-
-
Notifications
You must be signed in to change notification settings - Fork 51
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Add locking around sockets, since QUdpSocket isn't thread safe enough. #1122
base: master
Are you sure you want to change the base?
Conversation
Server side, this appears to be working fine. At least 74 and I were able to connect to anime-avatars, which is running 964dbbb right now. |
@@ -135,6 +138,8 @@ private slots: | |||
|
|||
Mutex _unreliableSequenceNumbersMutex; | |||
Mutex _connectionsHashMutex; | |||
Mutex _socketMutex; |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
since the problem is QUdpSocket (and WebRTCSocket?), should this locking logic be internal to NetworkSocket.h instead?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I think it makes more sense to keep it in udt::Socket.
NetworkSocket.h is an abstraction that could be used in a place where it's only ever used in a single thread and therefore not need locking. The problem happens because udt::Socket is called from multiple threads. And it also already deals with locking for other purposes.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
maybe, but it could just as easily be used in another multithreaded context and then we open ourselves up to more difficult to debug threading crashes 😄
at the very least, can we add something to NetworkSocket about this? maybe just a big comment, although if we could detect this and log a warning (just the first time it happens) that would also be great.
This fixes #1119
After this it's clear the networking code still needs work as in suboptimal conditions it may fail to connect to a domain or keep dropping off. But at least it doesn't crash.