[GHC] #12469: Memory fence on writes to MutVar/Array missing on ARM

[GHC]

#12469: Memory fence on writes to MutVar/Array missing on ARM
-------------------------------------+-------------------------------------
        Reporter:  rrnewton          |                Owner:
            Type:  bug               |               Status:  patch
        Priority:  normal            |            Milestone:  8.0.2
       Component:  Compiler          |              Version:  8.0.1
      Resolution:                    |             Keywords:  memory model
Operating System:  Unknown/Multiple  |         Architecture:
 Type of failure:  Incorrect result  |  Unknown/Multiple
  at runtime                         |            Test Case:
      Blocked By:                    |             Blocking:  12537
 Related Tickets:                    |  Differential Rev(s):  Phab:D2495
       Wiki Page:                    |  Phab:D2525
-------------------------------------+-------------------------------------

Comment (by trommler):

 Replying to [comment:20 erikd]:
 > @trommler, how are you triggering these errors?
 Unfortunately, I cannot directly observe these errors. I see them on
 Open(SUSE) Build Service where a build runs inside a qemu VM on a POWER8.
 The issue is very rare, 1700 packages build just fine and around 30 to 50
 fail. Most of the failures are in `Setup`, which we compile for each
 package, and very few (less than 3) fail with a panic. Rebuilding the
 package works with high probability.

 On my PowerMac quad core I see a segfault in `./Setup build -v -j4` which
 is what  is reported in #12537. In that ticket a compiler panic in
 `mkFastStringWith` is reported but I have not been able to provoke a panic
 on my machine. Is it worth trying some more runs and hope to observe a
 panic to rule out that qemu is doing something odd?

--
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/12469#comment:21>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler