Merge commit 'origin/master' into fedora/master

26 files changed, 1450 insertions, 306 deletions

diff --git a/sysdeps/i386/configure b/sysdeps/i386/configure
index ced0b31d0f..d1c4f7f501 100755
--- a/sysdeps/i386/configure
+++ b/sysdeps/i386/configure
@@ -1,13 +1,468 @@
+# Factoring default headers for most tests.
+ac_includes_default="\
+#include <stdio.h>
+#ifdef HAVE_SYS_TYPES_H
+# include <sys/types.h>
+#endif
+#ifdef HAVE_SYS_STAT_H
+# include <sys/stat.h>
+#endif
+#ifdef STDC_HEADERS
+# include <stdlib.h>
+# include <stddef.h>
+#else
+# ifdef HAVE_STDLIB_H
+#  include <stdlib.h>
+# endif
+#endif
+#ifdef HAVE_STRING_H
+# if !defined STDC_HEADERS && defined HAVE_MEMORY_H
+#  include <memory.h>
+# endif
+# include <string.h>
+#endif
+#ifdef HAVE_STRINGS_H
+# include <strings.h>
+#endif
+#ifdef HAVE_INTTYPES_H
+# include <inttypes.h>
+#endif
+#ifdef HAVE_STDINT_H
+# include <stdint.h>
+#endif
+#ifdef HAVE_UNISTD_H
+# include <unistd.h>
+#endif"
+
 # This file is generated from configure.in by Autoconf.  DO NOT EDIT!
  # Local configure fragment for sysdeps/i386.
 
 
-{ echo "$as_me:$LINENO: checking if gcc provides <cpuid.h>" >&5
-echo $ECHO_N "checking if gcc provides <cpuid.h>... $ECHO_C" >&6; }
-if test "${libc_cv_gcc_cpuid+set}" = set; then
+
+{ echo "$as_me:$LINENO: checking for grep that handles long lines and -e" >&5
+echo $ECHO_N "checking for grep that handles long lines and -e... $ECHO_C" >&6; }
+if test "${ac_cv_path_GREP+set}" = set; then
+  echo $ECHO_N "(cached) $ECHO_C" >&6
+else
+  # Extract the first word of "grep ggrep" to use in msg output
+if test -z "$GREP"; then
+set dummy grep ggrep; ac_prog_name=$2
+if test "${ac_cv_path_GREP+set}" = set; then
+  echo $ECHO_N "(cached) $ECHO_C" >&6
+else
+  ac_path_GREP_found=false
+# Loop through the user's path and test for each of PROGNAME-LIST
+as_save_IFS=$IFS; IFS=$PATH_SEPARATOR
+for as_dir in $PATH$PATH_SEPARATOR/usr/xpg4/bin
+do
+  IFS=$as_save_IFS
+  test -z "$as_dir" && as_dir=.
+  for ac_prog in grep ggrep; do
+  for ac_exec_ext in '' $ac_executable_extensions; do
+    ac_path_GREP="$as_dir/$ac_prog$ac_exec_ext"
+    { test -f "$ac_path_GREP" && $as_test_x "$ac_path_GREP"; } || continue
+    # Check for GNU ac_path_GREP and select it if it is found.
+  # Check for GNU $ac_path_GREP
+case `"$ac_path_GREP" --version 2>&1` in
+*GNU*)
+  ac_cv_path_GREP="$ac_path_GREP" ac_path_GREP_found=:;;
+*)
+  ac_count=0
+  echo $ECHO_N "0123456789$ECHO_C" >"conftest.in"
+  while :
+  do
+    cat "conftest.in" "conftest.in" >"conftest.tmp"
+    mv "conftest.tmp" "conftest.in"
+    cp "conftest.in" "conftest.nl"
+    echo 'GREP' >> "conftest.nl"
+    "$ac_path_GREP" -e 'GREP$' -e '-(cannot match)-' < "conftest.nl" >"conftest.out" 2>/dev/null || break
+    diff "conftest.out" "conftest.nl" >/dev/null 2>&1 || break
+    ac_count=`expr $ac_count + 1`
+    if test $ac_count -gt ${ac_path_GREP_max-0}; then
+      # Best one so far, save it but keep looking for a better one
+      ac_cv_path_GREP="$ac_path_GREP"
+      ac_path_GREP_max=$ac_count
+    fi
+    # 10*(2^10) chars as input seems more than enough
+    test $ac_count -gt 10 && break
+  done
+  rm -f conftest.in conftest.tmp conftest.nl conftest.out;;
+esac
+
+
+    $ac_path_GREP_found && break 3
+  done
+done
+
+done
+IFS=$as_save_IFS
+
+
+fi
+
+GREP="$ac_cv_path_GREP"
+if test -z "$GREP"; then
+  { { echo "$as_me:$LINENO: error: no acceptable $ac_prog_name could be found in $PATH$PATH_SEPARATOR/usr/xpg4/bin" >&5
+echo "$as_me: error: no acceptable $ac_prog_name could be found in $PATH$PATH_SEPARATOR/usr/xpg4/bin" >&2;}
+   { (exit 1); exit 1; }; }
+fi
+
+else
+  ac_cv_path_GREP=$GREP
+fi
+
+
+fi
+{ echo "$as_me:$LINENO: result: $ac_cv_path_GREP" >&5
+echo "${ECHO_T}$ac_cv_path_GREP" >&6; }
+ GREP="$ac_cv_path_GREP"
+
+
+{ echo "$as_me:$LINENO: checking for egrep" >&5
+echo $ECHO_N "checking for egrep... $ECHO_C" >&6; }
+if test "${ac_cv_path_EGREP+set}" = set; then
+  echo $ECHO_N "(cached) $ECHO_C" >&6
+else
+  if echo a | $GREP -E '(a|b)' >/dev/null 2>&1
+   then ac_cv_path_EGREP="$GREP -E"
+   else
+     # Extract the first word of "egrep" to use in msg output
+if test -z "$EGREP"; then
+set dummy egrep; ac_prog_name=$2
+if test "${ac_cv_path_EGREP+set}" = set; then
+  echo $ECHO_N "(cached) $ECHO_C" >&6
+else
+  ac_path_EGREP_found=false
+# Loop through the user's path and test for each of PROGNAME-LIST
+as_save_IFS=$IFS; IFS=$PATH_SEPARATOR
+for as_dir in $PATH$PATH_SEPARATOR/usr/xpg4/bin
+do
+  IFS=$as_save_IFS
+  test -z "$as_dir" && as_dir=.
+  for ac_prog in egrep; do
+  for ac_exec_ext in '' $ac_executable_extensions; do
+    ac_path_EGREP="$as_dir/$ac_prog$ac_exec_ext"
+    { test -f "$ac_path_EGREP" && $as_test_x "$ac_path_EGREP"; } || continue
+    # Check for GNU ac_path_EGREP and select it if it is found.
+  # Check for GNU $ac_path_EGREP
+case `"$ac_path_EGREP" --version 2>&1` in
+*GNU*)
+  ac_cv_path_EGREP="$ac_path_EGREP" ac_path_EGREP_found=:;;
+*)
+  ac_count=0
+  echo $ECHO_N "0123456789$ECHO_C" >"conftest.in"
+  while :
+  do
+    cat "conftest.in" "conftest.in" >"conftest.tmp"
+    mv "conftest.tmp" "conftest.in"
+    cp "conftest.in" "conftest.nl"
+    echo 'EGREP' >> "conftest.nl"
+    "$ac_path_EGREP" 'EGREP$' < "conftest.nl" >"conftest.out" 2>/dev/null || break
+    diff "conftest.out" "conftest.nl" >/dev/null 2>&1 || break
+    ac_count=`expr $ac_count + 1`
+    if test $ac_count -gt ${ac_path_EGREP_max-0}; then
+      # Best one so far, save it but keep looking for a better one
+      ac_cv_path_EGREP="$ac_path_EGREP"
+      ac_path_EGREP_max=$ac_count
+    fi
+    # 10*(2^10) chars as input seems more than enough
+    test $ac_count -gt 10 && break
+  done
+  rm -f conftest.in conftest.tmp conftest.nl conftest.out;;
+esac
+
+
+    $ac_path_EGREP_found && break 3
+  done
+done
+
+done
+IFS=$as_save_IFS
+
+
+fi
+
+EGREP="$ac_cv_path_EGREP"
+if test -z "$EGREP"; then
+  { { echo "$as_me:$LINENO: error: no acceptable $ac_prog_name could be found in $PATH$PATH_SEPARATOR/usr/xpg4/bin" >&5
+echo "$as_me: error: no acceptable $ac_prog_name could be found in $PATH$PATH_SEPARATOR/usr/xpg4/bin" >&2;}
+   { (exit 1); exit 1; }; }
+fi
+
+else
+  ac_cv_path_EGREP=$EGREP
+fi
+
+
+   fi
+fi
+{ echo "$as_me:$LINENO: result: $ac_cv_path_EGREP" >&5
+echo "${ECHO_T}$ac_cv_path_EGREP" >&6; }
+ EGREP="$ac_cv_path_EGREP"
+
+
+{ echo "$as_me:$LINENO: checking for ANSI C header files" >&5
+echo $ECHO_N "checking for ANSI C header files... $ECHO_C" >&6; }
+if test "${ac_cv_header_stdc+set}" = set; then
   echo $ECHO_N "(cached) $ECHO_C" >&6
 else
   cat >conftest.$ac_ext <<_ACEOF
+/* confdefs.h.  */
+_ACEOF
+cat confdefs.h >>conftest.$ac_ext
+cat >>conftest.$ac_ext <<_ACEOF
+/* end confdefs.h.  */
+#include <stdlib.h>
+#include <stdarg.h>
+#include <string.h>
+#include <float.h>
+
+int
+main ()
+{
+
+  ;
+  return 0;
+}
+_ACEOF
+rm -f conftest.$ac_objext
+if { (ac_try="$ac_compile"
+case "(($ac_try" in
+  *\"* | *\`* | *\\*) ac_try_echo=\$ac_try;;
+  *) ac_try_echo=$ac_try;;
+esac
+eval "echo \"\$as_me:$LINENO: $ac_try_echo\"") >&5
+  (eval "$ac_compile") 2>conftest.er1
+  ac_status=$?
+  grep -v '^ *+' conftest.er1 >conftest.err
+  rm -f conftest.er1
+  cat conftest.err >&5
+  echo "$as_me:$LINENO: \$? = $ac_status" >&5
+  (exit $ac_status); } && {
+	 test -z "$ac_c_werror_flag" ||
+	 test ! -s conftest.err
+       } && test -s conftest.$ac_objext; then
+  ac_cv_header_stdc=yes
+else
+  echo "$as_me: failed program was:" >&5
+sed 's/^/| /' conftest.$ac_ext >&5
+
+	ac_cv_header_stdc=no
+fi
+
+rm -f core conftest.err conftest.$ac_objext conftest.$ac_ext
+
+if test $ac_cv_header_stdc = yes; then
+  # SunOS 4.x string.h does not declare mem*, contrary to ANSI.
+  cat >conftest.$ac_ext <<_ACEOF
+/* confdefs.h.  */
+_ACEOF
+cat confdefs.h >>conftest.$ac_ext
+cat >>conftest.$ac_ext <<_ACEOF
+/* end confdefs.h.  */
+#include <string.h>
+
+_ACEOF
+if (eval "$ac_cpp conftest.$ac_ext") 2>&5 |
+  $EGREP "memchr" >/dev/null 2>&1; then
+  :
+else
+  ac_cv_header_stdc=no
+fi
+rm -f conftest*
+
+fi
+
+if test $ac_cv_header_stdc = yes; then
+  # ISC 2.0.2 stdlib.h does not declare free, contrary to ANSI.
+  cat >conftest.$ac_ext <<_ACEOF
+/* confdefs.h.  */
+_ACEOF
+cat confdefs.h >>conftest.$ac_ext
+cat >>conftest.$ac_ext <<_ACEOF
+/* end confdefs.h.  */
+#include <stdlib.h>
+
+_ACEOF
+if (eval "$ac_cpp conftest.$ac_ext") 2>&5 |
+  $EGREP "free" >/dev/null 2>&1; then
+  :
+else
+  ac_cv_header_stdc=no
+fi
+rm -f conftest*
+
+fi
+
+if test $ac_cv_header_stdc = yes; then
+  # /bin/cc in Irix-4.0.5 gets non-ANSI ctype macros unless using -ansi.
+  if test "$cross_compiling" = yes; then
+  :
+else
+  cat >conftest.$ac_ext <<_ACEOF
+/* confdefs.h.  */
+_ACEOF
+cat confdefs.h >>conftest.$ac_ext
+cat >>conftest.$ac_ext <<_ACEOF
+/* end confdefs.h.  */
+#include <ctype.h>
+#include <stdlib.h>
+#if ((' ' & 0x0FF) == 0x020)
+# define ISLOWER(c) ('a' <= (c) && (c) <= 'z')
+# define TOUPPER(c) (ISLOWER(c) ? 'A' + ((c) - 'a') : (c))
+#else
+# define ISLOWER(c) \
+		   (('a' <= (c) && (c) <= 'i') \
+		     || ('j' <= (c) && (c) <= 'r') \
+		     || ('s' <= (c) && (c) <= 'z'))
+# define TOUPPER(c) (ISLOWER(c) ? ((c) | 0x40) : (c))
+#endif
+
+#define XOR(e, f) (((e) && !(f)) || (!(e) && (f)))
+int
+main ()
+{
+  int i;
+  for (i = 0; i < 256; i++)
+    if (XOR (islower (i), ISLOWER (i))
+	|| toupper (i) != TOUPPER (i))
+      return 2;
+  return 0;
+}
+_ACEOF
+rm -f conftest$ac_exeext
+if { (ac_try="$ac_link"
+case "(($ac_try" in
+  *\"* | *\`* | *\\*) ac_try_echo=\$ac_try;;
+  *) ac_try_echo=$ac_try;;
+esac
+eval "echo \"\$as_me:$LINENO: $ac_try_echo\"") >&5
+  (eval "$ac_link") 2>&5
+  ac_status=$?
+  echo "$as_me:$LINENO: \$? = $ac_status" >&5
+  (exit $ac_status); } && { ac_try='./conftest$ac_exeext'
+  { (case "(($ac_try" in
+  *\"* | *\`* | *\\*) ac_try_echo=\$ac_try;;
+  *) ac_try_echo=$ac_try;;
+esac
+eval "echo \"\$as_me:$LINENO: $ac_try_echo\"") >&5
+  (eval "$ac_try") 2>&5
+  ac_status=$?
+  echo "$as_me:$LINENO: \$? = $ac_status" >&5
+  (exit $ac_status); }; }; then
+  :
+else
+  echo "$as_me: program exited with status $ac_status" >&5
+echo "$as_me: failed program was:" >&5
+sed 's/^/| /' conftest.$ac_ext >&5
+
+( exit $ac_status )
+ac_cv_header_stdc=no
+fi
+rm -f core *.core core.conftest.* gmon.out bb.out conftest$ac_exeext conftest.$ac_objext conftest.$ac_ext
+fi
+
+
+fi
+fi
+{ echo "$as_me:$LINENO: result: $ac_cv_header_stdc" >&5
+echo "${ECHO_T}$ac_cv_header_stdc" >&6; }
+if test $ac_cv_header_stdc = yes; then
+
+cat >>confdefs.h <<\_ACEOF
+#define STDC_HEADERS 1
+_ACEOF
+
+fi
+
+# On IRIX 5.3, sys/types and inttypes.h are conflicting.
+
+
+
+
+
+
+
+
+
+for ac_header in sys/types.h sys/stat.h stdlib.h string.h memory.h strings.h \
+		  inttypes.h stdint.h unistd.h
+do
+as_ac_Header=`echo "ac_cv_header_$ac_header" | $as_tr_sh`
+{ echo "$as_me:$LINENO: checking for $ac_header" >&5
+echo $ECHO_N "checking for $ac_header... $ECHO_C" >&6; }
+if { as_var=$as_ac_Header; eval "test \"\${$as_var+set}\" = set"; }; then
+  echo $ECHO_N "(cached) $ECHO_C" >&6
+else
+  cat >conftest.$ac_ext <<_ACEOF
+/* confdefs.h.  */
+_ACEOF
+cat confdefs.h >>conftest.$ac_ext
+cat >>conftest.$ac_ext <<_ACEOF
+/* end confdefs.h.  */
+$ac_includes_default
+
+#include <$ac_header>
+_ACEOF
+rm -f conftest.$ac_objext
+if { (ac_try="$ac_compile"
+case "(($ac_try" in
+  *\"* | *\`* | *\\*) ac_try_echo=\$ac_try;;
+  *) ac_try_echo=$ac_try;;
+esac
+eval "echo \"\$as_me:$LINENO: $ac_try_echo\"") >&5
+  (eval "$ac_compile") 2>conftest.er1
+  ac_status=$?
+  grep -v '^ *+' conftest.er1 >conftest.err
+  rm -f conftest.er1
+  cat conftest.err >&5
+  echo "$as_me:$LINENO: \$? = $ac_status" >&5
+  (exit $ac_status); } && {
+	 test -z "$ac_c_werror_flag" ||
+	 test ! -s conftest.err
+       } && test -s conftest.$ac_objext; then
+  eval "$as_ac_Header=yes"
+else
+  echo "$as_me: failed program was:" >&5
+sed 's/^/| /' conftest.$ac_ext >&5
+
+	eval "$as_ac_Header=no"
+fi
+
+rm -f core conftest.err conftest.$ac_objext conftest.$ac_ext
+fi
+ac_res=`eval echo '${'$as_ac_Header'}'`
+	       { echo "$as_me:$LINENO: result: $ac_res" >&5
+echo "${ECHO_T}$ac_res" >&6; }
+if test `eval echo '${'$as_ac_Header'}'` = yes; then
+  cat >>confdefs.h <<_ACEOF
+#define `echo "HAVE_$ac_header" | $as_tr_cpp` 1
+_ACEOF
+
+fi
+
+done
+
+
+if test "${ac_cv_header_cpuid_h+set}" = set; then
+  { echo "$as_me:$LINENO: checking for cpuid.h" >&5
+echo $ECHO_N "checking for cpuid.h... $ECHO_C" >&6; }
+if test "${ac_cv_header_cpuid_h+set}" = set; then
+  echo $ECHO_N "(cached) $ECHO_C" >&6
+fi
+{ echo "$as_me:$LINENO: result: $ac_cv_header_cpuid_h" >&5
+echo "${ECHO_T}$ac_cv_header_cpuid_h" >&6; }
+else
+  # Is the header compilable?
+{ echo "$as_me:$LINENO: checking cpuid.h usability" >&5
+echo $ECHO_N "checking cpuid.h usability... $ECHO_C" >&6; }
+cat >conftest.$ac_ext <<_ACEOF
+/* confdefs.h.  */
+_ACEOF
+cat confdefs.h >>conftest.$ac_ext
+cat >>conftest.$ac_ext <<_ACEOF
+/* end confdefs.h.  */
+$ac_includes_default
 #include <cpuid.h>
 _ACEOF
 rm -f conftest.$ac_objext
@@ -27,24 +482,103 @@ eval "echo \"\$as_me:$LINENO: $ac_try_echo\"") >&5
 	 test -z "$ac_c_werror_flag" ||
 	 test ! -s conftest.err
        } && test -s conftest.$ac_objext; then
-  libc_cv_gcc_cpuid=yes
+  ac_header_compiler=yes
 else
   echo "$as_me: failed program was:" >&5
 sed 's/^/| /' conftest.$ac_ext >&5
 
