Skip to content

perf(language_server): use Arc<RwLock> instead of Mutex for workspace workers #11328

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Conversation

Sysix
Copy link
Member

@Sysix Sysix commented May 27, 2025

Mutex will block others requests. RwLock + Arc is faster.
I asked ChatGPT about it:

🔒 So the Real Options Are:
✅ Arc<Mutex<Vec>>

Use this if:

All accesses (read and write) are relatively balanced or infrequent.

You want simplicity and don't expect much read contention.

✅ Arc<RwLock<Vec>>

Use this if:

Most access is reading, and you want to allow multiple concurrent reads.

Writes (e.g. adding/removing workers) are infrequent.

In this case, RwLock may improve performance under heavy read concurrency (e.g., lots of requests querying the current workspace list).

@github-actions github-actions bot added A-editor Area - Editor and Language Server C-performance Category - Solution not expected to change functional behavior, only performance labels May 27, 2025
Copy link
Member Author

Sysix commented May 27, 2025


How to use the Graphite Merge Queue

Add either label to this PR to merge it via the merge queue:

  • 0-merge - adds this PR to the back of the merge queue
  • hotfix - for urgent hot fixes, skip the queue and merge this PR next

You must have a Graphite account in order to use the merge queue. Sign up using this link.

An organization admin has enabled the Graphite Merge Queue in this repository.

Please do not merge from GitHub as this will restart CI on PRs being processed by the merge queue.

This stack of pull requests is managed by Graphite. Learn more about stacking.

Copy link

codspeed-hq bot commented May 27, 2025

CodSpeed Instrumentation Performance Report

Merging #11328 will not alter performance

Comparing 05-27-perf_language_server_use_arc_rwlock_instead_of_mutex_for_workspace_workers (0ed6c1a) with main (14cb3c7)

Summary

✅ 38 untouched benchmarks

@Sysix Sysix marked this pull request as ready for review May 27, 2025 17:49
@Sysix Sysix requested a review from camc314 as a code owner May 27, 2025 17:49
@camc314 camc314 added the 0-merge Merge with Graphite Merge Queue label May 27, 2025
Copy link
Contributor

camc314 commented May 27, 2025

Merge activity

…pace workers (#11328)

`Mutex` will block others requests. `RwLock` + `Arc` is faster.
I asked ChatGPT about it:

🔒 So the Real Options Are:
✅ Arc<Mutex<Vec<WorkspaceWorker>>>

Use this if:

    All accesses (read and write) are relatively balanced or infrequent.

    You want simplicity and don't expect much read contention.

✅ Arc<RwLock<Vec<WorkspaceWorker>>>

Use this if:

    Most access is reading, and you want to allow multiple concurrent reads.

    Writes (e.g. adding/removing workers) are infrequent.

In this case, RwLock may improve performance under heavy read concurrency (e.g., lots of requests querying the current workspace list).
@graphite-app graphite-app bot force-pushed the 05-27-perf_language_server_use_arc_rwlock_instead_of_mutex_for_workspace_workers branch from c5b74fd to 0ed6c1a Compare May 27, 2025 19:17
@graphite-app graphite-app bot merged commit 0ed6c1a into main May 27, 2025
25 checks passed
@graphite-app graphite-app bot deleted the 05-27-perf_language_server_use_arc_rwlock_instead_of_mutex_for_workspace_workers branch May 27, 2025 19:21
@graphite-app graphite-app bot removed the 0-merge Merge with Graphite Merge Queue label May 27, 2025
camc314 added a commit that referenced this pull request May 30, 2025
## [0.17.0] - 2025-05-30

- ead5309 linter: [**BREAKING**] Remove react from default plugin set
(#11382) (camc314)

### Features

- 2d25bd8 linter: Remove `unicorn/no-for-loop` over
`typescript/prefer-for-of` (#11354) (camc314)
- bbb7eb1 linter: Add auto-fix to react/forward-ref-uses-ref (#11342)
(yefan)
- 590c27b linter: Add auto-fix to unicorn/no-await-expression-member
(#11306) (yefan)
- 7824f01 linter: Implement suggestion for `jsx/no-useless-fragment`
(#10800) (Cam McHenry)
- 2083d33 linter/language_server: Add second editor suggestion for
`react/forward-ref-uses-ref` (#11375) (Sysix)

### Bug Fixes

- f6424dd linter: Reflect react plugin is disabled by default in cli
(#11397) (camc314)
- 25ecbfe linter: Remove use of `FrameworkFlags::React` to decide
whether rules should run (#11383) (camc314)
- 0d240e4 linter: False positive in react/exhaustive-deps with default
formal parameter (#11395) (camc314)
- c91697e linter: Fix panic in multi byte char in `TryFrom` aria
(#11350) (camc314)
- 9798ef1 linter: Stack overflow in no-async-endpoint-handlers (#11317)
(camc314)
- 348ad97 linter: Skip no-unused-vars on astro files (#11303) (camc314)
- 183d7f0 linter: Make `jest/no-large-snapshots` error easier to
comprehend (#11294) (Ulrich Stark)
- 4e606a5 linter: Improve `jest/no-large-snapshots` (#11291) (Ulrich
Stark)
- 14f790f linter: Improve `jest/no-restricted-matchers` (#11292) (Ulrich
Stark)
- a2c82be linter/block-scoped-var: Better diagnostic messages (#11290)
(DonIsaac)
- 19772e5 linter/no-unused-vars: Panic when variable is redeclared as
function in same scope (#11280) (Dunqing)
- 7af5bb1 oxc_language_server: Include save option for text document
sync capability (#11297) (Nicholas Rayburn)

### Performance

- 0ed6c1a language_server: Use `Arc<RwLock>` instead of `Mutex` for
workspace workers (#11328) (Sysix)
- a0ee946 linter: Reduce code size in `globals` (#11333) (shulaoda)
- c90c5e9 linter/no-unused-vars: Simplify checking self call usage
(#11281) (Dunqing)

### Documentation

- eae51ca linter: Clarify jsdoc/check-tag-names configuration (#11394)
(Rägnar O'ock)
- cd354d4 oxlint: Remove incorrect doc comment (#11326) (camc314)

### Refactor

- 42738f0 linter: Shorten code of match arms (#11389) (Ulrich Stark)
- 8a34447 linter: Improve `unicorn/text-encoding-identifier-case`
(#11386) (shulaoda)
- eaa605e linter: Avoid some `Arc::clone` in linter runtime (#11388)
(Boshen)
- 1cd8b9c linter: Fixes in `react/forward-ref-uses-ref` are suggestions
(#11376) (Sysix)
- 50ef691 linter: Add `diagnostics_with_multiple_fixes` to `LintContext`
(#11357) (Sysix)
- 606bb34 linter: Accept `PossibleFixes` instead of `Fix` for `Messages`
(#11295) (Sysix)
- 042a3f3 linter: Use `PossibleFixes` instead of `Option<Fix>` (#11284)
(Sysix)
- ffcfb46 linter: Improve `unicorn/throw-new-error` (#11364) (shulaoda)
- 8fb55c3 linter: Cleanup docs and simplify code of
`eslint/no-fallthrough` (#11331) (Ulrich Stark)
- e2f0f0a linter: Improve docs and simplify code of
`eslint/no-duplicate-imports` (#11320) (Ulrich Stark)
- b53b053 linter: Simplify accessing span of NameSpan (#11305) (Ulrich
Stark)
- 4bc2650 linter: Improve `eslint/no-void` (#11285) (shulaoda)

### Styling

- 49b664c linter: Remove needless newline in `declare_oxc_lint` (#11400)
(camc314)

### Testing

- c4f64aa linter: Explicitly disable correctness for clarity (#11327)
(camc314)
- a404b2c linter: `eslint/no-duplicate-imports` shouldn't report the
same span (#11324) (Ulrich Stark)

Co-authored-by: camc314 <[email protected]>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
A-editor Area - Editor and Language Server C-performance Category - Solution not expected to change functional behavior, only performance
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants