[llvm] [PowerPC] Do not generate `isel` instruction if target doesn't have this instruction (PR #72845)

[Kai Luo via llvm-commits]

bzEq wrote:

> is it necessary to keep pass expand-isel which also depends on hasISEL() in the ppc backend? And it would be good to have a perf test.

I don't see regression in SPEC2017 when `-mno-isel` and `-bplugin_opt:-mattr=-isel` are specified with PPCExpandISEL removed from the backend pipeline.
I'll investigate more with PPCExpandISEL removed and will post another patch to disable it if no further issue is found.

https://github.com/llvm/llvm-project/pull/72845