[tip:locking/core] locking/xchg/alpha: Remove superfluous memory barriers from the _local() variants

[tip-bot for Andrea Parri]

Commit-ID:  fbfcd0199170984bd3c2812e49ed0fe7b226959a
Gitweb: https://git.kernel.org/tip/fbfcd0199170984bd3c2812e49ed0fe7b226959a
Author: Andrea Parri 
AuthorDate: Tue, 27 Feb 2018 05:00:58 +0100
Committer:  Ingo Molnar 
CommitDate: Mon, 12 Mar 2018 10:59:03 +0100

locking/xchg/alpha: Remove superfluous memory barriers from the _local() 
variants

The following two commits:

  79d442461df74 ("locking/xchg/alpha: Clean up barrier usage by using smp_mb() 
in place of __ASM__MB")
  472e8c55cf662 ("locking/xchg/alpha: Fix xchg() and cmpxchg() memory ordering 
bugs")

... ended up adding unnecessary barriers to the _local() variants on Alpha,
which the previous code took care to avoid.

Fix them by adding the smp_mb() into the cmpxchg() macro rather than into the
cmpxchg() variants.

Reported-by: Will Deacon 
Signed-off-by: Andrea Parri 
Cc: Alan Stern 
Cc: Andrew Morton 
Cc: Ivan Kokshaysky 
Cc: Linus Torvalds 
Cc: Matt Turner 
Cc: Paul E. McKenney 
Cc: Peter Zijlstra 
Cc: Richard Henderson 
Cc: Thomas Gleixner 
Cc: linux-al...@vger.kernel.org
Fixes: 472e8c55cf662 ("locking/xchg/alpha: Fix xchg() and cmpxchg() memory 
ordering bugs")
Fixes: 79d442461df74 ("locking/xchg/alpha: Clean up barrier usage by using 
smp_mb() in place of __ASM__MB")
Link: 
http://lkml.kernel.org/r/1519704058-13430-1-git-send-email-parri.and...@gmail.com
Signed-off-by: Ingo Molnar 
---
 arch/alpha/include/asm/cmpxchg.h | 20 
 arch/alpha/include/asm/xchg.h| 27 ---
 2 files changed, 16 insertions(+), 31 deletions(-)

diff --git a/arch/alpha/include/asm/cmpxchg.h b/arch/alpha/include/asm/cmpxchg.h
index 8a2b331e43fe..6c7c39452471 100644
--- a/arch/alpha/include/asm/cmpxchg.h
+++ b/arch/alpha/include/asm/cmpxchg.h
@@ -38,19 +38,31 @@
 #define cmpxchg(type, args...) __cmpxchg ##type(args)
 #include 
 
+/*
+ * The leading and the trailing memory barriers guarantee that these
+ * operations are fully ordered.
+ */
 #define xchg(ptr, x)   \
 ({ \
+   __typeof__(*(ptr)) __ret;   \
__typeof__(*(ptr)) _x_ = (x);   \
-   (__typeof__(*(ptr))) __xchg((ptr), (unsigned long)_x_,  \
-sizeof(*(ptr)));   \
+   smp_mb();   \
+   __ret = (__typeof__(*(ptr)))\
+   __xchg((ptr), (unsigned long)_x_, sizeof(*(ptr)));  \
+   smp_mb();   \
+   __ret;  \
 })
 
 #define cmpxchg(ptr, o, n) \
 ({ \
+   __typeof__(*(ptr)) __ret;   \
__typeof__(*(ptr)) _o_ = (o);   \
__typeof__(*(ptr)) _n_ = (n);   \
-   (__typeof__(*(ptr))) __cmpxchg((ptr), (unsigned long)_o_,   \
-   (unsigned long)_n_, sizeof(*(ptr)));\
+   smp_mb();   \
+   __ret = (__typeof__(*(ptr))) __cmpxchg((ptr),   \
+   (unsigned long)_o_, (unsigned long)_n_, sizeof(*(ptr)));\
+   smp_mb();   \
+   __ret;  \
 })
 
 #define cmpxchg64(ptr, o, n)   \
diff --git a/arch/alpha/include/asm/xchg.h b/arch/alpha/include/asm/xchg.h
index e2b59fac5257..7adb80c6746a 100644
--- a/arch/alpha/include/asm/xchg.h
+++ b/arch/alpha/include/asm/xchg.h
@@ -12,10 +12,6 @@
  * Atomic exchange.
  * Since it can be used to implement critical sections
  * it must clobber "memory" (also for interrupts in UP).
- *
- * The leading and the trailing memory barriers guarantee that these
- * operations are fully ordered.
- *
  */
 
 static inline unsigned long
@@ -23,7 +19,6 @@ xchg(_u8, volatile char *m, unsigned long val)
 {
unsigned long ret, tmp, addr64;
 
-   smp_mb();
__asm__ __volatile__(
"   andnot  %4,7,%3\n"
"   insbl   %1,%4,%1\n"
@@ -38,7 +33,6 @@ xchg(_u8, volatile char *m, unsigned long val)
".previous"
: "=" (ret), "=" (val), "=" (tmp), "=" (addr64)
: "r" ((long)m), "1" (val) : 

[tip:locking/core] locking/xchg/alpha: Remove superfluous memory barriers from the _local() variants

[tip-bot for Andrea Parri]

