These two subclasses of X86TargetMachine are basically identical.
The *only* thing that's different is the constructor. And that *only*
differs in the is64Bit argument that it passes to the X86TargetMachine
constructor. Replacing the hard-coded 'true' or 'false' with
'Triple(TT).getArch()==Triple::x86_64' makes them *actually* identical.
Can we just ditch the subclasses, move the fields and methods that they
share into the X86TargetMachine class, and use that instead? Or am I
missing something?
In the patch I'm about to post to llvm-commits, I *stop* using the
X86_64TargetMachine subclass entirely, and use X86_32TargetMachine for
the 64-bit target. And the 16-bit target too. And nothing seems to have
broken AFAICT...
--- a/lib/Target/X86/X86TargetMachine.cpp
+++ b/lib/Target/X86/X86TargetMachine.cpp
@@ -24,8 +24,9 @@ using namespace llvm;
extern "C" void LLVMInitializeX86Target() {
// Register the target.
+ RegisterTargetMachine<X86_32TargetMachine> W(TheX86_16Target);
RegisterTargetMachine<X86_32TargetMachine> X(TheX86_32Target);
- RegisterTargetMachine<X86_64TargetMachine> Y(TheX86_64Target);
+ RegisterTargetMachine<X86_32TargetMachine> Y(TheX86_64Target);
}
@@ -74,7 +75,7 @@ X86_32TargetMachine::X86_32TargetMachine(const Target &T, Stri
const TargetOptions &Options,
Reloc::Model RM, CodeModel::Model CM,
CodeGenOpt::Level OL)
- : X86TargetMachine(T, TT, CPU, FS, Options, RM, CM, OL, false),
+ : X86TargetMachine(T, TT, CPU, FS, Options, RM, CM, OL, Triple(TT).getArch())
DL(computeDataLayout(*getSubtargetImpl())),
InstrInfo(*this),