Added CMs MultiThreadedExecutor as an arg to `on_init()` #2323

soham2560 · 2025-06-09T10:06:25Z

Brief

This addition is the first step towards solving #2141 (also #1732 in some way). The intention currently is to give the user access to the MultiThreadedExecutor so that they can add nodes to it and publish if necessary. A future PR will add a node by default with the hardware_components name

Upon further review, another crucial change is introduced in this PR, we are constructing three structs

HardwareComponentInterfaceParams
- To be passed in on_init() and parsed by user to get all relevant data (hardware_info and executor at this point)
HardwareComponentParams
- To be used for all other initialisation, loading calls up till resource manager
ResourceManagerParams
- To construct Resource Manager

How was this tested?

Unfortunately the demo examples aren't directly compatible due to the changed API so had to modify them a bit, but was able to add a node to the executor and check

Couple of points to note

This does mean we have to update the ros2_control_demos with the updated API
It is expected the user will lock and use the executor as needed
documentation is yet to be added, in particular about notifying the user to not abuse the executor (since they can use the cancel() call internally and cripple the services etc for the CM basically )

codecov · 2025-06-09T10:47:45Z

Codecov Report

Attention: Patch coverage is 87.28070% with 29 lines in your changes missing coverage. Please review.

Project coverage is 88.95%. Comparing base (f97a2cb) to head (c4b02f1).

Files with missing lines	Patch %	Lines
hardware_interface/src/resource_manager.cpp	70.00%	8 Missing and 4 partials ⚠️
...rface_testing/test/test_components/test_sensor.cpp	25.00%	2 Missing and 1 partial ⚠️
...rface_testing/test/test_components/test_system.cpp	25.00%	2 Missing and 1 partial ⚠️
controller_manager/src/controller_manager.cpp	50.00%	2 Missing ⚠️
...hardware_components/test_single_joint_actuator.cpp	0.00%	2 Missing ⚠️
...e_interface/src/mock_components/generic_system.cpp	75.00%	0 Missing and 1 partial ⚠️
...t_hardware_components/test_force_torque_sensor.cpp	50.00%	0 Missing and 1 partial ⚠️
.../test/test_hardware_components/test_imu_sensor.cpp	50.00%	0 Missing and 1 partial ⚠️
...ware_components/test_system_with_command_modes.cpp	50.00%	0 Missing and 1 partial ⚠️
...test_hardware_components/test_two_joint_system.cpp	50.00%	0 Missing and 1 partial ⚠️
... and 2 more

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #2323      +/-   ##
==========================================
- Coverage   88.97%   88.95%   -0.02%     
==========================================
  Files         143      145       +2     
  Lines       16608    16664      +56     
  Branches     1429     1432       +3     
==========================================
+ Hits        14777    14824      +47     
- Misses       1275     1281       +6     
- Partials      556      559       +3

Flag	Coverage Δ
unittests	`88.95% <87.28%> (-0.02%)`	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

Files with missing lines	Coverage Δ
...er_manager/test/controller_manager_test_common.hpp	`89.83% <100.00%> (ø)`
...er/test/test_controller_manager_with_namespace.cpp	`100.00% <100.00%> (ø)`
...t_controllers_chaining_with_controller_manager.cpp	`99.25% <100.00%> (+<0.01%)`	⬆️
..._interface/include/hardware_interface/actuator.hpp	`100.00% <ø> (ø)`
.../include/hardware_interface/actuator_interface.hpp	`90.96% <100.00%> (ø)`
...re_interface/include/hardware_interface/sensor.hpp	`100.00% <ø> (ø)`
...ce/include/hardware_interface/sensor_interface.hpp	`85.00% <100.00%> (ø)`
...re_interface/include/hardware_interface/system.hpp	`100.00% <ø> (ø)`
...ce/include/hardware_interface/system_interface.hpp	`84.26% <100.00%> (ø)`
...face/types/hardware_component_interface_params.hpp	`100.00% <100.00%> (ø)`
... and 23 more

🚀 New features to boost your workflow:

❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

saikishor

A partial review here

I think as this is an API breaking change, we should open the corresponding PRs in the packages like gz_ros2_control and also webots_ros2_control probably? We can discuss this part when we finish the final review of this PR

hardware_interface/include/hardware_interface/actuator_interface.hpp

hardware_interface/include/hardware_interface/sensor_interface.hpp

hardware_interface/include/hardware_interface/actuator_interface.hpp

hardware_interface/include/hardware_interface/system_interface.hpp

soham2560 · 2025-06-09T13:24:04Z

we should open the corresponding PRs in the packages like gz_ros2_control

@saikishor have added PR at ros-controls/gz_ros2_control#606

…ce.hpp Co-authored-by: Sai Kishor Kothakota <[email protected]>

….hpp Co-authored-by: Sai Kishor Kothakota <[email protected]>

…ce.hpp Co-authored-by: Sai Kishor Kothakota <[email protected]>

….hpp Co-authored-by: Sai Kishor Kothakota <[email protected]>

saikishor · 2025-06-09T14:07:40Z

we should open the corresponding PRs in the packages like gz_ros2_control

@saikishor have added PR at ros-controls/gz_ros2_control#606

Awesome. Sorry I missed that one

….hpp Co-authored-by: Sai Kishor Kothakota <[email protected]>

Co-authored-by: Sai Kishor Kothakota <[email protected]>

…t.cpp Co-authored-by: Sai Kishor Kothakota <[email protected]>

christophfroehlich

Please check the pre-commit failures (have you installed it properly?)

soham2560 · 2025-06-10T20:32:42Z

Please check the pre-commit failures (have you installed it properly?)

apologies, forgot to rerun pre-commit after the suggested changes

saikishor

The changes itself LGTM

soham2560 · 2025-06-11T11:20:08Z

after some discussion with @saikishor and @bmagyar we have decided to go for a Pimpl approach on this PR, putting in draft till then

soham2560 · 2025-06-15T12:31:36Z

@bmagyar @saikishor do let me know how this looks

saikishor

Thank you for the changes. They are very good

Some nitpicking and other recommended changes

saikishor · 2025-06-15T20:29:57Z

hardware_interface/include/hardware_interface/types/hardware_component_interface_params.hpp

+  explicit HardwareComponentInterfaceParams(
+    const hardware_interface::HardwareInfo & hardware_info_param,
+    rclcpp::Executor::WeakPtr executor_param)
+  : hardware_info(hardware_info_param), executor(executor_param)
+  {
+  }
+
+  explicit HardwareComponentInterfaceParams(hardware_interface::HardwareComponentParams & params)
+  : hardware_info(params.hardware_info), executor(params.executor)
+  {
+  }
+};


I'm not sure if it is better to have no constructors here, because at some point, we might break this one too. Instead, filling in the information individually per element is better. We can always add them back in future

I don't have a strong opinion here

refer #2323 (comment)

saikishor · 2025-06-15T20:31:03Z

hardware_interface/include/hardware_interface/types/hardware_component_params.hpp

+  explicit HardwareComponentParams(
+    const hardware_interface::HardwareInfo & hardware_info_param, rclcpp::Logger logger_param,
+    rclcpp::Clock::SharedPtr clock_param, rclcpp::Executor::WeakPtr executor_param)
+  : hardware_info(hardware_info_param),
+    logger(logger_param),
+    clock(clock_param),
+    executor(executor_param)
+  {
+  }


refer #2323 (comment)

hardware_interface/include/hardware_interface/types/resource_manager_params.hpp

saikishor · 2025-06-16T20:42:34Z

hardware_interface/include/hardware_interface/resource_manager.hpp

+    rclcpp::node_interfaces::NodeLoggingInterface::SharedPtr logger_interface,
+    rclcpp::Executor::SharedPtr executor);

  /// Default constructor for the Resource Manager.
-  explicit ResourceManager(rclcpp::Clock::SharedPtr clock, rclcpp::Logger logger);
+  explicit ResourceManager(
+    rclcpp::Clock::SharedPtr clock, rclcpp::Logger logger, rclcpp::Executor::SharedPtr executor);


With the internal structure in place, do we have to extend the other constructors?. For the newer features, they can directly use the newer struct right?

