feat: Enable the Apple fast datapath#3455
Merged
Merged
Conversation
larseggert
force-pushed
the
feat-apple-fast-datapath
branch
from
8e32c67 to
255bb97
Compare
Merging this PR will not alter performance
Comparing 
Footnotes
|
Codecov Report❌ Patch coverage is Additional details and impacted files@@ Coverage Diff @@
## main #3455 +/- ##
==========================================
- Coverage 94.29% 94.23% -0.07%
==========================================
Files 127 131 +4
Lines 38739 39081 +342
Branches 38739 39081 +342
==========================================
+ Hits 36529 36827 +298
- Misses 1371 1408 +37
- Partials 839 846 +7
Flags with carried forward coverage won't be shown. Click here to find out more.
|
larseggert
force-pushed
the
feat-apple-fast-datapath
branch
from
255bb97 to
127d9ff
Compare
larseggert
requested review from
KershawChang,
martinthomson and
mxinden
as code owners
Contributor
There was a problem hiding this comment.
Pull request overview
Enables Quinn’s Apple “fast UDP datapath” support across the workspace and adds runtime plumbing so consumers (notably Gecko) can opt-in to sendmsg_x/recvmsg_x, while neqo-bin enables it unconditionally as a canary.
Changes:
- Enable
quinn-udp/fast-apple-datapathat the workspace dependency level and remove per-crate feature wiring. - Add an Apple-only, process-wide runtime flag in
neqo-udpto opt sockets into the fast datapath at creation time. - Add
cfg(apple)aliasing toneqo-binand unconditionally enable the fast datapath in the CLI socket setup (plus clippy suppression for Apple builds).
Reviewed changes
Copilot reviewed 6 out of 8 changed files in this pull request and generated 3 comments.
Show a summary per file
| File | Description |
|---|---|
| neqo-udp/src/lib.rs | Adds runtime enablement flag + Apple socket setup hook; adjusts Apple receive buffer strategy. |
| neqo-udp/Cargo.toml | Switches cfg_aliases build-dep to workspace and removes the crate-level fast-path feature. |
| neqo-bin/src/udp.rs | Enables Apple fast datapath unconditionally for CLI sockets. |
| neqo-bin/src/bin/client.rs | Adds Apple-only clippy suppression for large_futures. |
| neqo-bin/build.rs | Introduces cfg(apple) alias for consistent platform gating. |
| neqo-bin/Cargo.toml | Adds cfg_aliases build-dependency; removes fast-path feature. |
| Cargo.toml | Enables quinn-udp’s fast-apple-datapath feature for the whole workspace and adds cfg_aliases to workspace deps. |
| Cargo.lock | Records updated dependency graph including cfg_aliases for neqo-bin. |
You can also share your feedback on Copilot code review. Take the survey.
larseggert
force-pushed
the
feat-apple-fast-datapath
branch
from
2fd6e8d to
6be99fa
Compare
Unconditionally in neqo-bin, but with plumbing so Gecko can probe and enable it conditionally.
larseggert
force-pushed
the
feat-apple-fast-datapath
branch
from
6be99fa to
f910d92
Compare
Collaborator
Author
|
The Gecko side of this is here: https://github.com/mozilla-firefox/firefox/compare/main...larseggert:firefox:feat-apple-probe-fast?expand=1 |
Merged
mxinden
reviewed
Mar 18, 2026
Collaborator
Author
|
@mxinden thanks, both good suggestions. I think I confused myself between the neqo and Gecko patches... |
Contributor
Failed Interop TestsQUIC Interop Runner, client vs. server, differences relative to
All resultsSucceeded Interop TestsQUIC Interop Runner, client vs. server neqo-pr as client
neqo-pr as server
Unsupported Interop TestsQUIC Interop Runner, client vs. server neqo-pr as client
neqo-pr as server
|
Contributor
Client/server transfer resultsPerformance differences relative to 4288d22. Transfer of 33554432 bytes over loopback, min. 100 runs. All unit-less numbers are in milliseconds.
Table above only shows statistically significant changes. See all results below. All resultsTransfer of 33554432 bytes over loopback, min. 100 runs. All unit-less numbers are in milliseconds.
Download data for |
Contributor
Benchmark resultsSignificant performance differences relative to 4288d22. transfer/1-conn/1-100mb-resp (aka. Download)/mtu-1504: 💚 Performance has improved by -2.0606%. time: [198.33 ms 198.74 ms 199.22 ms]
thrpt: [501.96 MiB/s 503.18 MiB/s 504.22 MiB/s]
change:
time: [-2.4011% -2.0606% -1.7069] (p = 0.00 < 0.05)
thrpt: [+1.7365% +2.1040% +2.4602]
Performance has improved.
Found 2 outliers among 100 measurements (2.00%)
1 (1.00%) high mild
1 (1.00%) high severetransfer/1-conn/10_000-parallel-1b-resp (aka. RPS)/mtu-1504: 💔 Performance has regressed by +2.7600%. time: [289.72 ms 291.60 ms 293.49 ms]
thrpt: [34.073 Kelem/s 34.294 Kelem/s 34.516 Kelem/s]
change:
time: [+1.8946% +2.7600% +3.6924] (p = 0.00 < 0.05)
thrpt: [-3.5609% -2.6859% -1.8594]
Performance has regressed.
Found 1 outliers among 100 measurements (1.00%)
1 (1.00%) high mildtransfer/1-conn/1-100mb-req (aka. Upload)/mtu-1504: 💚 Performance has improved by -1.2540%. time: [201.88 ms 202.26 ms 202.66 ms]
thrpt: [493.44 MiB/s 494.41 MiB/s 495.33 MiB/s]
change:
time: [-1.5152% -1.2540% -1.0002] (p = 0.00 < 0.05)
thrpt: [+1.0103% +1.2699% +1.5386]
Performance has improved.
Found 2 outliers among 100 measurements (2.00%)
2 (2.00%) high mildAll resultstransfer/1-conn/1-100mb-resp (aka. Download)/mtu-1504: 💚 Performance has improved by -2.0606%. time: [198.33 ms 198.74 ms 199.22 ms]
thrpt: [501.96 MiB/s 503.18 MiB/s 504.22 MiB/s]
change:
time: [-2.4011% -2.0606% -1.7069] (p = 0.00 < 0.05)
thrpt: [+1.7365% +2.1040% +2.4602]
Performance has improved.
Found 2 outliers among 100 measurements (2.00%)
1 (1.00%) high mild
1 (1.00%) high severetransfer/1-conn/10_000-parallel-1b-resp (aka. RPS)/mtu-1504: 💔 Performance has regressed by +2.7600%. time: [289.72 ms 291.60 ms 293.49 ms]
thrpt: [34.073 Kelem/s 34.294 Kelem/s 34.516 Kelem/s]
change:
time: [+1.8946% +2.7600% +3.6924] (p = 0.00 < 0.05)
thrpt: [-3.5609% -2.6859% -1.8594]
Performance has regressed.
Found 1 outliers among 100 measurements (1.00%)
1 (1.00%) high mildtransfer/1-conn/1-1b-resp (aka. HPS)/mtu-1504: No change in performance detected. time: [38.477 ms 38.650 ms 38.841 ms]
thrpt: [25.746 B/s 25.873 B/s 25.989 B/s]
change:
time: [-0.6576% -0.0257% +0.6295] (p = 0.94 > 0.05)
thrpt: [-0.6256% +0.0258% +0.6619]
No change in performance detected.
Found 9 outliers among 100 measurements (9.00%)
2 (2.00%) high mild
7 (7.00%) high severetransfer/1-conn/1-100mb-req (aka. Upload)/mtu-1504: 💚 Performance has improved by -1.2540%. time: [201.88 ms 202.26 ms 202.66 ms]
thrpt: [493.44 MiB/s 494.41 MiB/s 495.33 MiB/s]
change:
time: [-1.5152% -1.2540% -1.0002] (p = 0.00 < 0.05)
thrpt: [+1.0103% +1.2699% +1.5386]
Performance has improved.
Found 2 outliers among 100 measurements (2.00%)
2 (2.00%) high mildstreams/walltime/1-streams/each-1000-bytes: No change in performance detected. time: [588.43 µs 592.49 µs 598.15 µs]
change: [-0.9165% -0.1114% +0.9646] (p = 0.83 > 0.05)
No change in performance detected.
Found 10 outliers among 100 measurements (10.00%)
10 (10.00%) high severestreams/walltime/1000-streams/each-1-bytes: Change within noise threshold. time: [12.371 ms 12.390 ms 12.409 ms]
change: [-0.4910% -0.2644% -0.0694] (p = 0.02 < 0.05)
Change within noise threshold.streams/walltime/1000-streams/each-1000-bytes: Change within noise threshold. time: [44.994 ms 45.085 ms 45.220 ms]
change: [+0.0480% +0.3577% +0.7239] (p = 0.02 < 0.05)
Change within noise threshold.
Found 2 outliers among 100 measurements (2.00%)
1 (1.00%) high mild
1 (1.00%) high severetransfer/walltime/pacing-false/varying-seeds: Change within noise threshold. time: [77.627 ms 77.687 ms 77.755 ms]
change: [-3.0059% -2.8235% -2.6458] (p = 0.00 < 0.05)
Change within noise threshold.
Found 6 outliers among 100 measurements (6.00%)
1 (1.00%) low mild
4 (4.00%) high mild
1 (1.00%) high severetransfer/walltime/pacing-true/varying-seeds: Change within noise threshold. time: [79.386 ms 79.487 ms 79.631 ms]
change: [+0.0492% +0.2951% +0.5262] (p = 0.01 < 0.05)
Change within noise threshold.
Found 3 outliers among 100 measurements (3.00%)
3 (3.00%) high severetransfer/walltime/pacing-false/same-seed: Change within noise threshold. time: [77.597 ms 77.696 ms 77.845 ms]
change: [-2.6578% -2.4391% -2.2208] (p = 0.00 < 0.05)
Change within noise threshold.
Found 3 outliers among 100 measurements (3.00%)
2 (2.00%) high mild
1 (1.00%) high severetransfer/walltime/pacing-true/same-seed: Change within noise threshold. time: [79.298 ms 79.394 ms 79.537 ms]
change: [-0.3209% -0.1706% +0.0131] (p = 0.04 < 0.05)
Change within noise threshold.
Found 1 outliers among 100 measurements (1.00%)
1 (1.00%) high severeDownload data for |
Merged
mxinden
approved these changes
Mar 30, 2026
Collaborator
@larseggert FYI |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Unconditionally in neqo-bin, but with plumbing so Gecko can probe and enable it conditionally.