[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]
Re: [tor-bugs] #21312 [Obfuscation/Snowflake]: snowflake-client is pegged at 100% cpu
#21312: snowflake-client is pegged at 100% cpu
-----------------------------------+------------------------
Reporter: arlolra | Owner: (none)
Type: defect | Status: new
Priority: Medium | Milestone:
Component: Obfuscation/Snowflake | Version:
Severity: Normal | Resolution:
Keywords: | Actual Points:
Parent ID: | Points:
Reviewer: | Sponsor:
-----------------------------------+------------------------
Comment (by yawning):
Replying to [comment:15 arlolra]:
> For what it's worth (and probably not much),
`rtc::scoped_refptr<PeerConnectionFactoryInterface> pc_factory;` should be
automatically deallocated when the `Peer` goes out of scope, so isn't
exactly leaking. The culprit is `vector<rtc::scoped_refptr<Peer>>
localPeers;` which never releases the `Peer`s. See,
https://github.com/keroserene/go-webrtc/issues/35#issuecomment-233465318
I guess?
The Peer dtor needs to call delete on the created Threads, because the
PeerConnectionFactory's dtor will not if you explicitly provide one. It's
not immediately obvious to me what handles tearing down the SocketServer
either, but there's a limit to how much I want to look at the bindings or
the webrtc code.
I can't tell where worker_thread_ in the PeerConnectionFactory is supposed
to be terminated either, though the PeerConnectionFactory dtor appears to
shut down the signalling thread. At least if the worker_thread_ magically
gets terminated somehow (which it may) the code will only be leaking
memory and maybe some file descriptors.
But at this point I'm done wasting my time even thinking about this. Good
luck.
--
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/21312#comment:16>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online
_______________________________________________
tor-bugs mailing list
tor-bugs@xxxxxxxxxxxxxxxxxxxx
https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-bugs