Updated to fedora-glibc-20061025T1857cvs/fedora-glibc-2_5_90-1

13 files changed, 580 insertions, 217 deletions

diff --git a/sysdeps/generic/ldsodefs.h b/sysdeps/generic/ldsodefs.h
index b7d0f9b7e7..e9198bca05 100644
--- a/sysdeps/generic/ldsodefs.h
+++ b/sysdeps/generic/ldsodefs.h
@@ -38,6 +38,7 @@
 #include <bits/libc-lock.h>
 #include <hp-timing.h>
 #include <tls.h>
+#include <rtld-lowlevel.h>
 
 __BEGIN_DECLS
 
diff --git a/sysdeps/generic/rtld-lowlevel.h b/sysdeps/generic/rtld-lowlevel.h
new file mode 100644
index 0000000000..353fad28d4
--- /dev/null
+++ b/sysdeps/generic/rtld-lowlevel.h
@@ -0,0 +1 @@
+#error "Lowlevel primitives for ld.so not implemented"
diff --git a/sysdeps/generic/unsecvars.h b/sysdeps/generic/unsecvars.h
index a7378b742f..d5b8119c9c 100644
--- a/sysdeps/generic/unsecvars.h
+++ b/sysdeps/generic/unsecvars.h
@@ -18,6 +18,7 @@
   "LOCALDOMAIN\0"							      \
   "LOCPATH\0"								      \
   "MALLOC_TRACE\0"							      \
+  "NIS_PATH\0"								      \
   "NLSPATH\0"								      \
   "RESOLV_HOST_CONF\0"							      \
   "RES_OPTIONS\0"							      \
diff --git a/sysdeps/i386/i486/bits/atomic.h b/sysdeps/i386/i486/bits/atomic.h
index c748761758..a27734c37e 100644
--- a/sysdeps/i386/i486/bits/atomic.h
+++ b/sysdeps/i386/i486/bits/atomic.h
@@ -1,4 +1,4 @@
-/* Copyright (C) 2002, 2003, 2004 Free Software Foundation, Inc.
+/* Copyright (C) 2002, 2003, 2004, 2006 Free Software Foundation, Inc.
    This file is part of the GNU C Library.
    Contributed by Ulrich Drepper <drepper@redhat.com>, 2002.
 
@@ -18,6 +18,7 @@
    02111-1307 USA.  */
 
 #include <stdint.h>
+#include <tls.h>	/* For tcbhead_t.  */
 
 
 typedef int8_t atomic8_t;
@@ -76,6 +77,40 @@ typedef uintmax_t uatomic_max_t;
 		       : "r" (newval), "m" (*mem), "0" (oldval));	      \
      ret; })
 
+
+#define __arch_c_compare_and_exchange_val_8_acq(mem, newval, oldval) \
+  ({ __typeof (*mem) ret;						      \
+     __asm __volatile ("cmpl $0, %%gs:%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, %%gs:%P5\n\t"                                \
+                       "je 0f\n\t"                                            \
+                       "lock\n"                                               \
+                       "0:\tcmpxchgw %w2, %1"				      \
+		       : "=a" (ret), "=m" (*mem)			      \
+		       : "r" (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, %%gs:%P5\n\t"                                \
+                       "je 0f\n\t"                                            \
+                       "lock\n"                                               \
+                       "0:\tcmpxchgl %2, %1"				      \
+		       : "=a" (ret), "=m" (*mem)			      \
+		       : "r" (newval), "m" (*mem), "0" (oldval),	      \
+			 "i" (offsetof (tcbhead_t, multiple_threads)));	      \
+     ret; })
+
 /* XXX We do not really need 64-bit compare-and-exchange.  At least
    not in the moment.  Using it would mean causing portability
    problems since not many other 32-bit architectures have support for
@@ -85,6 +120,8 @@ typedef uintmax_t uatomic_max_t;
 #if 1
 # define __arch_compare_and_exchange_val_64_acq(mem, newval, oldval) \
   ({ __typeof (*mem) ret = *(mem); abort (); ret = (newval); ret = (oldval); })
+# define __arch_c_compare_and_exchange_val_64_acq(mem, newval, oldval) \
+  ({ __typeof (*mem) ret = *(mem); abort (); ret = (newval); ret = (oldval); })
 #else
 # ifdef __PIC__
 #  define __arch_compare_and_exchange_val_64_acq(mem, newval, oldval) \
@@ -100,6 +137,24 @@ typedef uintmax_t uatomic_max_t;
 					  & 0xffffffff),		      \
 			 "d" (((unsigned long long int) (oldval)) >> 32));    \
      ret; })
+
+#  define __arch_c_compare_and_exchange_val_64_acq(mem, newval, oldval) \
+  ({ __typeof (*mem) ret;						      \
+     __asm __volatile ("xchgl %2, %%ebx\n\t"				      \
+		       "cmpl $0, %%gs:%P7\n\t"				      \
+		       "je 0f\n\t"					      \
+		       "lock\n"						      \
+		       "0:\tcmpxchg8b %1\n\t"				      \
+		       "xchgl %2, %%ebx"				      \
+		       : "=A" (ret), "=m" (*mem)			      \
+		       : "DS" (((unsigned long long int) (newval))	      \
+			       & 0xffffffff),				      \
+			 "c" (((unsigned long long int) (newval)) >> 32),     \
+			 "m" (*mem), "a" (((unsigned long long int) (oldval)) \
+					  & 0xffffffff),		      \
+			 "d" (((unsigned long long int) (oldval)) >> 32),     \
+			 "i" (offsetof (tcbhead_t, multiple_threads)));	      \
+     ret; })
 # else
 #  define __arch_compare_and_exchange_val_64_acq(mem, newval, oldval) \
   ({ __typeof (*mem) ret;						      \
@@ -112,6 +167,22 @@ typedef uintmax_t uatomic_max_t;
 					  & 0xffffffff),		      \
 			 "d" (((unsigned long long int) (oldval)) >> 32));    \
      ret; })
+
+#  define __arch_c_compare_and_exchange_val_64_acq(mem, newval, oldval) \
+  ({ __typeof (*mem) ret;						      \
+     __asm __volatile ("cmpl $0, %%gs:%P7\n\t"				      \
+		       "je 0f\n\t"					      \
+		       "lock\n"						      \
+		       "0:\tcmpxchg8b %1"				      \
+		       : "=A" (ret), "=m" (*mem)			      \
+		       : "b" (((unsigned long long int) (newval))	      \
+			      & 0xffffffff),				      \
+			 "c" (((unsigned long long int) (newval)) >> 32),     \
+			 "m" (*mem), "a" (((unsigned long long int) (oldval)) \
+					  & 0xffffffff),		      \
+			 "d" (((unsigned long long int) (oldval)) >> 32),     \
+			 "i" (offsetof (tcbhead_t, multiple_threads)));	      \
+     ret; })
 # endif
 #endif
 
@@ -139,21 +210,24 @@ typedef uintmax_t uatomic_max_t;
      result; })
 
