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

[GHC]

#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):

 I've also tried to build m68k crosscompiler
 (using https://wiki.debian.org/M68k/sbuildQEMU to test produced binaries)
 and noticed GHC has the issue similar to:

     http://bugs.python.org/issue17237

 m68k aligns structs containing integers and pointers to 2 bytes
 (not 4 bytes). That makes StgClosure struct addresses look tagged
 even if they are not.

 An example of GHC's startup crash.
 Test checks if untagging macro works:

 {{{
 Core was generated by `/tmp/a +RTS -Ds -Di -Dw -DG -Dg -Db -DS -Dt -Dp -Da
 -Dl -Dm -Dz -Dc -Dr'.
 Program terminated with signal SIGSEGV, Segmentation fault.
 #0  0x80463b0a in LOOKS_LIKE_INFO_PTR_NOT_NULL (p=32858) at
 includes/rts/storage/ClosureMacros.h:248
 248         return (info->type != INVALID_OBJECT && info->type <
 N_CLOSURE_TYPES) ? rtsTrue : rtsFalse;
 (gdb) bt
 #0  0x80463b0a in LOOKS_LIKE_INFO_PTR_NOT_NULL (p=32858) at
 includes/rts/storage/ClosureMacros.h:248
 #1  0x80463b46 in LOOKS_LIKE_INFO_PTR (p=32858) at
 includes/rts/storage/ClosureMacros.h:253
 #2  0x80463b6c in LOOKS_LIKE_CLOSURE_PTR (p=0x805aac6e
 <stg_dummy_ret_closure>) at includes/rts/storage/ClosureMacros.h:258
 #3  0x80463e4c in initStorage () at rts/sm/Storage.c:121
 #4  0x8043ffb4 in hs_init_ghc (argc=0xf6ffebf0, argv=0xf6ffebf4,
 rts_config=...) at rts/RtsStartup.c:181
 #5  0x80455982 in hs_main (argc=1, argv=0xf6ffed54,
 main_closure=0x804b8f70 <ZCMain_main_closure>, rts_config=...)
     at rts/RtsMain.c:51
 #6  0x80003c1c in main ()
 }}}

 GHC assumes last 2 pointer bits are tags on 32-bit targets.
 But here 'stg_dummy_ret_closure' address violates the assumption:

     LOOKS_LIKE_CLOSURE_PTR (p=0x805aac6e <stg_dummy_ret_closure>)

 One more thing to tweak.

--
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/11395#comment:16>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler