.

2 files changed, 95 insertions, 246 deletions

diff --git a/sysdeps/x86_64/bits/atomic.h b/sysdeps/x86_64/bits/atomic.h
index 65d6b02008..133a68d192 100644
--- a/sysdeps/x86_64/bits/atomic.h
+++ b/sysdeps/x86_64/bits/atomic.h
@@ -1,4 +1,4 @@
-/* Copyright (C) 2002, 2003, 2004, 2006 Free Software Foundation, Inc.
+/* Copyright (C) 2002, 2003, 2004 Free Software Foundation, Inc.
    This file is part of the GNU C Library.
    Contributed by Ulrich Drepper <drepper@redhat.com>, 2002.
 
@@ -18,7 +18,6 @@
    02111-1307 USA.  */
 
 #include <stdint.h>
-#include <tls.h>	/* For tcbhead_t.  */
 
 
 typedef int8_t atomic8_t;
@@ -81,54 +80,8 @@ typedef uintmax_t uatomic_max_t;
   ({ __typeof (*mem) ret;						      \
      __asm __volatile (LOCK_PREFIX "cmpxchgq %q2, %1"			      \
 		       : "=a" (ret), "=m" (*mem)			      \
-		       : "r" ((long int) (newval)), "m" (*mem),		      \
-			 "0" ((long int) (oldval)));			      \
-     ret; })
-
-
-#define __arch_c_compare_and_exchange_val_8_acq(mem, newval, oldval) \
-  ({ __typeof (*mem) ret;						      \
-    __asm __volatile ("cmpl $0, %%fs:%P5\n\t"				      \
-		      "je 0f\n\t"					      \
-		      "lock\n"						      \
-		       "0:\tcmpxchgb %b2, %1"				      \
-		       : "=a" (ret), "=m" (*mem)			      \
-		       : "q" (newval), "m" (*mem), "0" (oldval),	      \
-			 "i" (offsetof (tcbhead_t, multiple_threads)));	      \
-     ret; })
-
-#define __arch_c_compare_and_exchange_val_16_acq(mem, newval, oldval) \
-  ({ __typeof (*mem) ret;						      \
-    __asm __volatile ("cmpl $0, %%fs:%P5\n\t"				      \
-		      "je 0f\n\t"					      \
-		      "lock\n"						      \
-		       "0:\tcmpxchgw %w2, %1"				      \
-		       : "=a" (ret), "=m" (*mem)			      \
-		       : "q" (newval), "m" (*mem), "0" (oldval),	      \
-			 "i" (offsetof (tcbhead_t, multiple_threads)));	      \
-     ret; })
-
-#define __arch_c_compare_and_exchange_val_32_acq(mem, newval, oldval) \
-  ({ __typeof (*mem) ret;						      \
-    __asm __volatile ("cmpl $0, %%fs:%P5\n\t"				      \
-		      "je 0f\n\t"					      \
-		      "lock\n"						      \
-		       "0:\tcmpxchgl %2, %1"				      \
-		       : "=a" (ret), "=m" (*mem)			      \
-		       : "q" (newval), "m" (*mem), "0" (oldval),	      \
-			 "i" (offsetof (tcbhead_t, multiple_threads)));	      \
-     ret; })
-
-#define __arch_c_compare_and_exchange_val_64_acq(mem, newval, oldval) \
-  ({ __typeof (*mem) ret;						      \
-     __asm __volatile ("cmpl $0, %%fs:%P5\n\t"				      \
-		       "je 0f\n\t"					      \
-		       "lock\n"						      \
-		       "0:\tcmpxchgq %q2, %1"				      \
-		       : "=a" (ret), "=m" (*mem)			      \
-		       : "q" ((long int) (newval)), "m" (*mem),		      \
-			 "0" ((long int)oldval),			      \
-			 "i" (offsetof (tcbhead_t, multiple_threads)));	      \
+		       : "r" ((long) (newval)), "m" (*mem),		      \
+			 "0" ((long) (oldval)));			      \
      ret; })
 
 
@@ -154,76 +107,49 @@ typedef uintmax_t uatomic_max_t;
      result; })
 
