On Thu, Apr 26, 2018 at 09:26:56PM -0700, Joel Fernandes wrote:
> 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.
>
> In the future, we can add a new may_sleep API which can use this
> infrastructure for callbacks that actually can sleep which will support
> Mathieu's usecase of blocking probes.
>
> Test: Tested idle and preempt/irq tracepoints.
Looks good overall! One question and a few comments below.
Thanx, Paul
> [1] https://patchwork.kernel.org/patch/10344297/
>
> Cc: Steven Rostedt <rost...@goodmis.org>
> Cc: Peter Zilstra <pet...@infradead.org>
> Cc: Ingo Molnar <mi...@redhat.com>
> Cc: Mathieu Desnoyers <mathieu.desnoy...@efficios.com>
> Cc: Tom Zanussi <tom.zanu...@linux.intel.com>
> Cc: Namhyung Kim <namhy...@kernel.org>
> Cc: Thomas Glexiner <t...@linutronix.de>
> Cc: Boqun Feng <boqun.f...@gmail.com>
> Cc: Paul McKenney <paul...@linux.vnet.ibm.com>
> Cc: Frederic Weisbecker <fweis...@gmail.com>
> Cc: Randy Dunlap <rdun...@infradead.org>
> Cc: Masami Hiramatsu <mhira...@kernel.org>
> Cc: Fenguang Wu <fengguang...@intel.com>
> Cc: Baohong Liu <baohong....@intel.com>
> Cc: Vedang Patel <vedang.pa...@intel.com>
> Cc: kernel-t...@android.com
> Signed-off-by: Joel Fernandes <joe...@google.com>
> ---
> include/linux/tracepoint.h | 37 +++++++++++++++++++++++++++++--------
> kernel/tracepoint.c | 10 +++++++++-
> 2 files changed, 38 insertions(+), 9 deletions(-)
>
> diff --git a/include/linux/tracepoint.h b/include/linux/tracepoint.h
> index c94f466d57ef..a1c1987de423 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
> @@ -77,6 +80,7 @@ int unregister_tracepoint_module_notifier(struct
> notifier_block *nb)
> */
> static inline void tracepoint_synchronize_unregister(void)
> {
> + synchronize_srcu(&tracepoint_srcu);
> synchronize_sched();
> }
>
> @@ -129,18 +133,26 @@ 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, preempt_on)
> \
> 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); \
> + if (preempt_on) { \
> + WARN_ON_ONCE(in_nmi()); /* no srcu from nmi */ \
Very good on this check, thank you!
> + idx = srcu_read_lock(&tracepoint_srcu); \
Hmmm... Do I need to create a _notrace variant of srcu_read_lock()
and srcu_read_unlock()?
> + it_func_ptr = srcu_dereference((tp)->funcs, \
> + &tracepoint_srcu); \
> + } else { \
> + rcu_read_lock_sched_notrace(); \
> + it_func_ptr = \
> + rcu_dereference_sched((tp)->funcs); \
> + } \
> + \
> if (it_func_ptr) { \
> do { \
> it_func = (it_func_ptr)->func; \
> @@ -148,12 +160,21 @@ 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 (preempt_on) \
> + srcu_read_unlock(&tracepoint_srcu, idx); \
> + else \
> + rcu_read_unlock_sched_notrace(); \
> } while (0)
>
> #ifndef MODULE
> +/*
> + * This is for tracepoints that may be called from an rcu idle path. To make
> + * sure we its safe in the idle path, use the srcu instead of sched-rcu by
> + * specifying the preempt_on flag below. This has the obvious effect that
> any
> + * callback expecting preemption to be disabled should explicitly do so since
> + * with srcu it'd run with preempt on.
> + */
> #define __DECLARE_TRACE_RCU(name, proto, args, cond, data_proto, data_args) \
> static inline void trace_##name##_rcuidle(proto) \
> { \
> diff --git a/kernel/tracepoint.c b/kernel/tracepoint.c
> index 671b13457387..b3b1d65a2460 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;
>
> @@ -67,11 +70,16 @@ 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);
Chaining them through the two callbacks works, good!
> +}
> +
> static inline void release_probes(struct tracepoint_func *old)
> {
> if (old) {
> --
> 2.17.0.441.gb46fe60e1d-goog
>