 
-#define atomic_exchange_and_add(mem, value) \
+#define __arch_exchange_and_add_body(lock, pfx, mem, value) \
   ({ __typeof (*mem) __result;						      \
      __typeof (value) __addval = (value);				      \
      if (sizeof (*mem) == 1)						      \
-       __asm __volatile (LOCK_PREFIX "xaddb %b0, %1"			      \
+       __asm __volatile (lock "xaddb %b0, %1"				      \
 			 : "=r" (__result), "=m" (*mem)			      \
-			 : "0" (__addval), "m" (*mem));			      \
+			 : "0" (__addval), "m" (*mem),			      \
+			   "i" (offsetof (tcbhead_t, multiple_threads)));     \
      else if (sizeof (*mem) == 2)					      \
-       __asm __volatile (LOCK_PREFIX "xaddw %w0, %1"			      \
+       __asm __volatile (lock "xaddw %w0, %1"				      \
 			 : "=r" (__result), "=m" (*mem)			      \
-			 : "0" (__addval), "m" (*mem));			      \
+			 : "0" (__addval), "m" (*mem),			      \
+			   "i" (offsetof (tcbhead_t, multiple_threads)));     \
      else if (sizeof (*mem) == 4)					      \
-       __asm __volatile (LOCK_PREFIX "xaddl %0, %1"			      \
+       __asm __volatile (lock "xaddl %0, %1"				      \
 			 : "=r" (__result), "=m" (*mem)			      \
-			 : "0" (__addval), "m" (*mem));			      \
+			 : "0" (__addval), "m" (*mem),			      \
+			   "i" (offsetof (tcbhead_t, multiple_threads)));     \
      else								      \
        {								      \
 	 __typeof (mem) __memp = (mem);					      \
@@ -161,41 +235,64 @@ typedef uintmax_t uatomic_max_t;
 	 __result = *__memp;						      \
 	 do								      \
 	   __tmpval = __result;						      \
-	 while ((__result = __arch_compare_and_exchange_val_64_acq	      \
+	 while ((__result = pfx##_compare_and_exchange_val_64_acq	      \
 		 (__memp, __result + __addval, __result)) == __tmpval);	      \
        }								      \
      __result; })
 
+#define atomic_exchange_and_add(mem, value) \
+  __arch_exchange_and_add_body (LOCK_PREFIX, __arch, mem, value)
+
+#define __arch_exchange_and_add_cprefix \
+  "cmpl $0, %%gs:%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, __arch_c,    \
+				mem, value)
+
+
+#define __arch_add_body(lock, pfx, mem, value) \
+  do {									      \
+    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 "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								      \
+      {									      \
+	__typeof (value) __addval = (value);				      \
+	__typeof (mem) __memp = (mem);					      \
+	__typeof (*mem) __oldval = *__memp;				      \
+	__typeof (*mem) __tmpval;					      \
+	do								      \
+	  __tmpval = __oldval;						      \
+	while ((__oldval = pfx##_compare_and_exchange_val_64_acq	      \
+		(__memp, __oldval + __addval, __oldval)) == __tmpval);	      \
+      }									      \
+  } while (0)
 
 #define atomic_add(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							      \
-	      {								      \
-		__typeof (value) __addval = (value);			      \
-		__typeof (mem) __memp = (mem);				      \
-		__typeof (*mem) __oldval = *__memp;			      \
-		__typeof (*mem) __tmpval;				      \
-		do							      \
-		  __tmpval = __oldval;					      \
-		while ((__oldval = __arch_compare_and_exchange_val_64_acq     \
-		       (__memp, __oldval + __addval, __oldval)) == __tmpval); \
-	      }								      \
-	    })
+  __arch_add_body (LOCK_PREFIX, __arch, mem, value)
+
+#define __arch_add_cprefix \
+  "cmpl $0, %%gs:%P3\n\tje 0f\n\tlock\n0:\t"
+
+#define catomic_add(mem, value) \
+  __arch_add_body (__arch_add_cprefix, __arch_c, mem, value)
 
 
 #define atomic_add_negative(mem, value) \
@@ -236,30 +333,42 @@ typedef uintmax_t uatomic_max_t;
      __result; })
 
