make memory-safe against concurrent closure/operations #53
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
Make badger memory-safe against concurrent operations and close.
We don't have to clone the value.
ghost
assigned
This was referenced Mar 8, 2019
magik6k
approved these changes
Mar 8, 2019
There was a problem hiding this comment.
Some operations can take single-digit microseconds in optimal conditions (https://ipfs.io/ipfs/Qmc168UrkY3ngPuKpSJtkhxS1r7hZJxcZzcw1Nfan1yZ5w/ds-badger/has-bsize/record-size-nsop.png), but even in this case this should be around 1% perf hit.
|
Out of curiosity, any particular reason we didn’t go with an atomic integer to control the closed state? |
|
That's basically what a RWMutex is. An atomic "readers" counter and a lock. We need some way for Close to wait for all readers to finish so we need some type of lock/channel for that and RWMutex is pretty efficient (atomic increment/decrement unless there's a writer waiting). |
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Outdated suggestions cannot be applied.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.
Ideally, this would be implemented down in badger itself but there appear to have been some performance concerns.
Why not just avoid ever trying to operate on a closed datastore? Well, we try our best but I'd still like to avoid segfaulting on error. Especially because these concurrent operations can be unimportant reads.
Thoughts? I haven't benchmarked this but I really doubt taking a rwmutex will cost us much (given how we generally use datastores, that is).