/* Common definitions.

  Copyright (C) 1986-2024 Free Software Foundation, Inc.

  This file is part of GDB.

  This program is free software; you can redistribute it and/or modify
  it under the terms of the GNU General Public License as published by
  the Free Software Foundation; either version 3 of the License, or
  (at your option) any later version.

  This program is distributed in the hope that it will be useful,
  but WITHOUT ANY WARRANTY; without even the implied warranty of
  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  GNU General Public License for more details.

  You should have received a copy of the GNU General Public License
  along with this program.  If not, see <http://www.gnu.org/licenses/>.  */

#ifndef GDBSUPPORT_COMMON_DEFS_H
#define GDBSUPPORT_COMMON_DEFS_H

#if defined (__SANITIZE_THREAD__) && defined (__GNUC__) \
 && !defined (__clang__) && __GNUC__ <= 13

/* Work around PR gcc/110799.  */
#pragma GCC optimize("-fno-hoist-adjacent-loads")
#endif

#include <gdbsupport/config.h>

#undef PACKAGE_NAME
#undef PACKAGE
#undef PACKAGE_VERSION
#undef PACKAGE_STRING
#undef PACKAGE_TARNAME

#include "gnulib/config.h"

/* From:
   https://www.gnu.org/software/gnulib/manual/html_node/stdint_002eh.html

  "On some hosts that predate C++11, when using C++ one must define
  __STDC_CONSTANT_MACROS to make visible the definitions of constant
  macros such as INTMAX_C, and one must define __STDC_LIMIT_MACROS to
  make visible the definitions of limit macros such as INTMAX_MAX.".

  And:
   https://www.gnu.org/software/gnulib/manual/html_node/inttypes_002eh.html

  "On some hosts that predate C++11, when using C++ one must define
  __STDC_FORMAT_MACROS to make visible the declarations of format
  macros such as PRIdMAX."

  Must do this before including any system header, since other system
  headers may include stdint.h/inttypes.h.  */
#define __STDC_CONSTANT_MACROS 1
#define __STDC_LIMIT_MACROS 1
#define __STDC_FORMAT_MACROS 1

/* Some distros enable _FORTIFY_SOURCE by default, which on occasion
  has caused build failures with -Wunused-result when a patch is
  developed on a distro that does not enable _FORTIFY_SOURCE.  We
  enable it here in order to try to catch these problems earlier;
  plus this seems like a reasonable safety measure.  The check for
  optimization is required because _FORTIFY_SOURCE only works when
  optimization is enabled.  If _FORTIFY_SOURCE is already defined,
  then we don't do anything.  Also, on MinGW, fortify requires
  linking to -lssp, and to avoid the hassle of checking for
  that and linking to it statically, we just don't define
  _FORTIFY_SOURCE there.  */

#if (!defined _FORTIFY_SOURCE && defined __OPTIMIZE__ && __OPTIMIZE__ > 0 \
    && !defined(__MINGW32__))
#define _FORTIFY_SOURCE 2
#endif

/* We don't support Windows versions before XP, so we define
  _WIN32_WINNT correspondingly to ensure the Windows API headers
  expose the required symbols.

  NOTE: this must be kept in sync with common.m4.  */
#if defined (__MINGW32__) || defined (__CYGWIN__)
# ifdef _WIN32_WINNT
#  if _WIN32_WINNT < 0x0501
#   undef _WIN32_WINNT
#   define _WIN32_WINNT 0x0501
#  endif
# else
#  define _WIN32_WINNT 0x0501
# endif
#endif  /* __MINGW32__ || __CYGWIN__ */

#include <stdarg.h>
#include <stdio.h>

/* Include both cstdlib and stdlib.h to ensure we have standard functions
  defined both in the std:: namespace and in the global namespace.  */
#include <cstdlib>
#include <stdlib.h>

#include <stddef.h>
#include <stdint.h>
#include <string.h>
#ifdef HAVE_STRINGS_H
#include <strings.h>
#endif
#include <errno.h>
#if HAVE_ALLOCA_H
#include <alloca.h>
#endif

#include "ansidecl.h"
/* This is defined by ansidecl.h, but we prefer gnulib's version.  On
  MinGW, gnulib might enable __USE_MINGW_ANSI_STDIO, which may or not
  require use of attribute gnu_printf instead of printf.  gnulib
  checks that at configure time.  Since _GL_ATTRIBUTE_FORMAT_PRINTF_STANDARD
  is compatible with ATTRIBUTE_PRINTF, simply use it.  */
#undef ATTRIBUTE_PRINTF
#define ATTRIBUTE_PRINTF _GL_ATTRIBUTE_FORMAT_PRINTF_STANDARD

/* This is defined by ansidecl.h, but we disable the attribute.

