[GHC] #7655: 7.6.2 Segmentation Fault/Bus Error in large exponentation

[GHC]

#7655: 7.6.2 Segmentation Fault/Bus Error in large exponentation
-------------------------------------+----------------------------------
              Reporter:  Doug310     |            Owner:
                  Type:  bug         |           Status:  infoneeded
              Priority:  normal      |        Milestone:  7.8.4
             Component:  GHCi        |          Version:  7.8.1-rc1
            Resolution:              |         Keywords:  exponentiation
      Operating System:  MacOS X     |     Architecture:  x86_64 (amd64)
       Type of failure:  GHCi crash  |       Difficulty:  Unknown
             Test Case:              |       Blocked By:
              Blocking:              |  Related Tickets:
Differential Revisions:              |
-------------------------------------+----------------------------------

Comment (by rwbarton):

 Actually a few things were not quite right in that. `--enable-
 alloca=reentrant` is not the same as `WANT_TMP_REENTRANT`! It means "use
 `WANT_TMP_ALLOCA` if you have `alloca`, otherwise use
 `WANT_TMP_REENTRANT`". So that explains how `WANT_TMP_ALLOCA` is getting
 set.

 Also, my examination of the Debian `libgmp.so.10` was overly hasty. It is
 actually built with `WANT_TMP_ALLOCA`. There are several allocation paths
 in `mpn_mul`, and some call `malloc` when the amount to be allocated
 exceeds a certain threshold, while others unconditionally allocate on the
 stack.

 Not sure what the best thing to do here is aside from resorting to
 `WANT_TMP_REENTRANT`, which I fear may carry a noticeable performance
 penalty.

--
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/7655#comment:32>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler