PM: runtime: Fix conditional guard definitions

[PATCH v1] PM: runtime: Fix conditional guard definitions

Posted by Rafael J. Wysocki 3 months, 2 weeks ago

From: Rafael J. Wysocki <rafael.j.wysocki@intel.com>

Since pm_runtime_get_active() returns 0 on success, all of the
DEFINE_GUARD_COND() macros in pm_runtime.h need the "_RET == 0"
condition at the end of the argument list or they would not work
correctly.

Fixes: 9a0abc39450a ("PM: runtime: Add auto-cleanup macros for "resume and get" operations")
Reported-by: kernel test robot <lkp@intel.com>
Link: https://lore.kernel.org/linux-pm/202510191529.BCyjKlLQ-lkp@intel.com/
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
---
 include/linux/pm_runtime.h |    8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

--- a/include/linux/pm_runtime.h
+++ b/include/linux/pm_runtime.h
@@ -629,13 +629,13 @@ DEFINE_GUARD(pm_runtime_active_auto, str
  * device.
  */
 DEFINE_GUARD_COND(pm_runtime_active, _try,
-		  pm_runtime_get_active(_T, RPM_TRANSPARENT))
+		  pm_runtime_get_active(_T, RPM_TRANSPARENT), _RET == 0)
 DEFINE_GUARD_COND(pm_runtime_active, _try_enabled,
-		  pm_runtime_resume_and_get(_T))
+		  pm_runtime_resume_and_get(_T), _RET == 0)
 DEFINE_GUARD_COND(pm_runtime_active_auto, _try,
-		  pm_runtime_get_active(_T, RPM_TRANSPARENT))
+		  pm_runtime_get_active(_T, RPM_TRANSPARENT), _RET == 0)
 DEFINE_GUARD_COND(pm_runtime_active_auto, _try_enabled,
-		  pm_runtime_resume_and_get(_T))
+		  pm_runtime_resume_and_get(_T), _RET == 0)
 
 /**
  * pm_runtime_put_sync - Drop device usage counter and run "idle check" if 0.

Re: [PATCH v1] PM: runtime: Fix conditional guard definitions

Posted by Dhruva Gole 3 months, 1 week ago

On Oct 20, 2025 at 17:03:28 +0200, Rafael J. Wysocki wrote:
> From: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
> 
> Since pm_runtime_get_active() returns 0 on success, all of the
> DEFINE_GUARD_COND() macros in pm_runtime.h need the "_RET == 0"
> condition at the end of the argument list or they would not work
> correctly.
> 
> Fixes: 9a0abc39450a ("PM: runtime: Add auto-cleanup macros for "resume and get" operations")
> Reported-by: kernel test robot <lkp@intel.com>
> Link: https://lore.kernel.org/linux-pm/202510191529.BCyjKlLQ-lkp@intel.com/
> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
> ---
>  include/linux/pm_runtime.h |    8 ++++----
>  1 file changed, 4 insertions(+), 4 deletions(-)
> 
> --- a/include/linux/pm_runtime.h
> +++ b/include/linux/pm_runtime.h
> @@ -629,13 +629,13 @@ DEFINE_GUARD(pm_runtime_active_auto, str
>   * device.
>   */
>  DEFINE_GUARD_COND(pm_runtime_active, _try,
> -		  pm_runtime_get_active(_T, RPM_TRANSPARENT))
> +		  pm_runtime_get_active(_T, RPM_TRANSPARENT), _RET == 0)
>  DEFINE_GUARD_COND(pm_runtime_active, _try_enabled,
> -		  pm_runtime_resume_and_get(_T))
> +		  pm_runtime_resume_and_get(_T), _RET == 0)
>  DEFINE_GUARD_COND(pm_runtime_active_auto, _try,
> -		  pm_runtime_get_active(_T, RPM_TRANSPARENT))
> +		  pm_runtime_get_active(_T, RPM_TRANSPARENT), _RET == 0)
>  DEFINE_GUARD_COND(pm_runtime_active_auto, _try_enabled,
> -		  pm_runtime_resume_and_get(_T))
> +		  pm_runtime_resume_and_get(_T), _RET == 0)

The 3-argument form automatically assumes success, so we were
essentially ignoring RET val. This seems correct now.

