linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Steven Rostedt <rostedt@goodmis.org>
To: Josh Triplett <josh@joshtriplett.org>
Cc: paulmck@linux.vnet.ibm.com, linux-kernel@vger.kernel.org,
	mingo@elte.hu, laijs@cn.fujitsu.com, dipankar@in.ibm.com,
	akpm@linux-foundation.org, mathieu.desnoyers@polymtl.ca,
	niv@us.ibm.com, tglx@linutronix.de, peterz@infradead.org,
	Valdis.Kletnieks@vt.edu, dhowells@redhat.com,
	eric.dumazet@gmail.com, darren@dvhart.com, fweisbec@gmail.com,
	sbw@mit.edu, patches@linaro.org,
	"Paul E. McKenney" <paul.mckenney@linaro.org>
Subject: Re: [PATCH tip/core/rcu 04/15] rcu: Permit RCU_NONIDLE() to be used from interrupt context
Date: Tue, 04 Sep 2012 19:23:51 -0400	[thread overview]
Message-ID: <1346801031.27919.39.camel@gandalf.local.home> (raw)
In-Reply-To: <20120904230834.GB11494@jtriplet-mobl1>

On Tue, 2012-09-04 at 16:08 -0700, Josh Triplett wrote:
> On Tue, Sep 04, 2012 at 06:51:22PM -0400, Steven Rostedt wrote:
> > On Tue, 2012-09-04 at 15:33 -0700, Paul E. McKenney wrote:
> > > On Fri, Aug 31, 2012 at 11:00:52AM -0700, Josh Triplett wrote:
> > > > On Thu, Aug 30, 2012 at 11:56:17AM -0700, Paul E. McKenney wrote:
> > > > > From: "Paul E. McKenney" <paul.mckenney@linaro.org>
> > > > > 
> > > > > There is a need to use RCU from interrupt context, but either before
> > > > > rcu_irq_enter() is called or after rcu_irq_exit() is called.  If the
> > > > > interrupt occurs from idle, then lockdep-RCU will complain about such
> > > > > uses, as they appear to be illegal uses of RCU from the idle loop.
> > > > > In other environments, RCU_NONIDLE() could be used to properly protect
> > > > > the use of RCU, but RCU_NONIDLE() currently cannot be invoked except
> > > > > from process context.
> > > > > 
> > > > > This commit therefore modifies RCU_NONIDLE() to permit its use more
> > > > > globally.
> > > > > 
> > > > > Reported-by: Steven Rostedt <rostedt@goodmis.org>
> > > > > Signed-off-by: Paul E. McKenney <paul.mckenney@linaro.org>
> > > > > Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
> > > > 
> > > > Something seems wrong about this.  The addition of EXPORT_SYMBOL_GPL
> > > > suggests that such interrupt handlers might live in modules.  In what
> > > > situation might a module interrupt handler get called from the idle
> > > > loop, before rcu_irq_enter or after rcu_irq_exit, and need to know that
> > > > when using RCU?
> > > 
> > > Drivers can be in modules, in which case their interrupt handlers will
> > > also be in the corresponding module.  I do agree that in most cases,
> > > the irq_enter() and irq_exit() hooks would be invoked by non-module code,
> > > but I do believe that I had to add those exports due to build failures.
> > > 
> > > Steven will let me know if I am confused on this point.
> > > 
> > 
> > You're not confused, the situation is confusing :-/
> > 
> > Because some trace events happen inside the idle loop after rcu has
> > "shutdown", we needed to create "trace_foo_rcuidle()" handlers that can
> > handle this condition. That is, for every trace_foo() static inline
> > (used at the tracepoint location), there exists a static inline
> > trace_foo_rcuidle(), that looks something like this:
> > 
> > static inline void trace_##name##_rcuidle(proto) {
> > 	if (static_key_false(&__tracepoint_##name.key)) { 
> > 		rcu_idle_exit();
> > 		__DO_TRACE();
> > 		rcu_idle_enter();
> > 	}
> > }
> > 
> > Although these calls are never used by module code, because they are
> > static inlines, they are still defined for all tracepoints, kernel
> > tracepoints as well as module tracepoints. And thus, need the export :-(
> 
> Fair enough.
> 
> What about having the tracepoint code generation detect when building as
> part of a module via defined(MODULE), and omit the unused _rcuidle
> versions in those cases?  That would avoid the need to export those
> functions at all.  Strawman patch (not tested):
> 
> diff --git a/include/linux/tracepoint.h b/include/linux/tracepoint.h
> index 802de56..41e1ef2 100644
> --- a/include/linux/tracepoint.h
> +++ b/include/linux/tracepoint.h
> @@ -136,6 +136,22 @@ static inline void tracepoint_synchronize_unregister(void)
>  		postrcu;						\
>  	} while (0)
>  
> +#ifdef MODULE
> +#define __DECLARE_TRACE_RCU(name, proto, args, cond, data_proto, data_args) \
> +	static inline void trace_##name##_rcuidle(proto)		\
> +	{								\
> +		if (static_key_false(&__tracepoint_##name.key))		\
> +			__DO_TRACE(&__tracepoint_##name,		\
> +				TP_PROTO(data_proto),			\
> +				TP_ARGS(data_args),			\
> +				TP_CONDITION(cond),			\
> +				rcu_idle_exit(),			\
> +				rcu_idle_enter());			\
> +	}
> +#else
> +#define __DECLARE_TRACE_RCU(name, proto, args, cond, data_proto, data_args)
> +#endif
> +

