From: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
To: Steven Rostedt <rostedt@goodmis.org>
Cc: linux-kernel@vger.kernel.org, Ingo Molnar <mingo@kernel.org>,
Andrew Morton <akpm@linux-foundation.org>
Subject: Re: [PATCH 6/5]rcu/tracing: Add rcu_disabled to denote when rcu_irq_enter() will not work
Date: Fri, 7 Apr 2017 09:42:20 -0700 [thread overview]
Message-ID: <20170407164220.GY1600@linux.vnet.ibm.com> (raw)
In-Reply-To: <20170407123516.0560f48c@gandalf.local.home>
On Fri, Apr 07, 2017 at 12:35:16PM -0400, Steven Rostedt wrote:
> From: "Steven Rostedt (VMware)" <rostedt@goodmis.org>
>
> Tracing uses rcu_irq_enter() as a way to make sure that RCU is watching when
> it needs to use rcu_read_lock() and friends. This is because tracing can
> happen as RCU is about to enter user space, or about to go idle, and RCU
> does not watch for RCU read side critical sections as it makes the
> transition.
>
> There is a small location within the RCU infrastructure that rcu_irq_enter()
> itself will not work. If tracing were to occur in that section it will break
> if it tries to use rcu_irq_enter().
>
> Originally, this happens with the stack_tracer, because it will call
> save_stack_trace when it encounters stack usage that is greater than any
> stack usage it had encountered previously. There was a case where that
> happened in the RCU section where rcu_irq_enter() did not work, and lockdep
> complained loudly about it. To fix it, stack tracing added a call to be
> disabled and RCU would disable stack tracing during the critical section
> that rcu_irq_enter() was inoperable. This solution worked, but there are
> other cases that use rcu_irq_enter() and it would be a good idea to let RCU
> give a way to let others know that rcu_irq_enter() will not work. For
> example, in trace events.
>
> Another helpful aspect of this change is that it also moves the per cpu
> variable called in the RCU critical section into a cache locale along with
> other RCU per cpu variables used in that same location.
>
> I'm keeping the stack_trace_disable() code, as that still could be used in
> the future by places that really need to disable it. And since it's only a
> static inline, it wont take up any kernel text if it is not used.
>
> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
Looks better, especially __this_cpu_inc() and __this_cpu_dec().
How about rcu_irq_enter_disabled instead of rcu_disabled? We aren't
really disabling all of RCU. ;-)
Thanx, Paul
> ---
> include/linux/rcupdate.h | 5 +++++
> kernel/rcu/tree.c | 18 ++++++++++++++++--
> kernel/trace/trace_stack.c | 8 ++++++++
> 3 files changed, 29 insertions(+), 2 deletions(-)
>
> diff --git a/include/linux/rcupdate.h b/include/linux/rcupdate.h
> index de88b33..d922ee7 100644
> --- a/include/linux/rcupdate.h
> +++ b/include/linux/rcupdate.h
> @@ -97,6 +97,7 @@ void do_trace_rcu_torture_read(const char *rcutorturename,
> unsigned long secs,
> unsigned long c_old,
> unsigned long c);
> +bool rcu_is_disabled(void);
> #else
> static inline void rcutorture_get_gp_data(enum rcutorture_type test_type,
> int *flags,
> @@ -113,6 +114,10 @@ static inline void rcutorture_record_test_transition(void)
> static inline void rcutorture_record_progress(unsigned long vernum)
> {
> }
> +bool rcu_is_disabled(void)
> +{
> + return false;
> +}
> #ifdef CONFIG_RCU_TRACE
> void do_trace_rcu_torture_read(const char *rcutorturename,
> struct rcu_head *rhp,
> diff --git a/kernel/rcu/tree.c b/kernel/rcu/tree.c
> index 8b4d273..727325f 100644
> --- a/kernel/rcu/tree.c
> +++ b/kernel/rcu/tree.c
> @@ -285,6 +285,20 @@ static DEFINE_PER_CPU(struct rcu_dynticks, rcu_dynticks) = {
> };
>
> /*
> + * There's a few places, currently just in the tracing infrastructure,
> + * that uses rcu_irq_enter() to make sure RCU is watching. But there's
> + * a small location where that will not even work. In those cases
> + * rcu_is_disabled() needs to be checked to make sure rcu_irq_enter()
> + * can be called.
> + */
> +static DEFINE_PER_CPU(bool, rcu_disabled);
> +
> +bool rcu_is_disabled(void)
> +{
> + return this_cpu_read(rcu_disabled);
> +}
> +
> +/*
> * Record entry into an extended quiescent state. This is only to be
> * called when not already in an extended quiescent state.
> */
> @@ -800,10 +814,10 @@ static void rcu_eqs_enter_common(bool user)
> do_nocb_deferred_wakeup(rdp);
> }
> rcu_prepare_for_idle();
> - stack_tracer_disable();
> + __this_cpu_inc(rcu_disabled);
> rdtp->dynticks_nesting = 0; /* Breaks tracing momentarily. */
> rcu_dynticks_eqs_enter(); /* After this, tracing works again. */
> - stack_tracer_enable();
> + __this_cpu_dec(rcu_disabled);
> rcu_dynticks_task_enter();
>
> /*
> diff --git a/kernel/trace/trace_stack.c b/kernel/trace/trace_stack.c
> index f2f02ff..aec2df1 100644
> --- a/kernel/trace/trace_stack.c
> +++ b/kernel/trace/trace_stack.c
> @@ -96,6 +96,14 @@ check_stack(unsigned long ip, unsigned long *stack)
> if (in_nmi())
> return;
>
> + /*
> + * There's a slight chance that we are tracing inside the
> + * RCU infrastructure, and rcu_irq_enter() will not work
> + * as expected.
> + */
> + if (unlikely(rcu_is_disabled()))
> + return;
> +
> local_irq_save(flags);
> arch_spin_lock(&stack_trace_max_lock);
>
> --
> 2.9.3
>
next prev parent reply other threads:[~2017-04-07 16:42 UTC|newest]
Thread overview: 39+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-04-07 14:01 [PATCH 0/5 v2] tracing: Add usecase of synchronize_rcu_tasks() and stack_tracer_disable() Steven Rostedt
2017-04-07 14:01 ` [PATCH 1/5 v2] ftrace: Add use of synchronize_rcu_tasks() with dynamic trampolines Steven Rostedt
2017-04-07 14:01 ` [PATCH 2/5 v2] tracing: Replace the per_cpu() with this_cpu() in trace_stack.c Steven Rostedt
2017-04-07 14:36 ` Paul E. McKenney
2017-04-07 14:48 ` Steven Rostedt
2017-04-07 15:08 ` Paul E. McKenney
2017-04-07 14:01 ` [PATCH 3/5 v2] tracing: Add stack_tracer_disable/enable() functions Steven Rostedt
2017-04-07 14:22 ` Steven Rostedt
2017-04-07 14:25 ` [PATCH 3/5 v2.1] " Steven Rostedt
2017-04-07 14:01 ` [PATCH 4/5 v2] tracing: Rename trace_active to disable_stack_tracer and inline its modification Steven Rostedt
2017-04-07 14:01 ` [PATCH 5/5 v2] rcu: Fix dyntick-idle tracing Steven Rostedt
2017-04-07 14:40 ` Paul E. McKenney
2017-04-07 14:53 ` Steven Rostedt
2017-04-07 15:09 ` Paul E. McKenney
2017-04-07 15:29 ` Steven Rostedt
2017-04-07 14:43 ` [PATCH 0/5 v2] tracing: Add usecase of synchronize_rcu_tasks() and stack_tracer_disable() Paul E. McKenney
2017-04-07 14:58 ` Steven Rostedt
2017-04-07 15:11 ` Paul E. McKenney
2017-04-07 15:28 ` Steven Rostedt
2017-04-07 16:35 ` [PATCH 6/5]rcu/tracing: Add rcu_disabled to denote when rcu_irq_enter() will not work Steven Rostedt
2017-04-07 16:42 ` Paul E. McKenney [this message]
2017-04-07 16:44 ` Steven Rostedt
2017-04-07 16:53 ` Paul E. McKenney
2017-04-07 17:03 ` [PATCH 6/5 v2] rcu/tracing: " Steven Rostedt
2017-04-07 17:15 ` Paul E. McKenney
2017-04-07 17:06 ` [PATCH 7/5] tracing: Make sure rcu_irq_enter() can work for trace_*_rcuidle() trace events Steven Rostedt
2017-04-07 17:15 ` Paul E. McKenney
2017-04-07 17:19 ` Mathieu Desnoyers
2017-04-07 17:26 ` Steven Rostedt
2017-04-07 17:32 ` Steven Rostedt
2017-04-07 17:49 ` Mathieu Desnoyers
2017-04-07 17:55 ` Steven Rostedt
2017-04-07 18:10 ` [PATCH 7/5 v3] " Steven Rostedt
2017-04-07 18:17 ` Mathieu Desnoyers
2017-04-07 19:41 ` [PATCH 7/5 v4] " Steven Rostedt
2017-04-07 19:43 ` Steven Rostedt
2017-04-10 17:11 ` Mathieu Desnoyers
2017-04-07 17:28 ` [PATCH 7/5] " Steven Rostedt
2017-04-07 17:48 ` [PATCH 7/5 v2] " Steven Rostedt
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20170407164220.GY1600@linux.vnet.ibm.com \
--to=paulmck@linux.vnet.ibm.com \
--cc=akpm@linux-foundation.org \
--cc=linux-kernel@vger.kernel.org \
--cc=mingo@kernel.org \
--cc=rostedt@goodmis.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).