[HWArith] Does it make sense to add an arithmetic negation unary op?

[qobilidop]

I've been reading https://circt.llvm.org/docs/Dialects/HWArith/RationaleHWArith/ and start to wonder if an arithmetic negation op also fits here?

Proposal for the type inference rules:

	Operand type	Result type
(U)	ui<a>	si<r>, r = a + 1
(S)	si<a>	si<r>, r = a + 1

This is essentially the existing subtraction rules with a == 0. But since it's said that 0-width values are not supported yet. Maybe it's useful to add arithmetic negation on its own?