Split the Add, Sub, and Mul instruction opcodes into separate
integer and floating-point opcodes, introducing
FAdd, FSub, and FMul.
For now, the AsmParser, BitcodeReader, and IRBuilder all preserve
backwards compatability, and the Core LLVM APIs preserve backwards
compatibility for IR producers. Most front-ends won't need to change
immediately.
This implements the first step of the plan outlined here:
http://nondot.org/sabre/LLVMNotes/IntegerOverflow.txt
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@72897 91177308-0d34-0410-b5e6-96231b3b80d8
diff --git a/test/CodeGen/Generic/print-arith-fp.ll b/test/CodeGen/Generic/print-arith-fp.ll
index 87aa1a0..1e27061 100644
--- a/test/CodeGen/Generic/print-arith-fp.ll
+++ b/test/CodeGen/Generic/print-arith-fp.ll
@@ -24,9 +24,9 @@
%b_s = getelementptr [8 x i8]* @b_str, i64 0, i64 0 ; <i8*> [#uses=1]
call i32 (i8*, ...)* @printf( i8* %a_s, double %a ) ; <i32>:1 [#uses=0]
call i32 (i8*, ...)* @printf( i8* %b_s, double %b ) ; <i32>:2 [#uses=0]
- %add_r = add double %a, %b ; <double> [#uses=1]
- %sub_r = sub double %a, %b ; <double> [#uses=1]
- %mul_r = mul double %a, %b ; <double> [#uses=1]
+ %add_r = fadd double %a, %b ; <double> [#uses=1]
+ %sub_r = fsub double %a, %b ; <double> [#uses=1]
+ %mul_r = fmul double %a, %b ; <double> [#uses=1]
%div_r = fdiv double %b, %a ; <double> [#uses=1]
%rem_r = frem double %b, %a ; <double> [#uses=1]
%add_s = getelementptr [12 x i8]* @add_str, i64 0, i64 0 ; <i8*> [#uses=1]