[X86][SSE] isHorizontalBinOp - use getTargetShuffleInputs helper (REAPPLIED)

Use the getTargetShuffleInputs helper for all shuffle decoding

Reapplied (after reversion in rGfa0aff6d) with fix+test for subvector splitting - we weren't accounting for peeking through bitcasts changing the vector element count of the shuffle sources.