GitLab

We were testing neq instead of eq

StgTVarWatchQueue contains the threads blocked on a TVar in order 
youngest first. The list has to be traversed backwards to unpark the threads 
oldest first.

This improves the fairness when using STM in some situations.

This should improve scaling when using atomicModifyIORef

  MERGE TO 6.10

  MERGE TO 6.10

For some unknown reason in schedulePostRunThread() we were always
passing capabilities[0] rather than the current Capability to
throwToSingleThreaded().  This caused all kinds of weird failures and
crashes in STM code when running on multiple processors.

MERGE TO 6.10

  MERGE TO 6.10

MERGE TO 6.10

It's no longer needed, as base no longer #includes it

MERGE TO 6.10