[GHC] #5824: ARM StgRun register clobber list is broken

[GHC]

#5824: ARM StgRun register clobber list is broken
-------------------------------+--------------------------------------------
    Reporter:  bgamari         |       Owner:  simonmar        
        Type:  bug             |      Status:  patch           
    Priority:  high            |   Milestone:  7.4.2           
   Component:  Runtime System  |     Version:  7.4.1-rc2       
    Keywords:                  |          Os:  Unknown/Multiple
Architecture:  arm             |     Failure:  None/Unknown    
  Difficulty:  Unknown         |    Testcase:                  
   Blockedby:                  |    Blocking:                  
     Related:                  |  
-------------------------------+--------------------------------------------

Comment(by kgardas):

 Hi,
 Joachim (I hope nomeata is Joachim, if not please complain), could you be
 so kind and verify for us, that Debian is using pure ARM code? I.e. what
 `gcc -v' claims? Is it compiling by default in ARM mode or in Thumb mode?
 If the former, then I see the reason for this complain. In fact Ben
 already experience it himself on Ubuntu which is using Thumb mode by
 default and claims that r7 is used for frame pointer by compiler in this
 case. If Debian is using ARM mode, then fp is r11 and the compiler should
 complain about it in the same way.
 So if all this is true, then this should solve my question why Debian and
 Ubuntu compilers behaves in different way here.
 Now, we do have another issue. In ARM mode (on Debian presumably) we
 should mark R7 as clobbered IMHO. I'm not sure if we should mark R11 as
 clobbered in Thumb mode yet. Anyway, marking R11 as clobbered on Thumb is
 less important than marking R7 as clobbered on ARM as Thumb is not going
 to use R11 anyway due to its limited register space. I'm going to give
 this idea a try since on my Panda I do have gcc 4.4.1 (from older ubuntu),
 which is compiling in ARM mode by default so this should duplicate
 Debian's issue -- if my idea is right here. I will keep you posted.

-- 
Ticket URL: <http://hackage.haskell.org/trac/ghc/ticket/5824#comment:18>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler