[GHC] #8578: Improvements to SpinLock implementation

[GHC]

#8578: Improvements to SpinLock implementation
-------------------------------------+-------------------------------------
        Reporter:  parcs             |                Owner:  (none)
            Type:  task              |               Status:  new
        Priority:  normal            |            Milestone:
       Component:  Runtime System    |              Version:  7.7
      Resolution:                    |             Keywords:
Operating System:  Unknown/Multiple  |         Architecture:
 Type of failure:  Runtime           |  Unknown/Multiple
  performance bug                    |            Test Case:
      Blocked By:                    |             Blocking:
 Related Tickets:                    |  Differential Rev(s):
       Wiki Page:                    |
-------------------------------------+-------------------------------------

Comment (by dfeuer):

 Since thoughtpolice mentioned them, I'm wondering if there are specific
 places we should look at as HLE (or perhaps RTM) candidates. I don't
 ''think'' we can rely on the system glibc to use them to implement mutexes
 unless we ask for that specifically in some fashion. I'd expect very small
 and short transactions (most but not all of what we do) to benefit. Has
 anyone looked into that yet?

-- 
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8578#comment:10>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler