[CIR][CIRGen] support builtin signbit

[PikachuHyA]

This patch adds support for the __builtin_signbit intrinsic. The intrinsic requires special handling for PowerPC; however, since ClangIR does not currently support PowerPC, this handling is omitted in this implementation.

[PikachuHyA]

handle PPC fp128

clangir/clang/lib/CodeGen/CGBuiltin.cpp

Lines 668 to 685 in b53784a

	if (Ty->isPPC_FP128Ty()) {
	// We want the sign bit of the higher-order double. The bitcast we just
	// did works as if the double-double was stored to memory and then
	// read as an i128. The "store" will put the higher-order double in the
	// lower address in both little- and big-Endian modes, but the "load"
	// will treat those bits as a different part of the i128: the low bits in
	// little-Endian, the high bits in big-Endian. Therefore, on big-Endian
	// we need to shift the high bits down to the low before truncating.
	Width >>= 1;
	if (CGF.getTarget().isBigEndian()) {
	Value *ShiftCst = llvm::ConstantInt::get(IntTy, Width);
	V = CGF.Builder.CreateLShr(V, ShiftCst);
	}
	// We are truncating value in order to extract the higher-order
	// double, which we will be using to extract the sign from.
	IntTy = llvm::IntegerType::get(C, Width);
	V = CGF.Builder.CreateTrunc(V, IntTy);
	}

[PikachuHyA]

PPC test case is here
https://github.com/llvm/clangir/blob/main/clang/test/Analysis/builtin_signbit.cpp

[Lancern]

LGTM in general, some minor nits inline

[bcardosolopes]

Looks mostly good, we still need to track missing PPC support, see comments inline. Thanks for working on this

[bcardosolopes]

Thanks for the updates, few more comments

[bcardosolopes]

My suggestion is for you to first tackle the non-fp80 case, and add unrecheable for that case, in a follow up PR you tackle all the fp80 related shenenigans in one go. How about that?

[PikachuHyA]

My suggestion is for you to first tackle the non-fp80 case, and add unrecheable for that case, in a follow up PR you tackle all the fp80 related shenenigans in one go. How about that?

Thanks for your suggestion! I agree that it makes sense to tackle the non-fp80 case first.

[bcardosolopes]

For next rounds: during LLVM lowering we're better off erroring out return op.emitError(...) - I'm gonna merge this anyway now cause I know you're gonna be working on this soon anyways! Thanks for the fixes.