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

[GHC]

#9125: int-to-float conversion broken on ARM
-------------------------------------+-------------------------------------
        Reporter:  Ansible           |                   Owner:
            Type:  bug               |                  Status:  infoneeded
        Priority:  high              |               Milestone:  7.10.1
       Component:  Compiler          |                 Version:  7.8.3
  (CodeGen)                          |                Keywords:
      Resolution:                    |            Architecture:  arm
Operating System:  Unknown/Multiple  |               Test Case:
 Type of failure:  Incorrect result  |                Blocking:
  at runtime                         |  Differential Revisions:
      Blocked By:                    |
 Related Tickets:                    |
-------------------------------------+-------------------------------------

Comment (by amurrayc):

 I tried again with the 7.10 rc candidate (7.10.0.20141222) with LLVM 3.4
 and got the same error as comment:36. Should this be a new ticket? The
 original problem seems solved in the simulator build by compiling with
 3.4.

 I left some more details a while back on the list
 https://www.haskell.org/pipermail/ghc-devs/2014-October/006521.html

 The error occurs while compiling `libraries/integer-
 simple/GHC/Integer/Type.hs`.

 I don't really speak ARM assembly but the error appears to be in the
 `doubleFromPositive` section. I note that the label LCPI102_0 refers to a
 point near the beginning of the file and that from
 http://infocenter.arm.com/help/topic/com.arm.doc.dui0489c/Bcfcdfdd.html it
 seems that the label in a vldr instruction needs to point somewhere within
 1KB of the current instruction.

--
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/9125#comment:39>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler