Fix deadlock in multithreaded excerpt serialisation - #34733
Draft
ajuncosa wants to merge 1 commit into
Draft
Conversation
…ed excerpt serialisation
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
The goal is to avoid side effects on the score during write/serialisation, so excerpt serialisation can safely run in separate threads.
There was a deadlock when saving scores that contained multi measure rests and some hidden parts. The save called layout in that case, which frees Systems/Pages, and the
EngravingItemdestructor callsScore::onElementDestructionwhich ends withscore->elementDestroyed().send(e). This send was cross-thread, and that channel has a main-thread receiver for every score, so the send must register a port on the main thread's slot (QueuePool::regPort), and this locks the same mutex that the main thread is holding while it's processing queued callbacks (inQueuePool::processMessages), and the save is one of the queued commands.Apart from the freezing
doLayout()call, the write was also modifying the score in other ways (start/endCmd(),undoChangeProperty(VISIBLE, true)), and all this routes to the master score's single shared TransactionManager, whoseendTransactioncallsm_masterScore->update(), relaying out the master from a worker thread.I have moved the unhide and relayout out of the writer.