[$] Four short stories about preempt_count()

The discussion started out as a straightforwardpatch set from Thomas Gleixner making a minor change to how preemptioncounting is handled. The resulting discussion quickly spread out to covera number of issues relevant to core-kernel development in surprisingly fewmessages; each of those topics merits a quick look, starting with how thepreemption counter itself works. Sometimes a simple count turns out to notbe as simple as it seems.