FAILED: patch "[PATCH] parisc: Remove unnecessary barriers from spinlock.h" failed to apply to 4.9-stable tree

The patch below does not apply to the 4.9-stable tree.
If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <[email protected]>.

thanks,

greg k-h

------------------ original commit in Linus's tree ------------------

From: John David Anglin <[email protected]>
Date: Sun, 12 Aug 2018 16:31:17 -0400
Subject: [PATCH] parisc: Remove unnecessary barriers from spinlock.h
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Now that mb() is an instruction barrier, it will slow performance if we issue unnecessary barriers.

The spinlock defines have a number of unnecessary barriers.  The __ldcw() define is both a hardware and compiler barrier.  The mb() barriers in the routines using __ldcw() serve no purpose.

The only barrier needed is the one in arch_spin_unlock().  We need to ensure all accesses are complete prior to releasing the lock.

index 6f84b6acc86e..8a63515f03bf 100644
@@ -20,7 +20,6 @@ static inline void arch_spin_lock_flags(arch_spinlock_t *x,
 {
 	volatile unsigned int *a;
 
-	mb();
 	a = __ldcw_align(x);
 	while (__ldcw(a) == 0)
 		while (*a == 0)

@@ -30,17 +29,16 @@ static inline void arch_spin_lock_flags(arch_spinlock_t *x,
 				local_irq_disable();
 			} else
 				cpu_relax();
-	mb();
 }
 #define arch_spin_lock_flags arch_spin_lock_flags
 
 static inline void arch_spin_unlock(arch_spinlock_t *x)
 {
 	volatile unsigned int *a;
-	mb();
+
 	a = __ldcw_align(x);
-	*a = 1;
 	mb();
+	*a = 1;
 }
 
 static inline int arch_spin_trylock(arch_spinlock_t *x)

@@ -48,10 +46,8 @@ static inline int arch_spin_trylock(arch_spinlock_t *x)
 	volatile unsigned int *a;
 	int ret;
 
-	mb();
 	a = __ldcw_align(x);
         ret = __ldcw(a) != 0;
-	mb();
 
 	return ret;
 }