eSPI improvements scratchpad

[nathanaelhuffman]

• Need to deal with in-band reset. There's a bunch of stuff that should be reset by this that isn't
• Need to hold UART data until CRC good or do we just push it through and let the upper level deal?
• Need to deal with clearing out fifos that have stuff in them still when cs goes away. This could happen for various error cases and will break the syncronization of the command and response stuff and make errant decodes.
• Need to wire up some kind of logger stuff to the debug, maybe with some filters
• Need to wire up statistics of varying types