The thing in my mind was that it seems odd to just extend one variant, maybe a seperate constructor not dependent on previous ones would be better what do you think?

saikishor · 2025-06-16T20:43:17Z

hardware_interface/include/hardware_interface/sensor.hpp

+    rclcpp::node_interfaces::NodeClockInterface::SharedPtr clock_interface,
+    rclcpp::Executor::WeakPtr executor);


same here. do we have to extend this one too?

saikishor · 2025-06-16T20:43:41Z

hardware_interface/include/hardware_interface/system.hpp

  const rclcpp_lifecycle::State & initialize(
    const HardwareInfo & system_info, rclcpp::Logger logger,
-    rclcpp::node_interfaces::NodeClockInterface::SharedPtr clock_interface);
+    rclcpp::node_interfaces::NodeClockInterface::SharedPtr clock_interface,
+    rclcpp::Executor::WeakPtr executor);


saikishor · 2025-06-16T21:32:54Z

hardware_interface/src/resource_manager.cpp

+    hardware_interface::HardwareComponentParams params(
+      individual_hardware_info, resource_storage_->rm_logger_, resource_storage_->rm_clock_,
+      executor_);


May be we should store the params in the class, so we could use it later? This can help us avoid any human errors

makes sense, should we do it for all the structs or just this one in your opinion? and I am assuming it should be stored in resource manager publicly?

saikishor · 2025-06-16T21:33:47Z

hardware_interface/test/test_component_interfaces.cpp

+  auto test_executor = std::make_shared<rclcpp::executors::SingleThreadedExecutor>();
+  auto state = sensor_hw.initialize(
+    mock_hw_info, node->get_logger(), node->get_node_clock_interface(), test_executor);


If we would have left the old method, we don't have to create all these changes

…anager_params.hpp Co-authored-by: Sai Kishor Kothakota <[email protected]>

…omponent_params.hpp Co-authored-by: Sai Kishor Kothakota <[email protected]>

bmagyar · 2025-06-17T21:24:31Z

hardware_interface/include/hardware_interface/actuator_interface.hpp

@@ -158,18 +163,23 @@ class ActuatorInterface : public rclcpp_lifecycle::node_interfaces::LifecycleNod
        info_.thread_priority);
      async_handler_->start_thread();
    }
-    return on_init(hardware_info);
+    return on_init(hardware_interface::HardwareComponentInterfaceParams(params));


not sure about it but if we did not mark the constructor explicit, this could be as short as

Suggested change

return on_init(hardware_interface::HardwareComponentInterfaceParams(params));

return on_init(params);

soham2560 added 2 commits June 9, 2025 09:39

propagated executor from cm to hw.on_init

6e4c3c8

updated tests to include executor

345c055

soham2560 marked this pull request as draft June 9, 2025 10:06

github-actions bot requested review from ARK3r, christophfroehlich, destogl, homalozoa, olivier-stasse, progtologist and VX792 June 9, 2025 10:06

reverted import_component changes

50b42a2

soham2560 marked this pull request as ready for review June 9, 2025 10:50

github-actions bot requested review from livanov93, mcbed, mhubii, moriarty, rosterloh and saikishor June 9, 2025 10:50

saikishor requested changes Jun 9, 2025

View reviewed changes

soham2560 mentioned this pull request Jun 9, 2025

Added support for executor as an arg to on_init() ros-controls/gz_ros2_control#606

Draft

soham2560 and others added 6 commits June 9, 2025 18:55

Update hardware_interface/include/hardware_interface/actuator_interfa…

12ba419

…ce.hpp Co-authored-by: Sai Kishor Kothakota <[email protected]>

Update hardware_interface/include/hardware_interface/sensor_interface…

96cb747

….hpp Co-authored-by: Sai Kishor Kothakota <[email protected]>

Update hardware_interface/include/hardware_interface/sensor_interface…

d098c20

….hpp Co-authored-by: Sai Kishor Kothakota <[email protected]>

Update hardware_interface/include/hardware_interface/actuator_interfa…

72f3253

…ce.hpp Co-authored-by: Sai Kishor Kothakota <[email protected]>

Update hardware_interface/include/hardware_interface/system_interface…

e0a72a4

….hpp Co-authored-by: Sai Kishor Kothakota <[email protected]>

Update hardware_interface/include/hardware_interface/system_interface…

3541ab9

….hpp Co-authored-by: Sai Kishor Kothakota <[email protected]>

soham2560 mentioned this pull request Jun 9, 2025

Added support for executor as an arg to on_init() ros-controls/ros2_control_demos#829

