[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:                    |
-------------------------------------+-------------------------------------
Changes (by simonpj):

 * cc: simonmar (added)


Comment:

 > It looks like to avoid this we will either need to teach the simplifier
 not to throw away otherwise dead continuations which contain some
 "important" primops

 Can you give an example to show what it is throwing away, and why that's
 bad?  I don't get it yet.  I have even forgotten why `touch#` exists.

 Copying Simon Marlow.

-- 
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/14346#comment:19>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler