Compromise for critical*()/cpu_critical*() recommit. Cleanup the interrupt

disablement assumptions in kern_fork.c by adding another API call,
cpu_critical_fork_exit().  Cleanup the td_savecrit field by moving it
from MI to MD.  Temporarily move cpu_critical*() from <arch>/include/cpufunc.h
to <arch>/<arch>/critical.c (stage-2 will clean this up).

Implement interrupt deferral for i386 that allows interrupts to remain
enabled inside critical sections.  This also fixes an IPI interlock bug,
and requires uses of icu_lock to be enclosed in a true interrupt disablement.

This is the stage-1 commit.  Stage-2 will occur after stage-1 has stabilized,
and will move cpu_critical*() into its own header file(s) + other things.
This commit may break non-i386 architectures in trivial ways.  This should
be temporary.

Reviewed by:	core
Approved by:	core

1 files changed, 6 insertions, 0 deletions

diff --git a/sys/amd64/include/mptable.h b/sys/amd64/include/mptable.h
index 4f891db..50c91c7 100644
--- a/sys/amd64/include/mptable.h
+++ b/sys/amd64/include/mptable.h
@@ -2310,6 +2310,9 @@ ap_init(void)
 /*
  * For statclock, we send an IPI to all CPU's to have them call this
  * function.
+ *
+ * WARNING! unpend() will call statclock_process() directly and skip this
+ * routine.
  */
 void
 forwarded_statclock(struct trapframe frame)
@@ -2341,6 +2344,9 @@ forward_statclock(void)
  * sched_lock if we could simply peek at the CPU to determine the user/kernel
  * state and call hardclock_process() on the CPU receiving the clock interrupt
  * and then just use a simple IPI to handle any ast's if needed.
+ *
+ * WARNING! unpend() will call hardclock_process() directly and skip this
+ * routine.
  */
 void
 forwarded_hardclock(struct trapframe frame)