[PATCH] KVM: Micro-optimize pvclock_clocksource_read

[Li RongQing]

Compare the return of atomic64_cmpxchg with previous save for last,
to reduce the number of while-loop iterations with atomic operations
from two to one in the most common situation

Original patch at:
https://kvm.vger.kernel.narkive.com/WraXedaQ/patch-kvm-pvclock-clocksource-read-while-loop-optimization

Signed-off-by: Li RongQing <lirongqing@xxxxxxxxx>
---
 arch/x86/kernel/pvclock.c |    5 +++--
 1 files changed, 3 insertions(+), 2 deletions(-)

diff --git a/arch/x86/kernel/pvclock.c b/arch/x86/kernel/pvclock.c
index eda37df..98d106f 100644
--- a/arch/x86/kernel/pvclock.c
+++ b/arch/x86/kernel/pvclock.c
@@ -67,8 +67,8 @@ u8 pvclock_read_flags(struct pvclock_vcpu_time_info *src)
 u64 pvclock_clocksource_read(struct pvclock_vcpu_time_info *src)
 {
 	unsigned version;
+	u64 last, save;
 	u64 ret;
-	u64 last;
 	u8 flags;
 
 	do {
@@ -104,8 +104,9 @@ u64 pvclock_clocksource_read(struct pvclock_vcpu_time_info *src)
 	do {
 		if (ret < last)
 			return last;
+		save = last;
 		last = atomic64_cmpxchg(&last_value, last, ret);
-	} while (unlikely(last != ret));
+	} while (unlikely(last != save));
 
 	return ret;
 }
-- 
1.7.1