[llvm-commits] CVS: llvm/lib/Target/X86/X86ISelPattern.cpp

[Chris Lattner]

Changes in directory llvm/lib/Target/X86:

X86ISelPattern.cpp updated: 1.20 -> 1.21
---
Log message:

Implement 8-bit multiply for X86.


---
Diffs of the changes:  (+6 -1)

Index: llvm/lib/Target/X86/X86ISelPattern.cpp
diff -u llvm/lib/Target/X86/X86ISelPattern.cpp:1.20 llvm/lib/Target/X86/X86ISelPattern.cpp:1.21
--- llvm/lib/Target/X86/X86ISelPattern.cpp:1.20	Sun Jan  9 14:20:29 2005
+++ llvm/lib/Target/X86/X86ISelPattern.cpp	Mon Jan 10 14:55:48 2005
@@ -1213,7 +1213,12 @@
     Tmp2 = SelectExpr(N.getOperand(1));
     switch (N.getValueType()) {
     default: assert(0 && "Cannot add this type!");
-    case MVT::i8: assert(0 && "FIXME: MUL i8 not implemented yet!");
+    case MVT::i8:
+      // Must use the MUL instruction, which forces use of AL.
+      BuildMI(BB, X86::MOV8rr, 1, X86::AL).addReg(Tmp1);
+      BuildMI(BB, X86::MUL8r, 1).addReg(Tmp2);
+      BuildMI(BB, X86::MOV8rr, 1, Result).addReg(X86::AL);
+      return Result;
     case MVT::i16: Opc = X86::IMUL16rr; break;
     case MVT::i32: Opc = X86::IMUL32rr; break;
     case MVT::f32: