[GHC] #4368: T4144(dyn) failing on x86/Linux

[GHC]

#4368: T4144(dyn) failing on x86/Linux
-------------------------------+--------------------------------------------
    Reporter:  simonmar        |        Owner:  benl         
        Type:  bug             |       Status:  merge        
    Priority:  highest         |    Milestone:  7.0.1        
   Component:  Compiler (NCG)  |      Version:  6.13         
    Keywords:                  |     Testcase:               
   Blockedby:                  |   Difficulty:               
          Os:  Linux           |     Blocking:               
Architecture:  x86             |      Failure:  Runtime crash
-------------------------------+--------------------------------------------
Changes (by benl):

  * status:  new => merge


Comment:

 Fixed by:
 {{{
 Tue Oct 12 18:54:14 PDT 2010  benl at ouroborus.net
  * RegAlloc: Track slot liveness over jumps in spill cleaner
 }}}

 Note that the spill cleaner should really be doing a proper data flow
 analysis, which would propagate information about what slots are being
 used across block boundaries. At the moment we're just doing a single
 "iteration" where we should really be iterating to a fixed point where no
 more spill/reload instructions can be erased. It still works, but will
 leave more spill/reloads behind than strictly needed. I'm not sure what
 the performance penalty is wrt a perfect clean, and my DPH benchmark
 programs are all currently broken... probably no worse than before though.

-- 
Ticket URL: <http://hackage.haskell.org/trac/ghc/ticket/4368#comment:4>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler