Block rtxTimer.stop() until timer routine end

[at-wat]

Fix data race during TestRtxTimer.

Reference issue

Fixes #129

[codecov]

Codecov Report

Merging #132 into master will decrease coverage by 0.01%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master     #132      +/-   ##
==========================================
- Coverage   79.02%   79.00%   -0.02%     
==========================================
  Files          43       43              
  Lines        2798     2806       +8     
==========================================
+ Hits         2211     2217       +6     
- Misses        453      454       +1     
- Partials      134      135       +1     

Flag	Coverage Δ
#go	79.00% <100.00%> (-0.02%)	⬇️
#wasm	69.99% <100.00%> (+0.08%)	⬆️

Impacted Files	Coverage Δ
rtx_timer.go	97.91% <100.00%> (+0.18%)	⬆️
association.go	83.09% <0.00%> (-0.18%)	⬇️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update dbc92ec...3384f55. Read the comment docs.

[enobufs]

The build error appears to be complaining about the race on the variable nCbs.
The error log found in #129:

 /home/travis/gopath/src/github.com/pion/sctp/rtx_timer_test.go:206 +0x521

I was able to reproduce the issue and fixed this problem locally by using atomic methods for nCbs. Let me send another PR.

[at-wat]

@enobufs Yes, accessing nCbs using atomic will also fix the race.

On the other hand, current master branch version has no way to know when the caller can assume that the handler will no longer be called. It may cause another race in real application not only the test code, IMO.

[enobufs]

Yes, I see. Looks like you are trying to address another problem and we could file another issue to address it later?
I am actually dealing with a race condition between t3RtxTimer.stop() and t3RtxTimer.start() related to pion/webrtc#1270. Let us get the new PR (#134) I have just submitted out of the way first.

[at-wat]

Seems not mandatory.