Fix GH-21730: Mt19937::__debugInfo() leaks state HashTable when the serialize callback fails#33
Closed
Fix GH-21730: Mt19937::__debugInfo() leaks state HashTable when the serialize callback fails#33
Conversation
…e serialize callback fails Mt19937::__debugInfo() allocates a temporary HashTable with array_init(&t), calls the engine's serialize callback, and then inserts t into the return value. If the callback returns false, the method throws and hits RETURN_THROWS() before inserting t, so the HashTable leaks. PcgOneseq128XslRr64 and Xoshiro256StarStar alias the same method and share the leak. Niels Dossche fixed the same pattern in __serialize() via phpGH-20383 (720e006). That cleanup didn't touch __debugInfo(). Apply the same reordering here: insert t into return_value first, then let the callback populate it. RETURN_THROWS() then unwinds the return value cleanly. The path is latent in stock PHP because the three built-in serialize callbacks (mt19937, pcg, xoshiro) all return true, so no user code reaches the leak today. I'm fixing it for symmetry with phpGH-20383 and to keep the pattern from regressing if a future engine grows a failing serialize path. Closes phpGH-21730
iliaal
force-pushed
the
fix/gh-21730-mt19937-debuginfo-leak
branch
from
e16347f to
6fb13bb
Compare
Owner
Author
|
Submitted upstream as php#21733 |
This file contains hidden or 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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Fixes php#21730.
Mt19937::__debugInfo()allocates a temporary HashTable witharray_init(&t)and then calls the engine'sserializecallback. If the callback returns false, the method throws "Engine serialize failed" and hitsRETURN_THROWS()before insertingt, so the HashTable leaks.PcgOneseq128XslRr64andXoshiro256StarStaralias the same method and share the leak.Niels Dossche fixed the same pattern in
__serialize()via php#20383 (720e006). That cleanup didn't touch__debugInfo(). The fix here is identical: inserttinto the return value before calling the callback, soRETURN_THROWS()unwinds it cleanly on failure.The leak path is latent in stock PHP because the three built-in
serializecallbacks all returntrue, so no user code reaches it today. I'm filing it for symmetry with php#20383 and to keep the pattern from regressing if a future engine grows a failing serialize path.The new phpt (
debuginfo_states.phpt) locks down the success-path output shape for all three engines, so future refactors of this code have something to break against. The leak itself can't be exercised from PHP without patching the source (covered in the linked issue). I verified the fix under valgrind with a forced-failure patch: the fixed build reports zero definitely-lost bytes.