Egad! More macros on top of macros! Yeah I know I'm the most guilty of
this, but it just seems to add one more indirection that I would like to
avoid.

>  /*
>   * Make sure the alignment of the structure in the __tracepoints section will
>   * not add unwanted padding between the beginning of the section and the
> @@ -151,16 +167,7 @@ static inline void tracepoint_synchronize_unregister(void)
>  				TP_ARGS(data_args),			\
>  				TP_CONDITION(cond),,);			\
>  	}								\
> -	static inline void trace_##name##_rcuidle(proto)		\
> -	{								\
> -		if (static_key_false(&__tracepoint_##name.key))		\
> -			__DO_TRACE(&__tracepoint_##name,		\
> -				TP_PROTO(data_proto),			\
> -				TP_ARGS(data_args),			\
> -				TP_CONDITION(cond),			\
> -				rcu_idle_exit(),			\
> -				rcu_idle_enter());			\
> -	}								\
> +	__DECLARE_TRACE_RCU(name, proto, args, cond, data_proto, data_args) \
>  	static inline int						\
>  	register_trace_##name(void (*probe)(data_proto), void *data)	\
>  	{								\
> 
> 
> If that doesn't work out, please consider adding an explicit comment
> saying why you exported the functions.
> 

Or, we could also add in include/linux/rcupdate.h:

#ifdef MODULE
static inline void rcu_idle_enter(void) {
	panic("Don't call me from modules");
}
static inline void rcu_idle_exit(void) {
	panic("Don't call me from modules");
}
#else
extern void rcu_idle_enter(void);
extern void rcu_idle_exit(void);
#endif



Hmm, if there ever happens to be a governor that can be loaded as a
module, and if it has a tracepoint, then it would require this too.

