This repository was archived by the owner on Jan 23, 2023. It is now read-only.
Fix CancellationTokenRegistration.Unregister race condition #27949
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
tarekgh
approved these changes
Nov 26, 2019
maryamariyan
added
area-System.Threading
Servicing-consider
leecow
added
Servicing-approved
|
Moved to Mar as per Tactics. Will need to be explicitly approved for Feb, if required. |
|
Approved for Feb |
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
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Outdated suggestions cannot be applied.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.
Port dotnet/runtime#309 to release/3.1
Fixes https://github.com/dotnet/coreclr/issues/22946
Description
If multiple threads race to Unregister and zero out a CancellationTokenRegistration field, we can end up null ref'ing due to a torn read/write on a struct resulting in passing around some inconsistent state. This pattern is possible as part of clean up in Task.Delay(timeout, cancellationToken), with the null ref then happening on a thread pool thread that results in the app crashing.
Customer Impact
Valid coding patterns, both in customer code and in our framework implementation, resulting in NullReferenceExceptions that may not be catchable. Back in March a customer reported this, and just recently @rynowak hit up against it in ASP.NET.
Regression?
Yes, from .NET Framework 4.8 and .NET Core 2.0. This was introduced in #12819 in 2.1 as part of overhauling the CancellationToken implementation.
Testing
Isolated a repro that quickly fails (within a second) prior to the fix and successfully runs for a long time after.
Risk
Minimal.
cc: @tarekgh, @kouvel, @rynowak
@danmosemsft