 
-#define __arch_exchange_and_add_body(lock, mem, value)			      \
+#define atomic_exchange_and_add(mem, value) \
   ({ __typeof (*mem) result;						      \
      if (sizeof (*mem) == 1)						      \
-       __asm __volatile (lock "xaddb %b0, %1"				      \
+       __asm __volatile (LOCK_PREFIX "xaddb %b0, %1"			      \
 			 : "=r" (result), "=m" (*mem)			      \
-			 : "0" (value), "m" (*mem),			      \
-			   "i" (offsetof (tcbhead_t, multiple_threads)));     \
+			 : "0" (value), "m" (*mem));			      \
      else if (sizeof (*mem) == 2)					      \
-       __asm __volatile (lock "xaddw %w0, %1"				      \
+       __asm __volatile (LOCK_PREFIX "xaddw %w0, %1"			      \
 			 : "=r" (result), "=m" (*mem)			      \
-			 : "0" (value), "m" (*mem),			      \
-			   "i" (offsetof (tcbhead_t, multiple_threads)));     \
+			 : "0" (value), "m" (*mem));			      \
      else if (sizeof (*mem) == 4)					      \
-       __asm __volatile (lock "xaddl %0, %1"				      \
+       __asm __volatile (LOCK_PREFIX "xaddl %0, %1"			      \
 			 : "=r" (result), "=m" (*mem)			      \
-			 : "0" (value), "m" (*mem),			      \
-			   "i" (offsetof (tcbhead_t, multiple_threads)));     \
+			 : "0" (value), "m" (*mem));			      \
      else								      \
-       __asm __volatile (lock "xaddq %q0, %1"				      \
+       __asm __volatile (LOCK_PREFIX "xaddq %q0, %1"			      \
 			 : "=r" (result), "=m" (*mem)			      \
-			 : "0" ((long) (value)), "m" (*mem),		      \
-			   "i" (offsetof (tcbhead_t, multiple_threads)));     \
+			 : "0" ((long) (value)), "m" (*mem));		      \
      result; })
 
-#define atomic_exchange_and_add(mem, value) \
-  __arch_exchange_and_add_body (LOCK_PREFIX, mem, value)
-
-#define __arch_exchange_and_add_cprefix \
-  "cmpl $0, %%fs:%P4\n\tje 0f\n\tlock\n0:\t"
-
-#define catomic_exchange_and_add(mem, value) \
-  __arch_exchange_and_add_body (__arch_exchange_and_add_cprefix, mem, value)
-
-
-#define __arch_add_body(lock, pfx, mem, value)				      \
-  do {									      \
-    if (__builtin_constant_p (value) && (value) == 1)			      \
-      pfx##_increment (mem);						      \
-    else if (__builtin_constant_p (value) && (value) == -1)		      \
-      pfx##_decrement (mem);						      \
-    else if (sizeof (*mem) == 1)					      \
-      __asm __volatile (lock "addb %b1, %0"				      \
-			: "=m" (*mem)					      \
-			: "ir" (value), "m" (*mem),			      \
-			  "i" (offsetof (tcbhead_t, multiple_threads)));      \
-    else if (sizeof (*mem) == 2)					      \
-      __asm __volatile (lock "addw %w1, %0"				      \
-			: "=m" (*mem)					      \
-			: "ir" (value), "m" (*mem),			      \
-			  "i" (offsetof (tcbhead_t, multiple_threads)));      \
-    else if (sizeof (*mem) == 4)					      \
-      __asm __volatile (lock "addl %1, %0"				      \
-			: "=m" (*mem)					      \
-			: "ir" (value), "m" (*mem),			      \
-			  "i" (offsetof (tcbhead_t, multiple_threads)));      \
-    else								      \
-      __asm __volatile (lock "addq %q1, %0"				      \
-			: "=m" (*mem)					      \
-			: "ir" ((long) (value)), "m" (*mem),		      \
-			  "i" (offsetof (tcbhead_t, multiple_threads)));      \
-  } while (0)
 
