[PATCH v1 3/3] powerpc: Make eh value more explicit when using lwarx
Christophe Leroy
christophe.leroy at csgroup.eu
Tue Aug 2 19:02:38 AEST 2022
Just like 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 at csgroup.eu>
---
arch/powerpc/include/asm/atomic.h | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/arch/powerpc/include/asm/atomic.h b/arch/powerpc/include/asm/atomic.h
index 853dc86864f4..0204e77613ec 100644
--- a/arch/powerpc/include/asm/atomic.h
+++ b/arch/powerpc/include/asm/atomic.h
@@ -140,6 +140,7 @@ 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"
@@ -150,7 +151,7 @@ arch_atomic_try_cmpxchg_lock(atomic_t *v, int *old, int new)
"\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), "i" (eh)
: "cr0", "memory");
if (unlikely(r != o))
--
2.36.1
More information about the Linuxppc-dev
mailing list