linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [PATCH v2] rcu: ftrace: avoid tracing a few functions executed in stop machine
@ 2022-04-23 10:28 Patrick Wang
  2022-04-23 16:36 ` Paul E. McKenney
  2022-04-25 14:56 ` Steven Rostedt
  0 siblings, 2 replies; 4+ messages in thread
From: Patrick Wang @ 2022-04-23 10:28 UTC (permalink / raw)
  To: paulmck, rostedt, frederic, quic_neeraju, josh,
	mathieu.desnoyers, jiangshanlai, joel
  Cc: rcu, linux-kernel, patrick.wang.shcn

Because of the change of stop machine implementation, there are functions
being called while waiting currently:

----------------------------------------------------------------
Former stop machine wait loop:
do {
    cpu_relax(); => macro
    ...
} while (curstate != STOPMACHINE_EXIT);
-----------------------------------------------------------------
Current stop machine wait loop:
do {
    stop_machine_yield(cpumask); => function (notraced)
    ...
    touch_nmi_watchdog(); => function (notraced, inside calls also notraced)
    ...
    rcu_momentary_dyntick_idle(); => function (notraced, inside calls traced)
} while (curstate != MULTI_STOP_EXIT);
------------------------------------------------------------------

These functions (including the calls inside) should be marked notrace to avoid
their codes being updated when they are being called. The calls inside
rcu_momentary_dyntick_idle() still remain traced, and will cause crash:

  rcu: INFO: rcu_preempt detected stalls on CPUs/tasks:
  rcu: 	1-...!: (0 ticks this GP) idle=14f/1/0x4000000000000000 softirq=3397/3397 fqs=0
  rcu: 	3-...!: (0 ticks this GP) idle=ee9/1/0x4000000000000000 softirq=5168/5168 fqs=0
  	(detected by 0, t=8137 jiffies, g=5889, q=2 ncpus=4)
  Task dump for CPU 1:
  task:migration/1     state:R  running task     stack:    0 pid:   19 ppid:     2 flags:0x00000000
  Stopper: multi_cpu_stop+0x0/0x18c <- stop_machine_cpuslocked+0x128/0x174
  Call Trace:
  Task dump for CPU 3:
  task:migration/3     state:R  running task     stack:    0 pid:   29 ppid:     2 flags:0x00000000
  Stopper: multi_cpu_stop+0x0/0x18c <- stop_machine_cpuslocked+0x128/0x174
  Call Trace:
  rcu: rcu_preempt kthread timer wakeup didn't happen for 8136 jiffies! g5889 f0x0 RCU_GP_WAIT_FQS(5) ->state=0x402
  rcu: 	Possible timer handling issue on cpu=2 timer-softirq=594
  rcu: rcu_preempt kthread starved for 8137 jiffies! g5889 f0x0 RCU_GP_WAIT_FQS(5) ->state=0x402 ->cpu=2
  rcu: 	Unless rcu_preempt kthread gets sufficient CPU time, OOM is now expected behavior.
  rcu: RCU grace-period kthread stack dump:
  task:rcu_preempt     state:I stack:    0 pid:   14 ppid:     2 flags:0x00000000
  Call Trace:
    schedule+0x56/0xc2
    schedule_timeout+0x82/0x184
    rcu_gp_fqs_loop+0x19a/0x318
    rcu_gp_kthread+0x11a/0x140
    kthread+0xee/0x118
    ret_from_exception+0x0/0x14
  rcu: Stack dump where RCU GP kthread last ran:
  Task dump for CPU 2:
  task:migration/2     state:R  running task     stack:    0 pid:   24 ppid:     2 flags:0x00000000
  Stopper: multi_cpu_stop+0x0/0x18c <- stop_machine_cpuslocked+0x128/0x174
  Call Trace:

Mark the calls inside rcu_momentary_dyntick_idle():
 rcu_preempt_deferred_qs()
 rcu_preempt_need_deferred_qs()
 rcu_preempt_deferred_qs_irqrestore()
as notrace to prevent this.

Signed-off-by: Patrick Wang <patrick.wang.shcn@gmail.com>
---
v1->v2:
 - Modify log message.

 kernel/rcu/tree_plugin.h | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/kernel/rcu/tree_plugin.h b/kernel/rcu/tree_plugin.h
index c8ba0fe17267..971bb6a00ede 100644
--- a/kernel/rcu/tree_plugin.h
+++ b/kernel/rcu/tree_plugin.h
@@ -460,7 +460,7 @@ static bool rcu_preempt_has_tasks(struct rcu_node *rnp)
  * be quite short, for example, in the case of the call from
  * rcu_read_unlock_special().
  */
-static void
+notrace static void
 rcu_preempt_deferred_qs_irqrestore(struct task_struct *t, unsigned long flags)
 {
 	bool empty_exp;
@@ -581,7 +581,7 @@ rcu_preempt_deferred_qs_irqrestore(struct task_struct *t, unsigned long flags)
  * is disabled.  This function cannot be expected to understand these
  * nuances, so the caller must handle them.
  */
-static bool rcu_preempt_need_deferred_qs(struct task_struct *t)
+notrace static bool rcu_preempt_need_deferred_qs(struct task_struct *t)
 {
 	return (__this_cpu_read(rcu_data.cpu_no_qs.b.exp) ||
 		READ_ONCE(t->rcu_read_unlock_special.s)) &&
@@ -595,7 +595,7 @@ static bool rcu_preempt_need_deferred_qs(struct task_struct *t)
  * evaluate safety in terms of interrupt, softirq, and preemption
  * disabling.
  */
-static void rcu_preempt_deferred_qs(struct task_struct *t)
+notrace static void rcu_preempt_deferred_qs(struct task_struct *t)
 {
 	unsigned long flags;
 
-- 
2.25.1


^ permalink raw reply related	[flat|nested] 4+ messages in thread
* Re: [PATCH v2] rcu: ftrace: avoid tracing a few functions executed in stop machine
  2022-04-23 10:28 [PATCH v2] rcu: ftrace: avoid tracing a few functions executed in stop machine Patrick Wang
@ 2022-04-23 16:36 ` Paul E. McKenney
  2022-04-25 14:56 ` Steven Rostedt
  1 sibling, 0 replies; 4+ messages in thread
From: Paul E. McKenney @ 2022-04-23 16:36 UTC (permalink / raw)
  To: Patrick Wang
  Cc: rostedt, frederic, quic_neeraju, josh, mathieu.desnoyers,
	jiangshanlai, joel, rcu, linux-kernel

On Sat, Apr 23, 2022 at 06:28:44PM +0800, Patrick Wang wrote:
> Because of the change of stop machine implementation, there are functions
> being called while waiting currently:
> 
> ----------------------------------------------------------------
> Former stop machine wait loop:
> do {
>     cpu_relax(); => macro
>     ...
> } while (curstate != STOPMACHINE_EXIT);
> -----------------------------------------------------------------
> Current stop machine wait loop:
> do {
>     stop_machine_yield(cpumask); => function (notraced)
>     ...
>     touch_nmi_watchdog(); => function (notraced, inside calls also notraced)
>     ...
>     rcu_momentary_dyntick_idle(); => function (notraced, inside calls traced)
> } while (curstate != MULTI_STOP_EXIT);
> ------------------------------------------------------------------
> 
> These functions (including the calls inside) should be marked notrace to avoid
> their codes being updated when they are being called. The calls inside
> rcu_momentary_dyntick_idle() still remain traced, and will cause crash:
> 
>   rcu: INFO: rcu_preempt detected stalls on CPUs/tasks:
>   rcu: 	1-...!: (0 ticks this GP) idle=14f/1/0x4000000000000000 softirq=3397/3397 fqs=0
>   rcu: 	3-...!: (0 ticks this GP) idle=ee9/1/0x4000000000000000 softirq=5168/5168 fqs=0
>   	(detected by 0, t=8137 jiffies, g=5889, q=2 ncpus=4)
>   Task dump for CPU 1:
>   task:migration/1     state:R  running task     stack:    0 pid:   19 ppid:     2 flags:0x00000000
>   Stopper: multi_cpu_stop+0x0/0x18c <- stop_machine_cpuslocked+0x128/0x174
>   Call Trace:
>   Task dump for CPU 3:
>   task:migration/3     state:R  running task     stack:    0 pid:   29 ppid:     2 flags:0x00000000
>   Stopper: multi_cpu_stop+0x0/0x18c <- stop_machine_cpuslocked+0x128/0x174
>   Call Trace:
>   rcu: rcu_preempt kthread timer wakeup didn't happen for 8136 jiffies! g5889 f0x0 RCU_GP_WAIT_FQS(5) ->state=0x402
>   rcu: 	Possible timer handling issue on cpu=2 timer-softirq=594
>   rcu: rcu_preempt kthread starved for 8137 jiffies! g5889 f0x0 RCU_GP_WAIT_FQS(5) ->state=0x402 ->cpu=2
>   rcu: 	Unless rcu_preempt kthread gets sufficient CPU time, OOM is now expected behavior.
>   rcu: RCU grace-period kthread stack dump:
>   task:rcu_preempt     state:I stack:    0 pid:   14 ppid:     2 flags:0x00000000
>   Call Trace:
>     schedule+0x56/0xc2
>     schedule_timeout+0x82/0x184
>     rcu_gp_fqs_loop+0x19a/0x318
>     rcu_gp_kthread+0x11a/0x140
>     kthread+0xee/0x118
>     ret_from_exception+0x0/0x14
>   rcu: Stack dump where RCU GP kthread last ran:
>   Task dump for CPU 2:
>   task:migration/2     state:R  running task     stack:    0 pid:   24 ppid:     2 flags:0x00000000
>   Stopper: multi_cpu_stop+0x0/0x18c <- stop_machine_cpuslocked+0x128/0x174
>   Call Trace:
> 
> Mark the calls inside rcu_momentary_dyntick_idle():
>  rcu_preempt_deferred_qs()
>  rcu_preempt_need_deferred_qs()
>  rcu_preempt_deferred_qs_irqrestore()
> as notrace to prevent this.
> 
> Signed-off-by: Patrick Wang <patrick.wang.shcn@gmail.com>

