Repository navigation
Conversation
Signed-off-by: hsgwa <hasegawa.isp@gmail.com>
Signed-off-by: hsgwa <hasegawa.isp@gmail.com>
* use regex for wildcard matching (ros2#1839) * use regex for wildcard matching Co-authored-by: Aaron Lipinski <aaron.lipinski@roboticsplus.co.nz> Signed-off-by: Chen Lihui <lihui.chen@sony.com> * use map to process the content of parameter file by order Signed-off-by: Chen Lihui <lihui.chen@sony.com> * add more test cases Signed-off-by: Chen Lihui <lihui.chen@sony.com> * try to not decrease the performance and make the param win last Signed-off-by: Chen Lihui <lihui.chen@sony.com> * update node name Signed-off-by: Chen Lihui <lihui.chen@sony.com> * update document comment Signed-off-by: Chen Lihui <lihui.chen@sony.com> * add more test for parameter_map_from Signed-off-by: Chen Lihui <lihui.chen@sony.com> Co-authored-by: Aaron Lipinski <aaron.lipinski@roboticsplus.co.nz> (cherry picked from commit 6dd3a03) * not to break ABI Signed-off-by: Chen Lihui <lihui.chen@sony.com> Signed-off-by: Chen Lihui <lihui.chen@sony.com> Co-authored-by: Chen Lihui <lihui.chen@sony.com>
* Add a test case a subscriber on a new executor with a callback group triggered to receive a message Signed-off-by: Chen Lihui <lihui.chen@sony.com> * fix flaky and not to use spin_some Signed-off-by: Chen Lihui <lihui.chen@sony.com> * update comment Signed-off-by: Chen Lihui <lihui.chen@sony.com> * update for not using anti-pattern source code Signed-off-by: Chen Lihui <lihui.chen@sony.com> * add a notify guard condition for callback group Co-authored-by: William Woodall <william@osrfoundation.org> Signed-off-by: Chen Lihui <lihui.chen@sony.com> * Notify guard condition of Node not to be used in Executor it is only for the waitset of GraphListener Signed-off-by: Chen Lihui <lihui.chen@sony.com> * put code in the try catch Signed-off-by: Chen Lihui <lihui.chen@sony.com> * defer to create guard condition Signed-off-by: Chen Lihui <lihui.chen@sony.com> * use context directly for the create function Signed-off-by: Chen Lihui <lihui.chen@sony.com> * cpplint Signed-off-by: Chen Lihui <lihui.chen@sony.com> * fix that some case might call add_node after shutdown Signed-off-by: Chen Lihui <lihui.chen@sony.com> * nitpick and fix some potential bug Signed-off-by: Chen Lihui <lihui.chen@sony.com> * add sanity check as some case might not create notify guard condition after shutdown Signed-off-by: Chen Lihui <lihui.chen@sony.com> * Cleanup includes. Signed-off-by: Chris Lalancette <clalancette@openrobotics.org> * remove destroy method make a callback group can only create one guard condition Signed-off-by: Chen Lihui <lihui.chen@sony.com> * remove limitation that guard condition can not be re-created in callback group Signed-off-by: Chen Lihui <lihui.chen@sony.com> Signed-off-by: Chen Lihui <lihui.chen@sony.com> Signed-off-by: Chris Lalancette <clalancette@openrobotics.org> Co-authored-by: William Woodall <william@osrfoundation.org> Co-authored-by: Chris Lalancette <clalancette@openrobotics.org> (cherry picked from commit d119157) Co-authored-by: Chen Lihui <lihui.chen@sony.com>
…2#2065) * fix nullptr dereference in prune_requests_older_than Signed-off-by: akela1101 <akela1101@gmail.com> * add tests for prune_requests_older_than Signed-off-by: akela1101 <akela1101@gmail.com> * Update rclcpp/test/rclcpp/test_client.cpp Co-authored-by: Chen Lihui <lihui.chen@sony.com> Signed-off-by: akela1101 <akela1101@gmail.com> Signed-off-by: akela1101 <akela1101@gmail.com> Co-authored-by: Chen Lihui <lihui.chen@sony.com> (cherry picked from commit 1ac37b6) Co-authored-by: andrei <akela1101@gmail.com>
backport of ros2#2132 Signed-off-by: Tomoya Fujita <Tomoya.Fujita@sony.com> Co-authored-by: Chen Lihui <lihui.chen@sony.com>
…2167) If the intraprocess buffer still has data after taking, re-trigger the guard condition to ensure that the executor will continue to service it, even if incoming publications stop. Signed-off-by: Michael Carroll <michael@openrobotics.org> (cherry picked from commit 5f9695a) Co-authored-by: Michael Carroll <michael@openrobotics.org>