Draft

soham2560 and others added 10 commits June 9, 2025 20:19

Update hardware_interface/include/hardware_interface/resource_manager…

574a12d

….hpp Co-authored-by: Sai Kishor Kothakota <[email protected]>

Update hardware_interface/include/hardware_interface/resource_manager…

f9ca601

….hpp Co-authored-by: Sai Kishor Kothakota <[email protected]>

Update hardware_interface/include/hardware_interface/resource_manager…

6045aeb

….hpp Co-authored-by: Sai Kishor Kothakota <[email protected]>

Update hardware_interface/include/hardware_interface/resource_manager…

1e18428

….hpp Co-authored-by: Sai Kishor Kothakota <[email protected]>

Update hardware_interface/include/hardware_interface/resource_manager…

eeec025

….hpp Co-authored-by: Sai Kishor Kothakota <[email protected]>

Update hardware_interface/include/hardware_interface/resource_manager…

4a29280

….hpp Co-authored-by: Sai Kishor Kothakota <[email protected]>

Update hardware_interface/include/hardware_interface/resource_manager…

0d5f0bf

….hpp Co-authored-by: Sai Kishor Kothakota <[email protected]>

Update hardware_interface/test/test_component_interfaces.cpp

a51ef08

Co-authored-by: Sai Kishor Kothakota <[email protected]>

Update hardware_interface/test/test_component_interfaces_custom_expor…

2037255

…t.cpp Co-authored-by: Sai Kishor Kothakota <[email protected]>

updated pointer assignment

e2af20f

christophfroehlich requested changes Jun 10, 2025

View reviewed changes

updated pre-commit changes

ee91dc4

soham2560 requested review from christophfroehlich and saikishor June 11, 2025 07:45

saikishor reviewed Jun 11, 2025

View reviewed changes

soham2560 marked this pull request as draft June 11, 2025 11:20

soham2560 added 3 commits June 15, 2025 10:54

added HardwareComponent and HardwareComponentInterface params

ad598c3

added ResourceManageParams propagation

70a4ccc

updated params doc and constructors

8ed94f1

soham2560 marked this pull request as ready for review June 15, 2025 12:31

saikishor requested changes Jun 16, 2025

View reviewed changes

soham2560 and others added 5 commits June 17, 2025 08:22

Update hardware_interface/include/hardware_interface/types/resource_m…

d2418dd

…anager_params.hpp Co-authored-by: Sai Kishor Kothakota <[email protected]>

Update hardware_interface/include/hardware_interface/types/resource_m…

0e9276f

…anager_params.hpp Co-authored-by: Sai Kishor Kothakota <[email protected]>

Update hardware_interface/include/hardware_interface/types/resource_m…

f626063

…anager_params.hpp Co-authored-by: Sai Kishor Kothakota <[email protected]>

Update hardware_interface/include/hardware_interface/types/resource_m…

2191977

…anager_params.hpp Co-authored-by: Sai Kishor Kothakota <[email protected]>

Update hardware_interface/include/hardware_interface/types/hardware_c…

c4b02f1

…omponent_params.hpp Co-authored-by: Sai Kishor Kothakota <[email protected]>

bmagyar reviewed Jun 17, 2025

View reviewed changes

		rclcpp::node_interfaces::NodeClockInterface::SharedPtr clock_interface,
		rclcpp::Executor::WeakPtr executor);

	return on_init(hardware_interface::HardwareComponentInterfaceParams(params));
	return on_init(params);

Added CMs MultiThreadedExecutor as an arg to on_init() #2323

Are you sure you want to change the base?

Added CMs MultiThreadedExecutor as an arg to on_init() #2323

Uh oh!

Conversation

soham2560 commented Jun 9, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Brief

How was this tested?

Uh oh!

codecov bot commented Jun 9, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

saikishor left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

soham2560 commented Jun 9, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

saikishor commented Jun 9, 2025

Uh oh!

christophfroehlich left a comment

Choose a reason for hiding this comment

Uh oh!

soham2560 commented Jun 10, 2025

Uh oh!

saikishor left a comment

Choose a reason for hiding this comment

Uh oh!

soham2560 commented Jun 11, 2025

Uh oh!

soham2560 commented Jun 15, 2025

Uh oh!

saikishor 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!

Uh oh!

Uh oh!

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!

Uh oh!

Added CMs MultiThreadedExecutor as an arg to `on_init()` #2323

Added CMs MultiThreadedExecutor as an arg to `on_init()` #2323

soham2560 commented Jun 9, 2025 •

edited

Loading

codecov bot commented Jun 9, 2025 •

edited

Loading

soham2560 commented Jun 9, 2025 •

edited

Loading