 
-#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							      \
-	      {								      \
-		__typeof (mem) __memp = (mem);				      \
-		__typeof (*mem) __oldval = *__memp;			      \
-		__typeof (*mem) __tmpval;				      \
-		do							      \
-		  __tmpval = __oldval;					      \
-		while ((__oldval = __arch_compare_and_exchange_val_64_acq     \
-		       (__memp, __oldval + 1, __oldval)) == __tmpval);	      \
-	      }								      \
-	    })
+#define __arch_increment_body(lock,  pfx, 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								      \
+      {									      \
+	__typeof (mem) __memp = (mem);					      \
+	__typeof (*mem) __oldval = *__memp;				      \
+	__typeof (*mem) __tmpval;					      \
+	do								      \
+	  __tmpval = __oldval;						      \
+	while ((__oldval = pfx##_compare_and_exchange_val_64_acq	      \
+		(__memp, __oldval + 1, __oldval)) == __tmpval);		      \
+      }									      \
+  } while (0)
+
+#define atomic_increment(mem) __arch_increment_body (LOCK_PREFIX, __arch, mem)
+
+#define __arch_increment_cprefix \
+  "cmpl $0, %%gs:%P2\n\tje 0f\n\tlock\n0:\t"
+
+#define catomic_increment(mem) \
+  __arch_increment_body (__arch_increment_cprefix, __arch_c, mem)
 
 
 #define atomic_increment_and_test(mem) \
@@ -281,30 +390,42 @@ typedef uintmax_t uatomic_max_t;
      __result; })
 
 
-#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							      \
-	      {								      \
-		__typeof (mem) __memp = (mem);				      \
-		__typeof (*mem) __oldval = *__memp;			      \
-		__typeof (*mem) __tmpval;				      \
-		do							      \
-		  __tmpval = __oldval;					      \
-		while ((__oldval = __arch_compare_and_exchange_val_64_acq     \
-		       (__memp, __oldval - 1, __oldval)) == __tmpval); 	      \
-	      }								      \
-	    })
+#define __arch_decrement_body(lock, pfx, 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								      \
+      {									      \
+	__typeof (mem) __memp = (mem);					      \
+	__typeof (*mem) __oldval = *__memp;				      \
+	__typeof (*mem) __tmpval;					      \
+	do								      \
+	  __tmpval = __oldval;						      \
+	while ((__oldval = pfx##_compare_and_exchange_val_64_acq	      \
+		(__memp, __oldval - 1, __oldval)) == __tmpval); 	      \
+      }									      \
+  } while (0)
+
+#define atomic_decrement(mem) __arch_decrement_body (LOCK_PREFIX, __arch, mem)
+
+#define __arch_decrement_cprefix \
+  "cmpl $0, %%gs:%P2\n\tje 0f\n\tlock\n0:\t"
+
+#define catomic_decrement(mem) \
+  __arch_decrement_body (__arch_decrement_cprefix, __arch_c, mem)
 
 
 #define atomic_decrement_and_test(mem) \
@@ -327,21 +448,22 @@ typedef uintmax_t uatomic_max_t;
 
 
 #define atomic_bit_set(mem, bit) \
-  (void) ({ if (sizeof (*mem) == 1)					      \
-	      __asm __volatile (LOCK_PREFIX "orb %b2, %0"		      \
-				: "=m" (*mem)				      \
-				: "m" (*mem), "ir" (1 << (bit)));	      \
-	    else if (sizeof (*mem) == 2)				      \
-	      __asm __volatile (LOCK_PREFIX "orw %w2, %0"		      \
-				: "=m" (*mem)				      \
-				: "m" (*mem), "ir" (1 << (bit)));	      \
-	    else if (sizeof (*mem) == 4)				      \
-	      __asm __volatile (LOCK_PREFIX "orl %2, %0"		      \
-				: "=m" (*mem)				      \
-				: "m" (*mem), "ir" (1 << (bit)));	      \
-	    else							      \
-	      abort ();							      \
-	    })
+  do {									      \
+    if (sizeof (*mem) == 1)						      \
+      __asm __volatile (LOCK_PREFIX "orb %b2, %0"			      \
+			: "=m" (*mem)					      \
+			: "m" (*mem), "ir" (1 << (bit)));		      \
+    else if (sizeof (*mem) == 2)					      \
+      __asm __volatile (LOCK_PREFIX "orw %w2, %0"			      \
+			: "=m" (*mem)					      \
+			: "m" (*mem), "ir" (1 << (bit)));		      \
+    else if (sizeof (*mem) == 4)					      \
+      __asm __volatile (LOCK_PREFIX "orl %2, %0"			      \
+			: "=m" (*mem)					      \
+			: "m" (*mem), "ir" (1 << (bit)));		      \
+    else								      \
+      abort ();								      \
+  } while (0)
 
