[PATCH v3] cpuidle: psci: Init cpuidle only for present CPUs

[PATCH v3] cpuidle: psci: Init cpuidle only for present CPUs

[Jacky Bai]

for_each_possible_cpu() is currently used to initialize cpuidle
in the PSCI cpuidle driver.

However, in cpu_dev_register_generic(), for_each_present_cpu()
is used to register CPU devices which means the CPU devices are
only registered for present CPUs and not all possible CPUs.

With nosmp or maxcpus=0, only the boot CPU is present, leading
to the failure:

  |  Failed to register cpuidle device for cpu1

Change for_each_possible_cpu() to for_each_present_cpu() in the
PSCI cpuidle driver to ensure it only registers cpuidle devices
for CPUs that are actually present.

Fixes: b0c69e1214bc ("drivers: base: Use present CPUs in GENERIC_CPU_DEVICES")
Reviewed-by: Dhruva Gole <d-gole@ti.com>
Reviewed-by: Sudeep Holla <sudeep.holla@arm.com>
Tested-by: Yuanjie Yang <quic_yuanjiey@quicinc.com>
Signed-off-by: Jacky Bai <ping.bai@nxp.com>
---
 - v2 changes:
  - Improve the changelog as suggested by Sudeep
---
 drivers/cpuidle/cpuidle-psci.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/cpuidle/cpuidle-psci.c b/drivers/cpuidle/cpuidle-psci.c
index 2562dc001fc1..00117e9b33e8 100644
--- a/drivers/cpuidle/cpuidle-psci.c
+++ b/drivers/cpuidle/cpuidle-psci.c
@@ -410,7 +410,7 @@ static int psci_cpuidle_probe(struct platform_device *pdev)
 	struct cpuidle_driver *drv;
 	struct cpuidle_device *dev;
 
-	for_each_possible_cpu(cpu) {
+	for_each_present_cpu(cpu) {
 		ret = psci_idle_init_cpu(&pdev->dev, cpu);
 		if (ret)
 			goto out_fail;
-- 
2.34.1

Re: [PATCH v3] cpuidle: psci: Init cpuidle only for present CPUs

[Ulf Hansson]

On Thu, 6 Mar 2025 at 07:17, Jacky Bai <ping.bai@nxp.com> wrote:
>
> for_each_possible_cpu() is currently used to initialize cpuidle
> in the PSCI cpuidle driver.
>
> However, in cpu_dev_register_generic(), for_each_present_cpu()
> is used to register CPU devices which means the CPU devices are
> only registered for present CPUs and not all possible CPUs.
>
> With nosmp or maxcpus=0, only the boot CPU is present, leading
> to the failure:
>
>   |  Failed to register cpuidle device for cpu1
>
> Change for_each_possible_cpu() to for_each_present_cpu() in the
> PSCI cpuidle driver to ensure it only registers cpuidle devices
> for CPUs that are actually present.
>
> Fixes: b0c69e1214bc ("drivers: base: Use present CPUs in GENERIC_CPU_DEVICES")
> Reviewed-by: Dhruva Gole <d-gole@ti.com>
> Reviewed-by: Sudeep Holla <sudeep.holla@arm.com>
> Tested-by: Yuanjie Yang <quic_yuanjiey@quicinc.com>
> Signed-off-by: Jacky Bai <ping.bai@nxp.com>

Is this problem specific to cpuidle-psci?

Others are using for_each_possible_cpu() when registering their cpuidle drivers.

Kind regards
Uffe

> ---
>  - v2 changes:
>   - Improve the changelog as suggested by Sudeep
> ---
>  drivers/cpuidle/cpuidle-psci.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/cpuidle/cpuidle-psci.c b/drivers/cpuidle/cpuidle-psci.c
> index 2562dc001fc1..00117e9b33e8 100644
> --- a/drivers/cpuidle/cpuidle-psci.c
> +++ b/drivers/cpuidle/cpuidle-psci.c
> @@ -410,7 +410,7 @@ static int psci_cpuidle_probe(struct platform_device *pdev)
>         struct cpuidle_driver *drv;
>         struct cpuidle_device *dev;
>
> -       for_each_possible_cpu(cpu) {
> +       for_each_present_cpu(cpu) {
>                 ret = psci_idle_init_cpu(&pdev->dev, cpu);
>                 if (ret)
>                         goto out_fail;
> --
> 2.34.1
>

