fix: increase Hyperlight I/O buffer for large hostfs writes - #81
Merged
Conversation
The Hyperlight SDK defaults input/output shared-memory stacks to 16 KiB each. The hostfs VFS layer chunks writes at 32 KiB, but after base64 encoding + JSON envelope + FlatBuffer framing, a single chunk occupies ~44 KiB — overflowing the 16 KiB output stack and causing hcall_push to fail with EIO on writes exceeding ~12 KiB. Set the I/O buffer size to 128 KiB (sufficient for any single-chunk RPC with headroom) and expose it as a configurable field on VmConfig and SandboxBuilder so embedders can tune it for their workload. Signed-off-by: danbugs <danilochiarlone@gmail.com>
There was a problem hiding this comment.
Pull request overview
This PR increases the default Hyperlight shared-memory I/O buffer sizes to prevent host function call payload overflows during large hostfs write RPCs, and exposes a knob so embedders can tune the buffer size.
Changes:
- Add
io_buffer_sizetoVmConfigand plumb it through to HyperlightSandboxConfigurationinput/output data sizes. - Add
SandboxBuilder::io_buffer_size(...)for configuring the shared-memory I/O buffers. - Add an integration test that writes a >32 KiB binary file via hostfs and verifies its contents on the host.
Reviewed changes
Copilot reviewed 2 out of 2 changed files in this pull request and generated 3 comments.
| File | Description |
|---|---|
| host/tests/pyhl_runtime.rs | Adds a runtime integration test covering large binary writes through hostfs. |
| host/src/lib.rs | Introduces configurable I/O buffer sizing and increases the default to 128 KiB. |
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
Comment on lines
528
to
532
| pub struct VmConfig { | ||
| pub heap_size: u64, | ||
| pub stack_size: u64, | ||
| pub io_buffer_size: usize, | ||
| } |
Contributor
Author
There was a problem hiding this comment.
Fixed — added #[non_exhaustive] to VmConfig.
Comment on lines
572
to
+576
| fn sandbox_config(&self) -> SandboxConfiguration { | ||
| let mut cfg = SandboxConfiguration::default(); | ||
| cfg.set_heap_size(self.heap_size); | ||
| cfg.set_input_data_size(self.io_buffer_size); | ||
| cfg.set_output_data_size(self.io_buffer_size); |
Contributor
Author
There was a problem hiding this comment.
Pre-existing issue — stack_size predates this PR. Tracked separately.
Comment on lines
+303
to
+310
| let timing = rt | ||
| .run_code("with open('/host/big.bin', 'wb') as f: f.write(b'x' * 40000)") | ||
| .unwrap(); | ||
| assert_eq!(timing.exit_code, 0); | ||
|
|
||
| let data = std::fs::read(tmp.join("big.bin")).unwrap(); | ||
| assert_eq!(data.len(), 40000); | ||
| assert!(data.iter().all(|&b| b == b'x')); |
Contributor
Author
There was a problem hiding this comment.
Fixed — using 40 * 1024 now.
- Mark VmConfig as #[non_exhaustive] to prevent semver breakage when new fields are added - Use explicit 40 KiB (40 * 1024) in the large binary write test to match the PR description and make the boundary condition clearer Signed-off-by: danbugs <danilochiarlone@gmail.com>
Signed-off-by: danbugs <danilochiarlone@gmail.com>
![github-actions[bot]](https://avatars.githubusercontent.com/in/15368?s=60&v=4){kind=small}
Contributor
There was a problem hiding this comment.
Linux Benchmarks
Details
| Benchmark suite | Current: dab60b0 | Previous: 20ccc2d | Ratio |
|---|---|---|---|
hello_world (median) |
20 ms |
20 ms |
1 |
pandas (median) |
110 ms |
110 ms |
1 |
density (per VM) |
7 MB |
7 MB |
1 |
snapshot (disk) |
385 MiB |
385 MiB |
1 |
This comment was automatically generated by workflow using github-action-benchmark.
Contributor
There was a problem hiding this comment.
Windows Benchmarks
Details
| Benchmark suite | Current: dab60b0 | Previous: 20ccc2d | Ratio |
|---|---|---|---|
hello_world (median) |
267 ms |
261 ms |
1.02 |
pandas (median) |
896 ms |
800 ms |
1.12 |
density (per VM) |
6 MB |
6 MB |
1 |
snapshot (disk) |
392 MiB |
392 MiB |
1 |
This comment was automatically generated by workflow using github-action-benchmark.
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.
Summary
io_buffer_sizeonVmConfigandSandboxBuilderso embedders can tune itruntime_filesystem_large_binary_writeintegration test (40 KiB binary blob)Test plan
runtime_filesystem_large_binary_writepassescargo checkclean