kernel: resolve static analysis false positives

At least one static analysis tool is flagging a potential NULL
derefence in sys_clock_announce()'s tick processing loop where the
routine 'first()' is concerned. In practice, this does not occur as
...

  1. The code in question is protected by a spinlock.
  2. 'first()' does not change the contents of anything.

The code has consequently been tweaked to prevent similar such false
positives in the future.

Signed-off-by: Peter Mitsis <peter.mitsis@intel.com>