[DAGCombine][X86][AArch64][AMDGPU] (x - y) + -1 -> add (xor y, -1), x fold.


[DAGCombine][X86][AArch64][AMDGPU] (x - y) + -1 -> add (xor y, -1), x fold. Try 2

This prevents regressions in next patch,
and somewhat recovers from the regression to AMDGPU test in D62223.

It is indeed not great that we leave vector decrement,
don't transform it into vector add all-ones..


This is a recommit, originally committed in rL361855, but reverted
to investigate test-suite compile-time hangs.

Reviewers: RKSimon, craig.topper, spatel, arsenm

Reviewed By: RKSimon, arsenm

Subscribers: kzhuravl, jvesely, wdng, nhaehnle, yaxunl, javed.absar, dstuttard, tpr, t-tye, kristof.beyls, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D62263