video: tegra: add utility function to compute framebuffer stride

video: tegra: add utility function to compute framebuffer stride

framebuffers will generally not be exactly width * bpp / 8 bytes wide;
on tegra, linearly-addressed framebuffers will generally be rounded
up so that the stride is a multiple of 16B (so that they are compatible
with rendering from the hardware engines), and tiled framebuffers
will be a multiple of the tile width (64B).

add a utility function to tegra_dc to compute the correct stride
given a width, bpp and pixel layout, and use this in set_par.

Change-Id: I803a55b49c12476f20d5644707899c3fe1336c2d
Signed-off-by: Gary King <gking@nvidia.com>