[PATCH v2 RESEND] x86: optimize memcpy_flushcache

[Mike Snitzer]

From: Mikulas Patocka <mpatocka@xxxxxxxxxx>
Subject: [PATCH v2] x86: optimize memcpy_flushcache

In the context of constant short length stores to persistent memory,
memcpy_flushcache suffers from a 2% performance degradation compared to
explicitly using the "movnti" instruction.

Optimize 4, 8, and 16 byte memcpy_flushcache calls to explicitly use the
movnti instruction with inline assembler.

Signed-off-by: Mikulas Patocka <mpatocka@xxxxxxxxxx>
Reviewed-by: Dan Williams <dan.j.williams@xxxxxxxxx>
Signed-off-by: Mike Snitzer <snitzer@xxxxxxxxxx>
---
 arch/x86/include/asm/string_64.h | 28 +++++++++++++++++++++++++++-
 arch/x86/lib/usercopy_64.c       |  4 ++--
 2 files changed, 29 insertions(+), 3 deletions(-)

diff --git a/arch/x86/include/asm/string_64.h b/arch/x86/include/asm/string_64.h
index 533f74c300c2..aaba83478cdc 100644
--- a/arch/x86/include/asm/string_64.h
+++ b/arch/x86/include/asm/string_64.h
@@ -147,7 +147,33 @@ memcpy_mcsafe(void *dst, const void *src, size_t cnt)
 
 #ifdef CONFIG_ARCH_HAS_UACCESS_FLUSHCACHE
 #define __HAVE_ARCH_MEMCPY_FLUSHCACHE 1
-void memcpy_flushcache(void *dst, const void *src, size_t cnt);
+void __memcpy_flushcache(void *dst, const void *src, size_t cnt);
+static __always_inline void memcpy_flushcache(void *dst, const void *src, size_t cnt)
+{
+	if (__builtin_constant_p(cnt)) {
+		switch (cnt) {
+		case 4:
+			asm volatile("movntil %1, %0"
+				     : "=m" (*(u32 *)dst)
+				     : "r" (*(u32 *)src));
+			return;
+		case 8:
+			asm volatile("movntiq %1, %0"
+				     : "=m" (*(u64 *)dst)
+				     : "r" (*(u64 *)src));
+			return;
+		case 16:
+			asm volatile("movntiq %1, %0"
+				     : "=m" (*(u64 *)dst)
+				     : "r" (*(u64 *)src));
+			asm volatile("movntiq %1, %0"
+				     : "=m" (*(u64 *)(dst + 8))
+				     : "r" (*(u64 *)(src + 8)));
+			return;
+		}
+	}
+	__memcpy_flushcache(dst, src, cnt);
+}
 #endif
 
 #endif /* __KERNEL__ */
diff --git a/arch/x86/lib/usercopy_64.c b/arch/x86/lib/usercopy_64.c
index 75d3776123cc..26f515aa3529 100644
--- a/arch/x86/lib/usercopy_64.c
+++ b/arch/x86/lib/usercopy_64.c
@@ -133,7 +133,7 @@ long __copy_user_flushcache(void *dst, const void __user *src, unsigned size)
 	return rc;
 }
 
-void memcpy_flushcache(void *_dst, const void *_src, size_t size)
+void __memcpy_flushcache(void *_dst, const void *_src, size_t size)
 {
 	unsigned long dest = (unsigned long) _dst;
 	unsigned long source = (unsigned long) _src;
@@ -196,7 +196,7 @@ void memcpy_flushcache(void *_dst, const void *_src, size_t size)
 		clean_cache_range((void *) dest, size);
 	}
 }
-EXPORT_SYMBOL_GPL(memcpy_flushcache);
+EXPORT_SYMBOL_GPL(__memcpy_flushcache);
 
 void memcpy_page_flushcache(char *to, struct page *page, size_t offset,
 		size_t len)
-- 
2.15.0