 
 #define atomic_bit_test_set(mem, bit) \
@@ -364,3 +486,51 @@ 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								      \
+      abort ();								      \
+  } 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								      \
+      abort ();								      \
+  } while (0)
+
+#define atomic_or(mem, mask) __arch_or_body (LOCK_PREFIX, mem, mask)
+
+#define __arch_or_cprefix \
+  "cmpl $0, %%gs:%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/mach/hurd/futimes.c b/sysdeps/mach/hurd/futimes.c
index ca687b8bdf..4628eee2c3 100644
--- a/sysdeps/mach/hurd/futimes.c
+++ b/sysdeps/mach/hurd/futimes.c
@@ -1,5 +1,5 @@
 /* futimes -- change access and modification times of open file.  Hurd version.
-   Copyright (C) 2002 Free Software Foundation, Inc.
+   Copyright (C) 2002, 2006 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
@@ -28,20 +28,24 @@
 int
 __futimes (int fd, const struct timeval tvp[2])
 {
-  struct timeval timevals[2];
+  union tv
+  {
+    struct timeval tv;
+    time_value_t tvt;
+  };
+  const union tv *u = (const union tv *) tvp;
+  union tv nulltv[2];
   error_t err;
 
   if (tvp == NULL)
     {
       /* Setting the number of microseconds to `-1' tells the
          underlying filesystems to use the current time.  */
-      timevals[1].tv_usec = timevals[0].tv_usec = (time_t)-1;
-      tvp = timevals;
+      nulltv[0].tvt.microseconds = nulltv[1].tvt.microseconds = -1;
+      u = nulltv;
     }
 
-  err = HURD_DPORT_USE (fd, __file_utimes (port,
-					   *(time_value_t *) &tvp[0],
-					   *(time_value_t *) &tvp[1]));
+  err = HURD_DPORT_USE (fd, __file_utimes (port, u[0].tvt, u[1].tvt));
   return err ? __hurd_dfail (fd, err) : 0;
 }
 weak_alias (__futimes, futimes)
diff --git a/sysdeps/mach/hurd/lutimes.c b/sysdeps/mach/hurd/lutimes.c
index cf89d8862f..bf5610c467 100644
--- a/sysdeps/mach/hurd/lutimes.c
+++ b/sysdeps/mach/hurd/lutimes.c
@@ -1,5 +1,5 @@
 /* lutimes -- change access and modification times of a symlink.  Hurd version.
-   Copyright (C) 2002 Free Software Foundation, Inc.
+   Copyright (C) 2002, 2006 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
@@ -28,7 +28,13 @@
 int
 __lutimes (const char *file, const struct timeval tvp[2])
 {
-  struct timeval timevals[2];
+  union tv
+  {
+    struct timeval tv;
+    time_value_t tvt;
+  };
+  const union tv *u = (const union tv *) tvp;
+  union tv nulltv[2];
   error_t err;
   file_t port;
 
@@ -36,15 +42,14 @@ __lutimes (const char *file, const struct timeval tvp[2])
     {
       /* Setting the number of microseconds to `-1' tells the
          underlying filesystems to use the current time.  */
-      timevals[1].tv_usec = timevals[0].tv_usec = (time_t)-1;
-      tvp = timevals;
+      nulltv[0].tvt.microseconds = nulltv[1].tvt.microseconds = -1;
+      u = nulltv;
     }
 
