diff options
| -rw-r--r-- | compiler/nativeGen/PPC/Ppr.hs | 8 |
1 files changed, 6 insertions, 2 deletions
diff --git a/compiler/nativeGen/PPC/Ppr.hs b/compiler/nativeGen/PPC/Ppr.hs index 08c02f01c6..9326813a78 100644 --- a/compiler/nativeGen/PPC/Ppr.hs +++ b/compiler/nativeGen/PPC/Ppr.hs @@ -764,10 +764,14 @@ pprInstr (FNEG reg1 reg2) = pprUnary (sLit "fneg") reg1 reg2 pprInstr (FCMP reg1 reg2) = hcat [ char '\t', - text "fcmpu\tcr0, ", + text "fcmpu\t0, ", -- Note: we're using fcmpu, not fcmpo -- The difference is with fcmpo, compare with NaN is an invalid operation. - -- We don't handle invalid fp ops, so we don't care + -- We don't handle invalid fp ops, so we don't care. + -- Morever, we use `fcmpu 0, ...` rather than `fcmpu cr0, ...` for + -- better portability since some non-GNU assembler (such as + -- IBM's `as`) tend not to support the symbolic register name cr0. + -- This matches the syntax that GCC seems to emit for PPC targets. pprReg reg1, text ", ", pprReg reg2 |