bitbang_spi: Add half-duplex optimizations

Currently, the core of bitbang_spi is a full-duplex SPI loop but in
practice this code is only ever used half-duplex. Spliting this code
into two half duplex loops allows us to optimize performance by reducing
communications and/or CPU pipeline stalls.

The speed up varies depending on how much the overhead of
getting/setting pins dominates execution time. For a USB bit bang driver
running on a 7th generation Core i5, the time to probe drops from ~7.7
seconds to ~6.7 seconds when this patch is applied.

Change-Id: I33b9f363716f651146c09113bda5fffe53b16738
Signed-off-by: Daniel Thompson <daniel.thompson@linaro.org>
Reviewed-on: https://review.coreboot.org/26947
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Nico Huber <nico.h@gmx.de>
1 file changed