[PATCH] PM / sleep: Fix broken builds without CONFIG_PM_SLEEP_DEBUG

[PATCH] PM / sleep: Fix broken builds without CONFIG_PM_SLEEP_DEBUG

[Viresh Kumar]

The variable 'wakeup_irq' is defined within #ifdef CONFIG_PM_SLEEP_DEBUG
and used outside of it. And that breaks kernel build:

/home/viresh/linux/drivers/base/power/wakeup.c:871: undefined reference to `wakeup_irq'
/home/viresh/drivers/base/power/wakeup.c:871: undefined reference to `wakeup_irq'

Fix it by defining the variable outside of the ifdef.

Fixes: d1e59c235322 ("PM / sleep: Report interrupt that caused system wakeup")
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
---
 kernel/power/main.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/kernel/power/main.c b/kernel/power/main.c
index 9880bf888a5b..f97a188f4ccc 100644
--- a/kernel/power/main.c
+++ b/kernel/power/main.c
@@ -235,6 +235,9 @@ late_initcall(pm_debugfs_init);
 
 #endif /* CONFIG_PM_SLEEP */
 
+/* IRQ number which causes system wakeup */
+unsigned int wakeup_irq;
+
 #ifdef CONFIG_PM_SLEEP_DEBUG
 /*
  * pm_print_times: print time taken by devices to suspend and resume.
@@ -273,7 +276,6 @@ static inline void pm_print_times_init(void)
 	pm_print_times_enabled = !!initcall_debug;
 }
 
-unsigned int wakeup_irq;
 static ssize_t pm_wakeup_irq_show(struct kobject *kobj,
 					struct kobj_attribute *attr,
 					char *buf)
-- 
2.4.0

Re: [PATCH] PM / sleep: Fix broken builds without CONFIG_PM_SLEEP_DEBUG

[Pavel Machek]

On Tue 2015-09-15 13:42:21, Viresh Kumar wrote:
> The variable 'wakeup_irq' is defined within #ifdef CONFIG_PM_SLEEP_DEBUG
> and used outside of it. And that breaks kernel build:
> 
> /home/viresh/linux/drivers/base/power/wakeup.c:871: undefined reference to `wakeup_irq'
> /home/viresh/drivers/base/power/wakeup.c:871: undefined reference to `wakeup_irq'
> 
> Fix it by defining the variable outside of the ifdef.
> 
> Fixes: d1e59c235322 ("PM / sleep: Report interrupt that caused system wakeup")
> Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>

Acked-by: Pavel Machek <pavel@ucw.cz>

-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

Re: [PATCH] PM / sleep: Fix broken builds without CONFIG_PM_SLEEP_DEBUG

[Rafael J. Wysocki]

On Tuesday, September 15, 2015 01:42:21 PM Viresh Kumar wrote:
> The variable 'wakeup_irq' is defined within #ifdef CONFIG_PM_SLEEP_DEBUG
> and used outside of it. And that breaks kernel build:
> 
> /home/viresh/linux/drivers/base/power/wakeup.c:871: undefined reference to `wakeup_irq'
> /home/viresh/drivers/base/power/wakeup.c:871: undefined reference to `wakeup_irq'
> 
> Fix it by defining the variable outside of the ifdef.
> 
> Fixes: d1e59c235322 ("PM / sleep: Report interrupt that caused system wakeup")
> Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>

I've applied the v11 of the Alexandra's patch that doesn't have this problem AFAICS.

> ---
>  kernel/power/main.c | 4 +++-
>  1 file changed, 3 insertions(+), 1 deletion(-)
> 
> diff --git a/kernel/power/main.c b/kernel/power/main.c
> index 9880bf888a5b..f97a188f4ccc 100644
> --- a/kernel/power/main.c
> +++ b/kernel/power/main.c
> @@ -235,6 +235,9 @@ late_initcall(pm_debugfs_init);
>  
>  #endif /* CONFIG_PM_SLEEP */
>  
> +/* IRQ number which causes system wakeup */
> +unsigned int wakeup_irq;
> +
>  #ifdef CONFIG_PM_SLEEP_DEBUG
>  /*
>   * pm_print_times: print time taken by devices to suspend and resume.
> @@ -273,7 +276,6 @@ static inline void pm_print_times_init(void)
>  	pm_print_times_enabled = !!initcall_debug;
>  }
>  
> -unsigned int wakeup_irq;
>  static ssize_t pm_wakeup_irq_show(struct kobject *kobj,
>  					struct kobj_attribute *attr,
>  					char *buf)
> 

-- 
I speak only for myself.
Rafael J. Wysocki, Intel Open Source Technology Center.

