Doip server initial code by VinaykumarRS1995 · Pull Request #4 · eclipse-opensovd/uds2sovd-proxy

VinaykumarRS1995 · 2026-02-25T12:51:36Z

Summary

Checklist

I have tested my changes locally
I have added or updated documentation
I have linked related issues or discussions
I have added or updated tests

arvindr19 · 2026-02-26T04:19:44Z

Is this code supposed to build? When I ran cargo check and cargo build, I encountered some errors. IMO, code should build successfully.

VinaykumarRS1995 · 2026-02-26T09:43:47Z

Is this code supposed to build? When I ran cargo check and cargo build, I encountered some errors. IMO, code should build successfully.

build is successful is now.

arvindr19 · 2026-02-26T10:03:37Z

src/doip/diagnostic_message.rs

+    pub fn parse(payload: &[u8]) -> Result<Self, Error> {
+        let header: [u8; 4] =
+            payload
+                .get(..4)


magic number? other places I could see used const
Not just here, please check below ones

Fixed all magic numbers across the codebase (named constants , protocol versions, and message sizes). Will push in the next commit.

arvindr19 · 2026-02-26T10:10:36Z

fix minor formatting inconsistencies
toml = "0.9.11+spec-1.1.0" causing warning, why not toml = "0.9" instead?

arvindr19

Only header_parser.rs has any tracing at all. Every other file has no tracing.

arvindr19 · 2026-02-26T13:50:38Z

doip-server/src/doip/header.rs

+/// when the `DoIP` entity cannot process a received message due to header-level issues.
+#[derive(Debug, Clone, Copy, PartialEq, Eq)]
+#[repr(u8)]
+pub enum GenericNackCode {


same defined in errors.rs
why so?

Fixed , re-exported now

arvindr19 · 2026-02-26T13:52:56Z

doip-server/src/doip/routing_activation.rs

+// Response codes per ISO 13400-2:2019 Table 25
+#[derive(Debug, Clone, Copy, PartialEq, Eq)]
+#[repr(u8)]
+pub enum ResponseCode {


duplicates in error.rs

Fixed , re-exported now.

arvindr19 · 2026-02-26T13:54:44Z

doip-server/src/doip/diagnostic_message.rs

+// Diagnostic message negative ack codes per ISO 13400-2:2019 Table 28
+#[derive(Debug, Clone, Copy, PartialEq, Eq)]
+#[repr(u8)]
+pub enum NackCode {


duplicates in error.rs

arvindr19 · 2026-02-26T15:38:23Z

src/doip/header_parser.rs

+                inverse_version: DEFAULT_PROTOCOL_VERSION_INV,
+                payload_type: payload_type as u16,
+                payload_length: u32::try_from(payload.len())
+                    .expect("DoIP payload exceeds u32::MAX (protocol limit is 4MB)"),


lib should never panic, return Result instead

Fixed, retruns Result now.

darkwisebear

I didn't go into details yet, but from what I see there are some general things that need cleanup before I will do that:

Consolidate error types and use thiserror everywhere.
Introduce a trait for message types and use everywhere
Deduplicate common parsing code

darkwisebear · 2026-02-26T20:04:03Z

src/doip/alive_check.rs

+use bytes::{BufMut, Bytes, BytesMut};
+
+#[derive(Debug, Clone, PartialEq, Eq)]
+pub enum Error {


This crate references thiserror. You could use it to declare this error in a more convenient way.

darkwisebear · 2026-02-26T20:10:58Z

src/doip/vehicle_id.rs

+use bytes::{BufMut, Bytes, BytesMut};
+
+#[derive(Debug, Clone, PartialEq, Eq)]
+pub enum Error {


A similar definition exists in alive_check.rs. Imo we should have a common definition of an error type for the entire module.

moved to mod.rs and shared across

darkwisebear · 2026-02-26T20:13:08Z

doip-server/src/doip/vehicle_id.rs

+
+// Vehicle Identification Request (0x0001) - no payload
+#[derive(Debug, Clone, PartialEq, Eq, Default)]
+pub struct Request;


The shape of this module is very similar to that of alive_check.rs. I think there should be a trait for message types so that there is an impl Request for VehicleIdRequest or the like.

trait defined

darkwisebear · 2026-02-26T20:37:18Z

src/doip/vehicle_id.rs

+        let eid: [u8; 6] = payload
+            .get(..Self::LEN)
+            .and_then(|s| s.try_into().ok())
+            .ok_or(Error::PayloadTooShort {
+                expected: Self::LEN,
+                actual: payload.len(),
+            })?;