-	libc_cv_gcc_cpuid=no
+	ac_header_compiler=no
 fi
 
 rm -f core conftest.err conftest.$ac_objext conftest.$ac_ext
+{ echo "$as_me:$LINENO: result: $ac_header_compiler" >&5
+echo "${ECHO_T}$ac_header_compiler" >&6; }
+
+# Is the header present?
+{ echo "$as_me:$LINENO: checking cpuid.h presence" >&5
+echo $ECHO_N "checking cpuid.h presence... $ECHO_C" >&6; }
+cat >conftest.$ac_ext <<_ACEOF
+/* confdefs.h.  */
+_ACEOF
+cat confdefs.h >>conftest.$ac_ext
+cat >>conftest.$ac_ext <<_ACEOF
+/* end confdefs.h.  */
+#include <cpuid.h>
+_ACEOF
+if { (ac_try="$ac_cpp conftest.$ac_ext"
+case "(($ac_try" in
+  *\"* | *\`* | *\\*) ac_try_echo=\$ac_try;;
+  *) ac_try_echo=$ac_try;;
+esac
+eval "echo \"\$as_me:$LINENO: $ac_try_echo\"") >&5
+  (eval "$ac_cpp conftest.$ac_ext") 2>conftest.er1
+  ac_status=$?
+  grep -v '^ *+' conftest.er1 >conftest.err
+  rm -f conftest.er1
+  cat conftest.err >&5
+  echo "$as_me:$LINENO: \$? = $ac_status" >&5
+  (exit $ac_status); } >/dev/null && {
+	 test -z "$ac_c_preproc_warn_flag$ac_c_werror_flag" ||
+	 test ! -s conftest.err
+       }; then
+  ac_header_preproc=yes
+else
+  echo "$as_me: failed program was:" >&5
+sed 's/^/| /' conftest.$ac_ext >&5
+
+  ac_header_preproc=no
 fi
-{ echo "$as_me:$LINENO: result: $libc_cv_gcc_cpuid" >&5
-echo "${ECHO_T}$libc_cv_gcc_cpuid" >&6; }
-if test $libc_cv_gcc_cpuid != yes; then
+
+rm -f conftest.err conftest.$ac_ext
+{ echo "$as_me:$LINENO: result: $ac_header_preproc" >&5
+echo "${ECHO_T}$ac_header_preproc" >&6; }
+
+# So?  What about this header?
+case $ac_header_compiler:$ac_header_preproc:$ac_c_preproc_warn_flag in
+  yes:no: )
+    { echo "$as_me:$LINENO: WARNING: cpuid.h: accepted by the compiler, rejected by the preprocessor!" >&5
+echo "$as_me: WARNING: cpuid.h: accepted by the compiler, rejected by the preprocessor!" >&2;}
+    { echo "$as_me:$LINENO: WARNING: cpuid.h: proceeding with the compiler's result" >&5
+echo "$as_me: WARNING: cpuid.h: proceeding with the compiler's result" >&2;}
+    ac_header_preproc=yes
+    ;;
+  no:yes:* )
+    { echo "$as_me:$LINENO: WARNING: cpuid.h: present but cannot be compiled" >&5
+echo "$as_me: WARNING: cpuid.h: present but cannot be compiled" >&2;}
+    { echo "$as_me:$LINENO: WARNING: cpuid.h:     check for missing prerequisite headers?" >&5
+echo "$as_me: WARNING: cpuid.h:     check for missing prerequisite headers?" >&2;}
+    { echo "$as_me:$LINENO: WARNING: cpuid.h: see the Autoconf documentation" >&5
+echo "$as_me: WARNING: cpuid.h: see the Autoconf documentation" >&2;}
+    { echo "$as_me:$LINENO: WARNING: cpuid.h:     section \"Present But Cannot Be Compiled\"" >&5
+echo "$as_me: WARNING: cpuid.h:     section \"Present But Cannot Be Compiled\"" >&2;}
+    { echo "$as_me:$LINENO: WARNING: cpuid.h: proceeding with the preprocessor's result" >&5
+echo "$as_me: WARNING: cpuid.h: proceeding with the preprocessor's result" >&2;}
+    { echo "$as_me:$LINENO: WARNING: cpuid.h: in the future, the compiler will take precedence" >&5
+echo "$as_me: WARNING: cpuid.h: in the future, the compiler will take precedence" >&2;}
+
+    ;;
+esac
+{ echo "$as_me:$LINENO: checking for cpuid.h" >&5
+echo $ECHO_N "checking for cpuid.h... $ECHO_C" >&6; }
+if test "${ac_cv_header_cpuid_h+set}" = set; then
+  echo $ECHO_N "(cached) $ECHO_C" >&6
+else
+  ac_cv_header_cpuid_h=$ac_header_preproc
+fi
+{ echo "$as_me:$LINENO: result: $ac_cv_header_cpuid_h" >&5
+echo "${ECHO_T}$ac_cv_header_cpuid_h" >&6; }
+
+fi
+if test $ac_cv_header_cpuid_h = yes; then
+  :
+else
   { { echo "$as_me:$LINENO: error: gcc must provide the <cpuid.h> header" >&5
 echo "$as_me: error: gcc must provide the <cpuid.h> header" >&2;}
    { (exit 1); exit 1; }; }
 fi
 
+
+
 { echo "$as_me:$LINENO: checking if -g produces usable source locations for assembler-with-cpp" >&5
 echo $ECHO_N "checking if -g produces usable source locations for assembler-with-cpp... $ECHO_C" >&6; }
 if test "${libc_cv_cpp_asm_debuginfo+set}" = set; then
diff --git a/sysdeps/i386/configure.in b/sysdeps/i386/configure.in
index 800f928fbd..12dceaf844 100644
--- a/sysdeps/i386/configure.in
+++ b/sysdeps/i386/configure.in
@@ -1,12 +1,8 @@
 GLIBC_PROVIDES dnl See aclocal.m4 in the top level source directory.
 # Local configure fragment for sysdeps/i386.
 
