Expiration should do nothing except disable time slicing #21345
Merged
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
We have a feature called "expiration" whose purpose is to prevent a concurrent update from being starved by higher priority events. If a lane is CPU-bound for too long, we finish the rest of the work synchronously without allowing further interruptions. In the current implementation, we do this in sort of a roundabout way: once a lane is determined to have expired, we entangle it with SyncLane and switch to the synchronous work loop. There are a few flaws with the approach. One is that SyncLane has a particular semantic meaning besides its non-yieldiness. For example, `flushSync` will force remaining Sync work to finish; currently, that also includes expired work, which isn't an intended behavior, but rather an artifact of the implementation. An event worse example is that passive effects triggered by a Sync update are flushed synchronously, before paint, so that its result is guaranteed to be observed by the next discrete event. But expired work has no such requirement: we're flushing expired effects before paint unnecessarily. Aside from the behaviorial implications, the current implementation has proven to be fragile: more than once, we've accidentally regressed performance due to a subtle change in how expiration is handled. This PR aims to radically simplify how we model starvation protection by scaling back the implementation as much as possible. In this new model, if a lane is expired, we disable time slicing. That's it. We don't entangle it with SyncLane. The only thing we do is skip the call to `shouldYield` in between each time slice. This is identical to how we model synchronous-by-default updates in React 18.
|
Comparing: 0f5ebf3...ac0cde2 Critical size changesIncludes critical production bundles, as well as any change greater than 2%:
Significant size changesIncludes any change greater than 0.2%: (No significant changes) |
rickhanlonii
approved these changes
Apr 24, 2021
acdlite
force-pushed
the
expiration-simplify
branch
from
2177c16
to
ef3816d
Compare
facebook-github-bot
added
CLA Signed
React Core Team
acdlite
force-pushed
the
expiration-simplify
branch
from
ef3816d
to
ac0cde2
Compare
facebook-github-bot
pushed a commit
to facebook/react-native
that referenced
this pull request
Apr 28, 2021
Summary: This sync includes the following changes: - **[9a2591681](facebook/react@9a2591681 )**: Fix export //<Sebastian Markbage>// - **[4a8deb083](facebook/react@4a8deb083 )**: Switch the isPrimaryRender flag based on the stream config ([#21357](facebook/react#21357)) //<Sebastian Markbåge>// - **[bd4f056a3](facebook/react@bd4f056a3 )**: [Fizz] Implement lazy components and nodes ([#21355](facebook/react#21355)) //<Sebastian Markbåge>// - **[fc33f12bd](facebook/react@fc33f12bd )**: Remove unstable scheduler/tracing API ([#20037](facebook/react#20037)) //<Brian Vaughn>// - **[721238394](facebook/react@721238394 )**: Enable strict effects mode for React Native Facebook builds ([#21354](facebook/react#21354)) //<Brian Vaughn>// - **[48740429b](facebook/react@48740429b )**: Expiration: Do nothing except disable time slicing ([#21345](facebook/react#21345)) //<Andrew Clark>// - **[0f5ebf366](facebook/react@0f5ebf366 )**: Delete unreferenced type ([#21343](facebook/react#21343)) //<Andrew Clark>// - **[9cd52b27f](facebook/react@9cd52b27f )**: Restore context after an error happens ([#21341](facebook/react#21341)) //<Sebastian Markbåge>// - **[ad091759a](facebook/react@ad091759a )**: Revert "Emit reactroot attribute on the first element we discover ([#21154](facebook/react#21154))" ([#21340](facebook/react#21340)) //<Sebastian Markbåge>// - **[709f94841](facebook/react@709f94841 )**: [Fizz] Add FB specific streaming API and build ([#21337](facebook/react#21337)) //<Sebastian Markbåge>// - **[e8cdce40d](facebook/react@e8cdce40d )**: Don't flush sync at end of discreteUpdates ([#21327](facebook/react#21327)) //<Andrew Clark>// - **[a15586001](facebook/react@a15586001 )**: Fix: Don't flush discrete at end of batchedUpdates ([#21229](facebook/react#21229)) //<Andrew Clark>// - **[89847bf6e](facebook/react@89847bf6e )**: Continuous updates should interrupt transitions ([#21323](facebook/react#21323)) //<Andrew Clark>// - **[ef37d55b6](facebook/react@ef37d55b6 )**: Use performConcurrentWorkOnRoot for "sync default" ([#21322](facebook/react#21322)) //<Andrew Clark>// Changelog: [General][Changed] - React Native sync for revisions a632f7d...2a7bb41 jest_e2e[run_all_tests] Reviewed By: JoshuaGross Differential Revision: D28063006 fbshipit-source-id: 7e3535f80961706863b6c2188ee44b5796b2f000
koto
pushed a commit
to koto/react
that referenced
this pull request
Jun 15, 2021
We have a feature called "expiration" whose purpose is to prevent a concurrent update from being starved by higher priority events. If a lane is CPU-bound for too long, we finish the rest of the work synchronously without allowing further interruptions. In the current implementation, we do this in sort of a roundabout way: once a lane is determined to have expired, we entangle it with SyncLane and switch to the synchronous work loop. There are a few flaws with the approach. One is that SyncLane has a particular semantic meaning besides its non-yieldiness. For example, `flushSync` will force remaining Sync work to finish; currently, that also includes expired work, which isn't an intended behavior, but rather an artifact of the implementation. An event worse example is that passive effects triggered by a Sync update are flushed synchronously, before paint, so that its result is guaranteed to be observed by the next discrete event. But expired work has no such requirement: we're flushing expired effects before paint unnecessarily. Aside from the behaviorial implications, the current implementation has proven to be fragile: more than once, we've accidentally regressed performance due to a subtle change in how expiration is handled. This PR aims to radically simplify how we model starvation protection by scaling back the implementation as much as possible. In this new model, if a lane is expired, we disable time slicing. That's it. We don't entangle it with SyncLane. The only thing we do is skip the call to `shouldYield` in between each time slice. This is identical to how we model synchronous-by-default updates in React 18.
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.
We have a feature called "expiration" whose purpose is to prevent a concurrent update from being starved by higher priority events. If a lane is CPU-bound for too long, we finish the rest of the work synchronously without allowing further interruptions.
In the current implementation, we do this in sort of a roundabout way: once a lane is determined to have expired, we entangle it with SyncLane and switch to the synchronous work loop.
There are a few flaws with the approach. One is that SyncLane has a particular semantic meaning besides its non-yieldiness. For example,
flushSyncwill force remaining Sync work to finish; currently, that also includes expired work, which isn't an intended behavior, but rather an artifact of the implementation.An even worse example is that passive effects triggered by a Sync update are flushed synchronously, before paint, so that its result is guaranteed to be observed by the next discrete event. But expired work has no such requirement: we're flushing expired effects before paint unnecessarily.
Aside from the behaviorial implications, the current implementation has proven to be fragile: more than once, we've accidentally regressed performance due to a subtle change in how expiration is handled.
This PR aims to radically simplify how we model starvation protection by scaling back the implementation as much as possible. In this new model, if a lane is expired, we disable time slicing. That's it. We don't entangle it with SyncLane. The only thing we do is skip the call to
shouldYieldin between each time slice. This is identical to how we model synchronous-by default updates in React 18.