- Shuffle canonicalization improved to reverse operands to match
  more architectural shuffles.
- Handles shuffles where the order of operands is reversed.
- Adds tests for non-canonical shuffles, and for swizzles.
- Improves TryMatchConcat method.
- Substantially rewrites shuffles on ia32 to better handle swizzles
  and fix bugs on reversed shuffles where source registers are
  overwritten.
- Adds Palignr macro-assembler instructions for ia32.

Bug: v8:6020
Change-Id: I8e43a1e7650057c66690af1504b67509a1437d75
Reviewed-on: https://chromium-review.googlesource.com/1070934
Commit-Queue: Bill Budge <bbudge@chromium.org>
Reviewed-by: Deepti Gandluri <gdeepti@chromium.org>
Reviewed-by: Martyn Capewell <martyn.capewell@arm.com>
Reviewed-by: Jaroslav Sevcik <jarin@chromium.org>
Cr-Commit-Position: refs/heads/master@{#53648}