But the first time someone tries that, it will panic with the above
code ;-)

-- Steve



  reply	other threads:[~2012-09-04 23:23 UTC|newest]

Thread overview: 86+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2012-08-30 18:56 [PATCH tip/core/rcu 0/15] Miscellaneous fixes Paul E. McKenney
2012-08-30 18:56 ` [PATCH tip/core/rcu 01/15] rcu: Add PROVE_RCU_DELAY to provoke difficult races Paul E. McKenney
2012-08-30 18:56   ` [PATCH tip/core/rcu 02/15] rcu: Pull TINY_RCU dyntick-idle tracing into non-idle region Paul E. McKenney
2012-08-31 16:53     ` Josh Triplett
2012-08-30 18:56   ` [PATCH tip/core/rcu 03/15] rcu: Properly initialize ->boost_tasks on CPU offline Paul E. McKenney
2012-08-31 17:56     ` Josh Triplett
2012-09-06 14:40     ` Peter Zijlstra
2012-09-06 20:58       ` Paul E. McKenney
2012-08-30 18:56   ` [PATCH tip/core/rcu 04/15] rcu: Permit RCU_NONIDLE() to be used from interrupt context Paul E. McKenney
2012-08-31 18:00     ` Josh Triplett
2012-09-04 22:33       ` Paul E. McKenney
2012-09-04 22:48         ` Josh Triplett
2012-09-04 22:51         ` Steven Rostedt
2012-09-04 23:08           ` Josh Triplett
2012-09-04 23:23             ` Steven Rostedt [this message]
2012-09-04 23:33               ` Josh Triplett
2012-09-04 23:43                 ` Paul E. McKenney
2012-09-06 18:54                   ` Josh Triplett
2012-09-06 19:54                     ` Steven Rostedt
2012-09-07  6:09                       ` Josh Triplett
2012-09-07 14:24                         ` Paul E. McKenney
2012-09-07 14:47                           ` Josh Triplett
2012-09-07 15:16                             ` Steven Rostedt
2012-09-12  1:07                               ` Paul E. McKenney
2012-09-12 14:13                                 ` Steven Rostedt
2012-09-12 15:03                                   ` Paul E. McKenney
2012-09-12 15:18                                     ` Steven Rostedt
2012-09-12 16:57                                       ` Paul E. McKenney
2012-09-04 23:46                 ` Steven Rostedt
2012-09-05  0:42                   ` Josh Triplett
2012-09-05  6:23                   ` [PATCH] trace: Don't declare trace_*_rcuidle functions in modules Josh Triplett
2012-09-05 14:26                     ` Mathieu Desnoyers
2012-09-05 16:36                     ` Paul E. McKenney
2012-09-06 19:49                     ` Steven Rostedt
2012-09-14  6:07                     ` [tip:core/rcu] trace: Don' t " tip-bot for Josh Triplett
2012-09-04 23:14           ` [PATCH tip/core/rcu 04/15] rcu: Permit RCU_NONIDLE() to be used from interrupt context Paul E. McKenney
2012-08-30 18:56   ` [PATCH tip/core/rcu 05/15] rcu: Improve boost selection when moving tasks to root rcu_node Paul E. McKenney
2012-08-31 18:09     ` Josh Triplett
2012-08-30 18:56   ` [PATCH tip/core/rcu 06/15] rcu: Make offline-CPU checking allow for indefinite delays Paul E. McKenney
2012-08-31 18:12     ` Josh Triplett
2012-08-30 18:56   ` [PATCH tip/core/rcu 07/15] rcu: Fix obsolete rcu_initiate_boost() header comment Paul E. McKenney
2012-08-31 18:13     ` Josh Triplett
2012-08-30 18:56   ` [PATCH tip/core/rcu 08/15] rcu: Apply for_each_rcu_flavor() to increment_cpu_stall_ticks() Paul E. McKenney
2012-08-31 18:15     ` Josh Triplett
2012-09-04 22:44       ` Paul E. McKenney
