This repository has been archived by the owner on Jan 6, 2020. It is now read-only.
Maid 22 bad alloc #14
Open
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
This reverts commit 10d96e7.
This reverts commit e2a1b08.
This reverts commit c6ff3c0.
This reverts commit db7750c.
This reverts commit 655cece.
…vaults on same machine" This reverts commit db43b5f.
This reverts commit 911e0db.
…e it is being used in favour of Transport::Close() dispatching the connection teardown onto the appropriate ASIO thread instead. Unfortunately this creates occasional deadlocks, so I disabled all ASIO multithreading, thus solving mutex brittleness in one fell swoop. If the performance hit isn't terrible I'm keeping ASIO single threaded.
|
Your other pull requests have been merged Niall - I guess that makes this one obsolete? |
|
It'll have to be redone I'd say. Anyway Peter is on it, I explained to him the cause and likely fixes. |
|
test this please. |
|
Builds Triggered; |
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
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Outdated suggestions cannot be applied.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.
Ok, this is a fairly breaking change to how RUDP works in that ASIO worker threads decreases to 1, and therefore all concurrency within RUDP ceases, and therefore the cause of this bug where Socket was being destroyed concurrently to its use in other threads goes away. The other change is I ripped out my previous remedial work in making Multiplexer reference count deleted, because it was obviously obscuring the problem rather than fixing it.
I soak tested this last night and it hanged exactly once which may or may not be important - out of 500 iterations, that's a hang rate of 0.2%, and could be a false positive. The removal of the remedial work has introduced lots of ThreadSanitiser fails, this is a good thing and I'll try to figure out the new cause. Various tests still sporadically fail, but this branch doesn't have my increased timeouts, so that is expected.
So, if you start seeing hangs after applying this patch, get back to me :) Otherwise it probably fixes more problems than it causes, and therefore is an improvement.