-AC_CACHE_CHECK([if gcc provides <cpuid.h>], libc_cv_gcc_cpuid, [dnl
-AC_COMPILE_IFELSE([#include <cpuid.h>], libc_cv_gcc_cpuid=yes,
-		  libc_cv_gcc_cpuid=no)])
-if test $libc_cv_gcc_cpuid != yes; then
-  AC_MSG_ERROR([gcc must provide the <cpuid.h> header])
-fi
+AC_HEADER_CHECK([cpuid.h], ,
+  [AC_MSG_ERROR([gcc must provide the <cpuid.h> header])])
 
 AC_CACHE_CHECK(if -g produces usable source locations for assembler-with-cpp,
 	       libc_cv_cpp_asm_debuginfo, [dnl
diff --git a/sysdeps/i386/i686/multiarch/Makefile b/sysdeps/i386/i686/multiarch/Makefile
index 33d98c36e6..e1553b284e 100644
--- a/sysdeps/i386/i686/multiarch/Makefile
+++ b/sysdeps/i386/i686/multiarch/Makefile
@@ -2,3 +2,14 @@ ifeq ($(subdir),csu)
 aux += init-arch
 gen-as-const-headers += ifunc-defines.sym
 endif
+
+ifeq ($(subdir),string)
+ifeq (yes,$(config-cflags-sse4))
+sysdep_routines += strcspn-c strpbrk-c strspn-c strstr-c strcasestr-c
+CFLAGS-strcspn-c.c += -msse4
+CFLAGS-strpbrk-c.c += -msse4
+CFLAGS-strspn-c.c += -msse4
+CFLAGS-strstr.c += -msse4
+CFLAGS-strcasestr.c += -msse4
+endif
+endif
diff --git a/sysdeps/i386/i686/multiarch/strcasestr-c.c b/sysdeps/i386/i686/multiarch/strcasestr-c.c
new file mode 100644
index 0000000000..0d52b0e47a
--- /dev/null
+++ b/sysdeps/i386/i686/multiarch/strcasestr-c.c
@@ -0,0 +1,2 @@
+#define __strcasestr_sse2 __strcasestr_ia32
+#include <sysdeps/x86_64/multiarch/strcasestr-c.c>
diff --git a/sysdeps/i386/i686/multiarch/strcasestr.c b/sysdeps/i386/i686/multiarch/strcasestr.c
new file mode 100644
index 0000000000..511bb29ede
--- /dev/null
+++ b/sysdeps/i386/i686/multiarch/strcasestr.c
@@ -0,0 +1 @@
+#include <sysdeps/x86_64/multiarch/strcasestr.c>
diff --git a/sysdeps/i386/i686/multiarch/strcspn-c.c b/sysdeps/i386/i686/multiarch/strcspn-c.c
new file mode 100644
index 0000000000..6d61e190a8
--- /dev/null
+++ b/sysdeps/i386/i686/multiarch/strcspn-c.c
@@ -0,0 +1,2 @@
+#define __strcspn_sse2 __strcspn_ia32
+#include <sysdeps/x86_64/multiarch/strcspn-c.c>
diff --git a/sysdeps/i386/i686/multiarch/strcspn.S b/sysdeps/i386/i686/multiarch/strcspn.S
new file mode 100644
index 0000000000..73e7eb45a8
--- /dev/null
+++ b/sysdeps/i386/i686/multiarch/strcspn.S
@@ -0,0 +1,114 @@
+/* Multiple versions of strcspn
+   Copyright (C) 2009 Free Software Foundation, Inc.
+   Contributed by Intel Corporation.
+   This file is part of the GNU C Library.
+
+   The GNU C Library is free software; you can redistribute it and/or
+   modify it under the terms of the GNU Lesser General Public
+   License as published by the Free Software Foundation; either
+   version 2.1 of the License, or (at your option) any later version.
+
+   The GNU C Library 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
+   Lesser General Public License for more details.
+
+   You should have received a copy of the GNU Lesser General Public
+   License along with the GNU C Library; if not, write to the Free
+   Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
+   02111-1307 USA.  */
+
+#include <config.h>
+
+#ifdef HAVE_SSE4_SUPPORT
+
+#include <sysdep.h>
+#include <ifunc-defines.h>
+
+#ifdef USE_AS_STRPBRK
+#define STRCSPN_SSE42	__strpbrk_sse42
+#define STRCSPN_IA32	__strpbrk_ia32
+#define __GI_STRCSPN	__GI_strpbrk
+#else
+#ifndef STRCSPN
+#define STRCSPN		strcspn
+#define STRCSPN_SSE42	__strcspn_sse42
+#define STRCSPN_IA32	__strcspn_ia32
+#define __GI_STRCSPN	__GI_strcspn
+#endif
+#endif
+
+/* Define multiple versions only for the definition in libc.  Don't
+   define multiple versions for strpbrk in static library since we
+   need strpbrk before the initialization happened.  */
+#if (defined SHARED || !defined USE_AS_STRPBRK) && !defined NOT_IN_libc
+# ifdef SHARED
+	.section	.gnu.linkonce.t.__i686.get_pc_thunk.bx,"ax",@progbits
+	.globl	__i686.get_pc_thunk.bx
+	.hidden	__i686.get_pc_thunk.bx
+	.p2align 4
+	.type	__i686.get_pc_thunk.bx,@function
+__i686.get_pc_thunk.bx:
+	movl	(%esp), %ebx
+	ret
+
+	.text
+ENTRY(STRCSPN)
+	.type	STRCSPN, @gnu_indirect_function
+	pushl	%ebx
+	cfi_adjust_cfa_offset (4)
+	cfi_rel_offset (ebx, 0)
+	call	__i686.get_pc_thunk.bx
+	addl	$_GLOBAL_OFFSET_TABLE_, %ebx
+	cmpl	$0, KIND_OFFSET+__cpu_features@GOTOFF(%ebx)
+	jne	1f
+	call	__init_cpu_features
+1:	leal	STRCSPN_IA32@GOTOFF(%ebx), %eax
+	testl	$(1<<20), CPUID_OFFSET+COMMON_CPUID_INDEX_1*CPUID_SIZE+CPUID_ECX_OFFSET+__cpu_features@GOTOFF(%ebx)
+	jz	2f
+	leal	STRCSPN_SSE42@GOTOFF(%ebx), %eax
+2:	popl	%ebx
+	cfi_adjust_cfa_offset (-4);
+	cfi_restore (ebx)
+	ret
+END(STRCSPN)
+# else
+	.text
+ENTRY(STRCSPN)
+	.type	STRCSPN, @gnu_indirect_function
+	cmpl	$0, KIND_OFFSET+__cpu_features
+	jne	1f
+	call	__init_cpu_features
+1:	leal	STRCSPN_IA32, %eax
+	testl	$(1<<20), CPUID_OFFSET+COMMON_CPUID_INDEX_1*CPUID_SIZE+CPUID_ECX_OFFSET+__cpu_features
+	jz	2f
+	leal	STRCSPN_SSE42, %eax
+2:	ret
+END(STRCSPN)
+# endif
+
+# undef ENTRY
+# define ENTRY(name) \
+	.type STRCSPN_IA32, @function; \
+	.globl STRCSPN_IA32; \
+	.p2align 4; \
+	STRCSPN_IA32: cfi_startproc; \
+	CALL_MCOUNT
+# undef END
+# define END(name) \
+	cfi_endproc; .size STRCSPN_IA32, .-STRCSPN_IA32
+# undef libc_hidden_builtin_def
+/* IFUNC doesn't work with the hidden functions in shared library since
+   they will be called without setting up EBX needed for PLT which is
+   used by IFUNC.  */
+# define libc_hidden_builtin_def(name) \
+	.globl __GI_STRCSPN; __GI_STRCSPN = STRCSPN_IA32
+#endif
+
+#endif /* HAVE_SSE4_SUPPORT */
+
+#ifdef USE_AS_STRPBRK
+#include "../../strpbrk.S"
+#else
+#include "../../strcspn.S"
+#endif
diff --git a/sysdeps/i386/i686/multiarch/strlen.S b/sysdeps/i386/i686/multiarch/strlen.S
new file mode 100644
index 0000000000..0c1e8646ff
--- /dev/null
+++ b/sysdeps/i386/i686/multiarch/strlen.S
@@ -0,0 +1,154 @@
+/* Multiple versions of strlen
+   Copyright (C) 2009 Free Software Foundation, Inc.
+   Contributed by Intel Corporation.
+   This file is part of the GNU C Library.
+
+   The GNU C Library is free software; you can redistribute it and/or
+   modify it under the terms of the GNU Lesser General Public
+   License as published by the Free Software Foundation; either
+   version 2.1 of the License, or (at your option) any later version.
+
+   The GNU C Library 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
+   Lesser General Public License for more details.
+
+   You should have received a copy of the GNU Lesser General Public
+   License along with the GNU C Library; if not, write to the Free
+   Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
+   02111-1307 USA.  */
+
+#include <sysdep.h>
+#include <ifunc-defines.h>
+
+/* Define multiple versions only for the definition in libc and for the
+   DSO.  In static binaries, we need strlen before the initialization
+   happened.  */
+#if defined SHARED && !defined NOT_IN_libc
+	.section	.gnu.linkonce.t.__i686.get_pc_thunk.bx,"ax",@progbits
+	.globl	__i686.get_pc_thunk.bx
+	.hidden	__i686.get_pc_thunk.bx
+	.p2align 4
+	.type	__i686.get_pc_thunk.bx,@function
+__i686.get_pc_thunk.bx:
+	movl	(%esp), %ebx
+	ret
+
+	.text
+ENTRY(strlen)
+	.type	strlen, @gnu_indirect_function
+	pushl	%ebx
+	cfi_adjust_cfa_offset (4)
+	cfi_rel_offset (ebx, 0)
+	call	__i686.get_pc_thunk.bx
+	addl	$_GLOBAL_OFFSET_TABLE_, %ebx
+	cmpl	$0, KIND_OFFSET+__cpu_features@GOTOFF(%ebx)
+	jne	1f
+	call	__init_cpu_features
+1:	leal	__strlen_ia32@GOTOFF(%ebx), %eax
+	testl	$(1<<26), CPUID_OFFSET+COMMON_CPUID_INDEX_1*CPUID_SIZE+CPUID_EDX_OFFSET+__cpu_features@GOTOFF(%ebx)
+	jz	2f
+	leal	__strlen_sse2@GOTOFF(%ebx), %eax
+2:	popl	%ebx
+	cfi_adjust_cfa_offset (-4);
+	cfi_restore (ebx)
+	ret
+END(strlen)
+
+#define CFI_POP(REG)						\
+  cfi_adjust_cfa_offset (-4);					\
+  cfi_restore (REG)
+
+#define RETURN		popl %esi; CFI_POP (esi); ret
+
+	.text
+ENTRY (__strlen_sse2)
+/*
+ * This implementation uses SSE instructions to compare up to 16 bytes
+ * at a time looking for the end of string (null char).
+ */
+	pushl	%esi
+	cfi_adjust_cfa_offset (4)
+	cfi_rel_offset (%esi, 0)
+	mov	8(%esp), %eax
+	mov	%eax, %ecx
+	pxor	%xmm0, %xmm0		/* 16 null chars */
+	mov	%eax, %esi
+	and	$15, %ecx
+	jz	1f			/* string is 16 byte aligned */
+
+	/*
+	* Unaligned case. Round down to 16-byte boundary before comparing
+	* 16 bytes for a null char. The code then compensates for any extra chars
+	* preceding the start of the string.
+	*/
+	and	$-16, %esi
+
+	pcmpeqb	(%esi), %xmm0
+	lea	16(%eax), %esi
+	pmovmskb %xmm0, %edx
+
+	shr	%cl, %edx		/* Compensate for bytes preceding the string */
+	test	%edx, %edx
+	jnz	2f
+	sub	%ecx, %esi		/* no null, adjust to next 16-byte boundary */
+	pxor	%xmm0, %xmm0		/* clear xmm0, may have been changed... */
+
+	.p2align 4
+1:					/* 16 byte aligned */
+	pcmpeqb	(%esi), %xmm0		/* look for null bytes */
+	pmovmskb %xmm0, %edx		/* move each byte mask of %xmm0 to edx */
+
+	add	$16, %esi		/* prepare to search next 16 bytes */
+	test	%edx, %edx		/* if no null byte, %edx must be 0 */
+	jnz	2f			/* found a null */
+
+	pcmpeqb	(%esi), %xmm0
+	pmovmskb %xmm0, %edx
+	add	$16, %esi
+	test	%edx, %edx
+	jnz	2f
+
+	pcmpeqb	(%esi), %xmm0
+	pmovmskb %xmm0, %edx
+	add	$16, %esi
+	test	%edx, %edx
+	jnz	2f
+
+	pcmpeqb	(%esi), %xmm0
+	pmovmskb %xmm0, %edx
+	add	$16, %esi
+	test	%edx, %edx
+	jz	1b
+
+2:
+	neg	%eax
+	lea	-16(%eax, %esi), %eax	/* calculate exact offset */
+	bsf	%edx, %ecx		/* Least significant 1 bit is index of null */
+	add	%ecx, %eax
+	popl	%esi
+	cfi_adjust_cfa_offset (-4)
+	cfi_restore (%esi)
+	ret
+
+END (__strlen_sse2)
+
+# undef ENTRY
+# define ENTRY(name) \
+	.type __strlen_ia32, @function; \
+	.globl __strlen_ia32; \
+	.p2align 4; \
+	__strlen_ia32: cfi_startproc; \
+	CALL_MCOUNT
+# undef END
+# define END(name) \
+	cfi_endproc; .size __strlen_ia32, .-__strlen_ia32
+# undef libc_hidden_builtin_def
+/* IFUNC doesn't work with the hidden functions in shared library since
+   they will be called without setting up EBX needed for PLT which is
+   used by IFUNC.  */
+# define libc_hidden_builtin_def(name) \
+	.globl __GI_strlen; __GI_strlen = __strlen_ia32
+#endif
+
+#include "../../i586/strlen.S"
diff --git a/sysdeps/i386/i686/multiarch/strpbrk-c.c b/sysdeps/i386/i686/multiarch/strpbrk-c.c
new file mode 100644
index 0000000000..5db62053b3
--- /dev/null
+++ b/sysdeps/i386/i686/multiarch/strpbrk-c.c
@@ -0,0 +1,2 @@
+#define __strpbrk_sse2 __strpbrk_ia32
+#include <sysdeps/x86_64/multiarch/strpbrk-c.c>
diff --git a/sysdeps/i386/i686/multiarch/strpbrk.S b/sysdeps/i386/i686/multiarch/strpbrk.S
new file mode 100644
index 0000000000..ed5bca6a94
--- /dev/null
+++ b/sysdeps/i386/i686/multiarch/strpbrk.S
@@ -0,0 +1,3 @@
+#define STRCSPN strpbrk
+#define USE_AS_STRPBRK
+#include "strcspn.S"
diff --git a/sysdeps/i386/i686/multiarch/strspn-c.c b/sysdeps/i386/i686/multiarch/strspn-c.c
new file mode 100644
index 0000000000..bea09dea71
--- /dev/null
+++ b/sysdeps/i386/i686/multiarch/strspn-c.c
@@ -0,0 +1,2 @@
+#define __strspn_sse2 __strspn_ia32
+#include <sysdeps/x86_64/multiarch/strspn-c.c>
diff --git a/sysdeps/i386/i686/multiarch/strspn.S b/sysdeps/i386/i686/multiarch/strspn.S
new file mode 100644
index 0000000000..f306d2d1fb
--- /dev/null
+++ b/sysdeps/i386/i686/multiarch/strspn.S
@@ -0,0 +1,95 @@
+/* Multiple versions of strspn
+   Copyright (C) 2009 Free Software Foundation, Inc.
+   Contributed by Intel Corporation.
+   This file is part of the GNU C Library.
+
+   The GNU C Library is free software; you can redistribute it and/or
+   modify it under the terms of the GNU Lesser General Public
+   License as published by the Free Software Foundation; either
+   version 2.1 of the License, or (at your option) any later version.
+
+   The GNU C Library 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
+   Lesser General Public License for more details.
+
+   You should have received a copy of the GNU Lesser General Public
+   License along with the GNU C Library; if not, write to the Free
+   Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
+   02111-1307 USA.  */
+
+#include <config.h>
+
+#ifdef HAVE_SSE4_SUPPORT
+
+#include <sysdep.h>
+#include <ifunc-defines.h>
+
+/* Define multiple versions only for the definition in libc.  */
+#ifndef NOT_IN_libc
+# ifdef SHARED
+	.section	.gnu.linkonce.t.__i686.get_pc_thunk.bx,"ax",@progbits
+	.globl	__i686.get_pc_thunk.bx
+	.hidden	__i686.get_pc_thunk.bx
+	.p2align 4
+	.type	__i686.get_pc_thunk.bx,@function
+__i686.get_pc_thunk.bx:
+	movl	(%esp), %ebx
+	ret
+
+	.text
+ENTRY(strspn)
+	.type	strspn, @gnu_indirect_function
+	pushl	%ebx
+	cfi_adjust_cfa_offset (4)
+	cfi_rel_offset (ebx, 0)
+	call	__i686.get_pc_thunk.bx
+	addl	$_GLOBAL_OFFSET_TABLE_, %ebx
+	cmpl	$0, KIND_OFFSET+__cpu_features@GOTOFF(%ebx)
+	jne	1f
+	call	__init_cpu_features
+1:	leal	__strspn_ia32@GOTOFF(%ebx), %eax
+	testl	$(1<<20), CPUID_OFFSET+COMMON_CPUID_INDEX_1*CPUID_SIZE+CPUID_ECX_OFFSET+__cpu_features@GOTOFF(%ebx)
+	jz	2f
+	leal	__strspn_sse42@GOTOFF(%ebx), %eax
+2:	popl	%ebx
+	cfi_adjust_cfa_offset (-4);
+	cfi_restore (ebx)
+	ret
+END(strspn)
+# else
+	.text
+ENTRY(strspn)
+	.type	strspn, @gnu_indirect_function
+	cmpl	$0, KIND_OFFSET+__cpu_features
+	jne	1f
+	call	__init_cpu_features
+1:	leal	__strspn_ia32, %eax
+	testl	$(1<<20), CPUID_OFFSET+COMMON_CPUID_INDEX_1*CPUID_SIZE+CPUID_ECX_OFFSET+__cpu_features
+	jz	2f
+	leal	__strspn_sse42, %eax
+2:	ret
+END(strspn)
+# endif
+
+# undef ENTRY
+# define ENTRY(name) \
+	.type __strspn_ia32, @function; \
+	.globl __strspn_ia32; \
+	.p2align 4
+	__strspn_ia32: cfi_startproc; \
+	CALL_MCOUNT
+# undef END
+# define END(name) \
+	cfi_endproc; .size __strspn_ia32, .-__strspn_ia32
+# undef libc_hidden_builtin_def
+/* IFUNC doesn't work with the hidden functions in shared library since
+   they will be called without setting up EBX needed for PLT which is
+   used by IFUNC.  */
+# define libc_hidden_builtin_def(name) \
+	.globl __GI_strspn; __GI_strspn = __strspn_ia32
+#endif
+
+#endif /* HAVE_SSE4_SUPPORT */
+
+#include "../../strspn.S"
diff --git a/sysdeps/i386/i686/multiarch/strstr-c.c b/sysdeps/i386/i686/multiarch/strstr-c.c
new file mode 100644
index 0000000000..7ef1157ce4
--- /dev/null
+++ b/sysdeps/i386/i686/multiarch/strstr-c.c
@@ -0,0 +1,12 @@
+#include "init-arch.h"
+
+#define STRSTR __strstr_ia32
+#undef libc_hidden_builtin_def
+#define libc_hidden_builtin_def(name) \
+  __hidden_ver1 (__strstr_ia32, __GI_strstr, __strstr_ia32);
+
+#include "string/strstr.c"
+
+extern char *__strstr_sse42 (const char *, const char *);
+
+libc_ifunc (strstr, HAS_SSE4_2 ? __strstr_sse42 : __strstr_ia32);
diff --git a/sysdeps/i386/i686/multiarch/strstr.c b/sysdeps/i386/i686/multiarch/strstr.c
new file mode 100644
index 0000000000..a97428c125
--- /dev/null
+++ b/sysdeps/i386/i686/multiarch/strstr.c
@@ -0,0 +1 @@
+#include <sysdeps/x86_64/multiarch/strstr.c>
diff --git a/sysdeps/x86_64/cacheinfo.c b/sysdeps/x86_64/cacheinfo.c
index f252fc2c6c..5b66c62eb3 100644
--- a/sysdeps/x86_64/cacheinfo.c
+++ b/sysdeps/x86_64/cacheinfo.c
@@ -516,13 +516,15 @@ init_cacheinfo (void)
           shared = handle_intel (_SC_LEVEL2_CACHE_SIZE, max_cpuid);
 	}
 