 #define atomic_add(mem, value) \
-  __arch_add_body (LOCK_PREFIX, atomic, mem, value)
-
-#define __arch_add_cprefix \
-  "cmpl $0, %%fs:%P3\n\tje 0f\n\tlock\n0:\t"
-
-#define catomic_add(mem, value) \
-  __arch_add_body (__arch_add_cprefix, catomic, mem, value)
+  (void) ({ if (__builtin_constant_p (value) && (value) == 1)		      \
+	      atomic_increment (mem);					      \
+	    else if (__builtin_constant_p (value) && (value) == 1)	      \
+	      atomic_decrement (mem);					      \
+	    else if (sizeof (*mem) == 1)				      \
+	      __asm __volatile (LOCK_PREFIX "addb %b1, %0"		      \
+				: "=m" (*mem)				      \
+				: "ir" (value), "m" (*mem));		      \
+	    else if (sizeof (*mem) == 2)				      \
+	      __asm __volatile (LOCK_PREFIX "addw %w1, %0"		      \
+				: "=m" (*mem)				      \
+				: "ir" (value), "m" (*mem));		      \
+	    else if (sizeof (*mem) == 4)				      \
+	      __asm __volatile (LOCK_PREFIX "addl %1, %0"		      \
+				: "=m" (*mem)				      \
+				: "ir" (value), "m" (*mem));		      \
+	    else							      \
+	      __asm __volatile (LOCK_PREFIX "addq %q1, %0"		      \
+				: "=m" (*mem)				      \
+				: "ir" ((long) (value)), "m" (*mem));	      \
+	    })
 
 
 #define atomic_add_negative(mem, value) \
@@ -268,37 +194,24 @@ typedef uintmax_t uatomic_max_t;
      __result; })
 
 
-#define __arch_increment_body(lock, mem) \
-  do {									      \
-    if (sizeof (*mem) == 1)						      \
-      __asm __volatile (lock "incb %b0"					      \
-			: "=m" (*mem)					      \
-			: "m" (*mem),					      \
-			  "i" (offsetof (tcbhead_t, multiple_threads)));      \
-    else if (sizeof (*mem) == 2)					      \
-      __asm __volatile (lock "incw %w0"					      \
-			: "=m" (*mem)					      \
-			: "m" (*mem),					      \
-			  "i" (offsetof (tcbhead_t, multiple_threads)));      \
-    else if (sizeof (*mem) == 4)					      \
-      __asm __volatile (lock "incl %0"					      \
-			: "=m" (*mem)					      \
-			: "m" (*mem),					      \
-			  "i" (offsetof (tcbhead_t, multiple_threads)));      \
-    else								      \
-      __asm __volatile (lock "incq %q0"					      \
-			: "=m" (*mem)					      \
-			: "m" (*mem),					      \
-			  "i" (offsetof (tcbhead_t, multiple_threads)));      \
-  } while (0)
-
-#define atomic_increment(mem) __arch_increment_body (LOCK_PREFIX, mem)
-
-#define __arch_increment_cprefix \
-  "cmpl $0, %%fs:%P2\n\tje 0f\n\tlock\n0:\t"
-
-#define catomic_increment(mem) \
-  __arch_increment_body (__arch_increment_cprefix, mem)
+#define atomic_increment(mem) \
+  (void) ({ if (sizeof (*mem) == 1)					      \
+	      __asm __volatile (LOCK_PREFIX "incb %b0"			      \
+				: "=m" (*mem)				      \
+				: "m" (*mem));				      \
+	    else if (sizeof (*mem) == 2)				      \
+	      __asm __volatile (LOCK_PREFIX "incw %w0"			      \
+				: "=m" (*mem)				      \
+				: "m" (*mem));				      \
+	    else if (sizeof (*mem) == 4)				      \
+	      __asm __volatile (LOCK_PREFIX "incl %0"			      \
+				: "=m" (*mem)				      \
+				: "m" (*mem));				      \
+	    else							      \
+	      __asm __volatile (LOCK_PREFIX "incq %q0"			      \
+				: "=m" (*mem)				      \
+				: "m" (*mem));				      \
+	    })
 
 
 #define atomic_increment_and_test(mem) \
@@ -322,37 +235,24 @@ typedef uintmax_t uatomic_max_t;
      __result; })
 
 
-#define __arch_decrement_body(lock, mem) \
-  do {									      \
-    if (sizeof (*mem) == 1)						      \
-      __asm __volatile (lock "decb %b0"					      \
-			: "=m" (*mem)					      \
-			: "m" (*mem),					      \
-			  "i" (offsetof (tcbhead_t, multiple_threads)));      \
-    else if (sizeof (*mem) == 2)					      \
-      __asm __volatile (lock "decw %w0"					      \
-			: "=m" (*mem)					      \
-			: "m" (*mem),					      \
-			  "i" (offsetof (tcbhead_t, multiple_threads)));      \
-    else if (sizeof (*mem) == 4)					      \
-      __asm __volatile (lock "decl %0"					      \
-			: "=m" (*mem)					      \
-			: "m" (*mem),					      \
-			  "i" (offsetof (tcbhead_t, multiple_threads)));      \
-    else								      \
-      __asm __volatile (lock "decq %q0"					      \
-			: "=m" (*mem)					      \
-			: "m" (*mem),					      \
-			  "i" (offsetof (tcbhead_t, multiple_threads)));      \
-  } while (0)
-
-#define atomic_decrement(mem) __arch_decrement_body (LOCK_PREFIX, mem)
-
-#define __arch_decrement_cprefix \
-  "cmpl $0, %%fs:%P2\n\tje 0f\n\tlock\n0:\t"
-
-#define catomic_decrement(mem) \
-  __arch_decrement_body (__arch_decrement_cprefix, mem)
+#define atomic_decrement(mem) \
+  (void) ({ if (sizeof (*mem) == 1)					      \
+	      __asm __volatile (LOCK_PREFIX "decb %b0"			      \
+				: "=m" (*mem)				      \
+				: "m" (*mem));				      \
+	    else if (sizeof (*mem) == 2)				      \
+	      __asm __volatile (LOCK_PREFIX "decw %w0"			      \
+				: "=m" (*mem)				      \
+				: "m" (*mem));				      \
+	    else if (sizeof (*mem) == 4)				      \
+	      __asm __volatile (LOCK_PREFIX "decl %0"			      \
+				: "=m" (*mem)				      \
+				: "m" (*mem));				      \
+	    else							      \
+	      __asm __volatile (LOCK_PREFIX "decq %q0"			      \
+				: "=m" (*mem)				      \
+				: "m" (*mem));				      \
+	    })
 
 
 #define atomic_decrement_and_test(mem) \
@@ -377,28 +277,27 @@ typedef uintmax_t uatomic_max_t;
 
 
 #define atomic_bit_set(mem, bit) \
-  do {									      \
-    if (sizeof (*mem) == 1)						      \
-      __asm __volatile (LOCK_PREFIX "orb %b2, %0"			      \
-			: "=m" (*mem)					      \
-			: "m" (*mem), "ir" (1L << (bit)));		      \
-    else if (sizeof (*mem) == 2)					      \
-      __asm __volatile (LOCK_PREFIX "orw %w2, %0"			      \
-			: "=m" (*mem)					      \
-			: "m" (*mem), "ir" (1L << (bit)));		      \
-    else if (sizeof (*mem) == 4)					      \
-      __asm __volatile (LOCK_PREFIX "orl %2, %0"			      \
-			: "=m" (*mem)					      \
-			: "m" (*mem), "ir" (1L << (bit)));		      \
-    else if (__builtin_constant_p (bit) && (bit) < 32)			      \
-      __asm __volatile (LOCK_PREFIX "orq %2, %0"			      \
-			: "=m" (*mem)					      \
-			: "m" (*mem), "i" (1L << (bit)));		      \
-    else								      \
-      __asm __volatile (LOCK_PREFIX "orq %q2, %0"			      \
-			: "=m" (*mem)					      \
-			: "m" (*mem), "r" (1UL << (bit)));		      \
-  } while (0)
+  (void) ({ if (sizeof (*mem) == 1)					      \
+	      __asm __volatile (LOCK_PREFIX "orb %b2, %0"		      \
+				: "=m" (*mem)				      \
+				: "m" (*mem), "ir" (1L << (bit)));	      \
+	    else if (sizeof (*mem) == 2)				      \
+	      __asm __volatile (LOCK_PREFIX "orw %w2, %0"		      \
+				: "=m" (*mem)				      \
+				: "m" (*mem), "ir" (1L << (bit)));	      \
+	    else if (sizeof (*mem) == 4)				      \
+	      __asm __volatile (LOCK_PREFIX "orl %2, %0"		      \
+				: "=m" (*mem)				      \
+				: "m" (*mem), "ir" (1L << (bit)));	      \
+	    else if (__builtin_constant_p (bit) && (bit) < 32)		      \
+	      __asm __volatile (LOCK_PREFIX "orq %2, %0"		      \
+				: "=m" (*mem)				      \
+				: "m" (*mem), "i" (1L << (bit)));	      \
+	    else							      \
+	      __asm __volatile (LOCK_PREFIX "orq %q2, %0"		      \
+				: "=m" (*mem)				      \
+				: "m" (*mem), "r" (1UL << (bit)));	      \
+	    })
 
