>   Right. Could someone tag the MP-branch before any kernel API
> changes from -current, diverging from release-1-6, get merged in?

I don't quite understand the point of this.

For one the MP branch has a pretty bad race condition right now, making all
softinterrupts stall on some machines after some time of activity. Since the
proposed fix is still being tested, tagging now doesn't make sense to me.

Furthermore I don't see why using 1.6 as a base for an experimental branch
would give you any benefits over doing what MP-UPDATING suggests.


Martin