+      unsigned int ebx_1;
+
 #ifdef USE_MULTIARCH
       eax = __cpu_features.cpuid[COMMON_CPUID_INDEX_1].eax;
-      ebx = __cpu_features.cpuid[COMMON_CPUID_INDEX_1].ebx;
+      ebx_1 = __cpu_features.cpuid[COMMON_CPUID_INDEX_1].ebx;
       ecx = __cpu_features.cpuid[COMMON_CPUID_INDEX_1].ecx;
       edx = __cpu_features.cpuid[COMMON_CPUID_INDEX_1].edx;
 #else
-      __cpuid (1, eax, ebx, ecx, edx);
+      __cpuid (1, eax, ebx_1, ecx, edx);
 #endif
 
 #ifndef DISABLE_PREFERRED_MEMORY_INSTRUCTION
@@ -554,14 +556,46 @@ init_cacheinfo (void)
 	    }
           while (((eax >> 5) & 0x7) != level);
 
-	  threads = ((eax >> 14) & 0x3ff) + 1;
+	  threads = (eax >> 14) & 0x3ff;
+
+	  /* If max_cpuid >= 11, THREADS is the maximum number of
+	      addressable IDs for logical processors sharing the
+	      cache, instead of the maximum number of threads
+	      sharing the cache.  */
+	  if (threads && max_cpuid >= 11)
+	    {
+	      /* Find the number of logical processors shipped in
+		 one core and apply count mask.  */
+	      i = 0;
+	      while (1)
+		{
+		  __cpuid_count (11, i++, eax, ebx, ecx, edx);
+
+		  int shipped = ebx & 0xff;
+		  int type = ecx & 0xff0;
+		  if (shipped == 0 || type == 0)
+		    break;
+		  else if (type == 0x200)
+		    {
+		      int count_mask;
+
+		      /* Compute count mask.  */
+		      asm ("bsr %1, %0"
+			   : "=r" (count_mask) : "g" (threads));
+		      count_mask = ~(-1 << (count_mask + 1));
+		      threads = (shipped - 1) & count_mask;
+		      break;
+		    }
+		}
+	    }
+	  threads += 1;
 	}
       else
         {
 	intel_bug_no_cache_info:
 	  /* Assume that all logical threads share the highest cache level.  */
 
-	  threads = (ebx >> 16) & 0xff;
+	  threads = (ebx_1 >> 16) & 0xff;
 	}
 
       /* Cap usage of highest cache level to the number of supported
diff --git a/sysdeps/x86_64/dl-trampoline.S b/sysdeps/x86_64/dl-trampoline.S
index 20da6956f1..f9c60ad5cf 100644
--- a/sysdeps/x86_64/dl-trampoline.S
+++ b/sysdeps/x86_64/dl-trampoline.S
@@ -146,247 +146,17 @@ L(have_avx):
 2:	movl	%eax, L(have_avx)(%rip)
 	cmpl	$0, %eax
 
-1:	js	L(no_avx1)
+1:	js	L(no_avx)
 
-	/* This is to support AVX audit modules.  */
-	vmovdqu %ymm0,		      (LR_VECTOR_OFFSET)(%rsp)
-	vmovdqu %ymm1, (LR_VECTOR_OFFSET +   VECTOR_SIZE)(%rsp)
-	vmovdqu %ymm2, (LR_VECTOR_OFFSET + VECTOR_SIZE*2)(%rsp)
-	vmovdqu %ymm3, (LR_VECTOR_OFFSET + VECTOR_SIZE*3)(%rsp)
-	vmovdqu %ymm4, (LR_VECTOR_OFFSET + VECTOR_SIZE*4)(%rsp)
-	vmovdqu %ymm5, (LR_VECTOR_OFFSET + VECTOR_SIZE*5)(%rsp)
-	vmovdqu %ymm6, (LR_VECTOR_OFFSET + VECTOR_SIZE*6)(%rsp)
-	vmovdqu %ymm7, (LR_VECTOR_OFFSET + VECTOR_SIZE*7)(%rsp)
+#  define RESTORE_AVX
+#  include "dl-trampoline.h"
 
