Fix bug related to resetting of KernelScheduler._currentlyRunningTopLevelOperation #3235
Conversation
This file contains 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
…evelOperation. _currentlyRunningTopLevelOperation was not being reset in the case where a top level operation would schedule child operations. In this case, _currentlyRunningOperation was correctly being reset when each child operation was completed. However, when the parent operation was completed, we would fail to reset _currentlyRunningTopLevelOperation since we were comparing _currentlyRunningTopLevelOperation with the value of _currentlyRunningOperation (and currentlyRunningOperation would have a value null because it was reset to null when the child operation was completed above). The above bug was breaking parenting (and command token assignment) for subsequent commands. Subsequent commands were being incorrectly parented to the stale value reflected in _currentlyRunningTopLevelOperation. And this in turn was leading to hangs. This was an unfortunate regression triggered by a previous change that fixed a race condition. The regression happened because were were missing a test for the above scenario. I have included some tests for this as part of the current commit.
jonsequitur
approved these changes
Oct 9, 2023
shyamnamboodiripad
added a commit
to shyamnamboodiripad/interactive
that referenced
this pull request
Oct 9, 2023
…evelOperation (dotnet#3235) `_currentlyRunningTopLevelOperation` was not being reset in the case where a top-level operation would schedule child operations. In this case, `_currentlyRunningOperation` was correctly being reset when each child operation was completed. However, when the parent operation was completed, we would fail to reset `_currentlyRunningTopLevelOperation` since we were comparing `_currentlyRunningTopLevelOperation` with the value of `_currentlyRunningOperation` (and `_currentlyRunningOperation` would have a value null at this point because it was reset to null when the child operation was completed above). The above bug was breaking parenting (and command token assignment) for subsequent commands. Subsequent commands were being incorrectly parented to the stale value reflected in `_currentlyRunningTopLevelOperation`. And this in turn was leading to hangs. This was an unfortunate regression triggered by the changes in dotnet#3222 that fixed a race condition. The regression happened because were missing tests that exercised the above scenario. I have included some tests for this as part of the current commit. Also add a skipped test for related issue dotnet#3236.
shyamnamboodiripad
added a commit
that referenced
this pull request
Oct 9, 2023
…rrentlyRunningTopLevelOperation (#3237) Cherry-picked from #3235. `_currentlyRunningTopLevelOperation` was not being reset in the case where a top-level operation would schedule child operations. In this case, `_currentlyRunningOperation` was correctly being reset when each child operation was completed. However, when the parent operation was completed, we would fail to reset `_currentlyRunningTopLevelOperation` since we were comparing `_currentlyRunningTopLevelOperation` with the value of `_currentlyRunningOperation` (and `_currentlyRunningOperation` would have a value null at this point because it was reset to null when the child operation was completed above). The above bug was breaking parenting (and command token assignment) for subsequent commands. Subsequent commands were being incorrectly parented to the stale value reflected in `_currentlyRunningTopLevelOperation`. And this in turn was leading to hangs. This was an unfortunate regression triggered by the changes in #3222 that fixed a race condition. The regression happened because were missing tests that exercised the above scenario. I have included some tests for this as part of the current commit. Also add a skipped test for related issue #3236.
colombod
added a commit
that referenced
this pull request
Nov 22, 2023
* update dependencies from arcade * update to dotnet 8 * skip csharpproject kernel tests due to different binlog format * skip docker build step * set tool feed to experimental * update arcade sdk (#2697) * upgrade projects to <TargetFramework>net8.0</TargetFramework> * fix api contract * try multi taget * Update arcade * remove multi targets * Fix new warnings reported from .NET8 SDK. * new projcts * [feature/dotnet8] Fix race in KernelScheduler (#3222) * Use KernelCommandResult.Events in place of Kernel.Events. * Fix race. * Enable more complete logging for commands and events in tests. * Add a test. * [feature/dotnet8] Fix bug related to resetting of KernelScheduler._currentlyRunningTopLevelOperation (#3237) Cherry-picked from #3235. `_currentlyRunningTopLevelOperation` was not being reset in the case where a top-level operation would schedule child operations. In this case, `_currentlyRunningOperation` was correctly being reset when each child operation was completed. However, when the parent operation was completed, we would fail to reset `_currentlyRunningTopLevelOperation` since we were comparing `_currentlyRunningTopLevelOperation` with the value of `_currentlyRunningOperation` (and `_currentlyRunningOperation` would have a value null at this point because it was reset to null when the child operation was completed above). The above bug was breaking parenting (and command token assignment) for subsequent commands. Subsequent commands were being incorrectly parented to the stale value reflected in `_currentlyRunningTopLevelOperation`. And this in turn was leading to hangs. This was an unfortunate regression triggered by the changes in #3222 that fixed a race condition. The regression happened because were missing tests that exercised the above scenario. I have included some tests for this as part of the current commit. Also add a skipped test for related issue #3236. * update_sdk (#3320) * fix * Suppress new warning reported from .NET8 SDK. * these files are no longer needed * address comment * revert change --------- Co-authored-by: Brett V. Forsgren <brettfo@microsoft.com> Co-authored-by: Shyam Namboodiripad <gnamboo@microsoft.com> Co-authored-by: Shyam N <shyamnamboodiripad@users.noreply.github.com>
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.
_currentlyRunningTopLevelOperationwas not being reset in the case where a top-level operation would schedule child operations. In this case,_currentlyRunningOperationwas correctly being reset when each child operation was completed. However, when the parent operation was completed, we would fail to reset_currentlyRunningTopLevelOperationsince we were comparing_currentlyRunningTopLevelOperationwith the value of_currentlyRunningOperation(and_currentlyRunningOperationwould have a value null because it was reset to null when the child operation was completed above).The above bug was breaking parenting (and command token assignment) for subsequent commands. Subsequent commands were being incorrectly parented to the stale value reflected in
_currentlyRunningTopLevelOperation. And this in turn was leading to hangs.This was an unfortunate regression triggered by #3222 that fixed a race condition. The regression happened because were missing a test for the above scenario. I have included some tests for this as part of the current commit.