 
 #define atomic_bit_test_set(mem, bit) \
@@ -423,56 +322,3 @@ typedef uintmax_t uatomic_max_t;
 
 
 #define atomic_delay() asm ("rep; nop")
-
-
-#define atomic_and(mem, mask) \
-  do {									      \
-    if (sizeof (*mem) == 1)						      \
-      __asm __volatile (LOCK_PREFIX "andb %1, %b0"			      \
-			: "=m" (*mem)					      \
-			: "ir" (mask), "m" (*mem));			      \
-    else if (sizeof (*mem) == 2)					      \
-      __asm __volatile (LOCK_PREFIX "andw %1, %w0"			      \
-			: "=m" (*mem)					      \
-			: "ir" (mask), "m" (*mem));			      \
-    else if (sizeof (*mem) == 4)					      \
-      __asm __volatile (LOCK_PREFIX "andl %1, %0"			      \
-			: "=m" (*mem)					      \
-			: "ir" (mask), "m" (*mem));			      \
-    else								      \
-      __asm __volatile (LOCK_PREFIX "andq %1, %q0"			      \
-			: "=m" (*mem)					      \
-			: "ir" (mask), "m" (*mem));			      \
-  } while (0)
-
-
-#define __arch_or_body(lock, mem, mask)					      \
-  do {									      \
-    if (sizeof (*mem) == 1)						      \
-      __asm __volatile (lock "orb %1, %b0"				      \
-			: "=m" (*mem)					      \
-			: "ir" (mask), "m" (*mem),			      \
-			  "i" (offsetof (tcbhead_t, multiple_threads)));      \
-    else if (sizeof (*mem) == 2)					      \
-      __asm __volatile (lock "orw %1, %w0"				      \
-			: "=m" (*mem)					      \
-			: "ir" (mask), "m" (*mem),			      \
-			  "i" (offsetof (tcbhead_t, multiple_threads)));      \
-    else if (sizeof (*mem) == 4)					      \
-      __asm __volatile (lock "orl %1, %0"				      \
-			: "=m" (*mem)					      \
-			: "ir" (mask), "m" (*mem),			      \
-			  "i" (offsetof (tcbhead_t, multiple_threads)));      \
-    else								      \
-      __asm __volatile (lock "orq %1, %q0"				      \
-			: "=m" (*mem)					      \
-			: "ir" (mask), "m" (*mem),			      \
-			  "i" (offsetof (tcbhead_t, multiple_threads)));      \
-  } while (0)
-
-#define atomic_or(mem, mask) __arch_or_body (LOCK_PREFIX, mem, mask)
-
-#define __arch_or_cprefix \
-  "cmpl $0, %%fs:%P3\n\tje 0f\n\tlock\n0:\t"
-
-#define catomic_or(mem, mask) __arch_or_body (__arch_or_cprefix, mem, mask)
diff --git a/sysdeps/x86_64/bits/byteswap.h b/sysdeps/x86_64/bits/byteswap.h
index e1c861c75f..ec2b17889d 100644
--- a/sysdeps/x86_64/bits/byteswap.h
+++ b/sysdeps/x86_64/bits/byteswap.h
@@ -1,5 +1,6 @@
 /* Macros to swap the order of bytes in integer values.
-   Copyright (C) 1997, 1998, 2000, 2002, 2003 Free Software Foundation, Inc.
+   Copyright (C) 1997, 1998, 2000, 2002, 2003, 2007
+   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
@@ -59,7 +60,9 @@
 # if __WORDSIZE == 64 || (defined __i486__ || defined __pentium__	      \
 			  || defined __pentiumpro__ || defined __pentium4__   \
 			  || defined __k8__ || defined __athlon__	      \
-			  || defined __k6__)
+			  || defined __k6__ || defined __nocona__	      \
+			  || defined __core2__ || defined __geode__	      \
+			  || defined __amdfam10__)
 /* To swap the bytes in a word the i486 processors and up provide the
    `bswap' opcode.  On i386 we have to use three instructions.  */
 #  define __bswap_32(x) \