This was broken in two places: the register count was using the non-
transposed matrix row count. The block layout encoder was also not
set to transpose matrices, which was causing incorrect packing info
to be calculated in link.
BUG=angleproject:1923
Change-Id: I89094aa116fad4bda15f018498f8637520f12bd4
Reviewed-on: https://chromium-review.googlesource.com/454876
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Olli Etuaho <oetuaho@nvidia.com>