-	/* Save xmm0-xmm7 registers to detect if any of them are
-	   changed by audit module.  */
-	vmovdqa %xmm0,		    (LR_SIZE)(%rsp)
-	vmovdqa %xmm1, (LR_SIZE +   XMM_SIZE)(%rsp)
-	vmovdqa %xmm2, (LR_SIZE + XMM_SIZE*2)(%rsp)
-	vmovdqa %xmm3, (LR_SIZE + XMM_SIZE*3)(%rsp)
-	vmovdqa %xmm4, (LR_SIZE + XMM_SIZE*4)(%rsp)
-	vmovdqa %xmm5, (LR_SIZE + XMM_SIZE*5)(%rsp)
-	vmovdqa %xmm6, (LR_SIZE + XMM_SIZE*6)(%rsp)
-	vmovdqa %xmm7, (LR_SIZE + XMM_SIZE*7)(%rsp)
-
-L(no_avx1):
-# endif
-
-	movq %rsp, %rcx		# La_x86_64_regs pointer to %rcx.
-	movq 48(%rbx), %rdx	# Load return address if needed.
-	movq 40(%rbx), %rsi	# Copy args pushed by PLT in register.
-	movq 32(%rbx), %rdi	# %rdi: link_map, %rsi: reloc_index
-	leaq 16(%rbx), %r8
-	call _dl_profile_fixup	# Call resolver.
-
-	movq %rax, %r11		# Save return value.
-
-	movq 8(%rbx), %rax	# Get back register content.
-	movq LR_RDX_OFFSET(%rsp), %rdx
-	movq  LR_R8_OFFSET(%rsp), %r8
-	movq  LR_R9_OFFSET(%rsp), %r9
-
-	movaps		    (LR_XMM_OFFSET)(%rsp), %xmm0
-	movaps	 (LR_XMM_OFFSET + XMM_SIZE)(%rsp), %xmm1
-	movaps (LR_XMM_OFFSET + XMM_SIZE*2)(%rsp), %xmm2
-	movaps (LR_XMM_OFFSET + XMM_SIZE*3)(%rsp), %xmm3
-	movaps (LR_XMM_OFFSET + XMM_SIZE*4)(%rsp), %xmm4
-	movaps (LR_XMM_OFFSET + XMM_SIZE*5)(%rsp), %xmm5
-	movaps (LR_XMM_OFFSET + XMM_SIZE*6)(%rsp), %xmm6
-	movaps (LR_XMM_OFFSET + XMM_SIZE*7)(%rsp), %xmm7
-
-# ifdef HAVE_AVX_SUPPORT
-	cmpl	$0, L(have_avx)(%rip)
-	js	L(no_avx2)
-
-	/* Check if any xmm0-xmm7 registers are changed by audit
-	   module.  */
-	vpcmpeqq (LR_SIZE)(%rsp), %xmm0, %xmm8
-	vpmovmskb %xmm8, %esi
-	cmpl $0xffff, %esi
-	jne 1f
-	vmovdqu			(LR_VECTOR_OFFSET)(%rsp), %ymm0
-
-1:	vpcmpeqq (LR_SIZE + XMM_SIZE)(%rsp), %xmm1, %xmm8
-	vpmovmskb %xmm8, %esi
-	cmpl $0xffff, %esi
-	jne 1f
-	vmovdqu	  (LR_VECTOR_OFFSET + VECTOR_SIZE)(%rsp), %ymm1
-
-1:	vpcmpeqq (LR_SIZE + XMM_SIZE*2)(%rsp), %xmm2, %xmm8
-	vpmovmskb %xmm8, %esi
-	cmpl $0xffff, %esi
-	jne 1f
-	vmovdqu (LR_VECTOR_OFFSET + VECTOR_SIZE*2)(%rsp), %ymm2
-
-1:	vpcmpeqq (LR_SIZE + XMM_SIZE*3)(%rsp), %xmm3, %xmm8
-	vpmovmskb %xmm8, %esi
-	cmpl $0xffff, %esi
-	jne 1f
-	vmovdqu (LR_VECTOR_OFFSET + VECTOR_SIZE*3)(%rsp), %ymm3
-
-1:	vpcmpeqq (LR_SIZE + XMM_SIZE*4)(%rsp), %xmm4, %xmm8
-	vpmovmskb %xmm8, %esi
-	cmpl $0xffff, %esi
-	jne 1f
-	vmovdqu (LR_VECTOR_OFFSET + VECTOR_SIZE*4)(%rsp), %ymm4
-
-1:	vpcmpeqq (LR_SIZE + XMM_SIZE*5)(%rsp), %xmm5, %xmm8
-	vpmovmskb %xmm8, %esi
-	cmpl $0xffff, %esi
-	jne 1f
-	vmovdqu (LR_VECTOR_OFFSET + VECTOR_SIZE*5)(%rsp), %ymm5
-
-1:	vpcmpeqq (LR_SIZE + XMM_SIZE*6)(%rsp), %xmm6, %xmm8
-	vpmovmskb %xmm8, %esi
-	cmpl $0xffff, %esi
-	jne 1f
-	vmovdqu (LR_VECTOR_OFFSET + VECTOR_SIZE*6)(%rsp), %ymm6
-
-1:	vpcmpeqq (LR_SIZE + XMM_SIZE*7)(%rsp), %xmm7, %xmm8
-	vpmovmskb %xmm8, %esi
-	cmpl $0xffff, %esi
-	jne 1f
-	vmovdqu (LR_VECTOR_OFFSET + VECTOR_SIZE*7)(%rsp), %ymm7
-
-L(no_avx2):
-1:
-# endif
-	movq 16(%rbx), %r10	# Anything in framesize?
-	testq %r10, %r10
-	jns 3f
-
-	/* There's nothing in the frame size, so there
-	   will be no call to the _dl_call_pltexit. */
-
-	/* Get back registers content.  */
-	movq LR_RCX_OFFSET(%rsp), %rcx
-	movq LR_RSI_OFFSET(%rsp), %rsi
-	movq LR_RDI_OFFSET(%rsp), %rdi
-
-	movq %rbx, %rsp
-	movq (%rsp), %rbx
-	cfi_restore(rbx)
-	cfi_def_cfa_register(%rsp)
-
-	addq $48, %rsp		# Adjust the stack to the return value
-				# (eats the reloc index and link_map)
-	cfi_adjust_cfa_offset(-48)
-	jmp *%r11		# Jump to function address.
-
-3:
-	cfi_adjust_cfa_offset(48)
-	cfi_rel_offset(%rbx, 0)
-	cfi_def_cfa_register(%rbx)
-
-	/* At this point we need to prepare new stack for the function
-	   which has to be called.  We copy the original stack to a
-	   temporary buffer of the size specified by the 'framesize'
-	   returned from _dl_profile_fixup */
-
-	leaq LR_RSP_OFFSET(%rbx), %rsi	# stack
-	addq $8, %r10
-	andq $0xfffffffffffffff0, %r10
-	movq %r10, %rcx
-	subq %r10, %rsp
-	movq %rsp, %rdi
-	shrq $3, %rcx
-	rep
-	movsq
-
-	movq 24(%rdi), %rcx	# Get back register content.
-	movq 32(%rdi), %rsi
-	movq 40(%rdi), %rdi
-
-	call *%r11
-
-	mov 24(%rbx), %rsp	# Drop the copied stack content
-
-	/* Now we have to prepare the La_x86_64_retval structure for the
-	   _dl_call_pltexit.  The La_x86_64_regs is being pointed by rsp now,
-	   so we just need to allocate the sizeof(La_x86_64_retval) space on
-	   the stack, since the alignment has already been taken care of. */
-# ifdef HAVE_AVX_SUPPORT
-	/* sizeof(La_x86_64_retval).  Need extra space for 2 SSE
-	   registers to detect if xmm0/xmm1 registers are changed
-	   by audit module.  */
-	subq $(LRV_SIZE + XMM_SIZE*2), %rsp
-# else
-	subq $LRV_SIZE, %rsp	# sizeof(La_x86_64_retval)
-# endif
-	movq %rsp, %rcx		# La_x86_64_retval argument to %rcx.
-
-	/* Fill in the La_x86_64_retval structure.  */
-	movq %rax, LRV_RAX_OFFSET(%rcx)
-	movq %rdx, LRV_RDX_OFFSET(%rcx)
-
-	movaps %xmm0, LRV_XMM0_OFFSET(%rcx)
-	movaps %xmm1, LRV_XMM1_OFFSET(%rcx)
-
-# ifdef HAVE_AVX_SUPPORT
-	cmpl	$0, L(have_avx)(%rip)
-	js	L(no_avx3)
-
-	/* This is to support AVX audit modules.  */
-	vmovdqu %ymm0, LRV_VECTOR0_OFFSET(%rcx)
-	vmovdqu %ymm1, LRV_VECTOR1_OFFSET(%rcx)
-
-	/* Save xmm0/xmm1 registers to detect if they are changed
-	   by audit module.  */
-	vmovdqa %xmm0,		  (LRV_SIZE)(%rcx)
-	vmovdqa %xmm1, (LRV_SIZE + XMM_SIZE)(%rcx)
-
-L(no_avx3):
-# endif
-
-	fstpt LRV_ST0_OFFSET(%rcx)
-	fstpt LRV_ST1_OFFSET(%rcx)
-
-	movq 24(%rbx), %rdx	# La_x86_64_regs argument to %rdx.
-	movq 40(%rbx), %rsi	# Copy args pushed by PLT in register.
-        movq 32(%rbx), %rdi	# %rdi: link_map, %rsi: reloc_index
-	call _dl_call_pltexit
-
-	/* Restore return registers.  */
-	movq LRV_RAX_OFFSET(%rsp), %rax
-	movq LRV_RDX_OFFSET(%rsp), %rdx
-
-	movaps LRV_XMM0_OFFSET(%rsp), %xmm0
-	movaps LRV_XMM1_OFFSET(%rsp), %xmm1
-
-# ifdef HAVE_AVX_SUPPORT
-	cmpl	$0, L(have_avx)(%rip)
-	js	L(no_avx4)
-
-	/* Check if xmm0/xmm1 registers are changed by audit module.  */
-	vpcmpeqq (LRV_SIZE)(%rsp), %xmm0, %xmm2
-	vpmovmskb %xmm2, %esi
-	cmpl $0xffff, %esi
-	jne 1f
-	vmovdqu LRV_VECTOR0_OFFSET(%rsp), %ymm0
-
-1:	vpcmpeqq (LRV_SIZE + XMM_SIZE)(%rsp), %xmm1, %xmm2
-	vpmovmskb %xmm2, %esi
-	cmpl $0xffff, %esi
-	jne 1f
-	vmovdqu LRV_VECTOR1_OFFSET(%rsp), %ymm1
-
-L(no_avx4):
-1:
+	.align 16
+L(no_avx):
 # endif
 
-	fldt LRV_ST1_OFFSET(%rsp)
-	fldt LRV_ST0_OFFSET(%rsp)
-
-	movq %rbx, %rsp
-	movq (%rsp), %rbx
-	cfi_restore(rbx)
-	cfi_def_cfa_register(%rsp)
-
-	addq $48, %rsp		# Adjust the stack to the return value
-				# (eats the reloc index and link_map)
-	cfi_adjust_cfa_offset(-48)
-	retq
+#  undef RESTORE_AVX
+#  include "dl-trampoline.h"
 
