lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date:   Sat, 14 Jul 2018 10:31:57 -0400 (EDT)
From:   Mathieu Desnoyers <mathieu.desnoyers@...icios.com>
To:     "Joel Fernandes, Google" <joel@...lfernandes.org>
Cc:     linux-kernel <linux-kernel@...r.kernel.org>,
        kernel-team <kernel-team@...roid.com>,
        Boqun Feng <boqun.feng@...il.com>,
        Byungchul Park <byungchul.park@....com>,
        Erick Reyes <erickreyes@...gle.com>,
        Ingo Molnar <mingo@...hat.com>,
        Julia Cartwright <julia@...com>,
        Masami Hiramatsu <mhiramat@...nel.org>,
        Namhyung Kim <namhyung@...nel.org>,
        "Paul E. McKenney" <paulmck@...ux.vnet.ibm.com>,
        Peter Zijlstra <peterz@...radead.org>,
        rostedt <rostedt@...dmis.org>,
        Thomas Gleixner <tglx@...utronix.de>,
        Todd Kjos <tkjos@...gle.com>,
        Tom Zanussi <tom.zanussi@...ux.intel.com>,
        Will Deacon <will.deacon@....com>
Subject: Re: [PATCH v10 2/3] tracepoint: Make rcuidle tracepoint callers use
 SRCU

----- On Jul 13, 2018, at 5:55 PM, Joel Fernandes, Google joel@...lfernandes.org wrote:

> From: "Joel Fernandes (Google)" <joel@...lfernandes.org>
> 
> In recent tests with IRQ on/off tracepoints, a large performance
> overhead ~10% is noticed when running hackbench. This is root caused to
> calls to rcu_irq_enter_irqson and rcu_irq_exit_irqson from the
> tracepoint code. Following a long discussion on the list [1] about this,
> we concluded that srcu is a better alternative for use during rcu idle.
> Although it does involve extra barriers, its lighter than the sched-rcu
> version which has to do additional RCU calls to notify RCU idle about
> entry into RCU sections.
> 
> In this patch, we change the underlying implementation of the
> trace_*_rcuidle API to use SRCU. This has shown to improve performance
> alot for the high frequency irq enable/disable tracepoints.
> 
> Test: Tested idle and preempt/irq tracepoints.
> 
> Here are some performance numbers:
> 
> With a run of the following 30 times on a single core x86 Qemu instance
> with 1GB memory:
> hackbench -g 4 -f 2 -l 3000
> 
> Completion times in seconds. CONFIG_PROVE_LOCKING=y.
> 
> No patches (without this series)
> Mean: 3.048
> Median: 3.025
> Std Dev: 0.064
> 
> With Lockdep using irq tracepoints with RCU implementation:
> Mean: 3.451   (-11.66 %)
> Median: 3.447 (-12.22%)
> Std Dev: 0.049
> 
> With Lockdep using irq tracepoints with SRCU implementation (this series):
> Mean: 3.020   (I would consider the improvement against the "without
>	       this series" case as just noise).
> Median: 3.013
> Std Dev: 0.033
> 
> [1] https://patchwork.kernel.org/patch/10344297/
> 
> Reviewed-by: Mathieu Desnoyers <mathieu.desnoyers@...icios.com>

I'm fine with the changes done since last iteration.

Acked-by: Mathieu Desnoyers <mathieu.desnoyers@...icios.com>

Thanks!

