When CONFIG_MKBP_EVENT is enabled, the current code is incorrect because
we have a race condition when sending a new event (we force first the
interrupt, then send the actual event content to the mkbp event
framework which forces again the interrupt level).
If the software still called EC_CMD_MKBP_STATE while CONFIG_MKBP_EVENT
is enabled, this will kill the interrupt as soon as the FIFO is empty
even though other events are pending in order to be backward compatible
with firmware using the interrupt has a hint when polling the keyboard.
Signed-off-by: Vincent Palatin <vpalatin@chromium.org>
BRANCH=none
BUG=chrome-os-partner:33194
TEST=make buildall
fiddle with keyboard on Oak.
Change-Id: Iafaf4174124934328c4a0172adeca651e5551f28
Reviewed-on: https://chromium-review.googlesource.com/274070
Reviewed-by: Randall Spangler <rspangler@chromium.org>
Commit-Queue: Vincent Palatin <vpalatin@chromium.org>
Trybot-Ready: Vincent Palatin <vpalatin@chromium.org>
Tested-by: Vincent Palatin <vpalatin@chromium.org>
ebc3b8b93c