2012-08-30 18:56   ` [PATCH tip/core/rcu 09/15] rcu: Avoid rcu_print_detail_task_stall_rnp() segfault Paul E. McKenney
2012-08-31 18:19     ` Josh Triplett
2012-09-04 22:46       ` Paul E. McKenney
2012-09-04 22:55         ` Josh Triplett
2012-08-30 18:56   ` [PATCH tip/core/rcu 10/15] rcu: Protect rcu_node accesses during CPU stall warnings Paul E. McKenney
2012-08-31 18:23     ` Josh Triplett
2012-09-04 22:51       ` Paul E. McKenney
2012-09-06 14:51     ` Peter Zijlstra
2012-09-06 21:01       ` Paul E. McKenney
2012-08-30 18:56   ` [PATCH tip/core/rcu 11/15] rcu: Avoid spurious RCU " Paul E. McKenney
2012-08-31 18:24     ` Josh Triplett
2012-09-06 14:56     ` Peter Zijlstra
2012-09-06 15:07       ` Steven Rostedt
2012-09-06 15:19         ` Peter Zijlstra
2012-09-06 21:03           ` Paul E. McKenney
2012-09-06 21:41             ` Steven Rostedt
2012-09-06 21:58               ` Paul E. McKenney
2012-09-06 22:05                 ` Steven Rostedt
2012-09-06 22:22                   ` Paul E. McKenney
2012-09-07  7:00                     ` Peter Zijlstra
2012-09-07 14:42                       ` Steven Rostedt
2012-08-30 18:56   ` [PATCH tip/core/rcu 12/15] rcu: Remove redundant memory barrier from __call_rcu() Paul E. McKenney
2012-08-31 18:30     ` Josh Triplett
2012-08-31 18:40       ` Josh Triplett
2012-08-30 18:56   ` [PATCH tip/core/rcu 13/15] rcu: Move TINY_PREEMPT_RCU away from raw_local_irq_save() Paul E. McKenney
2012-08-31 18:34     ` Josh Triplett
2012-09-04 23:03       ` Paul E. McKenney
2012-08-30 18:56   ` [PATCH tip/core/rcu 14/15] time: RCU permitted to stop idle entry via softirq Paul E. McKenney
2012-08-31 18:51     ` Josh Triplett
2012-09-06 15:12     ` Peter Zijlstra
2012-09-06 21:35       ` Paul E. McKenney
2012-09-06 21:57         ` Steven Rostedt
2012-09-06 22:11           ` Paul E. McKenney
2012-08-30 18:56   ` [PATCH tip/core/rcu 15/15] kmemleak: Replace list_for_each_continue_rcu with new interface Paul E. McKenney
2012-08-31 18:55     ` Josh Triplett
2012-09-04 23:41       ` Paul E. McKenney
2012-08-31 16:49   ` [PATCH tip/core/rcu 01/15] rcu: Add PROVE_RCU_DELAY to provoke difficult races Josh Triplett
2012-09-04 22:36     ` Paul E. McKenney
2012-09-06 14:38   ` Peter Zijlstra
2012-09-06 20:51     ` Paul E. McKenney
2012-09-07  6:54       ` Peter Zijlstra

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=1346801031.27919.39.camel@gandalf.local.home \
    --to=rostedt@goodmis.org \
    --cc=Valdis.Kletnieks@vt.edu \
    --cc=akpm@linux-foundation.org \
    --cc=darren@dvhart.com \
    --cc=dhowells@redhat.com \
    --cc=dipankar@in.ibm.com \
    --cc=eric.dumazet@gmail.com \
    --cc=fweisbec@gmail.com \
    --cc=josh@joshtriplett.org \
    --cc=laijs@cn.fujitsu.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mathieu.desnoyers@polymtl.ca \
    --cc=mingo@elte.hu \
    --cc=niv@us.ibm.com \
    --cc=patches@linaro.org \
    --cc=paul.mckenney@linaro.org \
    --cc=paulmck@linux.vnet.ibm.com \
    --cc=peterz@infradead.org \
    --cc=sbw@mit.edu \
    --cc=tglx@linutronix.de \
    --subject='Re: [PATCH tip/core/rcu 04/15] rcu: Permit RCU_NONIDLE() to be used from interrupt context' \
    /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

This is a public inbox, see mirroring instructions
on how to clone and mirror all data and code used for this inbox