Improve generic strspn performance

As for strcspn, this patch improves strspn performance using a much
faster algorithm.  It first constructs a 256-entry table based on
the accept string and then uses it as a lookup table for the
input string.  As for strcspn optimization, it is generally at least
10 times faster than the existing implementation on bench-strspn
on a few AArch64 implementations.

Also the string/bits/string2.h inlines make no longer sense, as current
implementation will already implement most of the optimizations.

Tested on x86_64, i686, and aarch64.

	* string/strspn.c (strcspn): Rewrite function.
	* string/bits/string2.h (strspn): Use __builtin_strcspn.
	(__strspn_c1): Remove inline function.
	(__strspn_c2): Likewise.
	(__strspn_c3): Likewise.
	* string/string-inlines.c
	[SHLIB_COMPAT(libc, GLIBC_2_1_1, GLIBC_2_24)] (__strspn_c1): Add
	compatibility symbol.
	[SHLIB_COMPAT(libc, GLIBC_2_1_1, GLIBC_2_24)] (__strspn_c2):
	Likewise.
	[SHLIB_COMPAT(libc, GLIBC_2_1_1, GLIBC_2_24)] (__strspn_c3):
	Likewise.

4 files changed, 97 insertions, 86 deletions

diff --git a/ChangeLog b/ChangeLog
index 0d522ad4c4..79e2e23398 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,19 @@
+2016-04-01  Adhemerval Zanella  <adhemerval.zanella@linaro.org>
+
+	* string/strspn.c (strcspn): Rewrite function.
+	* string/bits/string2.h (strspn): Use __builtin_strcspn.
+	(__strspn_c1): Remove inline function.
+	(__strspn_c2): Likewise.
+	(__strspn_c3): Likewise.
+	* string/string-inlines.c
+	[SHLIB_COMPAT(libc, GLIBC_2_1_1, GLIBC_2_24)] (__strspn_c1): Add
+	compatibility symbol.
+	[SHLIB_COMPAT(libc, GLIBC_2_1_1, GLIBC_2_24)] (__strspn_c2):
+	Likewise.
+	[SHLIB_COMPAT(libc, GLIBC_2_1_1, GLIBC_2_24)] (__strspn_c3):
+	Likewise.
+	* string/string-inlines.c: Include generic version.
+
 2016-04-01  Wilco Dijkstra  <wdijkstr@arm.com>
 	    Adhemerval Zanella  <adhemerval.zanella@linaro.org>
 
diff --git a/string/bits/string2.h b/string/bits/string2.h
index a8df0db14a..75a66a12e7 100644
--- a/string/bits/string2.h
+++ b/string/bits/string2.h
@@ -914,78 +914,10 @@ __stpcpy_small (char *__dest,
 
 /* Return the length of the initial segment of S which
    consists entirely of characters in ACCEPT.  */
-#if !defined _HAVE_STRING_ARCH_strspn || defined _FORCE_INLINES
-# ifndef _HAVE_STRING_ARCH_strspn
-#  if __GNUC_PREREQ (3, 2)
-#   define strspn(s, accept) \
-  __extension__								      \
-  ({ char __a0, __a1, __a2;						      \
-     (__builtin_constant_p (accept) && __string2_1bptr_p (accept)	      \
-      ? ((__builtin_constant_p (s) && __string2_1bptr_p (s))		      \
-	 ? __builtin_strspn (s, accept)					      \
-	 : ((__a0 = ((const char *) (accept))[0], __a0 == '\0')		      \
-	    ? ((void) (s), (size_t) 0)					      \
-	    : ((__a1 = ((const char *) (accept))[1], __a1 == '\0')	      \
-	       ? __strspn_c1 (s, __a0)					      \
-	       : ((__a2 = ((const char *) (accept))[2], __a2 == '\0')	      \
-		  ? __strspn_c2 (s, __a0, __a1)				      \
-		  : (((const char *) (accept))[3] == '\0'		      \
-		     ? __strspn_c3 (s, __a0, __a1, __a2)		      \
-		     : __builtin_strspn (s, accept))))))		      \
-      : __builtin_strspn (s, accept)); })
-#  else
-#   define strspn(s, accept) \
-  __extension__								      \
-  ({ char __a0, __a1, __a2;						      \
-     (__builtin_constant_p (accept) && __string2_1bptr_p (accept)	      \
-      ? ((__a0 = ((const char *) (accept))[0], __a0 == '\0')		      \
-	 ? ((void) (s), (size_t) 0)					      \
-	 : ((__a1 = ((const char *) (accept))[1], __a1 == '\0')		      \
-	    ? __strspn_c1 (s, __a0)					      \
-	    : ((__a2 = ((const char *) (accept))[2], __a2 == '\0')	      \
-	       ? __strspn_c2 (s, __a0, __a1)				      \
-	       : (((const char *) (accept))[3] == '\0'			      \
-		  ? __strspn_c3 (s, __a0, __a1, __a2)			      \
-		  : strspn (s, accept)))))				      \
-      : strspn (s, accept)); })
-#  endif
+#ifndef _HAVE_STRING_ARCH_strspn
+# if __GNUC_PREREQ (3, 2)
+#  define strspn(s, accept) __builtin_strspn (s, accept)
 # endif
-
-__STRING_INLINE size_t __strspn_c1 (const char *__s, int __accept);
-__STRING_INLINE size_t
-__strspn_c1 (const char *__s, int __accept)
-{
-  size_t __result = 0;
-  /* Please note that __accept never can be '\0'.  */
-  while (__s[__result] == __accept)
-    ++__result;
-  return __result;
-}
-
-__STRING_INLINE size_t __strspn_c2 (const char *__s, int __accept1,
-				    int __accept2);
-__STRING_INLINE size_t
-__strspn_c2 (const char *__s, int __accept1, int __accept2)
-{
-  size_t __result = 0;
-  /* Please note that __accept1 and __accept2 never can be '\0'.  */
-  while (__s[__result] == __accept1 || __s[__result] == __accept2)
-    ++__result;
-  return __result;
-}
-
-__STRING_INLINE size_t __strspn_c3 (const char *__s, int __accept1,
-				    int __accept2, int __accept3);
-__STRING_INLINE size_t
-__strspn_c3 (const char *__s, int __accept1, int __accept2, int __accept3)
-{
-  size_t __result = 0;
-  /* Please note that __accept1 to __accept3 never can be '\0'.  */
-  while (__s[__result] == __accept1 || __s[__result] == __accept2
-	 || __s[__result] == __accept3)
-    ++__result;
-  return __result;
-}
 #endif
 
