worker: move JoinThread() back into exit callback #31468
Closed
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
de2c68c moved this call to the destructor, under the assumption that that would essentially be equivalent to running it as part of the callback since the worker would be destroyed along with the callback. However, the actual code in `Environment::RunAndClearNativeImmediates()` comes with the subtlety that testing whether a JS exception has been thrown happens between the invocation of the callback and its destruction, leaving a possible exception from `JoinThread()` potentially unhandled (and unintentionally silenced through the `TryCatch`). This affected exceptions thrown from the `'exit'` event of the Worker, and made the `parallel/test-worker-message-type-unknown` test flaky, as the invalid message was sometimes only received during the Worker thread’s exit handler. Fix this by moving the `JoinThread()` call back to where it was before. Refs: nodejs#31386
Prevent mistakes like the one fixed by the previous commit by destroying the callback immediately after it has been called.
nodejs-github-bot
added
c++
|
CI: https://ci.nodejs.org/job/node-test-pull-request/28576/ (:heavy_check_mark:) 👍 this comment to approve fast-tracking in order to fix CI. |
4 tasks
cjihrig
approved these changes
Jan 22, 2020
gireeshpunathil
approved these changes
Jan 23, 2020
|
Landed in 8fb5fe2...31b31a3 |
addaleax
added a commit
that referenced
this pull request
Jan 23, 2020
de2c68c moved this call to the destructor, under the assumption that that would essentially be equivalent to running it as part of the callback since the worker would be destroyed along with the callback. However, the actual code in `Environment::RunAndClearNativeImmediates()` comes with the subtlety that testing whether a JS exception has been thrown happens between the invocation of the callback and its destruction, leaving a possible exception from `JoinThread()` potentially unhandled (and unintentionally silenced through the `TryCatch`). This affected exceptions thrown from the `'exit'` event of the Worker, and made the `parallel/test-worker-message-type-unknown` test flaky, as the invalid message was sometimes only received during the Worker thread’s exit handler. Fix this by moving the `JoinThread()` call back to where it was before. Refs: #31386 PR-URL: #31468 Reviewed-By: Colin Ihrig <cjihrig@gmail.com> Reviewed-By: Gireesh Punathil <gpunathi@in.ibm.com>
codebytere
pushed a commit
that referenced
this pull request
Feb 17, 2020
de2c68c moved this call to the destructor, under the assumption that that would essentially be equivalent to running it as part of the callback since the worker would be destroyed along with the callback. However, the actual code in `Environment::RunAndClearNativeImmediates()` comes with the subtlety that testing whether a JS exception has been thrown happens between the invocation of the callback and its destruction, leaving a possible exception from `JoinThread()` potentially unhandled (and unintentionally silenced through the `TryCatch`). This affected exceptions thrown from the `'exit'` event of the Worker, and made the `parallel/test-worker-message-type-unknown` test flaky, as the invalid message was sometimes only received during the Worker thread’s exit handler. Fix this by moving the `JoinThread()` call back to where it was before. Refs: #31386 PR-URL: #31468 Reviewed-By: Colin Ihrig <cjihrig@gmail.com> Reviewed-By: Gireesh Punathil <gpunathi@in.ibm.com>
Merged
|
Blocked on backport of #31386 - if that's not intended to go back i can update the label! |
MylesBorins
pushed a commit
to addaleax/node
that referenced
this pull request
Apr 1, 2020
de2c68c moved this call to the destructor, under the assumption that that would essentially be equivalent to running it as part of the callback since the worker would be destroyed along with the callback. However, the actual code in `Environment::RunAndClearNativeImmediates()` comes with the subtlety that testing whether a JS exception has been thrown happens between the invocation of the callback and its destruction, leaving a possible exception from `JoinThread()` potentially unhandled (and unintentionally silenced through the `TryCatch`). This affected exceptions thrown from the `'exit'` event of the Worker, and made the `parallel/test-worker-message-type-unknown` test flaky, as the invalid message was sometimes only received during the Worker thread’s exit handler. Fix this by moving the `JoinThread()` call back to where it was before. Refs: nodejs#31386 PR-URL: nodejs#31468 Reviewed-By: Colin Ihrig <cjihrig@gmail.com> Reviewed-By: Gireesh Punathil <gpunathi@in.ibm.com>
MylesBorins
pushed a commit
to addaleax/node
that referenced
this pull request
Apr 1, 2020
Prevent mistakes like the one fixed by the previous commit by destroying the callback immediately after it has been called. PR-URL: nodejs#31468 Refs: nodejs#31386 Reviewed-By: Colin Ihrig <cjihrig@gmail.com> Reviewed-By: Gireesh Punathil <gpunathi@in.ibm.com>
MylesBorins
pushed a commit
that referenced
this pull request
Apr 1, 2020
de2c68c moved this call to the destructor, under the assumption that that would essentially be equivalent to running it as part of the callback since the worker would be destroyed along with the callback. However, the actual code in `Environment::RunAndClearNativeImmediates()` comes with the subtlety that testing whether a JS exception has been thrown happens between the invocation of the callback and its destruction, leaving a possible exception from `JoinThread()` potentially unhandled (and unintentionally silenced through the `TryCatch`). This affected exceptions thrown from the `'exit'` event of the Worker, and made the `parallel/test-worker-message-type-unknown` test flaky, as the invalid message was sometimes only received during the Worker thread’s exit handler. Fix this by moving the `JoinThread()` call back to where it was before. Refs: #31386 Backport-PR-URL: #32301 PR-URL: #31468 Reviewed-By: Colin Ihrig <cjihrig@gmail.com> Reviewed-By: Gireesh Punathil <gpunathi@in.ibm.com>
Merged
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.
worker: move JoinThread() back into exit callback
de2c68c moved this call to
the destructor, under the assumption that that would essentially
be equivalent to running it as part of the callback since the
worker would be destroyed along with the callback.
However, the actual code in
Environment::RunAndClearNativeImmediates()comes with the subtletythat testing whether a JS exception has been thrown
happens between the invocation of the callback and its destruction,
leaving a possible exception from
JoinThread()potentiallyunhandled (and unintentionally silenced through the
TryCatch).This affected exceptions thrown from the
'exit'event of theWorker, and made the
parallel/test-worker-message-type-unknowntest flaky, as the invalid message was sometimes only received
during the Worker thread’s exit handler.
Fix this by moving the
JoinThread()call back to where it wasbefore.
Refs: #31386
src: harden running native
SetImmediate()s slightlyPrevent mistakes like the one fixed by the previous commit
by destroying the callback immediately after it has been called.
Checklist
make -j4 test(UNIX), orvcbuild test(Windows) passes