[v1,05/12] PM: sleep: stats: Use step_failures[0] as a counter of successful cycles

Message ID 3290637.44csPzL39Z@kreacher
State New
Headers
Series PM: sleep: Fix up suspend stats handling and clean up core code |

Commit Message

Rafael J. Wysocki Jan. 22, 2024, 11:29 a.m. UTC
  From: Rafael J. Wysocki <rafael.j.wysocki@intel.com>

The first (index 0) cell of the step_failures[] array in struct
suspend_stats introduced previously can be used as a counter of
successful suspend-resume cycles instead of the separate "success"
field in it, so do that.

While at it, change the type of the "fail" field in struct
suspend_stats to unsigned int, because it cannot be negative.

No intentional functional impact.

Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
---
 include/linux/suspend.h |    3 +--
 kernel/power/main.c     |    9 +++++----
 kernel/power/suspend.c  |    2 +-
 3 files changed, 7 insertions(+), 7 deletions(-)
  

Comments

Stanislaw Gruszka Jan. 25, 2024, 7:52 a.m. UTC | #1
On Mon, Jan 22, 2024 at 12:29:11PM +0100, Rafael J. Wysocki wrote:
> From: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
> 
> The first (index 0) cell of the step_failures[] array in struct
> suspend_stats introduced previously can be used as a counter of
> successful suspend-resume cycles instead of the separate "success"
> field in it, so do that.
> 
> While at it, change the type of the "fail" field in struct
> suspend_stats to unsigned int, because it cannot be negative.
> 
> No intentional functional impact.
> 
> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
> ---
>  include/linux/suspend.h |    3 +--
>  kernel/power/main.c     |    9 +++++----
>  kernel/power/suspend.c  |    2 +-
>  3 files changed, 7 insertions(+), 7 deletions(-)
> 
> Index: linux-pm/include/linux/suspend.h
> ===================================================================
> --- linux-pm.orig/include/linux/suspend.h
> +++ linux-pm/include/linux/suspend.h
> @@ -55,8 +55,7 @@ enum suspend_stat_step {
>  
>  struct suspend_stats {
>  	unsigned int step_failures[SUSPEND_NR_STEPS];
> -	int	success;
<snip>
> -		   suspend_stats.success, suspend_stats.fail);
> +	seq_printf(s, "success: %u\nfail: %u\n",
> +		   suspend_stats.step_failures[SUSPEND_NONE],
> +		   suspend_stats.fail);
>  
>  	for (step = SUSPEND_FREEZE; step < SUSPEND_NR_STEPS; step++)
>  		seq_printf(s, "failed_%s: %u\n", suspend_step_names[step],
> Index: linux-pm/kernel/power/suspend.c
> ===================================================================
> --- linux-pm.orig/kernel/power/suspend.c
> +++ linux-pm/kernel/power/suspend.c
> @@ -620,7 +620,7 @@ int pm_suspend(suspend_state_t state)
>  		suspend_stats.fail++;
>  		dpm_save_failed_errno(error);
>  	} else {
> -		suspend_stats.success++;
> +		suspend_stats.step_failures[SUSPEND_NONE]++;

This looks confusing for me. I think would be better keep
success field and just remove SUSPEND_NONE from the 
suspend_stat_step and suspend_stat_names. Actually do
not introduce it, SUSPEND_NONE does not seems to be necessary
(SUSPEND_FREEZE can be 0).