Such code could be factored out, as there is similar code also in alive_check.rs. I think there should be a submodule messages with common definitions and then types for each message.

darkwisebear · 2026-02-26T20:40:26Z

src/doip/header_parser.rs

+
+impl PayloadType {
+    #[must_use]
+    pub fn from_u16(value: u16) -> Option<Self> {


Why not impl TryFrom<u16> for PayloadType? This is more canonical than such a custom function.

yes, updated.

darkwisebear · 2026-02-26T20:44:15Z

src/server/config.rs

+            let bind = server
+                .bind_address
+                .as_deref()
+                .unwrap_or(DEFAULT_BIND_ADDRESS);


serde allows for specifying defaults inside the type definition. This would save you a lot of code that of the scheme "if value { copy } else { use default }".

Yes , it is updated now

arvindr19 · 2026-02-27T05:25:17Z

src/doip/alive_check.rs

@@ -0,0 +1,144 @@
+/*
+ * Copyright (c) 2025 The Contributors to Eclipse OpenSOVD (see CONTRIBUTORS)


year should be 2026, please check all the license headers

updated where needed

arvindr19 · 2026-03-02T06:23:08Z

Please run cargo fmt --check; it currently fails. Once the CI workflow PR is merged, this PR will need to be reworked, so it’s better to fix it beforehand.

lh-sag

Just had a brief look at the PR.

@VinaykumarRS1995 did you had a chance to evaluate any of these crates to ease development?

Adding @alexmohr and @theswiftfox to share their experience with some of the crates above.

lh-sag · 2026-03-04T06:35:06Z

Cargo.toml

+[dependencies]
+# Async runtime
+tokio = { version = "1.35", features = ["full"] }
+tokio-util = { version = "0.7", features = ["codec"] }
+
+# Serialization
+serde = { version = "1.0", features = ["derive"] }
+toml = "0.9"
+
+# Logging
+tracing = "0.1"
+tracing-subscriber = { version = "0.3", features = ["env-filter"] }
+
+# Error handling
+thiserror = "2.0"
+anyhow = "1.0"
+
+# Data structures
+bytes = "1.0"
+dashmap = "6.0"
+parking_lot = "0.12"
+
+# Utilities
+uuid = { version = "1.6", features = ["v4"] }
+clap = { version = "4.4", features = ["derive"] }


Always use default-features = false and explicitly add feature as needed. This reduced sbom and binary size often significantly.

lh-sag · 2026-03-04T06:38:57Z

src/uds/mod.rs

+// #[cfg(any(test, feature = "test-handlers"))]
+// pub mod dummy_handler;


Please remove commented out code.

lh-sag · 2026-03-04T06:45:54Z

src/server/config.rs

+    fn parse_vin(s: &str) -> anyhow::Result<[u8; 17]> {
+        let bytes = s.as_bytes();
+        if bytes.len() != 17 {
+            anyhow::bail!("VIN must be exactly 17 characters");


Optional: I would rather return a proper error type than a string.

lh-sag · 2026-03-04T07:02:51Z

Cargo.toml

+dashmap = "6.0"
+parking_lot = "0.12"


Can you please double check that all crates are used/referenced in your code?

alexmohr · 2026-03-04T10:16:43Z

Just had a brief look at the PR.

@VinaykumarRS1995 did you had a chance to evaluate any of these crates to ease development?
* https://github.com/samp-reston/doip-definitions

* https://github.com/samp-reston/doip-codec

* https://github.com/samp-reston/doip-sockets

* https://github.com/samp-reston/doip-server
Adding @alexmohr and @theswiftfox to share their experience with some of the crates above.

We're only using codec and definitions in the CDA as we had some issues with sockets. But using codecs and definitions would already prevent re-implemeting some things here. The sockets connections can be re-used from CDA. Currently both doip libraries we use are forked by @theswiftfox. I believe it would be a good idea to use her fork as a base to officially provide an eclipse fork we use in all of our projects.

floroks · 2026-03-05T08:26:34Z

I believe it would be a good idea to use her fork as a base to officially provide an eclipse fork we use in all of our projects.

As we've also run into some issues with the original code, I think we need should fork them into the eclipse-opensovd repo, so we can adapt them quickly to our needs. I'll put it on the agenda for the next architecture board.

stmuench · 2026-03-06T08:44:28Z

Just had a brief look at the PR.

