Re: [patch bfd]: Fix PR ld/12742 for windows targets

[Kai Tietz <ktietz70 at googlemail dot com>]

2012/3/10 Jeffrey Walton <noloader@gmail.com>:
> On Fri, Mar 9, 2012 at 6:18 PM, xunxun <xunxun1982@gmail.com> wrote:
>> On Sat, Mar 10, 2012 at 3:54 AM, Kai Tietz <ktietz70@googlemail.com> wrote:
>>> Hi,
>>>
>>> this patch adds compatible implementation of dlfcn API on Windows
>>> targets, if this header/API isn't present.
>>>
>>> ChangeLog
>>>
>>> 2012-03-09 ?Kai Tietz ?<ktietz@redhat.com>
>>>
>>> ? ? ? ?PR ld/12742
>>> ? ? ? ?* configure.in (AC_CHECK_HEADERS): Test for windows.h and dlfcn.h.
>>> ? ? ? ?* plugin.c: Guard include of dlfcn.h if HAVE_DLFCN_H is defined.
>>> ? ? ? ?Add windows.h header include if HAVE_WINDOWS_H is defined.
>>> ? ? ? ?(dlerror): New static function if windows variant is used instead
>>> ? ? ? ?of dlfcn.h.
>>> ? ? ? ?(dlclose): Likewise.
>>> ? ? ? ?(dlopen): Likewise.
>>> ? ? ? ?(dlsym): Likewise.
>>> ? ? ? ?* configure: Regenerated.
>>> ? ? ? ?* config.in: Regenerated.
>>>
>>> Tested for i686-w64-mingw32 and x86_64-w64-mingw32. ?Ok for apply?
>>>
>>> Regards,
>>> Kai
>>>
>>> Index: config.in
>>> ===================================================================
>>> RCS file: /cvs/src/src/bfd/config.in,v
>>> retrieving revision 1.49
>>> diff -u -p -r1.49 config.in
>>> --- config.in ? 12 May 2011 07:41:40 -0000 ? ? ?1.49
>>> +++ config.in ? 9 Mar 2012 19:38:18 -0000
>>> @@ -245,6 +245,9 @@
>>> ?/* Define if <sys/procfs.h> has win32_pstatus_t. */
>>> ?#undef HAVE_WIN32_PSTATUS_T
>>>
>>> +/* Define to 1 if you have the <windows.h> header file. */
>>> +#undef HAVE_WINDOWS_H
>>> +
>>> ?/* Define to 1 if you have the <zlib.h> header file. */
>>> ?#undef HAVE_ZLIB_H
>>>
>>> Index: configure
>>> ===================================================================
>>> RCS file: /cvs/src/src/bfd/configure,v
>>> retrieving revision 1.359
>>> diff -u -p -r1.359 configure
>>> --- configure ? 25 Feb 2012 19:51:31 -0000 ? ? ?1.359
>>> +++ configure ? 9 Mar 2012 19:38:21 -0000
>>> @@ -13508,6 +13508,22 @@ fi
>>>
>>> ?fi
>>>
>>> +
>>> +for ac_header in windows.h dlfcn.h
>>> +do :
>>> + ?as_ac_Header=`$as_echo "ac_cv_header_$ac_header" | $as_tr_sh`
>>> +ac_fn_c_check_header_mongrel "$LINENO" "$ac_header" "$as_ac_Header"
>>> "$ac_includes_default"
>>> +eval as_val=\$$as_ac_Header
>>> + ? if test "x$as_val" = x""yes; then :
>>> + ?cat >>confdefs.h <<_ACEOF
>>> +#define `$as_echo "HAVE_$ac_header" | $as_tr_cpp` 1
>>> +_ACEOF
>>> +
>>> +fi
>>> +
>>> +done
>>> +
>>> +
>>> ?{ $as_echo "$as_me:${as_lineno-$LINENO}: checking whether string.h
>>> and strings.h may both be included" >&5
>>> ?$as_echo_n "checking whether string.h and strings.h may both be
>>> included... " >&6; }
>>> ?if test "${gcc_cv_header_string+set}" = set; then :
>>> Index: configure.in
>>> ===================================================================
>>> RCS file: /cvs/src/src/bfd/configure.in,v
>>> retrieving revision 1.307
>>> diff -u -p -r1.307 configure.in
>>> --- configure.in ? ? ? ?25 Feb 2012 19:51:32 -0000 ? ? ?1.307
>>> +++ configure.in ? ? ? ?9 Mar 2012 19:38:21 -0000
>>> @@ -190,6 +190,9 @@ AC_CHECK_HEADERS(fcntl.h sys/file.h sys/
>>> ?GCC_HEADER_STDINT(bfd_stdint.h)
>>> ?AC_HEADER_TIME
>>> ?AC_HEADER_DIRENT
>>> +
>>> +AC_CHECK_HEADERS(windows.h dlfcn.h)
>>> +
>>> ?ACX_HEADER_STRING
>>> ?AC_CHECK_FUNCS(fcntl getpagesize setitimer sysconf fdopen getuid getgid fileno)
>>> ?AC_CHECK_FUNCS(strtoull)
>>> Index: plugin.c
>>> ===================================================================
>>> RCS file: /cvs/src/src/bfd/plugin.c,v
>>> retrieving revision 1.14
>>> diff -u -p -r1.14 plugin.c
>>> --- plugin.c ? ?11 Jul 2011 15:03:07 -0000 ? ? ?1.14
>>> +++ plugin.c ? ?9 Mar 2012 19:38:22 -0000
>>> @@ -25,7 +25,13 @@
>>> ?#if BFD_SUPPORTS_PLUGINS
>>>
>>> ?#include <assert.h>
>>> +#ifdef HAVE_DLFCN_H
>>> ?#include <dlfcn.h>
>>> +#elif defined (HAVE_WINDOWS_H)
>>> +#include <windows.h>
>>> +#else
>>> +#error Unknown how to handle dynamic-load-libraries.
>>> +#endif
>>> ?#include <stdarg.h>
>>> ?#include "plugin-api.h"
>>> ?#include "sysdep.h"
>>> @@ -34,6 +40,37 @@
>>> ?#include "libiberty.h"
>>> ?#include <dirent.h>
>>>
>>> +#if !defined (HAVE_DLFCN_H) && defined (HAVE_WINDOWS_H)
>>> +
>>> +#define RTLD_NOW 0 ? ? ?/* Dummy value. ?*/
>>> +
>>> +static void *
>>> +dlopen (const char *file, int mode ATTRIBUTE_UNUSED)
>>> +{
>>> + ?return LoadLibrary (file);
>>> +}
>>> +
>>> +static void *
>>> +dlsym (void *handle, const char *name)
>>> +{
>>> + ?return GetProcAddress (handle, name);
>>> +}
>>> +
>>> +static int ATTRIBUTE_UNUSED
>>> +dlclose (void *handle)
>>> +{
>>> + ?FreeLibrary (handle);
>>> + ?return 0;
>>> +}
>>> +
>>> +static const char *
>>> +dlerror (void)
>>> +{
>>> + ?return "Unable to load DLL.";
>>> +}
>>> +
>>> +#endif /* !defined (HAVE_DLFCN_H) && defined (HAVE_WINDOWS_H) ?*/
>>> +
>>> ?#define bfd_plugin_close_and_cleanup
>>> _bfd_generic_close_and_cleanup
>>> ?#define bfd_plugin_bfd_free_cached_info
>>> _bfd_generic_bfd_free_cached_info
>>> ?#define bfd_plugin_new_section_hook
>>> _bfd_generic_new_section_hook
>>
>> why to add static?
>>
>> And I think dlclose is
>> int ATTRIBUTE_UNUSED
>> dlclose (void *handle)
>> {
>> ?bool ret;
>> ?ret = FreeLibrary (handle);
>> ?ret = !ret;
>> ?return (int) ret;
>> }
>> The similar is better.
> BOOL, not bool.
> http://msdn.microsoft.com/en-us/library/windows/desktop/ms683152(v=vs.85).aspx.

And, what shall be here the difference?  I still don't get the point.
A BOOL is for windows a type of int, bool is char.  Interesting here
is that for both the value is either 1 or 0.  It makes no difference,
as result is autopromoted.

Kai