Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1363050 > unrolled thread

Re: [PATCH v3 4/4] nmi_backtrace: generate one-line reports for idle cpus

Back to article view | Back to linux.kernel

This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by below is the oldest one visible, not the original post.

Contents

  Re: [PATCH v3 4/4] nmi_backtrace: generate one-line reports for idle cpus "Rafael J. Wysocki" <rjw@rjwysocki.net> - 2016-03-22 23:30 +0100
    Re: [PATCH v3 4/4] nmi_backtrace: generate one-line reports for idle  cpus Peter Zijlstra <peterz@infradead.org> - 2016-03-22 23:50 +0100
      Re: [PATCH v3 4/4] nmi_backtrace: generate one-line reports for idle cpus "Rafael J. Wysocki" <rjw@rjwysocki.net> - 2016-03-23 01:50 +0100
        Re: [PATCH v3 4/4] nmi_backtrace: generate one-line reports for idle  cpus Peter Zijlstra <peterz@infradead.org> - 2016-03-23 09:00 +0100

#1363050 — Re: [PATCH v3 4/4] nmi_backtrace: generate one-line reports for idle cpus

On Tuesday, March 22, 2016 01:19:39 PM Chris Metcalf wrote:
> When doing an nmi backtrace of many cores, most of which are idle,
> the output is a little overwhelming and very uninformative.  Suppress
> messages for cpus that are idling when they are interrupted and just
> emit one line, "NMI backtrace for N skipped: idling at pc 0xNNN".
> 
> We do this by grouping all the cpuidle code together into a new
> .cpuidle.text section, and then checking the address of the
> interrupted PC to see if it lies within that section.
> 
> This commit suitably tags x86, arm64, and tile idle routines,
> and only adds in the minimal framework for other architectures.
> 
> Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
> Tested-by: Peter Zijlstra (Intel) <peterz@infradead.org>
> Signed-off-by: Chris Metcalf <cmetcalf@mellanox.com>
> ---

[cut]

> diff --git a/drivers/idle/intel_idle.c b/drivers/idle/intel_idle.c
> index cd4510a63375..924554f920fb 100644
> --- a/drivers/idle/intel_idle.c
> +++ b/drivers/idle/intel_idle.c
> @@ -725,8 +725,8 @@ static struct cpuidle_state avn_cstates[] = {
>   *
>   * Must be called under local_irq_disable().
>   */
> -static int intel_idle(struct cpuidle_device *dev,
> -		struct cpuidle_driver *drv, int index)
> +static __cpuidle int intel_idle(struct cpuidle_device *dev,
> +				struct cpuidle_driver *drv, int index)
>  {
>  	unsigned long ecx = 1; /* break on interrupt flag */
>  	struct cpuidle_state *state = &drv->states[index];

Well, what about intel_idle_freeze()?  Or do we not care?

And analogous stuff in processor_idle.c for that matter?

acpi_idle_enter()/acpi_idle_enter_freeze() plus stuff called by those?

Thanks,
Rafael

[toc] | [next] | [standalone]

#1363065 — Re: [PATCH v3 4/4] nmi_backtrace: generate one-line reports for idle cpus

On Tue, Mar 22, 2016 at 11:31:11PM +0100, Rafael J. Wysocki wrote:

> > diff --git a/drivers/idle/intel_idle.c b/drivers/idle/intel_idle.c
> > index cd4510a63375..924554f920fb 100644
> > --- a/drivers/idle/intel_idle.c
> > +++ b/drivers/idle/intel_idle.c
> > @@ -725,8 +725,8 @@ static struct cpuidle_state avn_cstates[] = {
> >   *
> >   * Must be called under local_irq_disable().
> >   */
> > -static int intel_idle(struct cpuidle_device *dev,
> > -		struct cpuidle_driver *drv, int index)
> > +static __cpuidle int intel_idle(struct cpuidle_device *dev,
> > +				struct cpuidle_driver *drv, int index)
> >  {
> >  	unsigned long ecx = 1; /* break on interrupt flag */
> >  	struct cpuidle_state *state = &drv->states[index];
> 
> Well, what about intel_idle_freeze()?  Or do we not care?

I argued against it; when you're suspended the NMI watchdog is stopped
too. Then again, you've more experience debugging that thing, so if
you think its useful its not much effort adding it.

> And analogous stuff in processor_idle.c for that matter?
> 
> acpi_idle_enter()/acpi_idle_enter_freeze() plus stuff called by those?

Ah, I only tagged acpi_processor_ffh_cstate_enter() because I went from
mwait_idle_with_hints(), I suppose acpi_safe_halt() and
acpi_idle_do_entry() itself for the INB method should cover it?

(This being one of the reasons I asked Chris to Cc you; you know this
stuff far better than I do)


---
 drivers/acpi/processor_idle.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/acpi/processor_idle.c b/drivers/acpi/processor_idle.c
index 175c86bee3a9..d5b11fff9e88 100644
--- a/drivers/acpi/processor_idle.c
+++ b/drivers/acpi/processor_idle.c
@@ -111,7 +111,7 @@ static const struct dmi_system_id processor_power_dmi_table[] = {
  * Callers should disable interrupts before the call and enable
  * interrupts after return.
  */
-static void acpi_safe_halt(void)
+__cpuidle static void acpi_safe_halt(void)
 {
 	if (!tif_need_resched()) {
 		safe_halt();
@@ -680,7 +680,7 @@ static int acpi_idle_bm_check(void)
  *
  * Caller disables interrupt before call and enables interrupt after return.
  */
-static void acpi_idle_do_entry(struct acpi_processor_cx *cx)
+__cpuidle static void acpi_idle_do_entry(struct acpi_processor_cx *cx)
 {
 	if (cx->entry_method == ACPI_CSTATE_FFH) {
 		/* Call into architectural FFH based C-state */

[toc] | [prev] | [next] | [standalone]

#1363143

On Tuesday, March 22, 2016 11:45:57 PM Peter Zijlstra wrote:
> On Tue, Mar 22, 2016 at 11:31:11PM +0100, Rafael J. Wysocki wrote:
> 
> > > diff --git a/drivers/idle/intel_idle.c b/drivers/idle/intel_idle.c
> > > index cd4510a63375..924554f920fb 100644
> > > --- a/drivers/idle/intel_idle.c
> > > +++ b/drivers/idle/intel_idle.c
> > > @@ -725,8 +725,8 @@ static struct cpuidle_state avn_cstates[] = {
> > >   *
> > >   * Must be called under local_irq_disable().
> > >   */
> > > -static int intel_idle(struct cpuidle_device *dev,
> > > -		struct cpuidle_driver *drv, int index)
> > > +static __cpuidle int intel_idle(struct cpuidle_device *dev,
> > > +				struct cpuidle_driver *drv, int index)
> > >  {
> > >  	unsigned long ecx = 1; /* break on interrupt flag */
> > >  	struct cpuidle_state *state = &drv->states[index];
> > 
> > Well, what about intel_idle_freeze()?  Or do we not care?
> 
> I argued against it; when you're suspended the NMI watchdog is stopped
> too.

Is it also stopped for suspend-to-idle?  I'm not sure about that.

Where do I need to look to find out?

> Then again, you've more experience debugging that thing, so if
> you think its useful its not much effort adding it.
> 
> > And analogous stuff in processor_idle.c for that matter?
> > 
> > acpi_idle_enter()/acpi_idle_enter_freeze() plus stuff called by those?
> 
> Ah, I only tagged acpi_processor_ffh_cstate_enter() because I went from
> mwait_idle_with_hints(), I suppose acpi_safe_halt() and
> acpi_idle_do_entry() itself for the INB method should cover it?

Yes, these two should be sufficient.

> (This being one of the reasons I asked Chris to Cc you; you know this
> stuff far better than I do)
> 
> ---
>  drivers/acpi/processor_idle.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/acpi/processor_idle.c b/drivers/acpi/processor_idle.c
> index 175c86bee3a9..d5b11fff9e88 100644
> --- a/drivers/acpi/processor_idle.c
> +++ b/drivers/acpi/processor_idle.c
> @@ -111,7 +111,7 @@ static const struct dmi_system_id processor_power_dmi_table[] = {
>   * Callers should disable interrupts before the call and enable
>   * interrupts after return.
>   */
> -static void acpi_safe_halt(void)
> +__cpuidle static void acpi_safe_halt(void)
>  {
>  	if (!tif_need_resched()) {
>  		safe_halt();
> @@ -680,7 +680,7 @@ static int acpi_idle_bm_check(void)
>   *
>   * Caller disables interrupt before call and enables interrupt after return.
>   */
> -static void acpi_idle_do_entry(struct acpi_processor_cx *cx)
> +__cpuidle static void acpi_idle_do_entry(struct acpi_processor_cx *cx)
>  {
>  	if (cx->entry_method == ACPI_CSTATE_FFH) {
>  		/* Call into architectural FFH based C-state */

[toc] | [prev] | [next] | [standalone]

#1363230 — Re: [PATCH v3 4/4] nmi_backtrace: generate one-line reports for idle cpus

On Wed, Mar 23, 2016 at 01:50:00AM +0100, Rafael J. Wysocki wrote:

> > > Well, what about intel_idle_freeze()?  Or do we not care?
> > 
> > I argued against it; when you're suspended the NMI watchdog is stopped
> > too.
> 
> Is it also stopped for suspend-to-idle?  I'm not sure about that.
> 
> Where do I need to look to find out?

Hmm I have memories of writing a patch to that effect when we were
starting with that suspend-to-idle stuff, because people didn't like
being woken up all the time.

But now that I look I cannot find it either..

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web