From 1974532ca5223adc1615e0f81c0cc330ce4a41ca Mon Sep 17 00:00:00 2001
From: Tsahi Zidenberg <tsahee@amazon.com>
Date: Mon, 22 Jun 2020 08:26:47 +0000
Subject: [PATCH v2] Support outline-atomics on aarch64

outline-atomics is a gcc compilation flag that enables runtime detection
of CPU support for atomic instructions.
Performance on CPUs that do support atomic instructions is improved,
while compatibility and performance on CPUs without atomic instructions
is not hurt.
---
 configure    | 93 ++++++++++++++++++++++++++++++++++++++++++++++++++++
 configure.ac |  4 +++
 2 files changed, 97 insertions(+)

diff --git a/configure b/configure
index 19a3cd09a0..75c1ffc038 100755
--- a/configure
+++ b/configure
@@ -6319,6 +6319,99 @@ fi
   if test -n "$NOT_THE_CFLAGS"; then
     CFLAGS="$CFLAGS -Wno-stringop-truncation"
   fi
+  if test x"$host_cpu" == x"aarch64"; then
+
+{ $as_echo "$as_me:${as_lineno-$LINENO}: checking whether ${CC} supports -moutline-atomics, for CFLAGS" >&5
+$as_echo_n "checking whether ${CC} supports -moutline-atomics, for CFLAGS... " >&6; }
+if ${pgac_cv_prog_CC_cflags__moutline_atomics+:} false; then :
+  $as_echo_n "(cached) " >&6
+else
+  pgac_save_CFLAGS=$CFLAGS
+pgac_save_CC=$CC
+CC=${CC}
+CFLAGS="${CFLAGS} -moutline-atomics"
+ac_save_c_werror_flag=$ac_c_werror_flag
+ac_c_werror_flag=yes
+cat confdefs.h - <<_ACEOF >conftest.$ac_ext
+/* end confdefs.h.  */
+
+int
+main ()
+{
+
+  ;
+  return 0;
+}
+_ACEOF
+if ac_fn_c_try_compile "$LINENO"; then :
+  pgac_cv_prog_CC_cflags__moutline_atomics=yes
+else
+  pgac_cv_prog_CC_cflags__moutline_atomics=no
+fi
+rm -f core conftest.err conftest.$ac_objext conftest.$ac_ext
+ac_c_werror_flag=$ac_save_c_werror_flag
+CFLAGS="$pgac_save_CFLAGS"
+CC="$pgac_save_CC"
+fi
+{ $as_echo "$as_me:${as_lineno-$LINENO}: result: $pgac_cv_prog_CC_cflags__moutline_atomics" >&5
+$as_echo "$pgac_cv_prog_CC_cflags__moutline_atomics" >&6; }
+if test x"$pgac_cv_prog_CC_cflags__moutline_atomics" = x"yes"; then
+  CFLAGS="${CFLAGS} -moutline-atomics"
+fi
+
+
+    { $as_echo "$as_me:${as_lineno-$LINENO}: checking whether ${CXX} supports -moutline-atomics, for CXXFLAGS" >&5
+$as_echo_n "checking whether ${CXX} supports -moutline-atomics, for CXXFLAGS... " >&6; }
+if ${pgac_cv_prog_CXX_cxxflags__moutline_atomics+:} false; then :
+  $as_echo_n "(cached) " >&6
+else
+  pgac_save_CXXFLAGS=$CXXFLAGS
+pgac_save_CXX=$CXX
+CXX=${CXX}
+CXXFLAGS="${CXXFLAGS} -moutline-atomics"
+ac_save_cxx_werror_flag=$ac_cxx_werror_flag
+ac_cxx_werror_flag=yes
+ac_ext=cpp
+ac_cpp='$CXXCPP $CPPFLAGS'
+ac_compile='$CXX -c $CXXFLAGS $CPPFLAGS conftest.$ac_ext >&5'
+ac_link='$CXX -o conftest$ac_exeext $CXXFLAGS $CPPFLAGS $LDFLAGS conftest.$ac_ext $LIBS >&5'
+ac_compiler_gnu=$ac_cv_cxx_compiler_gnu
+
+cat confdefs.h - <<_ACEOF >conftest.$ac_ext
+/* end confdefs.h.  */
+
+int
+main ()
+{
+
+  ;
+  return 0;
+}
+_ACEOF
+if ac_fn_cxx_try_compile "$LINENO"; then :
+  pgac_cv_prog_CXX_cxxflags__moutline_atomics=yes
+else
+  pgac_cv_prog_CXX_cxxflags__moutline_atomics=no
+fi
+rm -f core conftest.err conftest.$ac_objext conftest.$ac_ext
+ac_ext=c
+ac_cpp='$CPP $CPPFLAGS'
+ac_compile='$CC -c $CFLAGS $CPPFLAGS conftest.$ac_ext >&5'
+ac_link='$CC -o conftest$ac_exeext $CFLAGS $CPPFLAGS $LDFLAGS conftest.$ac_ext $LIBS >&5'
+ac_compiler_gnu=$ac_cv_c_compiler_gnu
+
+ac_cxx_werror_flag=$ac_save_cxx_werror_flag
+CXXFLAGS="$pgac_save_CXXFLAGS"
+CXX="$pgac_save_CXX"
+fi
+{ $as_echo "$as_me:${as_lineno-$LINENO}: result: $pgac_cv_prog_CXX_cxxflags__moutline_atomics" >&5
+$as_echo "$pgac_cv_prog_CXX_cxxflags__moutline_atomics" >&6; }
+if test x"$pgac_cv_prog_CXX_cxxflags__moutline_atomics" = x"yes"; then
+  CXXFLAGS="${CXXFLAGS} -moutline-atomics"
+fi
+
+
+  fi
 elif test "$ICC" = yes; then
   # Intel's compiler has a bug/misoptimization in checking for
   # division by NAN (NaN == 0), -mp1 fixes it, so add it to the CFLAGS.
diff --git a/configure.ac b/configure.ac
index 6b9d0487a8..434313db5c 100644
--- a/configure.ac
+++ b/configure.ac
@@ -538,6 +538,10 @@ if test "$GCC" = yes -a "$ICC" = no; then
   if test -n "$NOT_THE_CFLAGS"; then
     CFLAGS="$CFLAGS -Wno-stringop-truncation"
   fi
+  if test x"$host_cpu" == x"aarch64"; then
+    PGAC_PROG_CC_CFLAGS_OPT([-moutline-atomics])
+    PGAC_PROG_CXX_CFLAGS_OPT([-moutline-atomics])
+  fi
 elif test "$ICC" = yes; then
   # Intel's compiler has a bug/misoptimization in checking for
   # division by NAN (NaN == 0), -mp1 fixes it, so add it to the CFLAGS.
-- 
2.25.1