X-Git-Url: https://git.liburcu.org/?p=urcu.git;a=blobdiff_plain;f=urcu%2Farch%2Fppc.h;h=8a96dd9deb50626efe6a9a8b75e6cbab7c0dc628;hp=d7317bb327c72afa48f087f50251cbe311a2ff62;hb=999991c6e4600c410181baea65bda9f406464872;hpb=9d2614f07691a813a3c560a6c0bcd0a7be854ed5

diff --git a/urcu/arch/ppc.h b/urcu/arch/ppc.h
index d7317bb..8a96dd9 100644
--- a/urcu/arch/ppc.h
+++ b/urcu/arch/ppc.h
@@ -24,6 +24,7 @@
 
 #include <urcu/compiler.h>
 #include <urcu/config.h>
+#include <urcu/syscall-compat.h>
 
 #ifdef __cplusplus
 extern "C" {
@@ -32,27 +33,66 @@ extern "C" {
 /* Include size of POWER5+ L3 cache lines: 256 bytes */
 #define CAA_CACHE_LINE_SIZE	256
 
-#define cmm_mb()    asm volatile("sync":::"memory")
+#ifdef __NO_LWSYNC__
+#define LWSYNC_OPCODE	"sync\n"
+#else
+#define LWSYNC_OPCODE	"lwsync\n"
+#endif
+
+/*
+ * Use sync for all cmm_mb/rmb/wmb barriers because lwsync does not
+ * preserve ordering of cacheable vs. non-cacheable accesses, so it
+ * should not be used to order with respect to MMIO operations.  An
+ * eieio+lwsync pair is also not enough for cmm_rmb, because it will
+ * order cacheable and non-cacheable memory operations separately---i.e.
+ * not the latter against the former.
+ */
+#define cmm_mb()         __asm__ __volatile__ ("sync":::"memory")
+
+/*
+ * lwsync orders loads in cacheable memory with respect to other loads,
+ * and stores in cacheable memory with respect to other stores.
+ * Therefore, use it for barriers ordering accesses to cacheable memory
+ * only.
+ */
+#define cmm_smp_rmb()    __asm__ __volatile__ (LWSYNC_OPCODE:::"memory")
+#define cmm_smp_wmb()    __asm__ __volatile__ (LWSYNC_OPCODE:::"memory")
 
 #define mftbl()						\
+	__extension__					\
 	({ 						\
 		unsigned long rval;			\
-		asm volatile("mftbl %0" : "=r" (rval));	\
+		__asm__ __volatile__ ("mftbl %0" : "=r" (rval));	\
 		rval;					\
 	})
 
 #define mftbu()						\
+	__extension__					\
 	({						\
 		unsigned long rval;			\
-		asm volatile("mftbu %0" : "=r" (rval));	\
+		__asm__ __volatile__ ("mftbu %0" : "=r" (rval));	\
+		rval;					\
+	})
+
+#define mftb()						\
+	__extension__					\
+	({						\
+		unsigned long long rval;		\
+		__asm__ __volatile__ ("mftb %0" : "=r" (rval));		\
 		rval;					\
 	})
 
 typedef unsigned long long cycles_t;
 
-static inline cycles_t caa_get_cycles (void)
+#ifdef __powerpc64__
+static inline cycles_t caa_get_cycles(void)
 {
-	long h, l;
+	return (cycles_t) mftb();
+}
+#else
+static inline cycles_t caa_get_cycles(void)
+{
+	unsigned long h, l;
 
 	for (;;) {
 		h = mftbu();
@@ -63,6 +103,7 @@ static inline cycles_t caa_get_cycles (void)
 			return (((cycles_t) h) << 32) + l;
 	}
 }
+#endif
 
 #ifdef __cplusplus 
 }