Regards
Stanislaw
  
Rafael J. Wysocki Jan. 25, 2024, 3:11 p.m. UTC | #2
On Thu, Jan 25, 2024 at 8:52 AM Stanislaw Gruszka
<stanislaw.gruszka@linux.intel.com> wrote:
>
> On Mon, Jan 22, 2024 at 12:29:11PM +0100, Rafael J. Wysocki wrote:
> > From: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
> >
> > The first (index 0) cell of the step_failures[] array in struct
> > suspend_stats introduced previously can be used as a counter of
> > successful suspend-resume cycles instead of the separate "success"
> > field in it, so do that.
> >
> > While at it, change the type of the "fail" field in struct
> > suspend_stats to unsigned int, because it cannot be negative.
> >
> > No intentional functional impact.
> >
> > Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
> > ---
> >  include/linux/suspend.h |    3 +--
> >  kernel/power/main.c     |    9 +++++----
> >  kernel/power/suspend.c  |    2 +-
> >  3 files changed, 7 insertions(+), 7 deletions(-)
> >
> > Index: linux-pm/include/linux/suspend.h
> > ===================================================================
> > --- linux-pm.orig/include/linux/suspend.h
> > +++ linux-pm/include/linux/suspend.h
> > @@ -55,8 +55,7 @@ enum suspend_stat_step {
> >
> >  struct suspend_stats {
> >       unsigned int step_failures[SUSPEND_NR_STEPS];
> > -     int     success;
> <snip>
> > -                suspend_stats.success, suspend_stats.fail);
> > +     seq_printf(s, "success: %u\nfail: %u\n",
> > +                suspend_stats.step_failures[SUSPEND_NONE],
> > +                suspend_stats.fail);
> >
> >       for (step = SUSPEND_FREEZE; step < SUSPEND_NR_STEPS; step++)
> >               seq_printf(s, "failed_%s: %u\n", suspend_step_names[step],
> > Index: linux-pm/kernel/power/suspend.c
> > ===================================================================
> > --- linux-pm.orig/kernel/power/suspend.c
> > +++ linux-pm/kernel/power/suspend.c
> > @@ -620,7 +620,7 @@ int pm_suspend(suspend_state_t state)
> >               suspend_stats.fail++;
> >               dpm_save_failed_errno(error);
> >       } else {
> > -             suspend_stats.success++;
> > +             suspend_stats.step_failures[SUSPEND_NONE]++;
>
> This looks confusing for me. I think would be better keep
> success field and just remove SUSPEND_NONE from the
> suspend_stat_step and suspend_stat_names. Actually do
> not introduce it, SUSPEND_NONE does not seems to be necessary
> (SUSPEND_FREEZE can be 0).

OK

I'll need to rearrange the series for that somewhat except for the
first two patches.

I guess it's OK to retain the R-by tags?

Thanks for all of the reviews!
  
Stanislaw Gruszka Jan. 25, 2024, 5:28 p.m. UTC | #3
On Thu, Jan 25, 2024 at 04:11:08PM +0100, Rafael J. Wysocki wrote:
> On Thu, Jan 25, 2024 at 8:52 AM Stanislaw Gruszka
> <stanislaw.gruszka@linux.intel.com> wrote:
> >
> > On Mon, Jan 22, 2024 at 12:29:11PM +0100, Rafael J. Wysocki wrote:
> > > From: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
> > >
> > > The first (index 0) cell of the step_failures[] array in struct
> > > suspend_stats introduced previously can be used as a counter of
> > > successful suspend-resume cycles instead of the separate "success"
> > > field in it, so do that.
> > >
> > > While at it, change the type of the "fail" field in struct
> > > suspend_stats to unsigned int, because it cannot be negative.
> > >
> > > No intentional functional impact.
> > >
> > > Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
> > > ---
> > >  include/linux/suspend.h |    3 +--
> > >  kernel/power/main.c     |    9 +++++----
> > >  kernel/power/suspend.c  |    2 +-
> > >  3 files changed, 7 insertions(+), 7 deletions(-)
> > >
> > > Index: linux-pm/include/linux/suspend.h
> > > ===================================================================
> > > --- linux-pm.orig/include/linux/suspend.h
> > > +++ linux-pm/include/linux/suspend.h
> > > @@ -55,8 +55,7 @@ enum suspend_stat_step {
> > >
> > >  struct suspend_stats {
> > >       unsigned int step_failures[SUSPEND_NR_STEPS];
> > > -     int     success;
> > <snip>
> > > -                suspend_stats.success, suspend_stats.fail);
> > > +     seq_printf(s, "success: %u\nfail: %u\n",
> > > +                suspend_stats.step_failures[SUSPEND_NONE],
> > > +                suspend_stats.fail);
> > >
> > >       for (step = SUSPEND_FREEZE; step < SUSPEND_NR_STEPS; step++)
> > >               seq_printf(s, "failed_%s: %u\n", suspend_step_names[step],
> > > Index: linux-pm/kernel/power/suspend.c
> > > ===================================================================
> > > --- linux-pm.orig/kernel/power/suspend.c
> > > +++ linux-pm/kernel/power/suspend.c
> > > @@ -620,7 +620,7 @@ int pm_suspend(suspend_state_t state)
> > >               suspend_stats.fail++;
> > >               dpm_save_failed_errno(error);
> > >       } else {
> > > -             suspend_stats.success++;
> > > +             suspend_stats.step_failures[SUSPEND_NONE]++;
> >
> > This looks confusing for me. I think would be better keep
> > success field and just remove SUSPEND_NONE from the
> > suspend_stat_step and suspend_stat_names. Actually do
> > not introduce it, SUSPEND_NONE does not seems to be necessary
> > (SUSPEND_FREEZE can be 0).
> 
> OK
> 
> I'll need to rearrange the series for that somewhat except for the
> first two patches.

