In both stable and development branches of the kernel, when
reschedule_idle is called, it is presumed, that reschedule_idle will
release the spin_lock, thus call spin_unlock_irqrestore. This does not
happen when the kernels are compiled this __SMP__. Artur Skawina
<skawina@geocities.com> thinks this might delay servicing interrupts
unnecessarily. (thanks Artur for your response).
Here's the patch against 2.2.14-pre17:
diff -ruN linux-2.2.14-pre17/kernel/sched.c linux/kernel/sched.c
--- linux-2.2.14-pre17/kernel/sched.c Wed Oct 20 02:14:02 1999
+++ linux/kernel/sched.c Wed Dec 29 12:22:28 1999
@@ -325,6 +325,7 @@
tsk = current;
if (preemption_goodness(tsk, p, this_cpu) > 0)
tsk->need_resched = 1;
+ spin_unlock_irqrestore(&runqueue_lock, flags);
#endif
}
And here's the patch against 2.3.35-pre6:
--- linux-2.3.35-pre6/kernel/sched.c Fri Dec 10 23:57:11 1999
+++ linux-2.3.34-post/kernel/sched.c Wed Dec 29 12:34:05 1999
@@ -273,6 +273,7 @@
tsk = cpu_curr(this_cpu);
if (preemption_goodness(tsk, p, this_cpu) > 0)
tsk->need_resched = 1;
+ spin_unlock_irqrestore(&runqueue_lock, flags);
#endif
}
Can you put this in?
me
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/