Re: [PATCH] PM / sleep: Fix broken builds without CONFIG_PM_SLEEP_DEBUG

[Viresh Kumar]

On 16-09-15, 03:43, Rafael J. Wysocki wrote:
> On Tuesday, September 15, 2015 01:42:21 PM Viresh Kumar wrote:
> > The variable 'wakeup_irq' is defined within #ifdef CONFIG_PM_SLEEP_DEBUG
> > and used outside of it. And that breaks kernel build:
> > 
> > /home/viresh/linux/drivers/base/power/wakeup.c:871: undefined reference to `wakeup_irq'
> > /home/viresh/drivers/base/power/wakeup.c:871: undefined reference to `wakeup_irq'
> > 
> > Fix it by defining the variable outside of the ifdef.
> > 
> > Fixes: d1e59c235322 ("PM / sleep: Report interrupt that caused system wakeup")
> > Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
> 
> I've applied the v11 of the Alexandra's patch that doesn't have this problem AFAICS.

For the record, as we have talked on IRC, even the v11 patch suffers
from this problem and you will be fixing it.

-- 
viresh

Re: [PATCH] PM / sleep: Fix broken builds without CONFIG_PM_SLEEP_DEBUG

[Rafael J. Wysocki]

On Wednesday, September 16, 2015 07:30:23 AM Viresh Kumar wrote:
> On 16-09-15, 03:43, Rafael J. Wysocki wrote:
> > On Tuesday, September 15, 2015 01:42:21 PM Viresh Kumar wrote:
> > > The variable 'wakeup_irq' is defined within #ifdef CONFIG_PM_SLEEP_DEBUG
> > > and used outside of it. And that breaks kernel build:
> > > 
> > > /home/viresh/linux/drivers/base/power/wakeup.c:871: undefined reference to `wakeup_irq'
> > > /home/viresh/drivers/base/power/wakeup.c:871: undefined reference to `wakeup_irq'
> > > 
> > > Fix it by defining the variable outside of the ifdef.
> > > 
> > > Fixes: d1e59c235322 ("PM / sleep: Report interrupt that caused system wakeup")
> > > Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
> > 
> > I've applied the v11 of the Alexandra's patch that doesn't have this problem AFAICS.
> 
> For the record, as we have talked on IRC, even the v11 patch suffers
> from this problem and you will be fixing it.

Yes, it was slightly messed up.  Should be better now, though.

Thanks,
Rafael

Re: [PATCH] PM / sleep: Fix broken builds without CONFIG_PM_SLEEP_DEBUG

[Rafael J. Wysocki]

On Wednesday, September 16, 2015 04:48:37 AM Rafael J. Wysocki wrote:
> On Wednesday, September 16, 2015 07:30:23 AM Viresh Kumar wrote:
> > On 16-09-15, 03:43, Rafael J. Wysocki wrote:
> > > On Tuesday, September 15, 2015 01:42:21 PM Viresh Kumar wrote:
> > > > The variable 'wakeup_irq' is defined within #ifdef CONFIG_PM_SLEEP_DEBUG
> > > > and used outside of it. And that breaks kernel build:
> > > > 
> > > > /home/viresh/linux/drivers/base/power/wakeup.c:871: undefined reference to `wakeup_irq'
> > > > /home/viresh/drivers/base/power/wakeup.c:871: undefined reference to `wakeup_irq'
> > > > 
> > > > Fix it by defining the variable outside of the ifdef.
> > > > 
> > > > Fixes: d1e59c235322 ("PM / sleep: Report interrupt that caused system wakeup")
> > > > Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
> > > 
> > > I've applied the v11 of the Alexandra's patch that doesn't have this problem AFAICS.
> > 
> > For the record, as we have talked on IRC, even the v11 patch suffers
> > from this problem and you will be fixing it.
> 
> Yes, it was slightly messed up.  Should be better now, though.

And as a side note, for patches that are in bleeding-edge only and not in
something like linux-next, you don't need to bother anyone with fixes except
for me (and maybe the patch author for their education mostly).

And I either drop patches that cause build problems to happen in bleeding-edge
or fix them.  That's what bleeding-edge is for.

Thanks,
Rafael

Re: [PATCH] PM / sleep: Fix broken builds without CONFIG_PM_SLEEP_DEBUG

[Viresh Kumar]

On 16-09-15, 04:54, Rafael J. Wysocki wrote:
> > Yes, it was slightly messed up.  Should be better now, though.

Yeah, its fine now.

> And as a side note, for patches that are in bleeding-edge only and not in
> something like linux-next, you don't need to bother anyone with fixes except
> for me (and maybe the patch author for their education mostly).
> 
> And I either drop patches that cause build problems to happen in bleeding-edge
> or fix them.  That's what bleeding-edge is for.

Okay, will keep that in mind.

-- 
viresh