Unkillable KSE threaded proc

[Julian Elischer]

Andrew Gallatin wrote:
> Julian Elischer writes:
>  > >
>  > >Maybe this would be easier to debug if I disabled preemption?
>  > >
>  > 
>  > 
>  > I think that this would possibly GO AWAY of you disab;ed preemption. 
>  > which would make it very hard to debug :-)
>  > 
> 
> Yes and no.  You initially asked me to try in -current because of
> some changes you'd made to the exit code.  RELENG_5 (with the old
> exit code and no preemption) shows a different problem (proc is
> just not killable).    If the proc was killable without preemption,
> that would at least show your new code is better..

try the attached diff:

> 
> Drew

-------------- next part --------------
Index: sys/kern/kern_switch.c
===========================================================================
--- sys/kern/kern_switch.c	2004/09/14 08:14:48	#76
+++ sys/kern/kern_switch.c	2004/09/14 08:14:48
@@ -350,11 +389,10 @@
 			}
 			kg->kg_avail_opennings = 1;
 		}
-		kg->kg_avail_opennings--;
 		sched_add(td, flags);
 		return;
 	}
 
 	tda = kg->kg_last_assigned;
 	if ((kg->kg_avail_opennings <= 0) &&
 	(tda && (tda->td_priority > td->td_priority))) {
@@ -415,7 +449,6 @@
 			td2 = TAILQ_NEXT(tda, td_runq);
 			kg->kg_last_assigned = td2;
 		}
-		kg->kg_avail_opennings--;
 		sched_add(td2, flags);
 	} else {
 		CTR3(KTR_RUNQ, "setrunqueue: held: td%p kg%p pid%d",
Index: sys/kern/sched_4bsd.c
===========================================================================
--- sys/kern/sched_4bsd.c	2004/09/14 08:14:48	#64
+++ sys/kern/sched_4bsd.c	2004/09/14 08:14:48
@@ -1024,6 +1024,7 @@
 	}
 	if ((td->td_proc->p_flag & P_NOLOAD) == 0)
 		sched_tdcnt++;
+	td->td_ksegrp->kg_avail_opennings--;
 	runq_add(ke->ke_runq, ke);
 	ke->ke_ksegrp->kg_runq_kses++;
 	ke->ke_state = KES_ONRUNQ;
Index: sys/kern/sched_ule.c
===========================================================================
--- sys/kern/sched_ule.c	2004/09/14 08:14:48	#127
+++ sys/kern/sched_ule.c	2004/09/14 08:14:48
@@ -1773,6 +1773,7 @@
 		curthread->td_flags |= TDF_NEEDRESCHED;
 	if (preemptive && maybe_preempt(td))
 		return;
+	td->td_ksegrp->kg_avail_opennings--;
 	ke->ke_ksegrp->kg_runq_threads++;
 	ke->ke_state = KES_ONRUNQ;