Add signed/unsigned 8-bit and 16-bit Representations

src/ia32/lithium-ia32.cc

Index: src/ia32/lithium-ia32.cc
diff --git a/src/ia32/lithium-ia32.cc b/src/ia32/lithium-ia32.cc
index fdddef3f479a6eb3af46d21f3b98bb9e2d8877c0..f8cfb140018f0c72de8d648757415ec9f162e037 100644
--- a/src/ia32/lithium-ia32.cc
+++ b/src/ia32/lithium-ia32.cc
@@ -2422,7 +2422,8 @@ LInstruction* LChunkBuilder::DoStoreNamedField(HStoreNamedField* instr) {
       !(FLAG_track_double_fields && instr->field_representation().IsDouble());
 
   LOperand* val;
-  if (instr->field_representation().IsByte()) {
+  if (instr->field_representation().IsInteger8() ||
+      instr->field_representation().IsUInteger8()) {
     // mov_b requires a byte register (i.e. any of eax, ebx, ecx, edx).
     // Just force the value to be in eax and we're safe here.

[Yang, 2013/11/07 09:32:43]

Don't we have the same issue with 16-bit accesses?

How do we actually manage for x64?

[Benedikt Meurer, 2013/11/07 12:57:55]

In long mode, the REX prefix can be used to access the least significant 8-bit
of edi, esi, ebp and esp.

[danno, 2013/11/07 13:19:57]

No, this issue doesn't exist with 16-bit access. In x64, with a REX prefix all
registers can be used in a byte register context.

[danno, 2013/11/07 13:19:57]

Done.

     val = UseFixed(instr->value(), eax);