[llvm] Port `NVPTXTargetLowering::LowerCONCAT_VECTORS` to SelectionDAG (PR #120030)

[Justin Fargnoli via llvm-commits]

justinfargnoli wrote:

> It should still merge to the generic code. AMDGPU has essentially the same code in custom lowering (and I'm sure many out of tree targets do the same)

Happy to hear that @Esan5's hard work won't go to waste! 

@Esan5, if you'd prefer, feel free to split off the NVPTX aspect of this PR into a separate PR. I'll promptly approve and merge it. 

https://github.com/llvm/llvm-project/pull/120030