Reviewed-by: Dhruva Gole <d-gole@ti.com>

Best regards,
Dhruva Gole
Texas Instruments Incorporated

Re: [PATCH v1] PM: runtime: Fix conditional guard definitions

Posted by Farhan Ali 3 months, 2 weeks ago

On 10/20/2025 8:03 AM, Rafael J. Wysocki wrote:
> From: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
>
> Since pm_runtime_get_active() returns 0 on success, all of the
> DEFINE_GUARD_COND() macros in pm_runtime.h need the "_RET == 0"
> condition at the end of the argument list or they would not work
> correctly.
>
> Fixes: 9a0abc39450a ("PM: runtime: Add auto-cleanup macros for "resume and get" operations")
> Reported-by: kernel test robot <lkp@intel.com>
> Link: https://lore.kernel.org/linux-pm/202510191529.BCyjKlLQ-lkp@intel.com/
> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
> ---
>   include/linux/pm_runtime.h |    8 ++++----
>   1 file changed, 4 insertions(+), 4 deletions(-)
>
> --- a/include/linux/pm_runtime.h
> +++ b/include/linux/pm_runtime.h
> @@ -629,13 +629,13 @@ DEFINE_GUARD(pm_runtime_active_auto, str
>    * device.
>    */
>   DEFINE_GUARD_COND(pm_runtime_active, _try,
> -		  pm_runtime_get_active(_T, RPM_TRANSPARENT))
> +		  pm_runtime_get_active(_T, RPM_TRANSPARENT), _RET == 0)
>   DEFINE_GUARD_COND(pm_runtime_active, _try_enabled,
> -		  pm_runtime_resume_and_get(_T))
> +		  pm_runtime_resume_and_get(_T), _RET == 0)
>   DEFINE_GUARD_COND(pm_runtime_active_auto, _try,
> -		  pm_runtime_get_active(_T, RPM_TRANSPARENT))
> +		  pm_runtime_get_active(_T, RPM_TRANSPARENT), _RET == 0)
>   DEFINE_GUARD_COND(pm_runtime_active_auto, _try_enabled,
> -		  pm_runtime_resume_and_get(_T))
> +		  pm_runtime_resume_and_get(_T), _RET == 0)
>   
>   /**
>    * pm_runtime_put_sync - Drop device usage counter and run "idle check" if 0.
>
>
This does fix the issue for me mentioned here 
https://lore.kernel.org/all/25435d82-575d-495f-ae61-bd38570ff9ad@linux.ibm.com/

Feel free to add

Tested-by: Farhan Ali <alifm@linux.ibm.com>

Re: [PATCH v1] PM: runtime: Fix conditional guard definitions

Posted by dan.j.williams@intel.com 3 months, 2 weeks ago

Rafael J. Wysocki wrote:
> From: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
> 
> Since pm_runtime_get_active() returns 0 on success, all of the
> DEFINE_GUARD_COND() macros in pm_runtime.h need the "_RET == 0"
> condition at the end of the argument list or they would not work
> correctly.
> 
> Fixes: 9a0abc39450a ("PM: runtime: Add auto-cleanup macros for "resume and get" operations")
> Reported-by: kernel test robot <lkp@intel.com>
> Link: https://lore.kernel.org/linux-pm/202510191529.BCyjKlLQ-lkp@intel.com/
> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
> ---
>  include/linux/pm_runtime.h |    8 ++++----
>  1 file changed, 4 insertions(+), 4 deletions(-)
> 
> --- a/include/linux/pm_runtime.h
> +++ b/include/linux/pm_runtime.h
> @@ -629,13 +629,13 @@ DEFINE_GUARD(pm_runtime_active_auto, str
>   * device.
>   */
>  DEFINE_GUARD_COND(pm_runtime_active, _try,
> -		  pm_runtime_get_active(_T, RPM_TRANSPARENT))
> +		  pm_runtime_get_active(_T, RPM_TRANSPARENT), _RET == 0)
>  DEFINE_GUARD_COND(pm_runtime_active, _try_enabled,
> -		  pm_runtime_resume_and_get(_T))
> +		  pm_runtime_resume_and_get(_T), _RET == 0)
>  DEFINE_GUARD_COND(pm_runtime_active_auto, _try,
> -		  pm_runtime_get_active(_T, RPM_TRANSPARENT))
> +		  pm_runtime_get_active(_T, RPM_TRANSPARENT), _RET == 0)
>  DEFINE_GUARD_COND(pm_runtime_active_auto, _try_enabled,
> -		  pm_runtime_resume_and_get(_T))
> +		  pm_runtime_resume_and_get(_T), _RET == 0)

