sf: set common timeouts in seconds, not milliseconds

Since timeouts are only hit when there is a problem in the system, we
don't want to prematurely timeout on a functioning setup.  Thus having
low timeouts (in milliseconds) doesn't gain us anything in the production
case, but rather increases likely hood of causing problems where none
otherwise exist.

Signed-off-by: Mike Frysinger <vapier@gentoo.org>
CC: Haavard Skinnemoen <haavard.skinnemoen@atmel.com>