Android/Linux/rdrand: Use once_cell::race::OnceBool instead of LazyBool. #483
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
|
I don't think that adding a new dependency to save less than 10 lines of straightforward code is worthwhile. |
This is designed to accompany #481, which adds a more important |
briansmith
force-pushed
the
b/oncecell-lazy
branch
2 times, most recently
from
cdd0a1b
to
4f59687
Compare
briansmith
force-pushed
the
b/oncecell-lazy
branch
3 times, most recently
from
de49003
to
3370fb9
Compare
briansmith
changed the title
once_cell::race::OnceBool instead of defining our own LazyBool.
briansmith
force-pushed
the
b/oncecell-lazy
branch
from
3370fb9
to
7b2d78d
Compare
|
I updated the commit message (and the GitHub issue description) to emphasize the difference in semantics, before and after. |
Remove src/lazy.rs. `lazy::LazyBool` had "last to win the race" semantics. When multiple threads see an uninitialized `LazyBool`, all of them will calculate a value. As they finish, each one will overwrite the value set by the thread that finished previously. If two threads calculate different values for the boolean, then the value of the boolean can change during the period where the threads are racing. This doesn't seem to be a huge issue with the way it is currently used, but it is hard to reason about. `once_cell::race::OnceBool` has "first to win the race" semantics. When multiple threads see an uninitialized `OnceBool`, all of them will calculate a value. The first one to finish will write its value; the rest will have their work ignored. Thus there is never any change in the stored value at any point. This is much easier to reason about. The different semantics come down to the fact that once_cell uses `AtomicUsize::compare_exchange` whereas lazy.rs was using `AtomicUsize::store`.
briansmith
force-pushed
the
b/oncecell-lazy
branch
from
7b2d78d
to
6aaba2f
Compare
|
This PR became stale and it's probably better to remake it from the master branch. But before that it's probably worth to discuss in a separate issue whether we need this migration in the first place. |
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.
Remove src/lazy.rs.
lazy::LazyBoolhad "last to win the race" semantics. When multiple threads see an uninitializedLazyBool, all of them will calculate a value. As they finish, each one will overwrite the value set by the thread that finished previously. If two threads calculate different values for the boolean, then the value of the boolean can change during the period where the threads are racing. This doesn't seem to be a huge issue with the way it is currently used, but it is hard to reason about.once_cell::race::OnceBoolhas "first to win the race" semantics. When multiple threads see an uninitializedOnceBool, all of them will calculate a value. The first one to finish will write its value; the rest will have their work ignored. Thus there is never any change in the stored value at any point. This is much easier to reason about.The different semantics come down to the fact that once_cell uses
AtomicUsize::compare_exchangewhereas lazy.rs was usingAtomicUsize::store.