AVX2 variant reads next blocks while calculating current block.
Avoid reading past the end of data, by switching back to original,
for last blocks.

Fixes #15617.

Change-Id: I04fa2d83f1b47995117c77b4a3d403a7dff594d4
Reviewed-on: https://go-review.googlesource.com/23138Reviewed-by: Keith Randall <khr@golang.org>
Run-TryBot: Ilya Tocar <ilya.tocar@intel.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>