Commit-ID:  fbfcd0199170984bd3c2812e49ed0fe7b226959a
Gitweb: https://git.kernel.org/tip/fbfcd0199170984bd3c2812e49ed0fe7b226959a
Author: Andrea Parri 
AuthorDate: Tue, 27 Feb 2018 05:00:58 +0100
Committer:  Ingo Molnar 
CommitDate: Mon, 12 Mar 2018 10:59:03 +0100

locking/xchg/alpha: Remove superfluous memory barriers from the _local() 
variants

The following two commits:

  79d442461df74 ("locking/xchg/alpha: Clean up barrier usage by using smp_mb() 
in place of __ASM__MB")
  472e8c55cf662 ("locking/xchg/alpha: Fix xchg() and cmpxchg() memory ordering 
bugs")

... ended up adding unnecessary barriers to the _local() variants on Alpha,
which the previous code took care to avoid.

Fix them by adding the smp_mb() into the cmpxchg() macro rather than into the
cmpxchg() variants.

Reported-by: Will Deacon 
Signed-off-by: Andrea Parri 
Cc: Alan Stern 
Cc: Andrew Morton 
Cc: Ivan Kokshaysky 
Cc: Linus Torvalds 
Cc: Matt Turner 
Cc: Paul E. McKenney 
Cc: Peter Zijlstra 
Cc: Richard Henderson 
Cc: Thomas Gleixner 
Cc: linux-al...@vger.kernel.org
Fixes: 472e8c55cf662 ("locking/xchg/alpha: Fix xchg() and cmpxchg() memory 
ordering bugs")
Fixes: 79d442461df74 ("locking/xchg/alpha: Clean up barrier usage by using 
smp_mb() in place of __ASM__MB")
Link: 
http://lkml.kernel.org/r/1519704058-13430-1-git-send-email-parri.and...@gmail.com
Signed-off-by: Ingo Molnar 
---
 arch/alpha/include/asm/cmpxchg.h | 20 
 arch/alpha/include/asm/xchg.h| 27 ---
 2 files changed, 16 insertions(+), 31 deletions(-)

diff --git a/arch/alpha/include/asm/cmpxchg.h b/arch/alpha/include/asm/cmpxchg.h
index 8a2b331e43fe..6c7c39452471 100644
--- a/arch/alpha/include/asm/cmpxchg.h
+++ b/arch/alpha/include/asm/cmpxchg.h
@@ -38,19 +38,31 @@
 #define cmpxchg(type, args...) __cmpxchg ##type(args)
 #include 
 
+/*
+ * The leading and the trailing memory barriers guarantee that these
+ * operations are fully ordered.
+ */
 #define xchg(ptr, x)   \
 ({ \
+   __typeof__(*(ptr)) __ret;   \
__typeof__(*(ptr)) _x_ = (x);   \
-   (__typeof__(*(ptr))) __xchg((ptr), (unsigned long)_x_,  \
-sizeof(*(ptr)));   \
+   smp_mb();   \
+   __ret = (__typeof__(*(ptr)))\
+   __xchg((ptr), (unsigned long)_x_, sizeof(*(ptr)));  \
+   smp_mb();   \
+   __ret;  \
 })
 
 #define cmpxchg(ptr, o, n) \
 ({ \
+   __typeof__(*(ptr)) __ret;   \
__typeof__(*(ptr)) _o_ = (o);   \
__typeof__(*(ptr)) _n_ = (n);   \
-   (__typeof__(*(ptr))) __cmpxchg((ptr), (unsigned long)_o_,   \
-   (unsigned long)_n_, sizeof(*(ptr)));\
+   smp_mb();   \
+   __ret = (__typeof__(*(ptr))) __cmpxchg((ptr),   \
+   (unsigned long)_o_, (unsigned long)_n_, sizeof(*(ptr)));\
+   smp_mb();   \
+   __ret;  \
 })
 
 #define cmpxchg64(ptr, o, n)   \
diff --git a/arch/alpha/include/asm/xchg.h b/arch/alpha/include/asm/xchg.h
index e2b59fac5257..7adb80c6746a 100644
--- a/arch/alpha/include/asm/xchg.h
+++ b/arch/alpha/include/asm/xchg.h
@@ -12,10 +12,6 @@
  * Atomic exchange.
  * Since it can be used to implement critical sections
  * it must clobber "memory" (also for interrupts in UP).
- *
- * The leading and the trailing memory barriers guarantee that these
- * operations are fully ordered.
- *
  */
 
 static inline unsigned long
@@ -23,7 +19,6 @@ xchg(_u8, volatile char *m, unsigned long val)
 {
unsigned long ret, tmp, addr64;
 
-   smp_mb();
__asm__ __volatile__(
"   andnot  %4,7,%3\n"
"   insbl   %1,%4,%1\n"
@@ -38,7 +33,6 @@ xchg(_u8, volatile char *m, unsigned long val)
".previous"
: "=" (ret), "=" (val), "=" (tmp), "=" (addr64)
: "r" ((long)m), "1" (val) : "memory");
-   smp_mb();
 
return ret;
 }
@@ -48,7 +42,6 @@ xchg(_u16, volatile short *m, unsigned long val)
 {
unsigned long ret, tmp, addr64;
 
-   smp_mb();
__asm__ __volatile__(
"   andnot  %4,7,%3\n"
"   inswl   %1,%4,%1\n"
@@ -63,7 +56,6 @@ xchg(_u16,