rework the quarantine: it doesn't really need to be accounted.
authorPierre Habouzit <pierre.habouzit@intersec.com>
Sat, 14 Jan 2012 20:36:45 +0000 (21:36 +0100)
committerPierre Habouzit <pierre.habouzit@intersec.com>
Sat, 14 Jan 2012 20:39:07 +0000 (21:39 +0100)
The notion of quarantine is purely virtual and we only care about it at
WAKE time. It significantly simplifies the fastpath of our code, namely
__pwqr_sb_update_state.

Also allow the WAKE command to directly unpark threads if that fits with
the concurrency level. I don't really expect userland to really use that,
but it doesn't break anything and makes sense.

Signed-off-by: Pierre Habouzit <pierre.habouzit@intersec.com>

No differences found