   port = __file_name_lookup (file, O_NOLINK, 0);
   if (port == MACH_PORT_NULL)
     return -1;
-  err = __file_utimes (port,
-		       *(time_value_t *) &tvp[0], *(time_value_t *) &tvp[1]);
+  err = __file_utimes (port, u[0].tvt, u[1].tvt);
   __mach_port_deallocate (__mach_task_self (), port);
   if (err)
     return __hurd_fail (err);
diff --git a/sysdeps/mach/hurd/utimes.c b/sysdeps/mach/hurd/utimes.c
index 74f3a342f4..fdc1427feb 100644
--- a/sysdeps/mach/hurd/utimes.c
+++ b/sysdeps/mach/hurd/utimes.c
@@ -1,4 +1,5 @@
-/* Copyright (C) 1991-1995, 97, 99, 2000 Free Software Foundation, Inc.
+/* Copyright (C) 1991-1995,1997,1999,2000,2006
+	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
@@ -28,7 +29,13 @@ __utimes (file, tvp)
      const char *file;
      const struct timeval tvp[2];
 {
-  struct timeval timevals[2];
+  union tv
+  {
+    struct timeval tv;
+    time_value_t tvt;
+  };
+  const union tv *u = (const union tv *) tvp;
+  union tv nulltv[2];
   error_t err;
   file_t port;
 
@@ -36,19 +43,17 @@ __utimes (file, tvp)
     {
       /* Setting the number of microseconds to `-1' tells the
          underlying filesystems to use the current time.  */
-      timevals[1].tv_usec = timevals[0].tv_usec = (time_t)-1;
-      tvp = timevals;
+      nulltv[0].tvt.microseconds = nulltv[1].tvt.microseconds = -1;
+      u = nulltv;
     }
 
   port = __file_name_lookup (file, 0, 0);
   if (port == MACH_PORT_NULL)
     return -1;
-  err = __file_utimes (port,
-		       *(time_value_t *) &tvp[0], *(time_value_t *) &tvp[1]);
+  err = __file_utimes (port, u[0].tvt, u[1].tvt);
   __mach_port_deallocate (__mach_task_self (), port);
   if (err)
     return __hurd_fail (err);
   return 0;
 }
-
 weak_alias (__utimes, utimes)
diff --git a/sysdeps/posix/getaddrinfo.c b/sysdeps/posix/getaddrinfo.c
index 9387e20db2..84a6293365 100644
--- a/sysdeps/posix/getaddrinfo.c
+++ b/sysdeps/posix/getaddrinfo.c
@@ -1395,10 +1395,10 @@ rfc3484_sort (const void *p1, const void *p2)
     {
       if (!(a1->source_addr_flags & in6ai_homeaddress)
 	  && (a2->source_addr_flags & in6ai_homeaddress))
-	return -1;
+	return 1;
       if ((a1->source_addr_flags & in6ai_homeaddress)
 	  && !(a2->source_addr_flags & in6ai_homeaddress))
-	return 1;
+	return -1;
     }
 
   /* Rule 5: Prefer matching label.  */
@@ -1435,11 +1435,11 @@ rfc3484_sort (const void *p1, const void *p2)
   if (a1->got_source_addr)
     {
       if (!(a1->source_addr_flags & in6ai_temporary)
-	  && (a1->source_addr_flags & in6ai_temporary))
+	  && (a2->source_addr_flags & in6ai_temporary))
 	return -1;
       if ((a1->source_addr_flags & in6ai_temporary)
-	  && !(a1->source_addr_flags & in6ai_temporary))
-	return -1;
+	  && !(a2->source_addr_flags & in6ai_temporary))
+	return 1;
 