Thank you, Patrick!

Steve, are you good with this version?  If so, I will replace the older
version in -rcu with this version.

							Thanx, Paul

> ---
> v1->v2:
>  - Modify log message.
> 
>  kernel/rcu/tree_plugin.h | 6 +++---
>  1 file changed, 3 insertions(+), 3 deletions(-)
> 
> diff --git a/kernel/rcu/tree_plugin.h b/kernel/rcu/tree_plugin.h
> index c8ba0fe17267..971bb6a00ede 100644
> --- a/kernel/rcu/tree_plugin.h
> +++ b/kernel/rcu/tree_plugin.h
> @@ -460,7 +460,7 @@ static bool rcu_preempt_has_tasks(struct rcu_node *rnp)
>   * be quite short, for example, in the case of the call from
>   * rcu_read_unlock_special().
>   */
> -static void
> +notrace static void
>  rcu_preempt_deferred_qs_irqrestore(struct task_struct *t, unsigned long flags)
>  {
>  	bool empty_exp;
> @@ -581,7 +581,7 @@ rcu_preempt_deferred_qs_irqrestore(struct task_struct *t, unsigned long flags)
>   * is disabled.  This function cannot be expected to understand these
>   * nuances, so the caller must handle them.
>   */
> -static bool rcu_preempt_need_deferred_qs(struct task_struct *t)
> +notrace static bool rcu_preempt_need_deferred_qs(struct task_struct *t)
>  {
>  	return (__this_cpu_read(rcu_data.cpu_no_qs.b.exp) ||
>  		READ_ONCE(t->rcu_read_unlock_special.s)) &&
> @@ -595,7 +595,7 @@ static bool rcu_preempt_need_deferred_qs(struct task_struct *t)
>   * evaluate safety in terms of interrupt, softirq, and preemption
>   * disabling.
>   */
> -static void rcu_preempt_deferred_qs(struct task_struct *t)
> +notrace static void rcu_preempt_deferred_qs(struct task_struct *t)
>  {
>  	unsigned long flags;
>  
> -- 
> 2.25.1
> 

^ permalink raw reply	[flat|nested] 4+ messages in thread
* Re: [PATCH v2] rcu: ftrace: avoid tracing a few functions executed in stop machine
  2022-04-23 10:28 [PATCH v2] rcu: ftrace: avoid tracing a few functions executed in stop machine Patrick Wang
  2022-04-23 16:36 ` Paul E. McKenney
@ 2022-04-25 14:56 ` Steven Rostedt
  2022-04-26 10:32   ` patrick wang
  1 sibling, 1 reply; 4+ messages in thread
From: Steven Rostedt @ 2022-04-25 14:56 UTC (permalink / raw)
  To: Patrick Wang
  Cc: paulmck, frederic, quic_neeraju, josh, mathieu.desnoyers,
	jiangshanlai, joel, rcu, linux-kernel

On Sat, 23 Apr 2022 18:28:44 +0800
Patrick Wang <patrick.wang.shcn@gmail.com> wrote:

> Signed-off-by: Patrick Wang <patrick.wang.shcn@gmail.com>
> ---
> v1->v2:
>  - Modify log message.
> 
>  kernel/rcu/tree_plugin.h | 6 +++---
>  1 file changed, 3 insertions(+), 3 deletions(-)
> 
> diff --git a/kernel/rcu/tree_plugin.h b/kernel/rcu/tree_plugin.h
> index c8ba0fe17267..971bb6a00ede 100644
> --- a/kernel/rcu/tree_plugin.h
> +++ b/kernel/rcu/tree_plugin.h
> @@ -460,7 +460,7 @@ static bool rcu_preempt_has_tasks(struct rcu_node *rnp)
>   * be quite short, for example, in the case of the call from
>   * rcu_read_unlock_special().
>   */
> -static void
> +notrace static void

I'm fine with the change, but to be consistent with the rest of the kernel,
static needs to come before notrace.

Either:

 static notrace void

or even (some places have):

  static void notrace

but "static" should always be first.