 	cfi_endproc
 	.size _dl_runtime_profile, .-_dl_runtime_profile
diff --git a/sysdeps/x86_64/dl-trampoline.h b/sysdeps/x86_64/dl-trampoline.h
new file mode 100644
index 0000000000..5d49ed4408
--- /dev/null
+++ b/sysdeps/x86_64/dl-trampoline.h
@@ -0,0 +1,269 @@
+/* Partial PLT profile trampoline to save and restore x86-64 vector
+   registers.
+   Copyright (C) 2009 Free Software Foundation, Inc.
+   This file is part of the GNU C Library.
+
+   The GNU C Library is free software; you can redistribute it and/or
+   modify it under the terms of the GNU Lesser General Public
+   License as published by the Free Software Foundation; either
+   version 2.1 of the License, or (at your option) any later version.
+
+   The GNU C Library 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
+   Lesser General Public License for more details.
+
+   You should have received a copy of the GNU Lesser General Public
+   License along with the GNU C Library; if not, write to the Free
+   Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
+   02111-1307 USA.  */
+
+#ifdef RESTORE_AVX
+	/* This is to support AVX audit modules.  */
+	vmovdqu %ymm0,		      (LR_VECTOR_OFFSET)(%rsp)
+	vmovdqu %ymm1, (LR_VECTOR_OFFSET +   VECTOR_SIZE)(%rsp)
+	vmovdqu %ymm2, (LR_VECTOR_OFFSET + VECTOR_SIZE*2)(%rsp)
+	vmovdqu %ymm3, (LR_VECTOR_OFFSET + VECTOR_SIZE*3)(%rsp)
+	vmovdqu %ymm4, (LR_VECTOR_OFFSET + VECTOR_SIZE*4)(%rsp)
+	vmovdqu %ymm5, (LR_VECTOR_OFFSET + VECTOR_SIZE*5)(%rsp)
+	vmovdqu %ymm6, (LR_VECTOR_OFFSET + VECTOR_SIZE*6)(%rsp)
+	vmovdqu %ymm7, (LR_VECTOR_OFFSET + VECTOR_SIZE*7)(%rsp)
+
+	/* Save xmm0-xmm7 registers to detect if any of them are
+	   changed by audit module.  */
+	vmovdqa %xmm0,		    (LR_SIZE)(%rsp)
+	vmovdqa %xmm1, (LR_SIZE +   XMM_SIZE)(%rsp)
+	vmovdqa %xmm2, (LR_SIZE + XMM_SIZE*2)(%rsp)
+	vmovdqa %xmm3, (LR_SIZE + XMM_SIZE*3)(%rsp)
+	vmovdqa %xmm4, (LR_SIZE + XMM_SIZE*4)(%rsp)
+	vmovdqa %xmm5, (LR_SIZE + XMM_SIZE*5)(%rsp)
+	vmovdqa %xmm6, (LR_SIZE + XMM_SIZE*6)(%rsp)
+	vmovdqa %xmm7, (LR_SIZE + XMM_SIZE*7)(%rsp)
+#endif
+
+	movq %rsp, %rcx		# La_x86_64_regs pointer to %rcx.
+	movq 48(%rbx), %rdx	# Load return address if needed.
+	movq 40(%rbx), %rsi	# Copy args pushed by PLT in register.
+	movq 32(%rbx), %rdi	# %rdi: link_map, %rsi: reloc_index
+	leaq 16(%rbx), %r8
+	call _dl_profile_fixup	# Call resolver.
+
+	movq %rax, %r11		# Save return value.
+
+	movq 8(%rbx), %rax	# Get back register content.
+	movq LR_RDX_OFFSET(%rsp), %rdx
+	movq  LR_R8_OFFSET(%rsp), %r8
+	movq  LR_R9_OFFSET(%rsp), %r9
+
+	movaps		    (LR_XMM_OFFSET)(%rsp), %xmm0
+	movaps	 (LR_XMM_OFFSET + XMM_SIZE)(%rsp), %xmm1
+	movaps (LR_XMM_OFFSET + XMM_SIZE*2)(%rsp), %xmm2
+	movaps (LR_XMM_OFFSET + XMM_SIZE*3)(%rsp), %xmm3
+	movaps (LR_XMM_OFFSET + XMM_SIZE*4)(%rsp), %xmm4
+	movaps (LR_XMM_OFFSET + XMM_SIZE*5)(%rsp), %xmm5
+	movaps (LR_XMM_OFFSET + XMM_SIZE*6)(%rsp), %xmm6
+	movaps (LR_XMM_OFFSET + XMM_SIZE*7)(%rsp), %xmm7
+
+#ifdef RESTORE_AVX
+	/* Check if any xmm0-xmm7 registers are changed by audit
+	   module.  */
+	vpcmpeqq (LR_SIZE)(%rsp), %xmm0, %xmm8
+	vpmovmskb %xmm8, %esi
+	cmpl $0xffff, %esi
+	je 2f
+	vmovdqa	%xmm0, (LR_VECTOR_OFFSET)(%rsp)
+	jmp 1f
+2:	vmovdqu	(LR_VECTOR_OFFSET)(%rsp), %ymm0
+	vmovdqa	%xmm0, (LR_XMM_OFFSET)(%rsp)
+
+1:	vpcmpeqq (LR_SIZE + XMM_SIZE)(%rsp), %xmm1, %xmm8
+	vpmovmskb %xmm8, %esi
+	cmpl $0xffff, %esi
+	je 2f
+	vmovdqa	%xmm1, (LR_VECTOR_OFFSET + VECTOR_SIZE)(%rsp)
+	jmp 1f
+2:	vmovdqu	(LR_VECTOR_OFFSET + VECTOR_SIZE)(%rsp), %ymm1
+	vmovdqa	%xmm1, (LR_XMM_OFFSET + XMM_SIZE)(%rsp)
+
+1:	vpcmpeqq (LR_SIZE + XMM_SIZE*2)(%rsp), %xmm2, %xmm8
+	vpmovmskb %xmm8, %esi
+	cmpl $0xffff, %esi
+	je 2f
+	vmovdqa	%xmm2, (LR_VECTOR_OFFSET + VECTOR_SIZE*2)(%rsp)
+	jmp 1f
+2:	vmovdqu (LR_VECTOR_OFFSET + VECTOR_SIZE*2)(%rsp), %ymm2
+	vmovdqa	%xmm2, (LR_XMM_OFFSET + XMM_SIZE*2)(%rsp)
+
+1:	vpcmpeqq (LR_SIZE + XMM_SIZE*3)(%rsp), %xmm3, %xmm8
+	vpmovmskb %xmm8, %esi
+	cmpl $0xffff, %esi
+	je 2f
+	vmovdqa	%xmm3, (LR_VECTOR_OFFSET + VECTOR_SIZE*3)(%rsp)
+	jmp 1f
+2:	vmovdqu (LR_VECTOR_OFFSET + VECTOR_SIZE*3)(%rsp), %ymm3
+	vmovdqa	%xmm3, (LR_XMM_OFFSET + XMM_SIZE*3)(%rsp)
+
+1:	vpcmpeqq (LR_SIZE + XMM_SIZE*4)(%rsp), %xmm4, %xmm8
+	vpmovmskb %xmm8, %esi
+	cmpl $0xffff, %esi
+	je 2f
+	vmovdqa	%xmm4, (LR_VECTOR_OFFSET + VECTOR_SIZE*4)(%rsp)
+	jmp 1f
+2:	vmovdqu (LR_VECTOR_OFFSET + VECTOR_SIZE*4)(%rsp), %ymm4
+	vmovdqa	%xmm4, (LR_XMM_OFFSET + XMM_SIZE*4)(%rsp)
+
+1:	vpcmpeqq (LR_SIZE + XMM_SIZE*5)(%rsp), %xmm5, %xmm8
+	vpmovmskb %xmm8, %esi
+	cmpl $0xffff, %esi
+	je 2f
+	vmovdqa	%xmm5, (LR_VECTOR_OFFSET + VECTOR_SIZE*5)(%rsp)
+	jmp 1f
+2:	vmovdqu (LR_VECTOR_OFFSET + VECTOR_SIZE*5)(%rsp), %ymm5
+	vmovdqa	%xmm5, (LR_XMM_OFFSET + XMM_SIZE*5)(%rsp)
+
+1:	vpcmpeqq (LR_SIZE + XMM_SIZE*6)(%rsp), %xmm6, %xmm8
+	vpmovmskb %xmm8, %esi
+	cmpl $0xffff, %esi
+	je 2f
+	vmovdqa	%xmm6, (LR_VECTOR_OFFSET + VECTOR_SIZE*6)(%rsp)
+	jmp 1f
+2:	vmovdqu (LR_VECTOR_OFFSET + VECTOR_SIZE*6)(%rsp), %ymm6
+	vmovdqa	%xmm6, (LR_XMM_OFFSET + XMM_SIZE*6)(%rsp)
+
+1:	vpcmpeqq (LR_SIZE + XMM_SIZE*7)(%rsp), %xmm7, %xmm8
+	vpmovmskb %xmm8, %esi
+	cmpl $0xffff, %esi
+	je 2f
+	vmovdqa	%xmm7, (LR_VECTOR_OFFSET + VECTOR_SIZE*7)(%rsp)
+	jmp 1f
+2:	vmovdqu (LR_VECTOR_OFFSET + VECTOR_SIZE*7)(%rsp), %ymm7
+	vmovdqa	%xmm7, (LR_XMM_OFFSET + XMM_SIZE*7)(%rsp)
+
+1:
+#endif
+	movq 16(%rbx), %r10	# Anything in framesize?
+	testq %r10, %r10
+	jns 3f
+
+	/* There's nothing in the frame size, so there
+	   will be no call to the _dl_call_pltexit. */
+
+	/* Get back registers content.  */
+	movq LR_RCX_OFFSET(%rsp), %rcx
+	movq LR_RSI_OFFSET(%rsp), %rsi
+	movq LR_RDI_OFFSET(%rsp), %rdi
+
+	movq %rbx, %rsp
+	movq (%rsp), %rbx
+	cfi_restore(rbx)
+	cfi_def_cfa_register(%rsp)
+
+	addq $48, %rsp		# Adjust the stack to the return value
+				# (eats the reloc index and link_map)
+	cfi_adjust_cfa_offset(-48)
+	jmp *%r11		# Jump to function address.
+
+3:
+	cfi_adjust_cfa_offset(48)
+	cfi_rel_offset(%rbx, 0)
+	cfi_def_cfa_register(%rbx)
+
+	/* At this point we need to prepare new stack for the function
+	   which has to be called.  We copy the original stack to a
+	   temporary buffer of the size specified by the 'framesize'
+	   returned from _dl_profile_fixup */
+
+	leaq LR_RSP_OFFSET(%rbx), %rsi	# stack
+	addq $8, %r10
+	andq $0xfffffffffffffff0, %r10
+	movq %r10, %rcx
+	subq %r10, %rsp
+	movq %rsp, %rdi
+	shrq $3, %rcx
+	rep
+	movsq
+
+	movq 24(%rdi), %rcx	# Get back register content.
+	movq 32(%rdi), %rsi
+	movq 40(%rdi), %rdi
+
+	call *%r11
+
+	mov 24(%rbx), %rsp	# Drop the copied stack content
+
+	/* Now we have to prepare the La_x86_64_retval structure for the
+	   _dl_call_pltexit.  The La_x86_64_regs is being pointed by rsp now,
+	   so we just need to allocate the sizeof(La_x86_64_retval) space on
+	   the stack, since the alignment has already been taken care of. */
+# ifdef RESTORE_AVX
+	/* sizeof(La_x86_64_retval).  Need extra space for 2 SSE
+	   registers to detect if xmm0/xmm1 registers are changed
+	   by audit module.  */
+	subq $(LRV_SIZE + XMM_SIZE*2), %rsp
+# else
+	subq $LRV_SIZE, %rsp	# sizeof(La_x86_64_retval)
+# endif
+	movq %rsp, %rcx		# La_x86_64_retval argument to %rcx.
+
+	/* Fill in the La_x86_64_retval structure.  */
+	movq %rax, LRV_RAX_OFFSET(%rcx)
+	movq %rdx, LRV_RDX_OFFSET(%rcx)
+
+	movaps %xmm0, LRV_XMM0_OFFSET(%rcx)
+	movaps %xmm1, LRV_XMM1_OFFSET(%rcx)
+
+# ifdef RESTORE_AVX
+	/* This is to support AVX audit modules.  */
+	vmovdqu %ymm0, LRV_VECTOR0_OFFSET(%rcx)
+	vmovdqu %ymm1, LRV_VECTOR1_OFFSET(%rcx)
+
+	/* Save xmm0/xmm1 registers to detect if they are changed
+	   by audit module.  */
+	vmovdqa %xmm0,		  (LRV_SIZE)(%rcx)
+	vmovdqa %xmm1, (LRV_SIZE + XMM_SIZE)(%rcx)
+# endif
+
+	fstpt LRV_ST0_OFFSET(%rcx)
+	fstpt LRV_ST1_OFFSET(%rcx)
+
+	movq 24(%rbx), %rdx	# La_x86_64_regs argument to %rdx.
+	movq 40(%rbx), %rsi	# Copy args pushed by PLT in register.
+        movq 32(%rbx), %rdi	# %rdi: link_map, %rsi: reloc_index
+	call _dl_call_pltexit
+
+	/* Restore return registers.  */
+	movq LRV_RAX_OFFSET(%rsp), %rax
+	movq LRV_RDX_OFFSET(%rsp), %rdx
+
+	movaps LRV_XMM0_OFFSET(%rsp), %xmm0
+	movaps LRV_XMM1_OFFSET(%rsp), %xmm1
+
+# ifdef RESTORE_AVX
+	/* Check if xmm0/xmm1 registers are changed by audit module.  */
+	vpcmpeqq (LRV_SIZE)(%rsp), %xmm0, %xmm2
+	vpmovmskb %xmm2, %esi
+	cmpl $0xffff, %esi
+	jne 1f
+	vmovdqu LRV_VECTOR0_OFFSET(%rsp), %ymm0
+
+1:	vpcmpeqq (LRV_SIZE + XMM_SIZE)(%rsp), %xmm1, %xmm2
+	vpmovmskb %xmm2, %esi
+	cmpl $0xffff, %esi
+	jne 1f
+	vmovdqu LRV_VECTOR1_OFFSET(%rsp), %ymm1
+
+1:
+# endif
+
+	fldt LRV_ST1_OFFSET(%rsp)
+	fldt LRV_ST0_OFFSET(%rsp)
+
+	movq %rbx, %rsp
+	movq (%rsp), %rbx
+	cfi_restore(rbx)
+	cfi_def_cfa_register(%rsp)
+
+	addq $48, %rsp		# Adjust the stack to the return value
+				# (eats the reloc index and link_map)
+	cfi_adjust_cfa_offset(-48)
+	retq
diff --git a/sysdeps/x86_64/multiarch/Makefile b/sysdeps/x86_64/multiarch/Makefile
index b066402204..0ded3b3261 100644
--- a/sysdeps/x86_64/multiarch/Makefile
+++ b/sysdeps/x86_64/multiarch/Makefile
@@ -4,7 +4,7 @@ gen-as-const-headers += ifunc-defines.sym
 endif
 
 ifeq ($(subdir),string)
-sysdep_routines += stpncpy-c strncpy-c
+sysdep_routines += stpncpy-c strncpy-c strcmp-ssse3 strncmp-ssse3
 ifeq (yes,$(config-cflags-sse4))
 sysdep_routines += strcspn-c strpbrk-c strspn-c strstr-c strcasestr-c
 CFLAGS-strcspn-c.c += -msse4
diff --git a/sysdeps/x86_64/multiarch/rawmemchr.S b/sysdeps/x86_64/multiarch/rawmemchr.S
index d4f265f430..08fd8769fc 100644
--- a/sysdeps/x86_64/multiarch/rawmemchr.S
+++ b/sysdeps/x86_64/multiarch/rawmemchr.S
@@ -38,6 +38,7 @@ END(rawmemchr)
 strong_alias (rawmemchr, __rawmemchr)
 