       /* XXX Do we need to check anything beside temporary addresses?  */
     }
diff --git a/sysdeps/unix/sysv/linux/Versions b/sysdeps/unix/sysv/linux/Versions
index 7c015b1d62..137b44c2b8 100644
--- a/sysdeps/unix/sysv/linux/Versions
+++ b/sysdeps/unix/sysv/linux/Versions
@@ -126,6 +126,9 @@ libc {
   GLIBC_2.5 {
     splice; sync_file_range; tee; vmsplice;
   }
+  GLIBC_2.6 {
+    epoll_pwait;
+  }
   GLIBC_PRIVATE {
     # functions used in other libraries
     __syscall_rt_sigqueueinfo;
diff --git a/sysdeps/unix/sysv/linux/i386/sysdep.h b/sysdeps/unix/sysv/linux/i386/sysdep.h
index 90423d8434..5286676fc1 100644
--- a/sysdeps/unix/sysv/linux/i386/sysdep.h
+++ b/sysdeps/unix/sysv/linux/i386/sysdep.h
@@ -297,7 +297,7 @@
 			cfi_restore (edi); L(POPDI1):
 
 #define PUSHARGS_6	_PUSHARGS_6
-#define DOARGS_6	_DOARGS_6 (36)
+#define DOARGS_6	_DOARGS_6 (40)
 #define POPARGS_6	_POPARGS_6
 #define _PUSHARGS_6	pushl %ebp; cfi_adjust_cfa_offset (4); \
 			cfi_rel_offset (ebp, 0); L(PUSHBP1): _PUSHARGS_5
diff --git a/sysdeps/unix/sysv/linux/sys/epoll.h b/sysdeps/unix/sysv/linux/sys/epoll.h
index 68f173a04d..d8901f7d53 100644
--- a/sysdeps/unix/sysv/linux/sys/epoll.h
+++ b/sysdeps/unix/sysv/linux/sys/epoll.h
@@ -1,4 +1,4 @@
-/* Copyright (C) 2002, 2003, 2004, 2005 Free Software Foundation, Inc.
+/* Copyright (C) 2002, 2003, 2004, 2005, 2006 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
@@ -22,6 +22,14 @@
 #include <stdint.h>
 #include <sys/types.h>
 
+/* Get __sigset_t.  */
+#include <bits/sigset.h>
+
+#ifndef __sigset_t_defined
+# define __sigset_t_defined
+typedef __sigset_t sigset_t;
+#endif
+
 
 enum EPOLL_EVENTS
   {
@@ -105,6 +113,16 @@ extern int epoll_ctl (int __epfd, int __op, int __fd,
 extern int epoll_wait (int __epfd, struct epoll_event *__events,
 		       int __maxevents, int __timeout);
 
+
+/* Same as epoll_wait, but the thread's signal mask is temporarily
+   and atomically replaced with the one provided as parameter.
+
+   This function is a cancellation point and therefore not marked with
+   __THROW.  */
+extern int epoll_pwait (int __epfd, struct epoll_event *__events,
+			int __maxevents, int __timeout,
+			__const __sigset_t *__ss);
+
 __END_DECLS
 
 #endif /* sys/epoll.h */
diff --git a/sysdeps/unix/sysv/linux/syscalls.list b/sysdeps/unix/sysv/linux/syscalls.list
index fca29b26ff..5776673733 100644
--- a/sysdeps/unix/sysv/linux/syscalls.list
+++ b/sysdeps/unix/sysv/linux/syscalls.list
@@ -10,6 +10,7 @@ delete_module	EXTRA	delete_module	3	delete_module
 epoll_create	EXTRA	epoll_create	i:i	epoll_create
 epoll_ctl	EXTRA	epoll_ctl	i:iiip	epoll_ctl
 epoll_wait	EXTRA	epoll_wait	Ci:ipii	epoll_wait
+epoll_pwait	EXTRA	epoll_pwait	Ci:ipiipi	epoll_pwait
 fdatasync	-	fdatasync	i:i	fdatasync
 flock		-	flock		i:ii	__flock		flock
 fork		-	fork		i:	__libc_fork	__fork fork
diff --git a/sysdeps/x86_64/bits/atomic.h b/sysdeps/x86_64/bits/atomic.h
index 133a68d192..65d6b02008 100644
--- a/sysdeps/x86_64/bits/atomic.h
+++ b/sysdeps/x86_64/bits/atomic.h
@@ -1,4 +1,4 @@
-/* Copyright (C) 2002, 2003, 2004 Free Software Foundation, Inc.
+/* Copyright (C) 2002, 2003, 2004, 2006 Free Software Foundation, Inc.
    This file is part of the GNU C Library.
    Contributed by Ulrich Drepper <drepper@redhat.com>, 2002.
 
