[POWERPC] Cell timebase bug workaround
The Cell CPU timebase has an erratum. When reading the entire 64 bits
of the timebase with one mftb instruction, there is a handful of cycles
window during which one might read a value with the low order 32 bits
already reset to 0x00000000 but the high order bits not yet incremeted
by one. This fixes it by reading the timebase again until the low order
32 bits is no longer 0. That might introduce occasional latencies if
hitting mftb just at the wrong time, but no more than 70ns on a cell
blade, and that was considered acceptable.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Acked-by: Olof Johansson <olof@lixom.net>
Signed-off-by: Paul Mackerras <paulus@samba.org>
diff --git a/include/asm-powerpc/timex.h b/include/asm-powerpc/timex.h
index e3f08cf..92dedde 100644
--- a/include/asm-powerpc/timex.h
+++ b/include/asm-powerpc/timex.h
@@ -8,6 +8,7 @@
*/
#include <asm/cputable.h>
+#include <asm/reg.h>
#define CLOCK_TICK_RATE 1024000 /* Underlying HZ */
@@ -15,13 +16,11 @@
static inline cycles_t get_cycles(void)
{
+#ifdef __powerpc64__
+ return mftb();
+#else
cycles_t ret;
-#ifdef __powerpc64__
-
- __asm__ __volatile__("mftb %0" : "=r" (ret) : );
-
-#else
/*
* For the "cycle" counter we use the timebase lower half.
* Currently only used on SMP.
@@ -41,9 +40,8 @@
" .long 99b-98b\n"
".previous"
: "=r" (ret) : "i" (CPU_FTR_601));
-#endif
-
return ret;
+#endif
}
#endif /* __KERNEL__ */