Re: [PATCH v3] cpuidle: psci: Init cpuidle only for present CPUs

[Sudeep Holla]

On Thu, Mar 06, 2025 at 11:53:14AM +0100, Ulf Hansson wrote:
> On Thu, 6 Mar 2025 at 07:17, Jacky Bai <ping.bai@nxp.com> wrote:
> >
> > for_each_possible_cpu() is currently used to initialize cpuidle
> > in the PSCI cpuidle driver.
> >
> > However, in cpu_dev_register_generic(), for_each_present_cpu()
> > is used to register CPU devices which means the CPU devices are
> > only registered for present CPUs and not all possible CPUs.
> >
> > With nosmp or maxcpus=0, only the boot CPU is present, leading
> > to the failure:
> >
> >   |  Failed to register cpuidle device for cpu1
> >
> > Change for_each_possible_cpu() to for_each_present_cpu() in the
> > PSCI cpuidle driver to ensure it only registers cpuidle devices
> > for CPUs that are actually present.
> >
> > Fixes: b0c69e1214bc ("drivers: base: Use present CPUs in GENERIC_CPU_DEVICES")
> > Reviewed-by: Dhruva Gole <d-gole@ti.com>
> > Reviewed-by: Sudeep Holla <sudeep.holla@arm.com>
> > Tested-by: Yuanjie Yang <quic_yuanjiey@quicinc.com>
> > Signed-off-by: Jacky Bai <ping.bai@nxp.com>
> 
> Is this problem specific to cpuidle-psci?
> 
> Others are using for_each_possible_cpu() when registering their cpuidle drivers.
> 

Good point. I assumed as this was very old patch, only this is left without
conversion to using present_cpumask. Looks like there are many other drivers
that need similar change. May be PSCI is most commonly used ones these days
and hardly anyone tests with nosmp or maxcpus=1 on those platforms.

But yes, all the users of for_each_possible_cpu() need to move to
for_each_present_cpu() if they are relying on CPU device being registered.

-- 
Regards,
Sudeep

RE: [PATCH v3] cpuidle: psci: Init cpuidle only for present CPUs

[Jacky Bai]

> Subject: Re: [PATCH v3] cpuidle: psci: Init cpuidle only for present CPUs
> 
> On Thu, Mar 06, 2025 at 11:53:14AM +0100, Ulf Hansson wrote:
> > On Thu, 6 Mar 2025 at 07:17, Jacky Bai <ping.bai@nxp.com> wrote:
> > >
> > > for_each_possible_cpu() is currently used to initialize cpuidle in
> > > the PSCI cpuidle driver.
> > >
> > > However, in cpu_dev_register_generic(), for_each_present_cpu() is
> > > used to register CPU devices which means the CPU devices are only
> > > registered for present CPUs and not all possible CPUs.
> > >
> > > With nosmp or maxcpus=0, only the boot CPU is present, leading to
> > > the failure:
> > >
> > >   |  Failed to register cpuidle device for cpu1
> > >
> > > Change for_each_possible_cpu() to for_each_present_cpu() in the PSCI
> > > cpuidle driver to ensure it only registers cpuidle devices for CPUs
> > > that are actually present.
> > >
> > > Fixes: b0c69e1214bc ("drivers: base: Use present CPUs in
> > > GENERIC_CPU_DEVICES")
> > > Reviewed-by: Dhruva Gole <d-gole@ti.com>
> > > Reviewed-by: Sudeep Holla <sudeep.holla@arm.com>
> > > Tested-by: Yuanjie Yang <quic_yuanjiey@quicinc.com>
> > > Signed-off-by: Jacky Bai <ping.bai@nxp.com>
> >
> > Is this problem specific to cpuidle-psci?
> >
> > Others are using for_each_possible_cpu() when registering their cpuidle
> drivers.
> >
> 
> Good point. I assumed as this was very old patch, only this is left without
> conversion to using present_cpumask. Looks like there are many other drivers
> that need similar change. May be PSCI is most commonly used ones these
> days and hardly anyone tests with nosmp or maxcpus=1 on those platforms.
> 
> But yes, all the users of for_each_possible_cpu() need to move to
> for_each_present_cpu() if they are relying on CPU device being registered.
> 

Yes, there are several other cpuidle drivers should have the similar issue, also
some other driver like cpufreq also need to covert to use present mask.

I will do a further check and fix them all together.

BR
> --
> Regards,
> Sudeep