2.x: Fix concurrent clear in observeOn while output-fused

[akarnokd]

Backport of #6708

---

There was another cancel-clear race leading to NPE or infinite loop inside both observeOn implementations.

Related: #6676

[codecov]

Codecov Report

Merging #6710 into 2.x will increase coverage by 0.01%.
The diff coverage is 100%.

@@             Coverage Diff              @@
##                2.x    #6710      +/-   ##
============================================
+ Coverage     98.26%   98.28%   +0.01%     
+ Complexity     6350     6347       -3     
============================================
  Files           677      677              
  Lines         45526    45526              
  Branches       6333     6333              
============================================
+ Hits          44736    44743       +7     
+ Misses          247      243       -4     
+ Partials        543      540       -3

Impacted Files	Coverage Δ	Complexity Δ
...internal/operators/flowable/FlowableObserveOn.java	97.21% <100%> (ø)	3 <0> (ø)	⬇️
...rnal/operators/observable/ObservableObserveOn.java	100% <100%> (ø)	3 <0> (ø)	⬇️
...a/io/reactivex/internal/util/QueueDrainHelper.java	97.22% <0%> (-2.78%)	56% <0%> (-2%)
...tivex/internal/schedulers/TrampolineScheduler.java	96.1% <0%> (-2.6%)	6% <0%> (ø)
...rnal/operators/observable/ObservableSwitchMap.java	91.53% <0%> (-2.12%)	3% <0%> (ø)
...perators/mixed/ObservableSwitchMapCompletable.java	97.89% <0%> (-2.11%)	3% <0%> (ø)
.../internal/disposables/ListCompositeDisposable.java	98% <0%> (-2%)	34% <0%> (-1%)
...nternal/operators/observable/ObservableWindow.java	98% <0%> (-2%)	3% <0%> (ø)
...tivex/internal/operators/single/SingleTimeout.java	98.33% <0%> (-1.67%)	2% <0%> (ø)
...io/reactivex/subscribers/SerializedSubscriber.java	98.86% <0%> (-1.14%)	26% <0%> (-1%)
... and 18 more

---

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 3cb610f...276b14a. Read the comment docs.