I missed the detail that these are named "try" guards, but return
"-error" on failure. In all the existing cases of "try" vs "conditional"
acquire the polarity is different, e.g.:

	DEFINE_GUARD_COND(rwsem_read, _try, down_read_trylock(_T))
	DEFINE_GUARD_COND(rwsem_read, _intr, down_read_interruptible(_T), _RET == 0)

So, while this fix is correct, I wonder if a follow-on patch should
change the naming..., but I cannot think of sufficient replacement.

Reminder for me to be vigilant about this detail moving forward:

Reviewed-by: Dan Williams <dan.j.williams@intel.com>

Re: [PATCH v1] PM: runtime: Fix conditional guard definitions

Posted by Jonathan Cameron 3 months, 2 weeks ago

On Mon, 20 Oct 2025 17:03:28 +0200
"Rafael J. Wysocki" <rafael@kernel.org> wrote:

> From: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
> 
> Since pm_runtime_get_active() returns 0 on success, all of the
> DEFINE_GUARD_COND() macros in pm_runtime.h need the "_RET == 0"
> condition at the end of the argument list or they would not work
> correctly.
> 
> Fixes: 9a0abc39450a ("PM: runtime: Add auto-cleanup macros for "resume and get" operations")
> Reported-by: kernel test robot <lkp@intel.com>
> Link: https://lore.kernel.org/linux-pm/202510191529.BCyjKlLQ-lkp@intel.com/
> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Makes sense.

Reviewed-by: Jonathan Cameron <jonathan.cameron@huawei.com>

These macros are a bit awkward to get right - I'd forgotten the oddity that
conditional locks use != 0 to mean the lock was taken.


> ---
>  include/linux/pm_runtime.h |    8 ++++----
>  1 file changed, 4 insertions(+), 4 deletions(-)
> 
> --- a/include/linux/pm_runtime.h
> +++ b/include/linux/pm_runtime.h
> @@ -629,13 +629,13 @@ DEFINE_GUARD(pm_runtime_active_auto, str
>   * device.
>   */
>  DEFINE_GUARD_COND(pm_runtime_active, _try,
> -		  pm_runtime_get_active(_T, RPM_TRANSPARENT))
> +		  pm_runtime_get_active(_T, RPM_TRANSPARENT), _RET == 0)
>  DEFINE_GUARD_COND(pm_runtime_active, _try_enabled,
> -		  pm_runtime_resume_and_get(_T))
> +		  pm_runtime_resume_and_get(_T), _RET == 0)
>  DEFINE_GUARD_COND(pm_runtime_active_auto, _try,
> -		  pm_runtime_get_active(_T, RPM_TRANSPARENT))
> +		  pm_runtime_get_active(_T, RPM_TRANSPARENT), _RET == 0)
>  DEFINE_GUARD_COND(pm_runtime_active_auto, _try_enabled,
> -		  pm_runtime_resume_and_get(_T))
> +		  pm_runtime_resume_and_get(_T), _RET == 0)
>  
>  /**
>   * pm_runtime_put_sync - Drop device usage counter and run "idle check" if 0.
> 
> 
>