[AArch64] Change aarch64_neon_pmull{,64} intrinsic ISel through a new

SDNode.

How:
1) Add AArch64ISD::PMULL SDNode, and extend aarch64_neon_pmull intrinsic
   tablegen pattern for this SDNode.
2) For aarch64_neon_pmull64, canonicalize i64 operands to v1i64 vectors
   during legalization.
3) For {aarch64_neon_pmull, aarch64_neon_pmull64}, combine intrinsic to
   SDNode.

Why
1) Adding the SDNode makes it easier to canonicalize i64 inputs (required by
   aarch64_neon_pmull64) to vector inputs. Vector inputs carries lane
   information, which helps dag-combiner to combine nodes (e.g. rewrite to a
   better node to prepare for instruction selection) and instruction-selection
   to emit instructions that use higher-half inputs in place
   (i.e., no need to move lane 1 content to lane 0).
2) Using the SDNode for aarch64_neon_pmull64 is NFC, yet without this we
   have to move the definition of {PMULLv1i64, PMULLv2i64} out of its
   current group of records without gains.

Test cases are commented with what is being tested in
`aarch64-pmull2.ll` and `pmull-ldr-merge.ll` under directory
`llvm/test/CodeGen/AArch64`.

Differential Revision: https://reviews.llvm.org/D131047