%]	10

Max RSS (memory usage)

Results (primary 1.7%, secondary -7.5%)

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

	mean	range	count
Regressions ❌ (primary)	1.7%	[1.7%, 1.7%]	1
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-7.5%	[-7.5%, -7.5%]	1
All ❌✅ (primary)	1.7%	[1.7%, 1.7%]	1

Cycles

Results (primary -2.5%, secondary 2.6%)

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)	2.6%	[2.4%, 2.7%]	2
Improvements ✅ (primary)	-2.5%	[-2.5%, -2.4%]	2
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-2.5%	[-2.5%, -2.4%]	2

Binary size

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

Bootstrap: 482.657s -> 481.396s (-0.26%)
Artifact size: 397.96 MiB -> 397.80 MiB (-0.04%)

nnethercote added 5 commits February 15, 2026 13:07

Remove QueryContext.

32e6a1a

`rustc_query_system` has been reduced so much that it's no longer needed. This avoids a lot of indirection and abstraction.

De-genericize the dep graph.

1d83208

By removing the generic `D` parameter from `DepGraph`, `DepGraphData`, `CurrentDepGraph`, `SerializedDepGraph`, `SerializedNodeHeader`, and `EncoderState`.

Remove Deps.

414be2e

It's no longer needed.

Remove DepContext.

7c877d9

It's no longer needed now that we can access `TyCtxt` directly.

Break up DepsType.

fcea886

By moving most of it into `DepKind`, and changing two methods into free functions.

rustbot assigned Zalathar Feb 15, 2026

This comment has been minimized.

Sign in to view

rust-bors bot pushed a commit that referenced this pull request Feb 15, 2026

Auto merge of #152636 - nnethercote:big-cleanups, r=<try>

0ede0e7

Big query system cleanups

rustbot added the S-waiting-on-perf Status: Waiting on a perf run to be completed. label Feb 15, 2026

Zalathar reviewed Feb 15, 2026

View reviewed changes

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 15, 2026

rustbot added the perf-regression-triaged The performance regression has been triaged. label Feb 15, 2026

rust-bors bot added the S-waiting-on-bors Status: Waiting on bors to run and complete tests. Bors will change the label on completion. label Feb 15, 2026

rust-bors bot removed the S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. label Feb 15, 2026

This comment has been minimized.

Sign in to view

rust-bors bot added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. and removed S-waiting-on-bors Status: Waiting on bors to run and complete tests. Bors will change the label on completion. labels Feb 16, 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 16, 2026

Zalathar mentioned this pull request Feb 16, 2026

Rollup of 4 pull requests #152691

Closed

This comment has been minimized.

Sign in to view

rust-bors bot added merged-by-bors This PR was explicitly merged by bors. and removed S-waiting-on-bors Status: Waiting on bors to run and complete tests. Bors will change the label on completion. labels Feb 16, 2026

rust-bors bot merged commit fef627b into rust-lang:main Feb 16, 2026
13 checks passed

rustbot added this to the 1.95.0 milestone Feb 16, 2026

Zoxc mentioned this pull request Feb 16, 2026

Remove generics from DepGraph #151982

Closed

nnethercote deleted the big-cleanups branch February 16, 2026 08:59

Zalathar mentioned this pull request Feb 16, 2026

Remove QueryCtxt and trait HasDepContext #152704

Open

		/// Implements [`QueryContext`] for use by [`rustc_query_system`], since that
		/// crate does not have direct access to [`TyCtxt`].

Uh oh!

Conversation

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

Uh oh!

Uh oh!

rustbot commented Feb 15, 2026

Uh oh!

nnethercote commented Feb 15, 2026

Uh oh!

Zalathar commented Feb 15, 2026

Uh oh!

This comment has been minimized.

This comment has been minimized.

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

petrochenkov Feb 15, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Zalathar commented Feb 15, 2026

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Zalathar commented Feb 15, 2026

Uh oh!

rust-bors bot commented Feb 15, 2026

Uh oh!

This comment has been minimized.

rust-timer commented Feb 15, 2026

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

Uh oh!

nnethercote commented Feb 15, 2026

Uh oh!

nnethercote commented Feb 15, 2026

Uh oh!

rust-bors bot commented Feb 15, 2026

Uh oh!

nnethercote commented Feb 16, 2026

Uh oh!

This comment has been minimized.

rust-bors bot commented Feb 16, 2026

Uh oh!

Zalathar commented Feb 16, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Zalathar commented Feb 16, 2026

Uh oh!

This comment has been minimized.

rust-bors bot commented Feb 16, 2026

Uh oh!

Uh oh!

github-actions bot commented Feb 16, 2026

Test differences

Job duration changes

Uh oh!

Zoxc commented Feb 16, 2026

Uh oh!

rust-timer commented Feb 16, 2026

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

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

6 participants

nnethercote commented Feb 15, 2026 •

edited by rustbot

Loading

petrochenkov Feb 15, 2026 •

edited

Loading

Zalathar commented Feb 16, 2026 •

edited

Loading