fix: Possible multithreading issue when two scores may come in parallel they may not safely fill pointer values #799
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
…r changes. By boosterHandler not being serializable, it forces Serializable objects to declare it as transient
…rder of finalization so it is dangerous. I wonder how is it guaranteed that the memory will be cleared though
|
@JoanFM it looks like this needs to be merged with latest? I think I also see some code that has been merged already in the other PR. |
Yes, I am merging it |
|
|
||
| private def getNumClasses: Int = { | ||
| val numClassesOut = lightgbmlib.new_intp() | ||
| LightGBMUtils.validate( | ||
| lightgbmlib.LGBM_BoosterGetNumClasses(boosterPtr, numClassesOut), | ||
| "Booster NumClasses") | ||
| lightgbmlib.intp_value(numClassesOut) | ||
| val out = lightgbmlib.intp_value(numClassesOut) | ||
| lightgbmlib.delete_intp(numClassesOut) |
There was a problem hiding this comment.
nice catch to delete the pointer here!
imatiach-msft
approved these changes
Feb 10, 2020
|
/azp run |
|
Azure Pipelines successfully started running 1 pipeline(s). |
Codecov Report
@@ Coverage Diff @@
## master #799 +/- ##
=========================================
Coverage ? 43.83%
=========================================
Files ? 241
Lines ? 9704
Branches ? 529
=========================================
Hits ? 4254
Misses ? 5450
Partials ? 0
Continue to review full report at Codecov.
|
ocworld
pushed a commit
to AhnLab-OSS/mmlspark
that referenced
this pull request
Mar 24, 2020
…el they may not safely fill pointer values (microsoft#799) * Move transient var to transient lazy val * Try guaranteeing NativeLibrary is loaded everytime it is needed but only loaded once * Forgot some lightGBMlibConstants * Add freeing of resources, handle all memory in BoosterHandler * Remove serialization from BoosterHandler, some documentation and minor changes. By boosterHandler not being serializable, it forces Serializable objects to declare it as transient * Style forbigs implementing finalize method, Java does not guarantee order of finalization so it is dangerous. I wonder how is it guaranteed that the memory will be cleared though * Add finalize method to call free for Native C++ allocated memory * predict normal correct constant * Change implementation to a ThreadLocal based synchronization
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.
fix: Possible multithreading issue when two scores may come in parallel they may not safely fill pointer values
Two threads can fill up scoredDataOutPtr in parallel and mix up their values