@VinaykumarRS1995 did you had a chance to evaluate any of these crates to ease development?

https://github.com/samp-reston/doip-definitions

https://github.com/samp-reston/doip-codec

https://github.com/samp-reston/doip-sockets

https://github.com/samp-reston/doip-server

Adding @alexmohr and @theswiftfox to share their experience with some of the crates above.

@lh-sag personally, I would be reluctant with using creates where there is only exactly one maintainer and no updates in the github repo since quite some time. This is risk in my eyes. What's your opinion on this situation?

VinaykumarRS1995 · 2026-03-10T09:55:43Z

Just had a brief look at the PR.

@VinaykumarRS1995 did you had a chance to evaluate any of these crates to ease development?

https://github.com/samp-reston/doip-definitions

https://github.com/samp-reston/doip-codec

https://github.com/samp-reston/doip-sockets

https://github.com/samp-reston/doip-server

Adding @alexmohr and @theswiftfox to share their experience with some of the crates above.

The consideration for integrating DoIP crates is tracked in #9 - will revisit once they are officially part of the OpenSOVD project. Not blocking this PR.

darkwisebear · 2026-03-03T13:27:08Z

src/doip/alive_check.rs

+}
+
+impl DoipParseable for Response {
+    fn parse(payload: &[u8]) -> std::result::Result<Self, DoipError> {


Suggested change

fn parse(payload: &[u8]) -> std::result::Result<Self, DoipError> {

fn parse(payload: &[u8]) -> crate::DoipResult<Self> {

This repeats in other files, please also fix there.

Done. Added pub use error::DoipResult in lib.rs and updated all signatures across the codebase — no raw Result<_, DoipError> left anywhere outside error.rs.

darkwisebear · 2026-03-03T13:30:21Z

src/doip/diagnostic_message.rs

+// Diagnostic message positive ack codes per ISO 13400-2:2019 Table 27
+#[derive(Debug, Clone, Copy, PartialEq, Eq)]
+#[repr(u8)]
+pub enum AckCode {


Why would this need to be exposed? Isn't that simply signalling "no error", which would be encoded as Ok(...)?

AckCode::Acknowledged was just encoding "no error" as a type — removed it. Replaced with AckResult::Positive (no code needed) and AckResult::Negative(NackCode) for the failure case.

darkwisebear · 2026-03-03T13:31:39Z

src/doip/diagnostic_message.rs

+pub struct PositiveAck {
+    source_address: u16,
+    target_address: u16,
+    ack_code: AckCode,


That seems redundant. The pure existence of the instance indicates success.

Merged PositiveAck and NegativeAck into a single DiagnosticAck with result: AckResult (Positive | Negative(NackCode)).

darkwisebear · 2026-03-03T14:01:30Z

src/doip/diagnostic_message.rs

+/// # Wire Format
+/// Payload: SA(2) + TA(2) + nack_code(1) + optional previous_diag_data
+#[derive(Debug, Clone, PartialEq, Eq)]
+pub struct NegativeAck {


This looks exactly like PositiveAck. I have the feeling that it makes sense to merge both structs, and also the code, into one entity that simply represents what was received, no matter whether it was negative or positive.

Merged PositiveAck and NegativeAck into a single DiagnosticAck with result: AckResult (Positive | Negative(NackCode)).

bharatGoswami8 · 2026-03-17T05:25:27Z

doip-server/src/doip/alive_check.rs

+        Some(0)
+    }
+
+    fn write_to(&self, _buf: &mut BytesMut) {}


maybe add todo! ("...") or unimplemented!("...") so if this method then it will through error.

Done. Added comment: // Intentionally empty — Alive Check Request has zero-length payload per ISO 13400-2.

bharatGoswami8 · 2026-03-17T05:31:55Z

doip-server/src/doip/alive_check.rs

+}
+
+impl Response {
+    pub const LEN: usize = 2;


why this const on impl...

Added /// doc comments to all constants explaining wire-format size.

bharatGoswami8 · 2026-03-17T05:55:30Z

doip-server/src/doip/mod.rs

+    /// Override this to enable pre-allocated buffers in [`to_bytes`], avoiding
+    /// incremental `BytesMut` reallocations for large messages.
+    fn serialized_len(&self) -> Option<usize> {
+        None


Returning None in the default implementation doesn’t seem correct to me. If a user calls this method on a type that doesn’t override it, they still get a valid Option, which may be misleading.

Done, Changes are pushed

bharatGoswami8 · 2026-03-17T05:56:38Z

doip-server/src/doip/mod.rs

+        .and_then(|s| s.try_into().ok())
+        .ok_or_else(|| {
+            let e = too_short(payload, N);
+            warn!("{} parse failed: {}", context, e);


maybe change to error log.

