[GHC] #9125: int-to-float conversion broken on ARM

[GHC]

#9125: int-to-float conversion broken on ARM
-------------------------------------+-------------------------------------
              Reporter:  Ansible     |             Owner:
                  Type:  bug         |            Status:  new
              Priority:  highest     |         Milestone:  7.8.4
             Component:  Compiler    |           Version:  7.8.3
            Resolution:              |          Keywords:
Differential Revisions:              |  Operating System:  Unknown/Multiple
          Architecture:  arm         |   Type of failure:  Incorrect result
            Difficulty:  Unknown     |  at runtime
            Blocked By:              |         Test Case:
       Related Tickets:              |          Blocking:
-------------------------------------+-------------------------------------

Comment (by amurrayc):

 I'm still learning about Cmm and calling conventions &c. so forgive my
 stumblings.

 If I modify `stg_decodeFloatzuIntzh` so that the two temporary pointers
 are shifted by one word the function returns the correct value. That is
 {{{
     reserve 2 = tmp {

       mp_tmp1  = tmp + WDS(1);
       mp_tmp_w = tmp;
 }}}
 becomes
 {{{
     reserve 3 = tmp {

       mp_tmp1  = tmp + WDS(2);
       mp_tmp_w = tmp + WDS(1);
 }}}
 With this change everything seems to work correctly.  I'm confused though.
 The pointer that was getting clobbered by the argument value was `tmp +
 WDS(1)` but now `mp_tmp_w` i.e. the exponent is fine.

 Anyway, I hope this give someone wiser a clue.

--
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/9125#comment:6>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler