Shifting into the sign bit of a 32 bit int and then converting to a unsigned
type is less straight forward than just shifting an unsigned value.
opcodes/ChangeLog:
2016-05-31 Trevor Saunders <tbsaunde+binutils@tbsaunde.org>
* sh-opc.h (ARCH_SH_HAS_DSP): Make the shifted value an unsigned
constant.
+2016-05-31 Trevor Saunders <tbsaunde+binutils@tbsaunde.org>
+
+ * sh-opc.h (ARCH_SH_HAS_DSP): Make the shifted value an unsigned
+ constant.
+
2016-05-29 H.J. Lu <hongjiu.lu@intel.com>
PR gas/20145
#define arch_sh_no_co (1 << 28) /* Neither FPU nor DSP co-processor. */
#define arch_sh_sp_fpu (1 << 29) /* Single precision FPU. */
#define arch_sh_dp_fpu (1 << 30) /* Double precision FPU. */
-#define arch_sh_has_dsp (1 << 31)
+#define arch_sh_has_dsp (1u << 31)
#define arch_sh_co_mask MASK (28, 31)