-- Steve


>  rcu_preempt_deferred_qs_irqrestore(struct task_struct *t, unsigned long flags)
>  {
>  	bool empty_exp;
> @@ -581,7 +581,7 @@ rcu_preempt_deferred_qs_irqrestore(struct task_struct *t, unsigned long flags)
>   * is disabled.  This function cannot be expected to understand these
>   * nuances, so the caller must handle them.
>   */
> -static bool rcu_preempt_need_deferred_qs(struct task_struct *t)
> +notrace static bool rcu_preempt_need_deferred_qs(struct task_struct *t)
>  {
>  	return (__this_cpu_read(rcu_data.cpu_no_qs.b.exp) ||
>  		READ_ONCE(t->rcu_read_unlock_special.s)) &&
> @@ -595,7 +595,7 @@ static bool rcu_preempt_need_deferred_qs(struct task_struct *t)
>   * evaluate safety in terms of interrupt, softirq, and preemption
>   * disabling.
>   */
> -static void rcu_preempt_deferred_qs(struct task_struct *t)
> +notrace static void rcu_preempt_deferred_qs(struct task_struct *t)
>  {
>  	unsigned long flags;
>  


^ permalink raw reply	[flat|nested] 4+ messages in thread
* Re: [PATCH v2] rcu: ftrace: avoid tracing a few functions executed in stop machine
  2022-04-25 14:56 ` Steven Rostedt
@ 2022-04-26 10:32   ` patrick wang
  0 siblings, 0 replies; 4+ messages in thread
From: patrick wang @ 2022-04-26 10:32 UTC (permalink / raw)
  To: Steven Rostedt
  Cc: paulmck, frederic, quic_neeraju, josh, mathieu.desnoyers,
	jiangshanlai, joel, rcu, linux-kernel

On Mon, Apr 25, 2022 at 10:56 PM Steven Rostedt <rostedt@goodmis.org> wrote:
>
> On Sat, 23 Apr 2022 18:28:44 +0800
> Patrick Wang <patrick.wang.shcn@gmail.com> wrote:
>
> > Signed-off-by: Patrick Wang <patrick.wang.shcn@gmail.com>
> > ---
> > v1->v2:
> >  - Modify log message.
> >
> >  kernel/rcu/tree_plugin.h | 6 +++---
> >  1 file changed, 3 insertions(+), 3 deletions(-)
> >
> > diff --git a/kernel/rcu/tree_plugin.h b/kernel/rcu/tree_plugin.h
> > index c8ba0fe17267..971bb6a00ede 100644
> > --- a/kernel/rcu/tree_plugin.h
> > +++ b/kernel/rcu/tree_plugin.h
> > @@ -460,7 +460,7 @@ static bool rcu_preempt_has_tasks(struct rcu_node *rnp)
> >   * be quite short, for example, in the case of the call from
> >   * rcu_read_unlock_special().
> >   */
> > -static void
> > +notrace static void
>
> I'm fine with the change, but to be consistent with the rest of the kernel,
> static needs to come before notrace.
>
> Either:
>
>  static notrace void
>
> or even (some places have):
>
>   static void notrace
>
> but "static" should always be first.
>

Thank you for the detailed advice. I will move "notrace" to behind "static".

Thanks,
Patrick

>
>
> >  rcu_preempt_deferred_qs_irqrestore(struct task_struct *t, unsigned long flags)
> >  {
> >       bool empty_exp;
> > @@ -581,7 +581,7 @@ rcu_preempt_deferred_qs_irqrestore(struct task_struct *t, unsigned long flags)
> >   * is disabled.  This function cannot be expected to understand these
> >   * nuances, so the caller must handle them.
> >   */
> > -static bool rcu_preempt_need_deferred_qs(struct task_struct *t)
> > +notrace static bool rcu_preempt_need_deferred_qs(struct task_struct *t)
> >  {
> >       return (__this_cpu_read(rcu_data.cpu_no_qs.b.exp) ||
> >               READ_ONCE(t->rcu_read_unlock_special.s)) &&
> > @@ -595,7 +595,7 @@ static bool rcu_preempt_need_deferred_qs(struct task_struct *t)
> >   * evaluate safety in terms of interrupt, softirq, and preemption
> >   * disabling.
> >   */
> > -static void rcu_preempt_deferred_qs(struct task_struct *t)
> > +notrace static void rcu_preempt_deferred_qs(struct task_struct *t)
> >  {
> >       unsigned long flags;
> >
>

^ permalink raw reply	[flat|nested] 4+ messages in thread
end of thread, other threads:[~2022-04-26 10:40 UTC | newest]

Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2022-04-23 10:28 [PATCH v2] rcu: ftrace: avoid tracing a few functions executed in stop machine Patrick Wang
2022-04-23 16:36 ` Paul E. McKenney
2022-04-25 14:56 ` Steven Rostedt
2022-04-26 10:32   ` patrick wang

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).