[GHC] #14346: 8.2.1 regression: heap corruption after safe foreign calls

[GHC]

#14346: 8.2.1 regression: heap corruption after safe foreign calls
-------------------------------------+-------------------------------------
        Reporter:  andrewchen        |                Owner:  (none)
            Type:  bug               |               Status:  infoneeded
        Priority:  highest           |            Milestone:
       Component:  Runtime System    |              Version:  8.2.1
      Resolution:                    |             Keywords:
Operating System:  Unknown/Multiple  |         Architecture:
                                     |  Unknown/Multiple
 Type of failure:  Runtime crash     |            Test Case:
      Blocked By:                    |             Blocking:
 Related Tickets:                    |  Differential Rev(s):
       Wiki Page:                    |
-------------------------------------+-------------------------------------

Comment (by bgamari):

 On looking at this with fresh eyes, it seems that unfortunately my
 analysis from comment:11 is flawed; the `movq %rbx,$rdi` is completely
 correct. We spill to the callee-saved  `%rbx register before
 `suspendThread` and then more the value from `%rbx` to `%rdi`, which is
 where we expect the first argument to reside. The second spill is simply
 preserving `_u4RH`, which is still alive after the call to `test`.

 Back to the drawing board. I think now I'll focus on catching the issue
 earlier in execution; namely, when we first get the `value mismatch`
 message.

-- 
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/14346#comment:12>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler