[Bug target/40722] [4.5/4.6 Regression] ia32intrin.h defines of _rotl, _rotr conflict with target stdlib.h decls

[Bug target/40722] [4.5/4.6 Regression] ia32intrin.h defines of _rotl, _rotr conflict with target stdlib.h decls

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40722

Richard Guenther <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|4.5.2                       |4.5.3

--- Comment #17 from Richard Guenther <rguenth at gcc dot gnu.org> 2010-12-16 13:03:28 UTC ---
GCC 4.5.2 is being released, adjusting target milestone.

[Bug target/40722] [4.5/4.6 Regression] ia32intrin.h defines of _rotl, _rotr conflict with target stdlib.h decls

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40722

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |jakub at gcc dot gnu.org

--- Comment #18 from Jakub Jelinek <jakub at gcc dot gnu.org> 2011-02-03 09:07:12 UTC ---
(In reply to comment #8)
> (In reply to comment #7)
> > An updated patch is at
> > 
> > http://gcc.gnu.org/ml/gcc-patches/2010-03/msg00937.html
> > 
> 
> Patch is fine. It is absolutely necessary to support gcc's intrinsic headers
> for mingw. The fixinclude approach is one way to solve it and I am fine by it.
> For mingw-w64 we used the #pragma push/pop_macro feature to make sure we
> declare function without getting problems by this define in ia32intrin.h.
> 
> Kai

Why can't you just
_CRTIMP unsigned int __cdecl __MINGW_NOTHROW (_rotl)(unsigned int, int)
__MINGW_ATTRIB_CONST;
?
_rotl etc. from ia32intrin.h is a function-like macro, so the above prevents
the expansion.
Is this still a problem on *mingw and *mingw64?  I mean, if x86intrin.h can't
be included, it sounds like a serious problem for those that use this arch, and
it is open for more than a year.
I guess H.J.'s patch is just too dangerous, it assumes that _rotl is the first
declaration and _lrotr is the last one, I guess much better would be to ifdef
out
each declaration separately, then there wouldn't be issues with unmatched
#if/#endif.

[Bug target/40722] [4.5/4.6 Regression] ia32intrin.h defines of _rotl, _rotr conflict with target stdlib.h decls

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40722

--- Comment #19 from Jakub Jelinek <jakub at gcc dot gnu.org> 2011-02-03 10:05:58 UTC ---
Created attachment 23233
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=23233
gcc46-pr40722.patch

Modified H.J.'s patch which just adds ()s around _rotl/_rotr/_lrotl/_lrotr
names in their prototypes, all I've tested is the sed command on mingw.org and
mingw64 stdlib.h and eyeballed the differences.
Can anyone please test this both with mingw64 and mingw.org stdlib.h?
Thanks.

[Bug target/40722] [4.5/4.6 Regression] ia32intrin.h defines of _rotl, _rotr conflict with target stdlib.h decls

[ktietz at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40722

Kai Tietz <ktietz at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |RESOLVED
         Resolution|                            |INVALID

--- Comment #20 from Kai Tietz <ktietz at gcc dot gnu.org> 2011-02-03 12:29:18 UTC ---
So, as more as I think about it, come to the point that this fix-include for
stdlib.h isn't something good. The correct solution here is that mingw.org
fixes their stdlib.h header by () (as suggested by this fix-include sed
command, which by the way works as expected without clobbering ifdef/else
blocks).
The issue I see here is, that the local include directory (LOCAL_INCLUDE_DIR)
comes before fix-include in include order. So for a native configuration, this
won't solve anything as it won't be used at all.

So IMHO, this problem should be closed as invalid.

[Bug target/40722] New: ia32intrin.h defines of _rotl, _rotr conflict with target stdlib.h decls.h

[dannysmith at users dot sourceforge dot net]

These defines in ia32intrin.h 

#define _lrotl(a,b)             __rold((a), (b))
#define _lrotr(a,b)             __rord((a), (b))
...
#define _rotl(a,b)              __rold((a), (b))
#define _rotr(a,b)              __rord((a), (b))


conflict with mingw32 stdlib.h which
declares those names as functions

_CRTIMP unsigned int __cdecl __MINGW_NOTHROW _rotl(unsigned int, int)
__MINGW_ATTRIB_CONST;
_CRTIMP unsigned int __cdecl __MINGW_NOTHROW _rotr(unsigned int, int)
__MINGW_ATTRIB_CONST;
_CRTIMP unsigned long __cdecl __MINGW_NOTHROW _lrotl(unsigned long, int)
__MINGW_ATTRIB_CONST;
_CRTIMP unsigned long __cdecl __MINGW_NOTHROW _lrotr(unsigned long, int)
__MINGW_ATTRIB_CONST;


This is caught by  g++.dg/other/i386-[23456].C


-- 
           Summary: ia32intrin.h defines of _rotl, _rotr conflict with
                    target stdlib.h  decls.h
           Product: gcc
           Version: 4.5.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: target
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: dannysmith at users dot sourceforge dot net
 GCC build triplet: i686-pc-mingw32
  GCC host triplet: i686-pc-mingw32
GCC target triplet: i686-pc-mingw32


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40722

[Bug target/40722] [4.5/4.6 Regression] ia32intrin.h defines of _rotl, _rotr conflict with target stdlib.h decls

[rguenth at gcc dot gnu dot org]

------- Comment #16 from rguenth at gcc dot gnu dot org  2010-07-31 09:29 -------
GCC 4.5.1 is being released, adjusting target milestone.


-- 

rguenth at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|4.5.1                       |4.5.2


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40722