  Say a developer starts out with:
  ...
  extern void foo (void *ptr) __attribute__((nonnull (1)));
  void foo (void *ptr) {}
  ...
  with the idea in mind to catch:
  ...
  foo (nullptr);
  ...
  at compile time with -Werror=nonnull, and then adds:
  ...
   void foo (void *ptr) {
  +  gdb_assert (ptr != nullptr);
   }
  ...
  to catch:
  ...
  foo (variable_with_nullptr_value);
  ...
  at runtime as well.

  Said developer then verifies that the assert works (using -O0), and commits
  the code.

  Some other developer then checks out the code and accidentally writes some
  variant of:
  ...
  foo (variable_with_nullptr_value);
  ...
  and builds with -O2, and ... the assert doesn't trigger, because it's
  optimized away by gcc.

  There's no suppported recipe to prevent the assertion from being optimized
  away (other than: build with -O0, or remove the nonnull attribute).  Note
  that -fno-delete-null-pointer-checks does not help.  A patch was submitted
  to improve gcc documentation to point this out more clearly (
  https://gcc.gnu.org/pipermail/gcc-patches/2021-July/576218.html ).  The
  patch also mentions a possible workaround that obfuscates the pointer
  using:
  ...
   void foo (void *ptr) {
  +  asm ("" : "+r"(ptr));
     gdb_assert (ptr != nullptr);
   }
  ...
  but that still requires the developer to manually add this in all cases
  where that's necessary.

  A warning was added to detect the situation: -Wnonnull-compare, which does
  help in detecting those cases, but each new gcc release may indicate a new
  batch of locations that needs fixing, which means we've added a maintenance
  burden.

  We could try to deal with the problem more proactively by introducing a
  gdb_assert variant like:
  ...
  void gdb_assert_non_null (void *ptr) {
     asm ("" : "+r"(ptr));
     gdb_assert (ptr != nullptr);
   }
   void foo (void *ptr) {
     gdb_assert_nonnull (ptr);
   }
  ...
  and make it a coding style to use it everywhere, but again, maintenance
  burden.

  With all these things considered, for now we go with the solution with the
  least maintenance burden: disable the attribute, such that we reliably deal
  with it everywhere.  */
#undef ATTRIBUTE_NONNULL
#define ATTRIBUTE_NONNULL(m)

#define ATTRIBUTE_UNUSED_RESULT __attribute__ ((__warn_unused_result__))
#define ATTRIBUTE_USED __attribute__ ((__used__))

#include "libiberty.h"
#include "pathmax.h"
#include "gdb/signals.h"
#include "gdb_locale.h"
#include "ptid.h"
#include "common-types.h"
#include "common-utils.h"
#include "gdb_assert.h"
#include "errors.h"
#include "print-utils.h"
#include "common-debug.h"
#include "cleanups.h"
#include "common-exceptions.h"
#include "gdbsupport/poison.h"

/* Pull in gdb::unique_xmalloc_ptr.  */
#include "gdbsupport/gdb_unique_ptr.h"

/* sbrk on macOS is not useful for our purposes, since sbrk(0) always
  returns the same value.  brk/sbrk on macOS is just an emulation
  that always returns a pointer to a 4MB section reserved for
  that.  */

#if defined (HAVE_SBRK) && !__APPLE__
#define HAVE_USEFUL_SBRK 1
#endif

#endif /* GDBSUPPORT_COMMON_DEFS_H */

You are viewing proxied material from ftp.icm.edu.pl. The copyright of proxied material belongs to its original authors. Any comments or complaints in relation to proxied material should be directed to the original authors of the content concerned. Please see the disclaimer for more details.