[CIR] Add support for long double type

[Lancern]

This PR adds support for the long double type in C/C++. It includes a new CIR type !cir.long_double to represent the long double type. CIRGen and LLVMIR lowering support for the new type is also added.

Since the underlying floating point format used by a long double value is implementation-defined, the !cir.long_double type is parameterized to include information about the underlying floating point format. Specifically, a long double value may have one of the following formats:

1. IEEE-754 binary64 format (i.e. the same format used by double);
2. x87 80-bit floating point format;
3. IEEE-754 binary128 format;
4. PowerPC double double format.

This PR invents 3 more CIR types to represent the above floating-point formats, and !cir.long_double is parameterized by another CIR floating-point type which represents its underlying format:

• !cir.long_double<!cir.double> represents the 1st variant above;
• !cir.long_double<!cir.f80> represents the 2nd variant above;
• !cir.long_double<!cir.f128> represents the 3rd variant above;
• !cir.long_double<!cir.ppc_doubledouble> represents the 4th variant above.

[bcardosolopes]

Overall looks good to me, thanks! Comments inline.

@sitio-couto you might want to take a look in this PR, given our previous discussions on floating point support.

[Lancern]

According to:

clangir/clang/lib/CIR/CodeGen/TargetInfo.cpp

Lines 434 to 458 in 216b6d8

	switch (Triple.getArch()) {
	default:
	assert(false && "Target not yet supported!");
	case llvm::Triple::aarch64: {
	AArch64ABIInfo::ABIKind Kind = AArch64ABIInfo::AAPCS;
	assert(getTarget().getABI() == "aapcs" ||
	getTarget().getABI() == "darwinpcs" &&
	"Only Darwin supported for aarch64");
	Kind = AArch64ABIInfo::DarwinPCS;
	return SetCIRGenInfo(new AArch64TargetCIRGenInfo(genTypes, Kind));
	}
	case llvm::Triple::x86_64: {
	StringRef ABI = getTarget().getABI();
	X86AVXABILevel AVXLevel = (ABI == "avx512" ? X86AVXABILevel::AVX512
	: ABI == "avx" ? X86AVXABILevel::AVX
	: X86AVXABILevel::None);
	switch (Triple.getOS()) {
	default:
	assert(false && "OSType NYI");
	case llvm::Triple::Linux:
	return SetCIRGenInfo(new X86_64TargetCIRGenInfo(genTypes, AVXLevel));
	}
	}

Seems like that CIR now only supports x86_64-*-linux-* or aarch64-*-darwin-* target triples. For the former targets, long double always use !cir.f80 as the underlying format. For the latter, long double always use double as the underlying format. Thus for now I cannot add tests for !cir.f128 or !cir.ppc_doubledouble. So in the latest commit I removed these two types from the PR and only keeps a single new type !cir.f80. !cir.f128 and !cir.ppc_doubledouble may be added in the future when we add support for more targets.

[Lancern]

Rebased onto the latest main.

[bcardosolopes]

Awesome, LGTM. It would be nice to get some aliases if you have suggestions and would feel like adding them in a follow up PR.

[bcardosolopes]

I tried to solve the merge conflict over the interface but maybe that wasn't enough. Once you fix this I'll land!

[Lancern]

Rebased onto the latest main.