Done. Changed all parse-failure sites to error!. warn! kept only for recoverable NACK paths.

bharatGoswami8 · 2026-03-17T06:00:00Z

doip-server/src/doip/mod.rs

+        })
+}
+
+// Re-export core types and constants for convenient access.


Consider keeping the header at the start of the file. If it’s user-facing, re-exporting it via lib.rs could be a good option.

Done. Moved all pub use re-exports to top mod.rs

bharatGoswami8 · 2026-03-17T06:12:38Z

doip-server/src/server/session.rs

+pub struct SessionManager {
+    sessions: RwLock<HashMap<u64, Session>>,
+    addr_to_session: RwLock<HashMap<SocketAddr, u64>>,
+    next_id: RwLock<u64>,


Do we really need RwLock for next_id ?

lh-sag · 2026-03-18T07:47:35Z

In general I would like to see structured logs instead of strings. This is done in CDA and Core. Can you please apply this to your logs?

debug!("Session {} routing activated: tester_address=0x{:04X}", self.id, tester_address);

With structured logging (simplified for demonstration purpose):

debug!(session = self.id, tester_address, "routing activated");

lh-sag · 2026-03-18T07:54:28Z

doip-server/src/server/session.rs

+    sessions: RwLock<HashMap<u64, Session>>,
+    addr_to_session: RwLock<HashMap<SocketAddr, u64>>,
+    next_id: RwLock<u64>,


Do you need such fine granular locking? Instead you could have an inner session manager and have a single rwlock.

Lock contention might be longer but you have atomicity.

I did not looked further into detail but please consider this if it is applicable.

Done. Consolidated to single RwLock for atomic map consistency + AtomicU64 for lock-free ID allocation.

lh-sag · 2026-03-18T07:56:57Z

doip-server/src/error.rs

+pub use crate::doip::routing_activation::ResponseCode as RoutingActivationCode;
+
+/// Result type alias for `DoIP` operations
+pub type DoipResult<T> = std::result::Result<T, DoipError>;


Usually the alias is Result and this is how it is named across the opensovd code base. Please also check the with the diag lib guys about the naming.

Renamed to pub type Result.

lh-sag · 2026-03-18T08:01:49Z

src/doip/hearder_parser.rs

+pub enum ParseError {
+    InvalidHeader(String),
+    Io(io::Error),
+}
+
+impl std::fmt::Display for ParseError {
+    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
+        match self {
+            Self::InvalidHeader(msg) => write!(f, "Invalid header: {}", msg),
+            Self::Io(e) => write!(f, "IO error: {}", e),
+        }
+    }
+}
+
+impl std::error::Error for ParseError {}
+
+impl From<io::Error> for ParseError {
+    fn from(e: io::Error) -> Self {
+        Self::Io(e)
+    }
+}
+


You already use thiserror, please use it here as well if it is applicable.

Done — ParseError removed, all errors unified under DoipError with #[derive(thiserror::Error)].

- Renamed hearder_parser.rs to header_parser.rs (typo fix) - Added unit tests for header parsing - Updated MAX_DOIP_MESSAGE_SIZE to 4MB

- Consolidate errors into DoipError; add TryFrom<u8> on all enums - Strongly type activation_type field (ActivationType, not u8) - Add DoipParseable / DoipSerializable traits + DoipPayload dispatch enum - Override serialized_len() on all impls; fix array size magic numbers - Make ParseError pub(crate); add serde(default), tracing

- Split header_parser.rs into header.rs + codec.rs - Merge ParseError into DoipError (single error type) - Add parse_fixed_slice helper, remove duplicated parse patterns - Add DoipParseable / DoipSerializable traits - Make struct fields private, add public getters - Set default-features = false on all dependencies - Trim tokio features to io-util only - Remove dead test-handlers feature from Cargo.toml

Replace std::result::Result<_, DoipError> with crate::DoipResult across all files; add pub use error::DoipResult in lib.rs - remove redundant AckCode enum - remove redundant ack_code field from PositiveAck - merge PositiveAck + NegativeAck into DiagnosticAck with AckResult enum (Positive | Negative(NackCode))

- structured tracing logging; error! for parse failures - rename DoipResult → Result - single RwLock<SessionManagerInner> + AtomicU64 - merge DiagnosticAck + AckResult; remove AckCode - doc comments on all public API items

stmuench · 2026-03-20T11:03:58Z

doip-server/src/lib.rs

+/// UDS service layer – bridges DoIP transport to ISO 14229-1 request/response handling.
+pub mod uds;
+pub use error::DoipError;
+pub use error::Result;


@VinaykumarRS1995 how would this Result type get used on user side now? will there still be a prefix? and how would it look like?

Yes,the crate name acts as the natural prefix. Re-exported via pub use error::Result in lib.rs, it's accessible as:

// fully qualified
fn parse() -> doip_server::Result
{ ... }

// or with import
use doip_server::Result;
fn connect() -> Result { ... }

VinaykumarRS1995 force-pushed the doip-server-initial branch 2 times, most recently from a11f029 to c674ff3 Compare February 26, 2026 09:40

arvindr19 reviewed Feb 26, 2026

View reviewed changes

darkwisebear reviewed Feb 26, 2026

View reviewed changes

arvindr19 reviewed Feb 27, 2026

View reviewed changes

VinaykumarRS1995 force-pushed the doip-server-initial branch 2 times, most recently from 81e9ff5 to 9fa0c6e Compare March 2, 2026 05:49

VinaykumarRS1995 marked this pull request as ready for review March 2, 2026 09:23

VinaykumarRS1995 requested a review from darkwisebear March 2, 2026 09:23

lh-sag requested review from a team and removed request for darkwisebear March 4, 2026 06:32

lh-sag reviewed Mar 4, 2026

View reviewed changes

VinaykumarRS1995 requested a review from lh-sag March 11, 2026 05:18

darkwisebear reviewed Mar 12, 2026

View reviewed changes

VinaykumarRS1995 requested a review from darkwisebear March 13, 2026 02:24

VinaykumarRS1995 force-pushed the doip-server-initial branch from c8ccfbf to 0c12dc0 Compare March 17, 2026 05:39

bharatGoswami8 reviewed Mar 17, 2026

View reviewed changes

lh-sag reviewed Mar 18, 2026

View reviewed changes

VinaykumarRS1995 added 10 commits March 20, 2026 14:27

feat(doip): implement header parsing with tokio codec

904d427

test(doip): add unit tests for header_parser

fb4a095

- Renamed hearder_parser.rs to header_parser.rs (typo fix) - Added unit tests for header parsing - Updated MAX_DOIP_MESSAGE_SIZE to 4MB

feat(error): add DoIP error types and NRC codes

7f96215

feat(uds): add UdsHandler trait for pluggable backends

5a2cdbe

style: apply cargo fmt

05c030b

chore: move doip-server into Cargo workspace to resolve merge conflict

61798f0

fix(review): address review comments

3e77b11

- structured tracing logging; error! for parse failures - rename DoipResult → Result - single RwLock<SessionManagerInner> + AtomicU64 - merge DiagnosticAck + AckResult; remove AckCode - doc comments on all public API items

VinaykumarRS1995 force-pushed the doip-server-initial branch from 3ce6bcb to 3e77b11 Compare March 20, 2026 09:33

VinaykumarRS1995 requested a review from lh-sag March 20, 2026 09:37

stmuench reviewed Mar 20, 2026

View reviewed changes

VinaykumarRS1995 requested a review from stmuench March 25, 2026 13:29

		@@ -0,0 +1,144 @@
		/*
		* Copyright (c) 2025 The Contributors to Eclipse OpenSOVD (see CONTRIBUTORS)

		// #[cfg(any(test, feature = "test-handlers"))]
		// pub mod dummy_handler;

	fn parse(payload: &[u8]) -> std::result::Result<Self, DoipError> {
	fn parse(payload: &[u8]) -> crate::DoipResult<Self> {

		dashmap = "6.0"
		parking_lot = "0.12"

Conversation

VinaykumarRS1995 commented Feb 25, 2026

Summary

Checklist

Uh oh!

arvindr19 commented Feb 26, 2026

Uh oh!

VinaykumarRS1995 commented Feb 26, 2026

Uh oh!

arvindr19 Feb 26, 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!

Choose a reason for hiding this comment

Uh oh!

arvindr19 commented Feb 26, 2026

Uh oh!

arvindr19 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!

darkwisebear 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!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

arvindr19 commented Mar 2, 2026

Uh oh!

lh-sag 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!

alexmohr commented Mar 4, 2026

Uh oh!

floroks commented Mar 5, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

arvindr19 Feb 26, 2026 •

edited

Loading

floroks commented Mar 5, 2026 •

edited

Loading

VinaykumarRS1995 Mar 13, 2026 •

edited

Loading