feat: IPC binding for socom by lurtz · Pull Request #86 · eclipse-score/inc_someip_gateway

lurtz · 2026-04-15T12:42:16Z

Transport socom semantics over IPC to another process. It uses shared memory for payloads to avoid copying huge payloads.

TODO

transmit string "ipc identifier" at first connection
measure memory usage
fix flaky tests
check calls to MakeUnexpected(): seems to require static strings
reduce number of static casts

Transport socom semantics over IPC to another process. It uses shared memory for payloads to avoid copying huge payloads.

github-actions · 2026-04-15T12:43:57Z

License Check Results

🚀 The license check job ran with the Bazel command:

bazel run //:license-check

Status: ⚠️ Needs Review

Click to expand output

[License Check Output]
Extracting Bazel installation...
Starting local Bazel server (8.6.0) and connecting to it...
INFO: Invocation ID: 1d58cae4-7702-4cab-bf46-79a14f48b469
Computing main repo mapping: 
Computing main repo mapping: 
Computing main repo mapping: 
Computing main repo mapping: 
Loading: 
Loading: 3 packages loaded
Loading: 3 packages loaded
    currently loading: 
Loading: 3 packages loaded
    currently loading: 
Loading: 3 packages loaded
    currently loading: 
Loading: 3 packages loaded
    currently loading: 
Loading: 4 packages loaded
Analyzing: target //:license-check (4 packages loaded, 0 targets configured)
Analyzing: target //:license-check (4 packages loaded, 0 targets configured)

Analyzing: target //:license-check (44 packages loaded, 10 targets configured)

Analyzing: target //:license-check (93 packages loaded, 10 targets configured)

Analyzing: target //:license-check (147 packages loaded, 1071 targets configured)

Analyzing: target //:license-check (161 packages loaded, 3211 targets configured)

Analyzing: target //:license-check (165 packages loaded, 9394 targets configured)

Analyzing: target //:license-check (165 packages loaded, 9407 targets configured)

Analyzing: target //:license-check (172 packages loaded, 9557 targets configured)

Analyzing: target //:license-check (172 packages loaded, 9557 targets configured)

Analyzing: target //:license-check (176 packages loaded, 11564 targets configured)

Analyzing: target //:license-check (176 packages loaded, 11564 targets configured)

INFO: Analyzed target //:license-check (177 packages loaded, 13143 targets configured).
[11 / 17] [Prepa] Expanding template external/score_tooling+/dash/tool/formatters/_dash_format_converter_stage2_bootstrap.py [for tool]
INFO: From Generating Dash formatted dependency file ...:
INFO: Successfully converted 2 packages from Cargo.lock to bazel-out/k8-fastbuild/bin/formatted.txt
INFO: Found 1 target...
Target //:license.check.license_check up-to-date:
  bazel-bin/license.check.license_check
  bazel-bin/license.check.license_check.jar
INFO: Elapsed time: 211.473s, Critical Path: 0.33s
INFO: 17 processes: 5 disk cache hit, 12 internal.
INFO: Build completed successfully, 17 total actions
INFO: Running command line: bazel-bin/license.check.license_check ./formatted.txt <args omitted>
usage: org.eclipse.dash.licenses.cli.Main [-batch <int>] [-cd <url>]
       [-confidence <int>] [-ef <url>] [-excludeSources <sources>] [-help] [-lic
       <url>] [-project <shortname>] [-repo <url>] [-review] [-summary <file>]
       [-timeout <seconds>] [-token <token>]

github-actions · 2026-04-15T12:48:26Z

The created documentation from the pull request is available at: docu-html

massif did not work yet but I am on it

However some changes should be considered workarounds

The Error type does not own any string, only string_view. Thus it the error message given must be likely a static string

send() can fail whenever the peer disconnects

NEOatNHNG · 2026-04-17T13:56:52Z

+### Overall Statistics
+- **Total Heap Allocated**: 10,447,202 bytes (~10.0 MB)
+- **Peak Heap Usage**: 195,318 bytes (~191 KB)
+- **Peak Stack Usage**: 26,832 bytes (~26 KB)


This means that we allocate/de-allocate a lot, right? This we should look into and minimize. In a safety-critical system it would be favorable to allocate on startup and then don't have any heap operations until shutdown. That means total heap allocated should be more or less equal to peak heap usage.

That has not been my interpretation of the data. I looked more at what I got in graphs. E.g. this

I interpreted this that after startup nothing is allocated anymore.

I reworked the performance tests and memory profiling tests. There was a bug, which skipped an iteration if it failed to allocate shared memory, which can happen after all slots were consumed, but not freed. Now it looks more reasonable:

Towards your conclusion that there might be many small allocations and deallocations, I have now an idea where they come from. The socom::Payload is an interface class and shared_ptr is used for reference counting. Thus we see lots of shared_ptr<socom::Payload> constructions and destructions.

with latest changes allocations during event updates are gone. However due to the use of unordered_map and std::vector, there could still some happen after init and these have to be properly pre initialized

lurtz · 2026-06-08T11:22:35Z

Test code when only one side provides a service, but not the other one. Currently tests are always configured to be symmetric and even the same service is possible to be offered on both sides.

UPDATE Dealt with in 1a52ead

mariuswbr

Stumbled on this while trying to use the IPC binding in someipd and gatewayd.

… times from multiple threads

… loss

…ctors

…t shared memory configuration for method calls and event updates.

