arm64: update 32-bit kuser helpers to ARMv8

This patch updates the barrier semantics in the kuser helper functions
to take advantage of the ARMv8 additions to AArch32, which are
guaranteed to be available in situations where these functions will be
called.

Note that this slightly changes the cmpxchg functions in that they are
no longer necessarily full barriers if they return 1. However, the
documentation only states they include their own barriers "as needed",
not that they are obligated to act as a full barrier for the caller.
Signed-off-by: Robin Murphy <robin.murphy@arm.com>
Acked-by: Will Deacon <will.deacon@arm.com>
CC: Matthew Leach <matthew.leach@arm.com>
CC: Dave Martin <dave.martin@arm.com>
CC: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>

arch/arm64/kernel/kuser32.S

@@ -38,33 +38,30 @@ __kuser_cmpxchg64:			// 0xffff0f60
 	.inst	0xe92d00f0		//	push		{r4, r5, r6, r7}
 	.inst	0xe1c040d0		//	ldrd		r4, r5, [r0]
 	.inst	0xe1c160d0		//	ldrd		r6, r7, [r1]
-	.inst	0xf57ff05f		//	dmb		sy
-	.inst	0xe1b20f9f		// 1:	ldrexd		r0, r1, [r2]
+	.inst	0xe1b20e9f		// 1:	ldaexd		r0, r1, [r2]
 	.inst	0xe0303004		//	eors		r3, r0, r4
 	.inst	0x00313005		//	eoreqs		r3, r1, r5
-	.inst	0x01a23f96		//	strexdeq	r3, r6, [r2]
+	.inst	0x01a23e96		//	stlexdeq	r3, r6, [r2]
 	.inst	0x03330001		//	teqeq		r3, #1
 	.inst	0x0afffff9		//	beq		1b
-	.inst	0xf57ff05f		//	dmb		sy
 	.inst	0xe2730000		//	rsbs		r0, r3, #0
 	.inst	0xe8bd00f0		//	pop		{r4, r5, r6, r7}
 	.inst	0xe12fff1e		//	bx		lr
 
 	.align	5
 __kuser_memory_barrier:			// 0xffff0fa0
-	.inst	0xf57ff05f		//	dmb		sy
+	.inst	0xf57ff05b		//	dmb		ish
 	.inst	0xe12fff1e		//	bx		lr
 
 	.align	5
 __kuser_cmpxchg:			// 0xffff0fc0
-	.inst	0xf57ff05f		//	dmb		sy
-	.inst	0xe1923f9f		// 1:	ldrex		r3, [r2]
+	.inst	0xe1923e9f		// 1:	ldaex		r3, [r2]
 	.inst	0xe0533000		//	subs		r3, r3, r0
-	.inst	0x01823f91		//	strexeq	r3, r1, [r2]
+	.inst	0x01823e91		//	stlexeq		r3, r1, [r2]
 	.inst	0x03330001		//	teqeq		r3, #1
 	.inst	0x0afffffa		//	beq		1b
 	.inst	0xe2730000		//	rsbs		r0, r3, #0
-	.inst	0xeaffffef		//	b		<__kuser_memory_barrier>
+	.inst	0xe12fff1e		//	bx		lr
 
 	.align	5
 __kuser_get_tls:			// 0xffff0fe0