Index: sched_ule.4
===================================================================
RCS file: /home/ncvs/src/share/man/man4/sched_ule.4,v
retrieving revision 1.7
diff -u -r1.7 sched_ule.4
--- sched_ule.4	5 Jun 2007 01:31:04 -0000	1.7
+++ sched_ule.4	26 Sep 2007 15:38:57 -0000
@@ -36,22 +36,20 @@
 The
 .Nm
 scheduler
-is an experimental scheduler that provides a number of advanced scheduler
-features not present in
-.Xr sched_4bsd 4 ,
-the default system scheduler.
+provides a number of advanced scheduler features not present in
+.Xr sched_4bsd 4 .
 These features address SMP and interactivity and include:
 .Pp
 .Bl -bullet -compact -offset indent
 .It
 Thread CPU affinity.
 .It
-CPU topology awareness, including for hyper-threading.
+CPU topology awareness for hyper-threading.
 .It
-Per-CPU run queues.
+Per-CPU run queues and locks.
 .It
 Interactivity heuristics that detect interactive applications and schedules
-them preferentially under high load.
+them preferentially.
 .El
 .Pp
 The following sysctls are relevant to the operation of
@@ -59,15 +57,30 @@
 .Bl -tag -width indent
 .It Va kern.sched.name
 This read-only sysctl reports the name of the active scheduler.
+.It Va kern.sched.preempt_thresh
+Min priority for preemption, lower priorities have greater precedence.
+.It Va kern.sched.interact
+Base interactivity score threshold.  Higher values allow more expensive tasks
+to remain interactive.  Valid values are 0..100 and the final score is
+calculated by subtracting the process nice value.
+.It Va kern.sched.slice
+Slice size for timeshare threads in units of stathz, see kern.clockrate.
+.It Va kern.sched.topology
+Set to 1 when a topology has been specified by the MD code.
+.It Va kern.sched.steal_thresh
+Minimum load on a remote cpu before we'll migrate threads before idling.
+.It Va kern.sched.balance_secs
+Average frequence in seconds to run the long-term CPU load balancer.
+.It Va kern.sched.balance
+Enables or disables the long-term load balancer.
+.It Va kern.sched.affinity
+Number of hz ticks to keep thread affinity for.
+.It Va kern.sched.pick_pri
+Pick the target cpu based on priority rather than load.  Optimizes for
+throughput vs latency.
 .El
 .Pp
-The
-.Nm
-scheduler
-is disabled by default as it is experimental.
-See the
-.Sx BUGS
-section for more information.
+Some sysctls will be available only on systems supporting SMP.
 .Sh SEE ALSO
 .Xr sched_4bsd 4 ,
 .Xr sysctl 8