 
diff --git a/string/string-inlines.c b/string/string-inlines.c
index 83bdd6c185..754b31530f 100644
--- a/string/string-inlines.c
+++ b/string/string-inlines.c
@@ -71,4 +71,40 @@ __old_strcspn_c3 (const char *__s, int __reject1, int __reject2,
   return __result;
 }
 compat_symbol (libc, __old_strcspn_c3, __strcspn_c3, GLIBC_2_1_1);
+
+size_t
+__old_strspn_c1 (const char *__s, int __accept)
+{
+  size_t __result = 0;
+  /* Please note that __accept never can be '\0'.  */
+  while (__s[__result] == __accept)
+    ++__result;
+  return __result;
+}
+compat_symbol (libc, __old_strspn_c1, __strspn_c1, GLIBC_2_1_1);
+
+size_t
+__old_strspn_c2 (const char *__s, int __accept1, int __accept2)
+{
+  size_t __result = 0;
+  /* Please note that __accept1 and __accept2 never can be '\0'.  */
+  while (__s[__result] == __accept1 || __s[__result] == __accept2)
+    ++__result;
+  return __result;
+}
+compat_symbol (libc, __old_strspn_c2, __strspn_c2, GLIBC_2_1_1);
+
+size_t
+__old_strspn_c3 (const char *__s, int __accept1, int __accept2,
+		 int __accept3)
+{
+  size_t __result = 0;
+  /* Please note that __accept1 to __accept3 never can be '\0'.  */
+  while (__s[__result] == __accept1 || __s[__result] == __accept2
+	 || __s[__result] == __accept3)
+    ++__result;
+  return __result;
+}
+compat_symbol (libc, __old_strspn_c3, __strspn_c3, GLIBC_2_1_1);
+
 #endif
diff --git a/string/strspn.c b/string/strspn.c
index f0635c156a..f98340b1b2 100644
--- a/string/strspn.c
+++ b/string/strspn.c
@@ -16,32 +16,59 @@
    <http://www.gnu.org/licenses/>.  */
 
 #include <string.h>
+#include <stdint.h>
 
 #undef strspn
 #ifndef STRSPN
-#define STRSPN strspn
+# define STRSPN strspn
 #endif
 
 /* Return the length of the maximum initial segment
    of S which contains only characters in ACCEPT.  */
 size_t
-STRSPN (const char *s, const char *accept)
+STRSPN (const char *str, const char *accept)
 {
-  const char *p;
-  const char *a;
-  size_t count = 0;
-
-  for (p = s; *p != '\0'; ++p)
+  if (accept[0] == '\0')
+    return 0;
+  if (__glibc_unlikely (accept[1] == '\0'))
     {
-      for (a = accept; *a != '\0'; ++a)
-	if (*p == *a)
-	  break;
-      if (*a == '\0')
-	return count;
-      else
-	++count;
+      const char *a = str;
+      for (; *str == *accept; str++);
+      return str - a;
     }
 
-  return count;
+  /* Use multiple small memsets to enable inlining on most targets.  */
+  unsigned char table[256];
+  unsigned char *p = memset (table, 0, 64);
+  memset (p + 64, 0, 64);
+  memset (p + 128, 0, 64);
+  memset (p + 192, 0, 64);
+
+  unsigned char *s = (unsigned char*) accept;
+  /* Different from strcspn it does not add the NULL on the table
+     so can avoid check if str[i] is NULL, since table['\0'] will
+     be 0 and thus stopping the loop check.  */
+  do
+    p[*s++] = 1;
+  while (*s);
+
+  s = (unsigned char*) str;
+  if (!p[s[0]]) return 0;
+  if (!p[s[1]]) return 1;
+  if (!p[s[2]]) return 2;
+  if (!p[s[3]]) return 3;
+
+  s = (unsigned char *) ((uintptr_t)(s) & ~3);
+  unsigned int c0, c1, c2, c3;
+  do {
+      s += 4;
+      c0 = p[s[0]];
+      c1 = p[s[1]];
+      c2 = p[s[2]];
+      c3 = p[s[3]];
+  } while ((c0 & c1 & c2 & c3) != 0);
+
+  size_t count = s - (unsigned char *) str;
+  return (c0 & c1) == 0 ? count + c0 : count + c2 + 2;
 }
 libc_hidden_builtin_def (strspn)