… mw-com-as-socom-bridge

feat: IPC binding for socom

bc57be9

Transport socom semantics over IPC to another process. It uses shared memory for payloads to avoid copying huge payloads.

lurtz added 18 commits April 15, 2026 14:19

inject message_passing Connection into client

fa04b24

inject message_passing Server into server

902fc45

simplify ipc gateway factories

72aa4ae

Send client identifier when connecting to server

d9b46e0

add UID and GID to client information

0079349

Memory profile

cc8867f

massif did not work yet but I am on it

add massif profile

3c37f15

better massif memory profile

859beee

cleanup shared memory allocation

77f3a59

Make concurrent test execution a little bit more stable

303c69c

Make tests more stable

cc715c5

However some changes should be considered workarounds

remove some workarounds for stable tests

a8dcd9d

ensure unique filenames in /dev/shm

1a933c2

stability fixes

2761d11

Fix use after free at Unexpected construction

82b984b

The Error type does not own any string, only string_view. Thus it the error message given must be likely a static string

do not assert on failed send()

fad6a58

send() can fail whenever the peer disconnects

fix unstable test: send Request_service after successful connection

c7fbd9c

Merge remote-tracking branch 'origin/main' into mw-com-as-socom-bridge

3b05b19

NEOatNHNG reviewed Apr 17, 2026

View reviewed changes

lurtz added 8 commits April 20, 2026 12:39

Merge remote-tracking branch 'origin/main' into mw-com-as-socom-bridge

4903906

Document why inplace_vector is not chosen

aa03b2a

Merge remote-tracking branch 'origin/main' into mw-com-as-socom-bridge

b03101a

Wait in benchmarks for free memory slots

f0a0948

rerun memory profiling with dedicated application

66e0dea

rerun benchmark

d04796e

Change Fixed_size_container::size to size_t type

d8bc116

Merge remote-tracking branch 'origin/main' into mw-com-as-socom-bridge

184784d

lurtz and others added 3 commits June 2, 2026 09:24

use newer communication module with public message_passing

d300016

Merge remote-tracking branch 'origin/main' into mw-com-as-socom-bridge

6e0a468

Merge branch 'main' into mw-com-as-socom-bridge

4e4d8a5

mariuswbr suggested changes Jun 10, 2026

View reviewed changes

Comment thread src/gateway_ipc_binding/src/binding_base.cpp

Comment thread src/gateway_ipc_binding/src/binding_base.cpp Outdated

lurtz and others added 25 commits June 11, 2026 19:19

Merge remote-tracking branch 'origin/main' into mw-com-as-socom-bridge

26c8e70

bump MODULE.bazel.lock

a51188b

remove done TODO

b9f66e7

remove race when calling Gateway_ipc_binding_server::start() multiple…

cf3c621

… times from multiple threads

Test and fix service connection setup at late connect

323ec73

test cleanup

127b4b3

test IPC destruction with one side connected

1ed211b

Test and fix IPC destruction with connected service

195e5ef

move setup into the fixture

cf00e0b

Client_connectors state change callback is called upon IPC connection…

796cb6d

… loss

Test restarting IPC connection works

f521873

test IPC server restart with active service running

3655213

check at socom::Runtime already the number of registered client conne…

9723cc0

…ctors

Merge remote-tracking branch 'origin/main' into mw-com-as-socom-bridge

dca42ac

Support IPC client destruction and reconnect

6472764

Stability fixes

3e3322a

Fix data race at IPC client destruction

2d1d679

Run m_channel->Stop() in tight loop in IPC client destructor

5d51f25

socom::Payload API for shrinking payloads

a82795d

Test different services offered on client and server side and documen…

1a52ead

…t shared memory configuration for method calls and event updates.

Merge branch 'main' into mw-com-as-socom-bridge

f24f390

Merge remote-tracking branch 'origin/main' into mw-com-as-socom-bridge

3d826ea

Merge remote-tracking branch 'elektrobit/mw-com-as-socom-bridge' into…

8c0e30a

… mw-com-as-socom-bridge

Remove generated profiling

d2c82fb

Merge branch 'main' into mw-com-as-socom-bridge

bb94084

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

feat: IPC binding for socom#86

feat: IPC binding for socom#86
lurtz wants to merge 87 commits into
eclipse-score:mainfrom
elektrobit-contrib:mw-com-as-socom-bridge

lurtz commented Apr 15, 2026 •

edited

Loading

Uh oh!

github-actions Bot commented Apr 15, 2026 •

edited

Loading

Uh oh!

github-actions Bot commented Apr 15, 2026

Uh oh!

Uh oh!

NEOatNHNG Apr 17, 2026

Uh oh!

lurtz Apr 17, 2026

Uh oh!

lurtz Apr 22, 2026

Uh oh!

lurtz Apr 28, 2026

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

lurtz commented Jun 8, 2026 •

edited

Loading

Uh oh!

mariuswbr left a comment

Uh oh!

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

Uh oh!

Conversation

lurtz commented Apr 15, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

TODO

Uh oh!

github-actions Bot commented Apr 15, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

License Check Results

Uh oh!

github-actions Bot commented Apr 15, 2026

Uh oh!

Uh oh!

NEOatNHNG Apr 17, 2026

Choose a reason for hiding this comment

Uh oh!

lurtz Apr 17, 2026

Choose a reason for hiding this comment

Uh oh!

lurtz Apr 22, 2026

Choose a reason for hiding this comment

Uh oh!

lurtz Apr 28, 2026

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

lurtz commented Jun 8, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

mariuswbr left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

lurtz commented Apr 15, 2026 •

edited

Loading

github-actions Bot commented Apr 15, 2026 •

edited

Loading

lurtz commented Jun 8, 2026 •

edited

Loading