[WinEH] Delete addFnAttr("no-frame-pointer-elim") which seems no longer needed

[MaskRay]

Summary

It was added in rL238619.

[rnk]

lgtm

[pengfei]

In D71862#1795756, @rnk wrote:

lgtm

Does funclet not need rbp to access parent local variables now?

[rnk]

In D71862#1795766, @pengfei wrote:

In D71862#1795756, @rnk wrote:

lgtm

Does funclet not need rbp to access parent local variables now?

No, it does need to go via ebp/rbp, but this attribute appears to no longer be required to achieve that.

[pengfei]

In D71862#1795767, @rnk wrote:

In D71862#1795766, @pengfei wrote:

In D71862#1795756, @rnk wrote:

lgtm

Does funclet not need rbp to access parent local variables now?

No, it does need to go via ebp/rbp, but this attribute appears to no longer be required to achieve that.

Got it, thanks!

llvm/lib/Target/X86/X86WinEHState.cpp

bool WinEHStatePass::runOnFunction(Function &F) {
 
   Type *Int8PtrType = Type::getInt8PtrTy(TheModule->getContext());
   SetJmp3 = TheModule->getOrInsertFunction(
       "_setjmp3", FunctionType::get(
                       Type::getInt32Ty(TheModule->getContext()),
                       {Int8PtrType, Type::getInt32Ty(TheModule->getContext())},
                       /*isVarArg=*/true));
 
-  // Disable frame pointer elimination in this function.
-  // FIXME: Do the nested handlers need to keep the parent ebp in ebp, or can we
-  // use an arbitrary register?
-  F.addFnAttr("no-frame-pointer-elim", "true");
-
   emitExceptionRegistrationRecord(&F);
 
   // The state numbers calculated here in IR must agree with what we calculate
   // later on for the MachineFunction. In particular, if an IR pass deletes an
   // unreachable EH pad after this point before machine CFG construction, we
   // will be in trouble. If this assumption is ever broken, we should turn the
   // numbers into an immutable analysis pass.
   WinEHFuncInfo FuncInfo;