@@ -18,6 +18,7 @@
    02111-1307 USA.  */
 
 #include <stdint.h>
+#include <tls.h>	/* For tcbhead_t.  */
 
 
 typedef int8_t atomic8_t;
@@ -80,8 +81,54 @@ typedef uintmax_t uatomic_max_t;
   ({ __typeof (*mem) ret;						      \
      __asm __volatile (LOCK_PREFIX "cmpxchgq %q2, %1"			      \
 		       : "=a" (ret), "=m" (*mem)			      \
-		       : "r" ((long) (newval)), "m" (*mem),		      \
-			 "0" ((long) (oldval)));			      \
+		       : "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)));	      \
      ret; })
 
 
@@ -107,49 +154,76 @@ typedef uintmax_t uatomic_max_t;
      result; })
 
 
-#define atomic_exchange_and_add(mem, value) \
+#define __arch_exchange_and_add_body(lock, mem, value)			      \
   ({ __typeof (*mem) result;						      \
      if (sizeof (*mem) == 1)						      \
-       __asm __volatile (LOCK_PREFIX "xaddb %b0, %1"			      \
+       __asm __volatile (lock "xaddb %b0, %1"				      \
 			 : "=r" (result), "=m" (*mem)			      \
-			 : "0" (value), "m" (*mem));			      \
+			 : "0" (value), "m" (*mem),			      \
+			   "i" (offsetof (tcbhead_t, multiple_threads)));     \
      else if (sizeof (*mem) == 2)					      \
-       __asm __volatile (LOCK_PREFIX "xaddw %w0, %1"			      \
+       __asm __volatile (lock "xaddw %w0, %1"				      \
 			 : "=r" (result), "=m" (*mem)			      \
-			 : "0" (value), "m" (*mem));			      \
+			 : "0" (value), "m" (*mem),			      \
+			   "i" (offsetof (tcbhead_t, multiple_threads)));     \
      else if (sizeof (*mem) == 4)					      \
-       __asm __volatile (LOCK_PREFIX "xaddl %0, %1"			      \
+       __asm __volatile (lock "xaddl %0, %1"				      \
 			 : "=r" (result), "=m" (*mem)			      \
-			 : "0" (value), "m" (*mem));			      \
+			 : "0" (value), "m" (*mem),			      \
+			   "i" (offsetof (tcbhead_t, multiple_threads)));     \
      else								      \
-       __asm __volatile (LOCK_PREFIX "xaddq %q0, %1"			      \
+       __asm __volatile (lock "xaddq %q0, %1"				      \
 			 : "=r" (result), "=m" (*mem)			      \
-			 : "0" ((long) (value)), "m" (*mem));		      \
+			 : "0" ((long) (value)), "m" (*mem),		      \
+			   "i" (offsetof (tcbhead_t, multiple_threads)));     \
      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) \
-  (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));	      \
-	    })
+  __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)
 
 
 #define atomic_add_negative(mem, value) \
@@ -194,24 +268,37 @@ typedef uintmax_t uatomic_max_t;
      __result; })
 
 
-#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 __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_and_test(mem) \
@@ -235,24 +322,37 @@ typedef uintmax_t uatomic_max_t;
      __result; })
 
 
-#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 __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_and_test(mem) \
@@ -277,27 +377,28 @@ typedef uintmax_t uatomic_max_t;
 
 
 #define atomic_bit_set(mem, bit) \
-  (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)));	      \
-	    })
+  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)
 
 
 #define atomic_bit_test_set(mem, bit) \
@@ -322,3 +423,56 @@ 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)