[llvm] [AMDGPU] Improve codegen for copysign(x, fneg(y)) (PR #207178)

[Jay Foad via llvm-commits]

jayfoad wrote:

> the visitBITCAST convert the `bitcast(fcopysign(const, fneg(x))` to `and(bitcast(fneg(x)), signbit)` before any target specific combine can see the fcopysign node. Once folded we cannot match any AMDGPU specific combine, custom lowering or Isel pattern

Why couldn't you write an isel pattern that matches that?

https://github.com/llvm/llvm-project/pull/207178