Manage DataModel::Provider lifecycle in DeviceControllerFactory #42272
Conversation
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
|
|
![gemini-code-assist[bot]](https://avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
Contributor
There was a problem hiding this comment.
Code Review
This pull request correctly addresses a dangling pointer issue in the InteractionModelEngine singleton by ensuring the DataModelProvider is properly shut down and reset. This is a crucial fix for ensuring hermetic behavior in unit tests. The added test case effectively validates this fix.
tcarmelveilleux
approved these changes
Dec 3, 2025
|
PR #42272: Size comparison from 078ba11 to f404cad Full report (21 builds for bl602, bl702, bl702l, cc13x4_26x4, cc32xx, psoc6, qpg, realtek, stm32)
|
|
PR #42272: Size comparison from 078ba11 to 73b6449 Full report (34 builds for bl602, bl702, bl702l, cc13x4_26x4, cc32xx, efr32, esp32, nxp, psoc6, qpg, realtek, stm32, telink)
|
|
PR #42272: Size comparison from 078ba11 to 97c267e Full report (34 builds for bl602, bl702, bl702l, cc13x4_26x4, cc32xx, efr32, esp32, nxp, psoc6, qpg, realtek, stm32, telink)
|
|
PR #42272: Size comparison from 078ba11 to 7594633 Full report (35 builds for bl602, bl702, bl702l, cc13x4_26x4, cc32xx, efr32, esp32, nrfconnect, nxp, psoc6, qpg, realtek, stm32, telink)
|
bzbarsky-apple
requested changes
Dec 5, 2025
src/app/InteractionModelEngine.cpp
Outdated
| // mpFabricTable = nullptr; | ||
| // mpExchangeMgr = nullptr; | ||
|
|
||
| if (mDataModelProvider) |
Contributor
There was a problem hiding this comment.
Please see discussion in #42149
Since IME startup is not what starts the data model provider, it's weird to have shutdown shutting it down.... We need to figure out whether the startup needs to move or whether the clearing of the data model provider needs to happen in the same scope as the setting (i.e. in the relevant test scope) as needed.
Contributor
Author
There was a problem hiding this comment.
Ack. Thank you for the reference to the discussion.
This naive approach was based on InteractionModelEngine::SetDataModelProvider calling Shutdown on the old DataModel::Provider.
I reached out offline to @andy31415 for guidance.
Contributor
There was a problem hiding this comment.
Yeah, I think we all agree the current semantics are bad.... If this is blocking something, we could land it as a stopgap, but long-term we really don't want shutdown shutting down things that init/startup did not start up.
Contributor
Author
There was a problem hiding this comment.
This is currently blocking testing with code-driven data model provider, which in turn blocks the use of code-driven data model provider.
That said, this can't land as-is for the same reasons you mentioned. Will need to wait for additional discussions with @andy31415 to clarify intended use.
... but long-term we really don't want shutdown shutting down things that init/startup did not start up.
Would you mind clarifying what you mean by this? As far as I can see InteractionModelEngine::SetDataModelProvider does initialize the DataModel::Provider (via call to Startup).
Contributor
Author
There was a problem hiding this comment.
Oh, I guess you mean InteractionModelEngine::Init/Shutdown ... fair enough.
Contributor
Author
There was a problem hiding this comment.
I guess for my uses, calling SetDataModelProvider with a nullptr may unblock work.
Contributor
There was a problem hiding this comment.
For the controller factory, if init sets to non-null, shutdown can set to null ... we should generally be symmetric.
dnicoara
changed the title
|
PR #42272: Size comparison from ba864e9 to fb240d3 Full report (5 builds for cc32xx, realtek, stm32)
|
* On Shutdown call InteractionModelEngine::SetDataModelProvider with nullptr * Cache DataModel::Provider within DeviceControllerFactory to allow it it reinitialize.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
5 participants
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.
Manage DataModel::Provider lifecycle in DeviceControllerFactory
nullptr
it reinitialize.
Summary
InteractionModelEngine is a global singleton and does not get re-instantiated for each test. DataModelProvider delegate is never cleaned up during Shutdown, thus the pointer leaks across tests. During a second test InteractionModelEngine::SetDataModelProvider() will attempt to call Shutdown on the existing pointer. If DataModelProvider is local to the test this will crash as the old DataModelProvider would have been deleted.
Testing
CI
Readability checklist
The checklist below will help the reviewer finish PR review in time and keep the
code readable:
descriptive
“When in Rome…”
rule (coding style)
See: Pull Request Guidelines