Mathieu

> Cleaned-up-by: Peter Zijlstra <peterz@...radead.org>
> Signed-off-by: Joel Fernandes (Google) <joel@...lfernandes.org>
> ---
> include/linux/tracepoint.h | 45 +++++++++++++++++++++++++++++++-------
> kernel/tracepoint.c        | 16 +++++++++++++-
> 2 files changed, 52 insertions(+), 9 deletions(-)
> 
> diff --git a/include/linux/tracepoint.h b/include/linux/tracepoint.h
> index 19a690b559ca..97e1d365a817 100644
> --- a/include/linux/tracepoint.h
> +++ b/include/linux/tracepoint.h
> @@ -15,6 +15,7 @@
>  */
> 
> #include <linux/smp.h>
> +#include <linux/srcu.h>
> #include <linux/errno.h>
> #include <linux/types.h>
> #include <linux/cpumask.h>
> @@ -33,6 +34,8 @@ struct trace_eval_map {
> 
> #define TRACEPOINT_DEFAULT_PRIO	10
> 
> +extern struct srcu_struct tracepoint_srcu;
> +
> extern int
> tracepoint_probe_register(struct tracepoint *tp, void *probe, void *data);
> extern int
> @@ -75,10 +78,16 @@ int unregister_tracepoint_module_notifier(struct
> notifier_block *nb)
>  * probe unregistration and the end of module exit to make sure there is no
>  * caller executing a probe when it is freed.
>  */
> +#ifdef CONFIG_TRACEPOINTS
> static inline void tracepoint_synchronize_unregister(void)
> {
> +	synchronize_srcu(&tracepoint_srcu);
> 	synchronize_sched();
> }
> +#else
> +static inline void tracepoint_synchronize_unregister(void)
> +{ }
> +#endif
> 
> #ifdef CONFIG_HAVE_SYSCALL_TRACEPOINTS
> extern int syscall_regfunc(void);
> @@ -129,18 +138,34 @@ extern void syscall_unregfunc(void);
>  * as "(void *, void)". The DECLARE_TRACE_NOARGS() will pass in just
>  * "void *data", where as the DECLARE_TRACE() will pass in "void *data, proto".
>  */
> -#define __DO_TRACE(tp, proto, args, cond, rcucheck)			\
> +#define __DO_TRACE(tp, proto, args, cond, rcuidle)			\
> 	do {								\
> 		struct tracepoint_func *it_func_ptr;			\
> 		void *it_func;						\
> 		void *__data;						\
> +		int __maybe_unused idx = 0;				\
> 									\
> 		if (!(cond))						\
> 			return;						\
> -		if (rcucheck)						\
> -			rcu_irq_enter_irqson();				\
> -		rcu_read_lock_sched_notrace();				\
> -		it_func_ptr = rcu_dereference_sched((tp)->funcs);	\
> +									\
> +		/* srcu can't be used from NMI */			\
> +		if (rcuidle && in_nmi())				\
> +			WARN_ON_ONCE(1);				\
> +									\
> +		/* keep srcu and sched-rcu usage consistent */		\
> +		preempt_disable_notrace();				\
> +									\
> +		/*							\
> +		 * For rcuidle callers, use srcu since sched-rcu	\
> +		 * doesn't work from the idle path.			\
> +		 */							\
> +		if (rcuidle)						\
> +			idx = srcu_read_lock_notrace(&tracepoint_srcu);	\
> +		else							\
> +			rcu_read_lock_sched_notrace();			\
> +									\
> +		it_func_ptr = rcu_dereference_raw((tp)->funcs);		\
> +									\
> 		if (it_func_ptr) {					\
> 			do {						\
> 				it_func = (it_func_ptr)->func;		\
> @@ -148,9 +173,13 @@ extern void syscall_unregfunc(void);
> 				((void(*)(proto))(it_func))(args);	\
> 			} while ((++it_func_ptr)->func);		\
> 		}							\
> -		rcu_read_unlock_sched_notrace();			\
> -		if (rcucheck)						\
> -			rcu_irq_exit_irqson();				\
> +									\
> +		if (rcuidle)						\
> +			srcu_read_unlock_notrace(&tracepoint_srcu, idx);\
> +		else							\
> +			rcu_read_unlock_sched_notrace();		\
> +									\
> +		preempt_enable_notrace();				\
> 	} while (0)
> 
> #ifndef MODULE
> diff --git a/kernel/tracepoint.c b/kernel/tracepoint.c
> index 6dc6356c3327..955148d91b74 100644
> --- a/kernel/tracepoint.c
> +++ b/kernel/tracepoint.c
> @@ -31,6 +31,9 @@
> extern struct tracepoint * const __start___tracepoints_ptrs[];
> extern struct tracepoint * const __stop___tracepoints_ptrs[];
> 
> +DEFINE_SRCU(tracepoint_srcu);
> +EXPORT_SYMBOL_GPL(tracepoint_srcu);
> +
> /* Set to 1 to enable tracepoint debug output */
> static const int tracepoint_debug;
> 
> @@ -67,16 +70,27 @@ static inline void *allocate_probes(int count)
> 	return p == NULL ? NULL : p->probes;
> }
> 
> -static void rcu_free_old_probes(struct rcu_head *head)
> +static void srcu_free_old_probes(struct rcu_head *head)
> {
> 	kfree(container_of(head, struct tp_probes, rcu));
> }
> 
> +static void rcu_free_old_probes(struct rcu_head *head)
> +{
> +	call_srcu(&tracepoint_srcu, head, srcu_free_old_probes);
> +}
> +
> static inline void release_probes(struct tracepoint_func *old)
> {
> 	if (old) {
> 		struct tp_probes *tp_probes = container_of(old,
> 			struct tp_probes, probes[0]);
> +		/*
> +		 * Tracepoint probes are protected by both sched RCU and SRCU,
> +		 * by calling the SRCU callback in the sched RCU callback we
> +		 * cover both cases. So let us chain the SRCU and sched RCU
> +		 * callbacks to wait for both grace periods.
> +		 */
> 		call_rcu_sched(&tp_probes->rcu, rcu_free_old_probes);
> 	}
> }
> --
> 2.18.0.203.gfac676dfb9-goog

-- 
Mathieu Desnoyers
EfficiOS Inc.
http://www.efficios.com

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