perf, ARM: Modify kuser rmb() call to compile for Thumb-2

The Thumb-2 instruction set does not provide an encoding
for sub pc, r0, #95 as present in the rmb() definition used
by perf. This results in compilation failure when using a
compiler targetting an instruction set other than ARM.

This patch redefines rmb() for ARM by casting the address
of the kuser helper to a function pointer, therefore getting
the compiler to take care of making the call.

Patch taken against tip/master.
Signed-off-by: Will Deacon <will.deacon@arm.com>
Cc: Russell King - ARM Linux <linux@arm.linux.org.uk>
Cc: Jamie Iles <jamie.iles@picochip.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
LKML-Reference: <1267616878-2154-1-git-send-email-will.deacon@arm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>

tools/perf/perf.h

@@ -65,9 +65,7 @@
  * Use the __kuser_memory_barrier helper in the CPU helper page. See
  * arch/arm/kernel/entry-armv.S in the kernel source for details.
  */
-#define rmb()		asm volatile("mov r0, #0xffff0fff; mov lr, pc;" \
-				     "sub pc, r0, #95" ::: "r0", "lr", "cc", \
-				     "memory")
+#define rmb()		((void(*)(void))0xffff0fa0)()
 #define cpu_relax()	asm volatile("":::"memory")
 #endif