Improve ethread atomics

The ethread atomics API now also provide double word size atomics.
Double word size atomics are implemented using native atomic
instructions on x86 (when the cmpxchg8b instruction is available)
and on x86_64 (when the cmpxchg16b instruction is available). On
other hardware where 32-bit atomics or word size atomics are
available, an optimized fallback is used; otherwise, a spinlock,
or a mutex based fallback is used.

The ethread library now performs runtime tests for presence of
hardware features, such as for example SSE2 instructions, instead
of requiring this to be determined at compile time.

There are now functions implementing each atomic operation with the
following implied memory barrier semantics: none, read, write,
acquire, release, and full. Some of the operation-barrier
combinations aren't especially useful. But instead of filtering
useful ones out, and potentially miss a useful one, we implement
them all.

A much smaller set of functionality for native atomics are required
to be implemented than before. More or less only cmpxchg and a
membar macro are required to be implemented for each atomic size.
Other functions will automatically be constructed from these. It is,
of course, often wise to implement more that this if possible from a
performance perspective.

2 files changed, 28 insertions, 16 deletions

diff --git a/erts/lib_src/win/ethr_event.c b/erts/lib_src/win/ethr_event.c
index 68f093f49c..bc2f635c26 100644
--- a/erts/lib_src/win/ethr_event.c
+++ b/erts/lib_src/win/ethr_event.c
@@ -1,7 +1,7 @@
 /*
  * %CopyrightBegin%
  *
- * Copyright Ericsson AB 2009-2010. All Rights Reserved.
+ * Copyright Ericsson AB 2009-2011. All Rights Reserved.
  *
  * The contents of this file are subject to the Erlang Public License,
  * Version 1.1, (the "License"); you may not use this file except in
@@ -28,13 +28,10 @@
 
 /* --- Windows implementation of thread events ------------------------------ */
 
-#pragma intrinsic(_InterlockedExchangeAdd)
-#pragma intrinsic(_InterlockedCompareExchange)
-
 int
 ethr_event_init(ethr_event *e)
 {
-    e->state = ETHR_EVENT_OFF__;
+    ethr_atomic32_init(&e->state, ETHR_EVENT_OFF__);
     e->handle = CreateEvent(NULL, FALSE, FALSE, NULL);
     if (e->handle == INVALID_HANDLE_VALUE)
 	return ethr_win_get_errno__();
@@ -63,7 +60,6 @@ ethr_event_reset(ethr_event *e)
 static ETHR_INLINE int
 wait(ethr_event *e, int spincount)
 {
-    LONG state;
     DWORD code;
     int sc, res, until_yield = ETHR_YIELD_AFTER_BUSY_LOOPS;
 
@@ -73,13 +69,9 @@ wait(ethr_event *e, int spincount)
     sc = spincount;
 
     while (1) {
-	long on;
+	ethr_sint32_t state;
 	while (1) {
-#if ETHR_READ_AND_SET_WITHOUT_INTERLOCKED_OP__
-	    state = e->state;
-#else
-	    state = _InterlockedExchangeAdd(&e->state, (LONG) 0);
-#endif
+	    state = ethr_atomic32_read(&e->state);
 	    if (state == ETHR_EVENT_ON__)
 		return 0;
 	    if (sc == 0)
@@ -95,9 +87,9 @@ wait(ethr_event *e, int spincount)
 	}
 
 	if (state != ETHR_EVENT_OFF_WAITER__) {
-	    state = _InterlockedCompareExchange(&e->state,
-						ETHR_EVENT_OFF_WAITER__,
-						ETHR_EVENT_OFF__);
+	    state = ethr_atomic32_cmpxchg(&e->state,
+					  ETHR_EVENT_OFF_WAITER__,
+					  ETHR_EVENT_OFF__);
 	    if (state == ETHR_EVENT_ON__)
 		return 0;
 	    ETHR_ASSERT(state == ETHR_EVENT_OFF__);
diff --git a/erts/lib_src/win/ethread.c b/erts/lib_src/win/ethread.c
index 789a360b11..3abda6de4c 100644
--- a/erts/lib_src/win/ethread.c
+++ b/erts/lib_src/win/ethread.c
@@ -1,7 +1,7 @@
 /*
  * %CopyrightBegin%
  *
- * Copyright Ericsson AB 2010. All Rights Reserved.
+ * Copyright Ericsson AB 2010-2011. All Rights Reserved.
  *
  * The contents of this file are subject to the Erlang Public License,
  * Version 1.1, (the "License"); you may not use this file except in
@@ -35,6 +35,7 @@
 #include <winerror.h>
 #include <stdio.h>
 #include <limits.h>
+#include <intrin.h>
 
 #define ETHR_INLINE_FUNC_NAME_(X) X ## __
 #define ETHREAD_IMPL__
@@ -158,6 +159,25 @@ ethr_abort__(void)
 #endif
 }
 
+#if defined(ETHR_X86_RUNTIME_CONF__)
+
+#pragma intrinsic(__cpuid)
+
+void
+ethr_x86_cpuid__(int *eax, int *ebx, int *ecx, int *edx)
+{
+    int CPUInfo[4];
+
+    __cpuid(CPUInfo, *eax);
+
+    *eax = CPUInfo[0];
+    *ebx = CPUInfo[1];
+    *ecx = CPUInfo[2];
+    *edx = CPUInfo[3];
+}
+
+#endif /* ETHR_X86_RUNTIME_CONF__ */
+
 /*
  * ----------------------------------------------------------------------------
  * Exported functions