 
+	.section .text.sse4.2,"ax",@progbits
 	.align 	16
 	.type	__rawmemchr_sse42, @function
 __rawmemchr_sse42:
diff --git a/sysdeps/x86_64/multiarch/strcmp-ssse3.S b/sysdeps/x86_64/multiarch/strcmp-ssse3.S
new file mode 100644
index 0000000000..98cecb8942
--- /dev/null
+++ b/sysdeps/x86_64/multiarch/strcmp-ssse3.S
@@ -0,0 +1,3 @@
+#define USE_SSSE3 1
+#define STRCMP __strcmp_ssse3
+#include "../strcmp.S"
diff --git a/sysdeps/x86_64/multiarch/strcmp.S b/sysdeps/x86_64/multiarch/strcmp.S
index 1a315737af..05adf1e2e6 100644
--- a/sysdeps/x86_64/multiarch/strcmp.S
+++ b/sysdeps/x86_64/multiarch/strcmp.S
@@ -34,6 +34,7 @@
 	mov	%r9, %r11
 
 #define STRCMP_SSE42	__strncmp_sse42
+#define STRCMP_SSSE3	__strncmp_ssse3
 #define STRCMP_SSE2	__strncmp_sse2
 #define __GI_STRCMP	__GI_strncmp
 #else
@@ -41,6 +42,7 @@
 #ifndef STRCMP
 #define STRCMP		strcmp
 #define STRCMP_SSE42	__strcmp_sse42
+#define STRCMP_SSSE3	__strcmp_ssse3
 #define STRCMP_SSE2	__strcmp_sse2
 #define __GI_STRCMP	__GI_strcmp
 #endif
@@ -60,10 +62,14 @@ ENTRY(STRCMP)
 	cmpl	$0, __cpu_features+KIND_OFFSET(%rip)
 	jne	1f
 	call	__init_cpu_features
-1:	leaq	STRCMP_SSE2(%rip), %rax
-	testl	$(1<<20), __cpu_features+CPUID_OFFSET+COMMON_CPUID_INDEX_1*CPUID_SIZE+CPUID_ECX_OFFSET(%rip)
-	jz	2f
+1:
 	leaq	STRCMP_SSE42(%rip), %rax
+	testl	$(1<<20), __cpu_features+CPUID_OFFSET+COMMON_CPUID_INDEX_1*CPUID_SIZE+CPUID_ECX_OFFSET(%rip)
+	jnz	2f
+	leaq	STRCMP_SSSE3(%rip), %rax
+	testl	$(1<<9), __cpu_features+CPUID_OFFSET+COMMON_CPUID_INDEX_1*CPUID_SIZE+CPUID_ECX_OFFSET(%rip)
+	jnz	2f
+	leaq	STRCMP_SSE2(%rip), %rax
 2:	ret
 END(STRCMP)
 
diff --git a/sysdeps/x86_64/multiarch/strcspn-c.c b/sysdeps/x86_64/multiarch/strcspn-c.c
index 4512267d3f..daeebe1bf5 100644
--- a/sysdeps/x86_64/multiarch/strcspn-c.c
+++ b/sysdeps/x86_64/multiarch/strcspn-c.c
@@ -86,11 +86,13 @@ STRCSPN_SSE42 (const char *s, const char *a)
 
   const char *aligned;
   __m128i mask;
+  /* Fake initialization.  gcc otherwise will warn.  */
+  asm ("" : "=xm" (mask));
   int offset = (int) ((size_t) a & 15);
   if (offset != 0)
     {
       /* Load masks.  */
-      aligned = (const char *) ((size_t) a & 0xfffffffffffffff0L);
+      aligned = (const char *) ((size_t) a & -16L);
       __m128i mask0 = _mm_load_si128 ((__m128i *) aligned);
 
       switch (offset)
@@ -229,7 +231,7 @@ STRCSPN_SSE42 (const char *s, const char *a)
   if (offset != 0)
     {
       /* Check partial string.  */
-      aligned = (const char *) ((size_t) s & 0xfffffffffffffff0L);
+      aligned = (const char *) ((size_t) s & -16L);
       __m128i value = _mm_load_si128 ((__m128i *) aligned);
 
       switch (offset)
diff --git a/sysdeps/x86_64/multiarch/strlen.S b/sysdeps/x86_64/multiarch/strlen.S
index 82b03ccc28..4342c6cdab 100644
--- a/sysdeps/x86_64/multiarch/strlen.S
+++ b/sysdeps/x86_64/multiarch/strlen.S
@@ -40,6 +40,7 @@ ENTRY(strlen)
 END(strlen)
 
 
+	.section .text.sse4.2,"ax",@progbits
 	.align 	16
 	.type	__strlen_sse42, @function
 __strlen_sse42:
diff --git a/sysdeps/x86_64/multiarch/strncmp-ssse3.S b/sysdeps/x86_64/multiarch/strncmp-ssse3.S
new file mode 100644
index 0000000000..a320a3e949
--- /dev/null
+++ b/sysdeps/x86_64/multiarch/strncmp-ssse3.S
@@ -0,0 +1,4 @@
+#define USE_SSSE3 1
+#define STRCMP __strncmp_ssse3
+#define USE_AS_STRNCMP
+#include "../strcmp.S"
diff --git a/sysdeps/x86_64/multiarch/strspn-c.c b/sysdeps/x86_64/multiarch/strspn-c.c
index 5b99f0d383..be9e8ac0a8 100644
--- a/sysdeps/x86_64/multiarch/strspn-c.c
+++ b/sysdeps/x86_64/multiarch/strspn-c.c
@@ -68,7 +68,7 @@ __strspn_sse42 (const char *s, const char *a)
   if (offset != 0)
     {
       /* Load masks.  */
-      aligned = (const char *) ((size_t) a & 0xfffffffffffffff0L);
+      aligned = (const char *) ((size_t) a & -16L);
       __m128i mask0 = _mm_load_si128 ((__m128i *) aligned);
 
       switch (offset)
@@ -207,7 +207,7 @@ __strspn_sse42 (const char *s, const char *a)
   if (offset != 0)
     {
       /* Check partial string.  */
-      aligned = (const char *) ((size_t) s & 0xfffffffffffffff0L);
+      aligned = (const char *) ((size_t) s & -16L);
       __m128i value = _mm_load_si128 ((__m128i *) aligned);
 
       switch (offset)
diff --git a/sysdeps/x86_64/strcmp.S b/sysdeps/x86_64/strcmp.S
index 340a64ba35..650ec173b6 100644
--- a/sysdeps/x86_64/strcmp.S
+++ b/sysdeps/x86_64/strcmp.S
@@ -51,7 +51,12 @@
 # endif
 #endif
 
+#ifndef USE_SSSE3
 	.text
+#else
+        .section .text.ssse3,"ax",@progbits
+#endif
+
 ENTRY (BP_SYM (STRCMP))
 #ifdef NOT_IN_libc
 /* Simple version since we can't use SSE registers in ld.so.  */
@@ -244,9 +249,13 @@ LABEL(gobble_ashr_1):
 	movdqa	(%rdi, %rcx), %xmm2
 	movdqa	%xmm2, %xmm4		 /* store for next cycle */
 
+#ifndef USE_SSSE3
 	psrldq	$1, %xmm3
 	pslldq	$15, %xmm2
 	por	%xmm3, %xmm2		/* merge into one 16byte value */
+#else
+	palignr	$1, %xmm3, %xmm2	/* merge into one 16byte value */
+#endif
 
 	pcmpeqb	%xmm1, %xmm0
 	pcmpeqb	%xmm2, %xmm1
@@ -269,9 +278,13 @@ LABEL(gobble_ashr_1):
 	movdqa	(%rdi, %rcx), %xmm2
 	movdqa	%xmm2, %xmm4		/* store for next cycle */
 
+#ifndef USE_SSSE3
 	psrldq	$1, %xmm3
-	pslldq 	$15, %xmm2
+	pslldq	$15, %xmm2
 	por	%xmm3, %xmm2		/* merge into one 16byte value */
+#else
+	palignr	$1, %xmm3, %xmm2	/* merge into one 16byte value */
+#endif
 
 	pcmpeqb	%xmm1, %xmm0
 	pcmpeqb	%xmm2, %xmm1
@@ -363,9 +376,13 @@ LABEL(gobble_ashr_2):
 	movdqa	(%rdi, %rcx), %xmm2
 	movdqa	%xmm2, %xmm4
 
+#ifndef USE_SSSE3
 	psrldq	$2, %xmm3
 	pslldq	$14, %xmm2
-	por	%xmm3, %xmm2
+	por	%xmm3, %xmm2		/* merge into one 16byte value */
+#else
+	palignr	$2, %xmm3, %xmm2	/* merge into one 16byte value */
+#endif
 
 	pcmpeqb	%xmm1, %xmm0
 	pcmpeqb	%xmm2, %xmm1
@@ -389,9 +406,13 @@ LABEL(gobble_ashr_2):
 	movdqa	(%rdi, %rcx), %xmm2
 	movdqa	%xmm2, %xmm4
 
+#ifndef USE_SSSE3
 	psrldq	$2, %xmm3
-	pslldq 	$14, %xmm2
-	por	%xmm3, %xmm2
+	pslldq	$14, %xmm2
+	por	%xmm3, %xmm2		/* merge into one 16byte value */
+#else
+	palignr	$2, %xmm3, %xmm2	/* merge into one 16byte value */
+#endif
 
 	pcmpeqb	%xmm1, %xmm0
 	pcmpeqb	%xmm2, %xmm1
@@ -477,9 +498,13 @@ LABEL(gobble_ashr_3):
 	movdqa	(%rdi, %rcx), %xmm2
 	movdqa	%xmm2, %xmm4
 
+#ifndef USE_SSSE3
 	psrldq	$3, %xmm3
 	pslldq	$13, %xmm2
-	por	%xmm3, %xmm2
+	por	%xmm3, %xmm2		/* merge into one 16byte value */
+#else
+	palignr	$3, %xmm3, %xmm2	/* merge into one 16byte value */
+#endif
 
 	pcmpeqb	%xmm1, %xmm0
 	pcmpeqb	%xmm2, %xmm1
@@ -503,9 +528,13 @@ LABEL(gobble_ashr_3):
 	movdqa	(%rdi, %rcx), %xmm2
 	movdqa	%xmm2, %xmm4
 
+#ifndef USE_SSSE3
 	psrldq	$3, %xmm3
-	pslldq 	$13, %xmm2
-	por	%xmm3, %xmm2
+	pslldq	$13, %xmm2
+	por	%xmm3, %xmm2		/* merge into one 16byte value */
+#else
+	palignr	$3, %xmm3, %xmm2	/* merge into one 16byte value */
+#endif
 
 	pcmpeqb	%xmm1, %xmm0
 	pcmpeqb	%xmm2, %xmm1
@@ -591,9 +620,13 @@ LABEL(gobble_ashr_4):
 	movdqa	(%rdi, %rcx), %xmm2
 	movdqa	%xmm2, %xmm4
 
+#ifndef USE_SSSE3
 	psrldq	$4, %xmm3
 	pslldq	$12, %xmm2
-	por	%xmm3, %xmm2
+	por	%xmm3, %xmm2		/* merge into one 16byte value */
+#else
+	palignr	$4, %xmm3, %xmm2	/* merge into one 16byte value */
+#endif
 
 	pcmpeqb	%xmm1, %xmm0
 	pcmpeqb	%xmm2, %xmm1
@@ -617,9 +650,13 @@ LABEL(gobble_ashr_4):
 	movdqa	(%rdi, %rcx), %xmm2
 	movdqa	%xmm2, %xmm4
 
+#ifndef USE_SSSE3
 	psrldq	$4, %xmm3
-	pslldq 	$12, %xmm2
-	por	%xmm3, %xmm2
+	pslldq	$12, %xmm2
+	por	%xmm3, %xmm2		/* merge into one 16byte value */
+#else
+	palignr	$4, %xmm3, %xmm2	/* merge into one 16byte value */
+#endif
 
 	pcmpeqb	%xmm1, %xmm0
 	pcmpeqb	%xmm2, %xmm1
@@ -705,9 +742,13 @@ LABEL(gobble_ashr_5):
 	movdqa	(%rdi, %rcx), %xmm2
 	movdqa	%xmm2, %xmm4
 
+#ifndef USE_SSSE3
 	psrldq	$5, %xmm3
 	pslldq	$11, %xmm2
-	por	%xmm3, %xmm2
+	por	%xmm3, %xmm2		/* merge into one 16byte value */
+#else
+	palignr	$5, %xmm3, %xmm2	/* merge into one 16byte value */
+#endif
 
 	pcmpeqb	%xmm1, %xmm0
 	pcmpeqb	%xmm2, %xmm1
@@ -731,9 +772,13 @@ LABEL(gobble_ashr_5):
 	movdqa	(%rdi, %rcx), %xmm2
 	movdqa	%xmm2, %xmm4
 
+#ifndef USE_SSSE3
 	psrldq	$5, %xmm3
-	pslldq 	$11, %xmm2
-	por	%xmm3, %xmm2
+	pslldq	$11, %xmm2
+	por	%xmm3, %xmm2		/* merge into one 16byte value */
+#else
+	palignr	$5, %xmm3, %xmm2	/* merge into one 16byte value */
+#endif
 
 	pcmpeqb	%xmm1, %xmm0
 	pcmpeqb	%xmm2, %xmm1
@@ -819,9 +864,13 @@ LABEL(gobble_ashr_6):
 	movdqa	(%rdi, %rcx), %xmm2
 	movdqa	%xmm2, %xmm4
 
+#ifndef USE_SSSE3
 	psrldq	$6, %xmm3
 	pslldq	$10, %xmm2
-	por	%xmm3, %xmm2
+	por	%xmm3, %xmm2		/* merge into one 16byte value */
+#else
+	palignr	$6, %xmm3, %xmm2	/* merge into one 16byte value */
+#endif
 
 	pcmpeqb	%xmm1, %xmm0
 	pcmpeqb	%xmm2, %xmm1
@@ -845,9 +894,13 @@ LABEL(gobble_ashr_6):
 	movdqa	(%rdi, %rcx), %xmm2
 	movdqa	%xmm2, %xmm4
 
+#ifndef USE_SSSE3
 	psrldq	$6, %xmm3
-	pslldq 	$10, %xmm2
-	por	%xmm3, %xmm2
+	pslldq	$10, %xmm2
+	por	%xmm3, %xmm2		/* merge into one 16byte value */
+#else
+	palignr	$6, %xmm3, %xmm2	/* merge into one 16byte value */
+#endif
 
 	pcmpeqb	%xmm1, %xmm0
 	pcmpeqb	%xmm2, %xmm1
@@ -933,9 +986,13 @@ LABEL(gobble_ashr_7):
 	movdqa	(%rdi, %rcx), %xmm2
 	movdqa	%xmm2, %xmm4
 
+#ifndef USE_SSSE3
 	psrldq	$7, %xmm3
 	pslldq	$9, %xmm2
-	por	%xmm3, %xmm2
+	por	%xmm3, %xmm2		/* merge into one 16byte value */
+#else
+	palignr	$7, %xmm3, %xmm2	/* merge into one 16byte value */
+#endif
 
 	pcmpeqb	%xmm1, %xmm0
 	pcmpeqb	%xmm2, %xmm1
@@ -959,9 +1016,13 @@ LABEL(gobble_ashr_7):
 	movdqa	(%rdi, %rcx), %xmm2
 	movdqa	%xmm2, %xmm4
 
+#ifndef USE_SSSE3
 	psrldq	$7, %xmm3
-	pslldq 	$9, %xmm2
-	por	%xmm3, %xmm2
+	pslldq	$9, %xmm2
+	por	%xmm3, %xmm2		/* merge into one 16byte value */
+#else
+	palignr	$7, %xmm3, %xmm2	/* merge into one 16byte value */
+#endif
 
 	pcmpeqb	%xmm1, %xmm0
 	pcmpeqb	%xmm2, %xmm1
@@ -1047,9 +1108,13 @@ LABEL(gobble_ashr_8):
 	movdqa	(%rdi, %rcx), %xmm2
 	movdqa	%xmm2, %xmm4
 
+#ifndef USE_SSSE3
 	psrldq	$8, %xmm3
 	pslldq	$8, %xmm2
-	por	%xmm3, %xmm2
+	por	%xmm3, %xmm2		/* merge into one 16byte value */
+#else
+	palignr	$8, %xmm3, %xmm2	/* merge into one 16byte value */
+#endif
 
 	pcmpeqb	%xmm1, %xmm0
 	pcmpeqb	%xmm2, %xmm1
@@ -1073,9 +1138,13 @@ LABEL(gobble_ashr_8):
 	movdqa	(%rdi, %rcx), %xmm2
 	movdqa	%xmm2, %xmm4
 
+#ifndef USE_SSSE3
 	psrldq	$8, %xmm3
-	pslldq 	$8, %xmm2
-	por	%xmm3, %xmm2
+	pslldq	$8, %xmm2
+	por	%xmm3, %xmm2		/* merge into one 16byte value */
+#else
+	palignr	$8, %xmm3, %xmm2	/* merge into one 16byte value */
+#endif
 
 	pcmpeqb	%xmm1, %xmm0
 	pcmpeqb	%xmm2, %xmm1
@@ -1161,9 +1230,13 @@ LABEL(gobble_ashr_9):
 	movdqa	(%rdi, %rcx), %xmm2
 	movdqa	%xmm2, %xmm4
 
+#ifndef USE_SSSE3
 	psrldq	$9, %xmm3
 	pslldq	$7, %xmm2
-	por	%xmm3, %xmm2
+	por	%xmm3, %xmm2		/* merge into one 16byte value */
+#else
+	palignr	$9, %xmm3, %xmm2	/* merge into one 16byte value */
+#endif
 
 	pcmpeqb	%xmm1, %xmm0
 	pcmpeqb	%xmm2, %xmm1
@@ -1187,9 +1260,13 @@ LABEL(gobble_ashr_9):
 	movdqa	(%rdi, %rcx), %xmm2
 	movdqa	%xmm2, %xmm4
 
+#ifndef USE_SSSE3
 	psrldq	$9, %xmm3
-	pslldq 	$7, %xmm2
-	por	%xmm3, %xmm2
+	pslldq	$7, %xmm2
+	por	%xmm3, %xmm2		/* merge into one 16byte value */
+#else
+	palignr	$9, %xmm3, %xmm2	/* merge into one 16byte value */
+#endif
 
 	pcmpeqb	%xmm1, %xmm0
 	pcmpeqb	%xmm2, %xmm1
@@ -1275,9 +1352,13 @@ LABEL(gobble_ashr_10):
 	movdqa	(%rdi, %rcx), %xmm2
 	movdqa	%xmm2, %xmm4
 
+#ifndef USE_SSSE3
 	psrldq	$10, %xmm3
 	pslldq	$6, %xmm2
-	por	%xmm3, %xmm2
+	por	%xmm3, %xmm2		/* merge into one 16byte value */
+#else
+	palignr	$10, %xmm3, %xmm2	/* merge into one 16byte value */
+#endif
 
 	pcmpeqb	%xmm1, %xmm0
 	pcmpeqb	%xmm2, %xmm1
@@ -1301,9 +1382,13 @@ LABEL(gobble_ashr_10):
 	movdqa	(%rdi, %rcx), %xmm2
 	movdqa	%xmm2, %xmm4
 
+#ifndef USE_SSSE3
 	psrldq	$10, %xmm3
-	pslldq 	$6, %xmm2
-	por	%xmm3, %xmm2
+	pslldq	$6, %xmm2
+	por	%xmm3, %xmm2		/* merge into one 16byte value */
+#else
+	palignr	$10, %xmm3, %xmm2	/* merge into one 16byte value */
+#endif
 
 	pcmpeqb	%xmm1, %xmm0
 	pcmpeqb	%xmm2, %xmm1
@@ -1389,9 +1474,13 @@ LABEL(gobble_ashr_11):
 	movdqa	(%rdi, %rcx), %xmm2
 	movdqa	%xmm2, %xmm4
 
+#ifndef USE_SSSE3
 	psrldq	$11, %xmm3
 	pslldq	$5, %xmm2
-	por	%xmm3, %xmm2
+	por	%xmm3, %xmm2		/* merge into one 16byte value */
+#else
+	palignr	$11, %xmm3, %xmm2	/* merge into one 16byte value */
+#endif
 
 	pcmpeqb	%xmm1, %xmm0
 	pcmpeqb	%xmm2, %xmm1
@@ -1415,9 +1504,13 @@ LABEL(gobble_ashr_11):
 	movdqa	(%rdi, %rcx), %xmm2
 	movdqa	%xmm2, %xmm4
 
+#ifndef USE_SSSE3
 	psrldq	$11, %xmm3
-	pslldq 	$5, %xmm2
-	por	%xmm3, %xmm2
+	pslldq	$5, %xmm2
+	por	%xmm3, %xmm2		/* merge into one 16byte value */
+#else
+	palignr	$11, %xmm3, %xmm2	/* merge into one 16byte value */
+#endif
 
 	pcmpeqb	%xmm1, %xmm0
 	pcmpeqb	%xmm2, %xmm1
@@ -1503,9 +1596,13 @@ LABEL(gobble_ashr_12):
 	movdqa	(%rdi, %rcx), %xmm2
 	movdqa	%xmm2, %xmm4
 
+#ifndef USE_SSSE3
 	psrldq	$12, %xmm3
 	pslldq	$4, %xmm2
-	por	%xmm3, %xmm2
+	por	%xmm3, %xmm2		/* merge into one 16byte value */
+#else
+	palignr	$12, %xmm3, %xmm2	/* merge into one 16byte value */
+#endif
 
 	pcmpeqb	%xmm1, %xmm0
 	pcmpeqb	%xmm2, %xmm1
@@ -1529,9 +1626,13 @@ LABEL(gobble_ashr_12):
 	movdqa	(%rdi, %rcx), %xmm2
 	movdqa	%xmm2, %xmm4
 
+#ifndef USE_SSSE3
 	psrldq	$12, %xmm3
-	pslldq 	$4, %xmm2
-	por	%xmm3, %xmm2
+	pslldq	$4, %xmm2
+	por	%xmm3, %xmm2		/* merge into one 16byte value */
+#else
+	palignr	$12, %xmm3, %xmm2	/* merge into one 16byte value */
+#endif
 
 	pcmpeqb	%xmm1, %xmm0
 	pcmpeqb	%xmm2, %xmm1
@@ -1617,9 +1718,13 @@ LABEL(gobble_ashr_13):
 	movdqa	(%rdi, %rcx), %xmm2
 	movdqa	%xmm2, %xmm4
 
+#ifndef USE_SSSE3
 	psrldq	$13, %xmm3
 	pslldq	$3, %xmm2
-	por	%xmm3, %xmm2
+	por	%xmm3, %xmm2		/* merge into one 16byte value */
+#else
+	palignr	$13, %xmm3, %xmm2	/* merge into one 16byte value */
+#endif
 
 	pcmpeqb	%xmm1, %xmm0
 	pcmpeqb	%xmm2, %xmm1
@@ -1643,9 +1748,13 @@ LABEL(gobble_ashr_13):
 	movdqa	(%rdi, %rcx), %xmm2
 	movdqa	%xmm2, %xmm4
 
+#ifndef USE_SSSE3
 	psrldq	$13, %xmm3
-	pslldq 	$3, %xmm2
-	por	%xmm3, %xmm2
+	pslldq	$3, %xmm2
+	por	%xmm3, %xmm2		/* merge into one 16byte value */
+#else
+	palignr	$13, %xmm3, %xmm2	/* merge into one 16byte value */
+#endif
 
 	pcmpeqb	%xmm1, %xmm0
 	pcmpeqb	%xmm2, %xmm1
@@ -1731,9 +1840,13 @@ LABEL(gobble_ashr_14):
 	movdqa	(%rdi, %rcx), %xmm2
 	movdqa	%xmm2, %xmm4
 
+#ifndef USE_SSSE3
 	psrldq	$14, %xmm3
 	pslldq	$2, %xmm2
-	por	%xmm3, %xmm2
+	por	%xmm3, %xmm2		/* merge into one 16byte value */
+#else
+	palignr	$14, %xmm3, %xmm2	/* merge into one 16byte value */
+#endif
 
 	pcmpeqb	%xmm1, %xmm0
 	pcmpeqb	%xmm2, %xmm1
@@ -1757,9 +1870,13 @@ LABEL(gobble_ashr_14):
 	movdqa	(%rdi, %rcx), %xmm2
 	movdqa	%xmm2, %xmm4
 
+#ifndef USE_SSSE3
 	psrldq	$14, %xmm3
-	pslldq 	$2, %xmm2
-	por	%xmm3, %xmm2
+	pslldq	$2, %xmm2
+	por	%xmm3, %xmm2		/* merge into one 16byte value */
+#else
+	palignr	$14, %xmm3, %xmm2	/* merge into one 16byte value */
+#endif
 
 	pcmpeqb	%xmm1, %xmm0
 	pcmpeqb	%xmm2, %xmm1
@@ -1847,9 +1964,13 @@ LABEL(gobble_ashr_15):
 	movdqa	(%rdi, %rcx), %xmm2
 	movdqa	%xmm2, %xmm4
 
+#ifndef USE_SSSE3
 	psrldq	$15, %xmm3
 	pslldq	$1, %xmm2
-	por	%xmm3, %xmm2
+	por	%xmm3, %xmm2		/* merge into one 16byte value */
+#else
+	palignr	$15, %xmm3, %xmm2	/* merge into one 16byte value */
+#endif
 
 	pcmpeqb	%xmm1, %xmm0
 	pcmpeqb	%xmm2, %xmm1
@@ -1873,9 +1994,13 @@ LABEL(gobble_ashr_15):
 	movdqa	(%rdi, %rcx), %xmm2
 	movdqa	%xmm2, %xmm4
 
+#ifndef USE_SSSE3
 	psrldq	$15, %xmm3
-	pslldq 	$1, %xmm2
-	por	%xmm3, %xmm2
+	pslldq	$1, %xmm2
+	por	%xmm3, %xmm2		/* merge into one 16byte value */
+#else
+	palignr	$15, %xmm3, %xmm2	/* merge into one 16byte value */
+#endif
 
 	pcmpeqb	%xmm1, %xmm0
 	pcmpeqb	%xmm2, %xmm1