…os2#2219) (ros2#2223) ros2#2167 (comment) Signed-off-by: Tomoya Fujita <Tomoya.Fujita@sony.com> (cherry picked from commit fe2e0e4) Co-authored-by: Tomoya Fujita <Tomoya.Fujita@sony.com>
* change publish tp def Signed-off-by: yamasaki <114902604+ymski@users.noreply.github.com> * added publisher_handle Signed-off-by: yamasaki <114902604+ymski@users.noreply.github.com> --------- Signed-off-by: yamasaki <114902604+ymski@users.noreply.github.com>
Merge commit 'ae8b033ae01601475104bf9e4c2350b98c9560c4' into work_for_update_humble_tracepoint_added
Merge commit '9171122eae940504200889b2a7b12344fa7c0c8b' into work_for_update_humble_tracepoint_added
Merge commit '00ef09cbf34563db27b9aac9b0206eb142cf3995' into work_for_update_humble_tracepoint_added
Merge commit '52327dd3a3b0fc47446ce86ecaf16fdde4fc4a60' into work_for_update_humble_tracepoint_added
…nt_added chore: merge from humble 16.0.5
…ric subscription (tier4#7) * added initialization tracepoint to generic_subscription Signed-off-by: yamasaki <114902604+ymski@users.noreply.github.com> * added callback_register Signed-off-by: yamasaki <114902604+ymski@users.noreply.github.com> * added callback_start/end Signed-off-by: yamasaki <114902604+ymski@users.noreply.github.com> * added subscription-side trace point Signed-off-by: ymski <114902604+ymski@users.noreply.github.com> * add rclcpp_publish to generic_publisher Signed-off-by: ymski <114902604+ymski@users.noreply.github.com> * removed unused dispatch_subscription_callback trace point Signed-off-by: ymski <114902604+ymski@users.noreply.github.com> * Update rclcpp/src/rclcpp/generic_subscription.cpp Co-authored-by: isp-uetsuki <35490433+isp-uetsuki@users.noreply.github.com> * fix messege addr Signed-off-by: ymski <114902604+ymski@users.noreply.github.com> * remove redundant includes Signed-off-by: ymski <114902604+ymski@users.noreply.github.com> --------- Signed-off-by: yamasaki <114902604+ymski@users.noreply.github.com> Signed-off-by: ymski <114902604+ymski@users.noreply.github.com> Co-authored-by: isp-uetsuki <35490433+isp-uetsuki@users.noreply.github.com>
nago-7563
marked this pull request as draft
October 29, 2026 11:10
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.
Description
This PR increases the service queue depth of
~/_container/load_nodeand~/_container/unload_nodeinComponentManagerfrom the default 10 to 200.The same issue was reported upstream in ros2#2362 and fixed in ros2#2363 (backported to Humble in ros2#2380). The
humble_tracepoint_addedbranch is based on 16.0.5 and does not include this fix (the fix is in ros2/rclcpp Humble since 16.0.8).Instead of cherry-picking ros2#2380, the change is written to match the current code of
rclcpp_components/src/component_manager.cppon the ros2/rclcpphumblebranch. For this part, that code is identical to ros2#2380.list_nodesis not changed, as in upstream.Problem
With the default service QoS (KEEP_LAST, depth 10), when more than 10 LoadNode requests are pending in a single-threaded component container, the oldest requests are overwritten in the DDS reader history and silently dropped. The DDS layer has already ACKed them, so the client (launch_ros) waits forever and the node is never loaded.
In our environment (Autoware + AWSIM, ROS 2 Humble, CycloneDDS, built with CARET),
component_state_monitor.launch.pysends 15 LoadNode requests in parallel to/system/component_state_monitor/containerat startup. Under CPU load (e.g., with a limited number of cores or a lowered CPU frequency), sometopic_state_monitornodes were not loaded, and Autoware could not enter autonomous mode. Without CARET, the same setup works because the stock Humblerclcpp_componentsalready has the fix.We confirmed with a CycloneDDS discovery trace that the request reader of
load_nodehashistory=0:10(KEEP_LAST, 10) before this change andhistory=0:200after.Reproduction
Same as ros2#2362: one
component_containerand 30LoadComposableNodesofcomposition::Talkersent in parallel; count loaded nodes withros2 component listafter 15 s.setenv_caret.bash), before this change/opt/ros/humble)With this change applied, the
topic_state_monitorloading failure no longer occurs in our Autoware startup tests, including runs with restricted CPU resources.Related