From 9efb74da4b972c69b54ca91d0c83e5078f3b99b0 Mon Sep 17 00:00:00 2001
From: Samy Al Bahra <sbahra@repnop.org>
Date: Thu, 21 May 2015 12:22:37 -0400
Subject: [PATCH] ck_hp: Implement ck_hp_set_fence.

On TSO architectures, this relies on atomic ordering guarantees
rather than a full barrier. On Pentium M, this results in
approximately 30% improvement in latency for stack.
---
 include/ck_hp.h | 15 +++++++++++++++
 1 file changed, 15 insertions(+)

diff --git a/include/ck_hp.h b/include/ck_hp.h
index d032f72..c7d8cfb 100644
--- a/include/ck_hp.h
+++ b/include/ck_hp.h
@@ -28,6 +28,7 @@
 #define CK_HP_H
 
 #include <ck_cc.h>
+#include <ck_md.h>
 #include <ck_pr.h>
 #include <ck_stack.h>
 
@@ -81,6 +82,20 @@ ck_hp_set(struct ck_hp_record *record, unsigned int i, void *pointer)
 	return;
 }
 
+CK_CC_INLINE static void
+ck_hp_set_fence(struct ck_hp_record *record, unsigned int i, void *pointer)
+{
+
+#ifdef CK_MD_TSO
+	ck_pr_fas_ptr(&record->pointers[i], pointer);
+#else
+	ck_pr_store_ptr(&record->pointers[i], pointer);
+	ck_pr_fence_memory();
+#endif
+
+	return;
+}
+
 CK_CC_INLINE static void
 ck_hp_clear(struct ck_hp_record *record)
 {