arm64: Avoid masking "old" for LSE cmpxchg() implementation

The CAS instructions implicitly access only the relevant bits of the "old" argument, so there is no need for explicit masking via type-casting as there is in the LL/SC implementation. Move the casting into the LL/SC code and remove it altogether for the LSE implementation. Signed-off-by: Will Deacon <[email protected]> Signed-off-by: Adam W. Willis <[email protected]> Signed-off-by: Forenche <[email protected]> Signed-off-by: Carlos Ayrton Lopez Arroyo <[email protected]> Signed-off-by: clarencelol <[email protected]> Signed-off-by: Anush02198 <[email protected]>
akeromi · Jun 30, 2021 · 2d6963f · 2d6963f
1 parent d8afdf1
commit 2d6963f
Show file tree

Hide file tree

Showing 3 changed files with 12 additions and 4 deletions.
diff --git a/arch/arm64/include/asm/atomic_ll_sc.h b/arch/arm64/include/asm/atomic_ll_sc.h
@@ -257,6 +257,14 @@ __LL_SC_PREFIX(__cmpxchg_case_##name##sz(volatile void *ptr,		\
 	unsigned long tmp;						\
 	u##sz oldval;							\
 									\
+	/*								\
+	 * Sub-word sizes require explicit casting so that the compare  \
+	 * part of the cmpxchg doesn't end up interpreting non-zero	\
+	 * upper bits of the register containing "old".			\
+	 */								\
+	if (sz < 32)							\
+		old = (u##sz)old;					\
+									\
 	asm volatile(							\
 	"	prfm	pstl1strm, %[v]\n"				\
 	"1:	ld" #acq "xr" #sfx "\t%" #w "[oldval], %[v]\n"		\

diff --git a/arch/arm64/include/asm/atomic_lse.h b/arch/arm64/include/asm/atomic_lse.h
@@ -509,11 +509,11 @@ static inline long atomic64_dec_if_positive(atomic64_t *v)
 
 #define __CMPXCHG_CASE(w, sfx, name, sz, mb, cl...)			\
 static inline u##sz __cmpxchg_case_##name##sz(volatile void *ptr,	\
-					      unsigned long old,	\
+					      u##sz old,		\
 					      u##sz new)		\
 {									\
 	register unsigned long x0 asm ("x0") = (unsigned long)ptr;	\
-	register unsigned long x1 asm ("x1") = old;			\
+	register u##sz x1 asm ("x1") = old;				\
 	register u##sz x2 asm ("x2") = new;				\
 									\
 	asm volatile(							\

diff --git a/arch/arm64/include/asm/cmpxchg.h b/arch/arm64/include/asm/cmpxchg.h
@@ -123,9 +123,9 @@ static __always_inline unsigned long __cmpxchg##sfx(volatile void *ptr,	\
 {									\
 	switch (size) {							\
 	case 1:								\
-		return __cmpxchg_case##sfx##_8(ptr, (u8)old, new);	\
+		return __cmpxchg_case##sfx##_8(ptr, old, new);		\
 	case 2:								\
-		return __cmpxchg_case##sfx##_16(ptr, (u16)old, new);	\
+		return __cmpxchg_case##sfx##_16(ptr, old, new);		\
 	case 4:								\
 		return __cmpxchg_case##sfx##_32(ptr, old, new);		\
 	case 8:								\