arm64: cmpxchg: truncate sub-word signed types before comparison

When performing a cmpxchg operation on a signed sub-word type (e.g. s8),
we need to ensure that the upper register bits of the "old" value used
for comparison are zeroed, otherwise we may erroneously fail the cmpxchg
which may even be interpreted as success by the caller (if the compiler
performs the truncation as part of its check). This has been observed
in mod_state, where negative values where causing problems with
this_cpu_cmpxchg.

This patch fixes the issue by explicitly casting 8-bit and 16-bit "old"
values using unsigned types in our cmpxchg wrappers. 32-bit types can be
left alone, since the underlying asm makes use of W registers in this
case.

Reported-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>

arch/arm64/include/asm/cmpxchg.h

@@ -122,9 +122,9 @@ static inline unsigned long __cmpxchg(volatile void *ptr, unsigned long old,
 {
 	switch (size) {
 	case 1:
-		return __cmpxchg_case_1(ptr, old, new);
+		return __cmpxchg_case_1(ptr, (u8)old, new);
 	case 2:
-		return __cmpxchg_case_2(ptr, old, new);
+		return __cmpxchg_case_2(ptr, (u16)old, new);
 	case 4:
 		return __cmpxchg_case_4(ptr, old, new);
 	case 8:
@@ -141,9 +141,9 @@ static inline unsigned long __cmpxchg_mb(volatile void *ptr, unsigned long old,
 {
 	switch (size) {
 	case 1:
-		return __cmpxchg_case_mb_1(ptr, old, new);
+		return __cmpxchg_case_mb_1(ptr, (u8)old, new);
 	case 2:
-		return __cmpxchg_case_mb_2(ptr, old, new);
+		return __cmpxchg_case_mb_2(ptr, (u16)old, new);
 	case 4:
 		return __cmpxchg_case_mb_4(ptr, old, new);
 	case 8: