Re: [GHC] #11395: The via-C code generation backend is incompatible with gcc 5.3.1 on m68k (ELF)

10 Jan 2016

      #11395: The via-C code generation backend is incompatible with gcc 5.3.1 on m68k
(ELF)
--------------------------------------------+------------------------------
        Reporter:  mkarcher                 |                Owner:
            Type:  bug                      |               Status:  new
        Priority:  normal                   |            Milestone:
       Component:  Compiler                 |              Version:  7.10.2
      Resolution:                           |             Keywords:
Operating System:  Linux                    |         Architecture:  m68k
 Type of failure:  GHC doesn't work at all  |            Test Case:
      Blocked By:                           |             Blocking:
 Related Tickets:                           |  Differential Rev(s):
       Wiki Page:                           |
--------------------------------------------+------------------------------

Comment (by slyfox):

 To get generated code we can run:
 {{{
 "inplace/bin/ghc-stage1" \
     -static  -H64m -O1 \
     -optc-flto=9 -optc-ffat-lto-objects \
     -optl-flto=9 -optl-ffat-lto-objects \
     -optc-ggdb2 -optl-ggdb2 \
     -j4 \
     -optc-Werror=implicit-function-declaration -optc-Werror=overflow \
     -Wall  \
     -lbfd \
     -Iincludes -Iincludes/dist -Iincludes/dist-derivedconstants/header \
     -Iincludes/dist-ghcconstants/header -Irts -Irts/dist/build \
     -DCOMPILING_RTS -this-package-key rts \
     -optc-DNOSMP -dcmm-lint \
     -i -irts -irts/dist/build \
     -irts/dist/build/autogen \
     -Irts/dist/build -Irts/dist/build/autogen \
     -O2 \
     -C rts/Exception.cmm \
     -o rts/dist/build/Exception.hc
 }}}

 '''rts/dist/build/Exception.hc''' contains the following declarations:
 {{{
 /* (incomplete) symbol declaraion */
 EF_(raiseExceptionHelper);
 ...
 /* exact prototype dectaration, assignment and call: */
 {
     W_ (*ghcFunPtr)(void *, void *, void *);
     ghcFunPtr = ((W_ (*)(void *, void *, void *))raiseExceptionHelper);
     _c2y = ghcFunPtr((void *)(W_)BaseReg, (void *)(W_)CurrentTSO, (void
 *)_c2z);;}
 }}}

 It comes from '''rts/Exception.cmm''':
 {{{
 retry_pop_stack:
     SAVE_THREAD_STATE();
     (frame_type) = ccall raiseExceptionHelper(BaseReg "ptr", CurrentTSO
 "ptr", exception "ptr");
     LOAD_THREAD_STATE();
 }}}
 As we see unannotated return type defaults to '''W_''' ('''StgWord''',
 non-pointer type).

 So the question is:

 Does gcc ignore type coercion given in '''ghcFunPtr =''' assignment and
 uses original declaration?
 gcc's '''-fdump-*-all''' options might hint at where gcc loses information
 about argument type.

 Could it be real (but similar) error happens somewhere else?

--
Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/11395#comment:3
GHC http://www.haskell.org/ghc/
The Glasgow Haskell Compiler