Add 9p fs to LiteBox by CvvT · Pull Request #663 · microsoft/litebox

CvvT · 2026-02-11T19:05:06Z

This PR adds implementation of 9p fs that uses 9P2000.L protocol. For testing, I used diod as the remote server. The implementation of 9p client was largely migrated from VSBox.

Note that different from other fs, this only support blocking operations as it relies on network. Not sure if it breaks any assumption that LiteBox has.

Currently it only allows issuing one call over network at a time and waiting for it to complete. We should allow concurrent requests. I will fix it in the next PR.

github-actions · 2026-02-11T22:22:19Z

🤖 SemverChecks 🤖 ⚠️ Potential breaking API changes detected ⚠️

Click for details

--- failure enum_variant_added: enum variant added on exhaustive enum ---

Description:
A publicly-visible enum without #[non_exhaustive] has a new variant.
        ref: https://doc.rust-lang.org/cargo/reference/semver.html#enum-variant-new
       impl: https://github.com/obi1kenobi/cargo-semver-checks/tree/v0.46.0/src/lints/enum_variant_added.ron

Failed in:
  variant MigrationError:Io in /home/runner/work/litebox/litebox/litebox/src/fs/layered.rs:427
  variant TruncateError:Io in /home/runner/work/litebox/litebox/litebox/src/fs/errors.rs:95

--- failure method_parameter_count_changed: pub method parameter count changed ---

Description:
A publicly-visible method now takes a different number of parameters, not counting the receiver (self) parameter.
        ref: https://doc.rust-lang.org/cargo/reference/semver.html#fn-change-arity
       impl: https://github.com/obi1kenobi/cargo-semver-checks/tree/v0.46.0/src/lints/method_parameter_count_changed.ron

Failed in:
  litebox::fs::nine_p::FileSystem::new now takes 5 parameters instead of 1, in /home/runner/work/litebox/litebox/litebox/src/fs/nine_p/mod.rs:313

--- failure trait_requires_more_generic_type_params: trait now requires more generic type parameters ---

Description:
A trait now requires more generic type parameters than it used to. Uses of this trait that supplied the previously-required number of generic types will be broken. To fix this, consider supplying default values for newly-added generic types.
        ref: https://doc.rust-lang.org/cargo/reference/semver.html#trait-new-parameter-no-default
       impl: https://github.com/obi1kenobi/cargo-semver-checks/tree/v0.46.0/src/lints/trait_requires_more_generic_type_params.ron

Failed in:
  trait FileSystem (1 -> 2 required generic types) in /home/runner/work/litebox/litebox/litebox/src/fs/nine_p/mod.rs:277

--- failure type_requires_more_generic_type_params: type now requires more generic type parameters ---

Description:
A type now requires more generic type parameters than it used to. Uses of this type that supplied the previously-required number of generic types will be broken. To fix this, consider supplying default values for newly-added generic types.
        ref: https://doc.rust-lang.org/cargo/reference/semver.html#trait-new-parameter-no-default
       impl: https://github.com/obi1kenobi/cargo-semver-checks/tree/v0.46.0/src/lints/type_requires_more_generic_type_params.ron

Failed in:
  Struct FileSystem (1 -> 2 required generic types) in /home/runner/work/litebox/litebox/litebox/src/fs/nine_p/mod.rs:277

wdcui

I went through the code and left some comments. Thank you for adding 9p support in LiteBox!

wdcui · 2026-02-12T22:24:54Z

litebox/src/fs/layered.rs

        match self.lower.file_status(path) {
            Ok(stat) => Ok(stat.file_type),
            Err(FileStatusError::ClosedFd) => unreachable!(),
+            Err(FileStatusError::Io) => Err(PathError::NoSuchFileOrDirectory),


Io is a pretty general erorr name. Why would it imply NoSuchFileOrDirectory?

wdcui · 2026-02-12T22:27:25Z

litebox/src/fs/errors.rs

    #[error("error when truncating: {0}")]
    TruncateError(#[from] TruncateError),
+    #[error("I/O error")]
+    Io,


Should we interpret Io error as network failure instead of the file or directory not exist?

wdcui · 2026-02-12T22:53:07Z

litebox/src/fs/nine_p/mod.rs

+        }
+    }
+
+    /// Walk to the parent of a path and return the parent fid and the name


Nit: "the name of the child"?

wdcui · 2026-02-12T22:55:08Z

litebox/src/fs/nine_p/mod.rs

+            super::FileStatus {
+                file_type,
+                mode: super::Mode::from_bits_truncate(attr.stat.mode),
+                size: usize::try_from(attr.stat.size).expect("file size exceeds usize"),


expect() will cause litebox to crash, right? Should we avoid that?

wdcui · 2026-02-12T23:03:06Z

litebox/src/fs/nine_p/mod.rs

+
+        let bytes_read = self.client.read(fid, read_offset as u64, buf)?;
+
+        // Update offset if not using explicit offset


Shouldn't we always update the offset stored in the descriptor table?

wdcui · 2026-02-12T23:28:59Z

litebox/src/fs/nine_p/fcall.rs

+/// Tagged 9P message
+#[derive(Clone, Debug)]
+pub(super) struct TaggedFcall<'a> {
+    pub(super) tag: u16,


What is the tag for?

wdcui · 2026-02-12T23:31:17Z

litebox/src/fs/nine_p/fcall.rs

+        encode_fcall(buf, tag, fcall)?;
+
+        // Write the size at the beginning
+        let size = u32::try_from(buf.len()).expect("buffer length exceeds u32");


Should we return the error instead of panic?

wdcui · 2026-02-12T23:32:17Z

litebox/src/fs/nine_p/fcall.rs

+}
+
+// ============================================================================
+// Encoding functions


Can we simplify the encoding/decoding to serde of structs?

wdcui · 2026-02-12T23:35:48Z

litebox/src/fs/nine_p/transport.rs

+//! 9P transport layer abstraction
+//!
+//! This module defines traits for reading and writing 9P protocol messages over
+//! an underlying transport (e.g., TCP socket, virtio-9p, etc.).


Where is the network integration?

wdcui · 2026-02-12T23:37:27Z

litebox/src/fs/nine_p/tests.rs

+// ---------------------------------------------------------------------------
+
+/// A wrapper around `TcpStream` that implements the litebox 9P transport traits.
+struct TcpTransport {


Why do we have both implementations and tests in this file?

jaybosamiya-ms · 2026-02-12T23:38:15Z

litebox/src/fs/nine_p/mod.rs

+//!
+//! # Submodules
+//!
+//! The 9P implementation is split into several submodules:
+//! - `fcall` - Protocol message definitions and encoding/decoding
+//! - `transport` - Transport layer traits and message I/O
+//! - `client` - High-level 9P client for protocol operations


Minor comment (I have not reviewed the code, just was looking through the docs): we might wish to skip this from the doc-strings and change it to a regular comment. Anything in the doc-strings (here, module-level) effectively behaves like a spec. The details most of these submodules (which are not necessarily public) is not useful/needed to callers of this module.

CvvT added 16 commits February 11, 2026 10:39

add 9p fs

f532ff2

fix 9p fs

e4ba56a

fix error code

79afcd0

add test

0a83435

use 9p in linux shim

c80537c

fix fcall

cc36869

remove clone_static

17d79df

remove some tests

5a9cede

install diod

8c4d643

revert

a078095

fix fmt

b44c382

refactor code

6bbb2a4

add debug

333c449

update comment

5cd4fc9

revert change

d5e5adc

add error handling

28d252a

CvvT marked this pull request as ready for review February 11, 2026 22:23

wdcui reviewed Feb 12, 2026

View reviewed changes

jaybosamiya-ms reviewed Feb 12, 2026

View reviewed changes


		let bytes_read = self.client.read(fid, read_offset as u64, buf)?;

		// Update offset if not using explicit offset

Conversation

CvvT commented Feb 11, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

github-actions bot commented Feb 11, 2026

Uh oh!

wdcui left a comment

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!

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!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

CvvT commented Feb 11, 2026 •

edited

Loading