Allow repeated waitms commands without resetting

At present issuing several 'waitms 1000' commands immediately after each
other trips the watchdog. Add a watchdog reload to avoid this.

Also document the behaviour in the command help.

BUG=b:72542719
BRANCH=none
TEST=manually on grunt, pasting these three lines in:
  waitms 1000
  waitms 1000
  waitms 1000
and see that it does not reset now.
Change-Id: I453708299e4e26c1bbdb5fc406f26e916e7389af
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/955927
Reviewed-by: Edward Hill <ecgh@chromium.org>
Reviewed-by: Martin Roth <martinroth@chromium.org>

common/timer.c

@@ -13,6 +13,7 @@
 #include "util.h"
 #include "task.h"
 #include "timer.h"
+#include "watchdog.h"
 
 #define TIMER_SYSJUMP_TAG 0x4d54  /* "TM" */
 
@@ -265,13 +266,25 @@ static int command_wait(int argc, char **argv)
 	if (*e)
 		return EC_ERROR_PARAM1;
 
+	/*
+	 * Waiting for too long (e.g. 3s) will cause the EC to reset due to a
+	 * watchdog timeout. This is intended behaviour and is in fact used by
+	 * a FAFT test to check that the watchdog timer is working.
+	 */
 	udelay(i * 1000);
 
+	/*
+	 * Reload the watchdog so that issuing multiple small waitms commands
+	 * quickly one after the other will not cause a reset.
+	 */
+	watchdog_reload();
+
 	return EC_SUCCESS;
 }
+/* Typically a large delay (e.g. 3s) will cause a reset */
 DECLARE_CONSOLE_COMMAND(waitms, command_wait,
 			"msec",
-			"Busy-wait for msec");
+			"Busy-wait for msec (large delays will reset)");
 #endif
 
 #ifdef CONFIG_CMD_FORCETIME