arm64: add missing conversion to __wsum in ip_fast_csum()

ARM64 implementation of ip_fast_csum() do most of the work
in 128 or 64 bit and call csum_fold() to finalize. csum_fold()
itself take a __wsum argument, to insure that this value is
always a 32bit native-order value.

Fix this by adding the sadly needed '__force' to cast the native
'sum' to the type '__wsum'.
Signed-off-by: Luc Van Oostenryck <luc.vanoostenryck@gmail.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>

arch/arm64/include/asm/checksum.h

@@ -42,7 +42,7 @@ static inline __sum16 ip_fast_csum(const void *iph, unsigned int ihl)
 	} while (--ihl);
 
 	sum += ((sum >> 32) | (sum << 32));
-	return csum_fold(sum >> 32);
+	return csum_fold((__force u32)(sum >> 32));
 }
 #define ip_fast_csum ip_fast_csum