powerpc: Make eh value more explicit when using lwarx

Just like the first patch of this series, define a local 'eh' in order
to make the code clearer.

And IS_ENABLED() returns either 1 or 0 so no need to do
IS_ENABLED(CONFIG_PPC64) ? 1 : 0.

Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu>
[mpe: Use symbolic names, use 'n' constraint per Segher]
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/629befaa2d05e2922346e58a383886510d6af55a.1659430931.git.christophe.leroy@csgroup.eu

arch/powerpc/include/asm/atomic.h

@@ -140,17 +140,18 @@ static __always_inline bool
 arch_atomic_try_cmpxchg_lock(atomic_t *v, int *old, int new)
 {
 	int r, o = *old;
+	unsigned int eh = IS_ENABLED(CONFIG_PPC64);
 
 	__asm__ __volatile__ (
-"1:	lwarx	%0,0,%2,%5	# atomic_try_cmpxchg_acquire		\n"
+"1:	lwarx	%0,0,%2,%[eh]	# atomic_try_cmpxchg_acquire		\n"
 "	cmpw	0,%0,%3							\n"
 "	bne-	2f							\n"
 "	stwcx.	%4,0,%2							\n"
 "	bne-	1b							\n"
 "\t"	PPC_ACQUIRE_BARRIER "						\n"
 "2:									\n"
 	: "=&r" (r), "+m" (v->counter)
-	: "r" (&v->counter), "r" (o), "r" (new), "i" (IS_ENABLED(CONFIG_PPC64) ? 1 : 0)
+	: "r" (&v->counter), "r" (o), "r" (new), [eh] "n" (eh)
 	: "cr0", "memory");
 
 	if (unlikely(r != o))