Re: Priority Inheritance Test (Real-Time Preemption)

[Ingo Molnar]

* Ingo Molnar <mingo@xxxxxxx> wrote:

> it can produce such a flow on SMP, or if you add in a third non-RT
> task (task-C). Agreed?

here's the 4-task flow. I've simplified the operations to make it easier
to overview: "L1-success means" 'task locked lock1 and got it'. 
"L1-wait" means it tried to get lock1 but has to wait. "RT-prio" means a
non-RT task got boosted to RT priority. "old-prio" means priority got
restored to the original non-RT value. "UL-1" means 'unlocked lock1'.

	task-A		task-B		task-C		task-RT
	-------------------------------------------------------
	L2-success
			L1-success
					L1-wait
					.		L2-wait
					.		boost-A
	RT-prio				.		.
	L1-wait				.		.
	boost-B				.		.
	.		RT-prio		.		.
	.		[ 1 ms ]	.		.
	.		UL-1		.		.
	.		!RT-prio	.		.
	get-L1				.		.
	[ 1 ms ]			.		.
	UL-1				.		.
					get-L1		.
	UL-2						.
	old-prio					.
							get-L2
							L1-wait
							boost-C
					RT-prio		.
					[ 1 msec ]	.
					UL-1		.
					old-prio	.
							get-L1
							[ success ]

this is a 3 milliseconds worst-case. Ok?

	Ingo
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/