[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20181214084639.f4jnunlalkx2hkz6@linutronix.de>
Date: Fri, 14 Dec 2018 09:46:40 +0100
From: Sebastian Andrzej Siewior <bigeasy@...utronix.de>
To: Benjamin Bouvier <benjamin.bouvier@...ia.com>
Cc: mingo@...hat.com, peterz@...radead.org, rostedt@...dmis.org,
u.kleine-koenig@...gutronix.de, pkondeti@...eaurora.org,
alexander.sverdlin@...ia.com, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] sched/debug: Add tracepoint for RT throttling
On 2018-12-14 10:27:41 [+0200], Benjamin Bouvier wrote:
> As there is currently no information on which RT task is throttled, add a
> new tracepoint for debug purpose. By the way improve printk debugging
> message by adding task name, pid and cpu for which the throttle has been
> executed.
Throttling is for _all_ RT class applications. The task you print might
be innocent and getting the blame. The interesting part is to see all
the RT tasks (not sure about DL) since the last time, the timer was
started on the CPU.
Sebastian
Powered by blists - more mailing lists