Make Shutdown work correctly

[RedFlames]

I wrote this based off the exceptions that happen when you call /api/shutdown

Somehow this error still gets logged, despite everything shutting down correctly...

(01/16/2023 17:12:07) [CRI] [netplus] Error in thread pool thread 3: System.Net.Sockets.SocketException (10004): A blocking operation was interrupted by a call to WSACancelBlockingCall.
   at System.Net.Sockets.Socket.ReceiveFrom(Byte[] buffer, Int32 offset, Int32 size, SocketFlags socketFlags, EndPoint& remoteEP)
   at System.Net.Sockets.Socket.ReceiveFrom(Byte[] buffer, EndPoint& remoteEP)
   at Celeste.Mod.CelesteNet.Server.UDPReceiverRole.Worker.StartWorker(Socket socket, CancellationToken token) in C:\Program Files (x86)\Steam\steamapps\common\Celeste\Mods\CelesteNet_dev\CelesteNet.Server\ConPlus\UDPReceiverRole.cs:line 29
   at Celeste.Mod.CelesteNet.Server.MultipleSocketBinderRole.RoleWorker.StartWorker(CancellationToken token) in C:\Program Files (x86)\Steam\steamapps\common\Celeste\Mods\CelesteNet_dev\CelesteNet.Server\ConPlus\MultipleSocketBinderRole.cs:line 18
   at Celeste.Mod.CelesteNet.Server.NetPlusThread.ThreadLoop() in C:\Program Files (x86)\Steam\steamapps\common\Celeste\Mods\CelesteNet_dev\CelesteNet.Server\NetPlus\ThreadPool.cs:line 196

I have no idea if this breaks anything in how the server actually runs; I only went by testing /api/shutdown 🙂

• moved the RWLock.Dispose calls outside of a use of the lock
• made RWLock.Inner private because noone should be touching that directly anyways
• properly cancel threads on shutdown(?)

[RedFlames]

I made this PR the last time we had a catastrophic lock disposal situation, I guess. Some kinda really rare race... this may be good to merge and see what it does on the server 🤔
Anyways TODO this line needs changing because it shouldn't default to VVV log level (honestly nothing ever should? Why is that the implicit level?)