I wouldn't mind to skip this change and just remove SUSPEND_NONE
in separate patch.

> I guess it's OK to retain the R-by tags?

Yes, is OK.
 
Regards
Stanislaw
  

Patch

Index: linux-pm/include/linux/suspend.h
===================================================================
--- linux-pm.orig/include/linux/suspend.h
+++ linux-pm/include/linux/suspend.h
@@ -55,8 +55,7 @@  enum suspend_stat_step {
 
 struct suspend_stats {
 	unsigned int step_failures[SUSPEND_NR_STEPS];
-	int	success;
-	int	fail;
+	unsigned int fail;
 #define	REC_FAILED_NUM	2
 	int	last_failed_dev;
 	char	failed_devs[REC_FAILED_NUM][40];
Index: linux-pm/kernel/power/main.c
===================================================================
--- linux-pm.orig/kernel/power/main.c
+++ linux-pm/kernel/power/main.c
@@ -339,8 +339,7 @@  static ssize_t _name##_show(struct kobje
 }								\
 static struct kobj_attribute _name = __ATTR_RO(_name)
 
-suspend_attr(success, "%d\n");
-suspend_attr(fail, "%d\n");
+suspend_attr(fail, "%u\n");
 suspend_attr(last_hw_sleep, "%llu\n");
 suspend_attr(total_hw_sleep, "%llu\n");
 suspend_attr(max_hw_sleep, "%llu\n");
@@ -354,6 +353,7 @@  static ssize_t _name##_show(struct kobje
 }								\
 static struct kobj_attribute _name = __ATTR_RO(_name)
 
+suspend_step_attr(success, SUSPEND_NONE);
 suspend_step_attr(failed_freeze, SUSPEND_FREEZE);
 suspend_step_attr(failed_prepare, SUSPEND_PREPARE);
 suspend_step_attr(failed_suspend, SUSPEND_SUSPEND);
@@ -458,8 +458,9 @@  static int suspend_stats_show(struct seq
 	last_step = suspend_stats.last_failed_step + REC_FAILED_NUM - 1;
 	last_step %= REC_FAILED_NUM;
 
-	seq_printf(s, "success: %d\nfail: %d\n",
-		   suspend_stats.success, suspend_stats.fail);
+	seq_printf(s, "success: %u\nfail: %u\n",
+		   suspend_stats.step_failures[SUSPEND_NONE],
+		   suspend_stats.fail);
 
 	for (step = SUSPEND_FREEZE; step < SUSPEND_NR_STEPS; step++)
 		seq_printf(s, "failed_%s: %u\n", suspend_step_names[step],
Index: linux-pm/kernel/power/suspend.c
===================================================================
--- linux-pm.orig/kernel/power/suspend.c
+++ linux-pm/kernel/power/suspend.c
@@ -620,7 +620,7 @@  int pm_suspend(suspend_state_t state)
 		suspend_stats.fail++;
 		dpm_save_failed_errno(error);
 	} else {
-		suspend_stats.success++;
+		suspend_stats.step_failures[SUSPEND_NONE]++;
 	}
 	pr_info("suspend exit\n");
 	return error;