* [PATCH v3] cpufreq: brcmstb-avs: fix imbalance of cpufreq policy refcount
@ 2020-01-19 7:09 qiwuchen55
2020-01-20 5:32 ` Viresh Kumar
0 siblings, 1 reply; 9+ messages in thread
From: qiwuchen55 @ 2020-01-19 7:09 UTC (permalink / raw)
To: mmayer, rjw, viresh.kumar, f.fainelli
Cc: bcm-kernel-feedback-list, linux-pm, linux-arm-kernel,
linux-kernel, chenqiwu
From: chenqiwu <chenqiwu@xiaomi.com>
brcm_avs_cpufreq_get() calls cpufreq_cpu_get() to get the cpufreq policy,
meanwhile, it also increments the kobject reference count to mark it busy.
However, a corresponding call of cpufreq_cpu_put() is ignored to decrement
the kobject reference count back, which may lead to a potential stuck risk
that the cpuhp thread deadly waits for dropping of kobject refcount when
cpufreq policy free.
For fixing this bug, cpufreq_get_policy() is referenced to do a proper
cpufreq_cpu_get()/cpufreq_cpu_put() and fill a policy copy for the user.
If the policy return NULL, we just return 0 to hit the code path of
cpufreq_driver->get.
Signed-off-by: chenqiwu <chenqiwu@xiaomi.com>
---
drivers/cpufreq/brcmstb-avs-cpufreq.c | 12 ++++++++++--
1 file changed, 10 insertions(+), 2 deletions(-)
diff --git a/drivers/cpufreq/brcmstb-avs-cpufreq.c b/drivers/cpufreq/brcmstb-avs-cpufreq.c
index 77b0e5d..ee0d404 100644
--- a/drivers/cpufreq/brcmstb-avs-cpufreq.c
+++ b/drivers/cpufreq/brcmstb-avs-cpufreq.c
@@ -452,8 +452,16 @@ static bool brcm_avs_is_firmware_loaded(struct private_data *priv)
static unsigned int brcm_avs_cpufreq_get(unsigned int cpu)
{
- struct cpufreq_policy *policy = cpufreq_cpu_get(cpu);
- struct private_data *priv = policy->driver_data;
+ struct cpufreq_policy policy;
+ struct private_data *priv;
+
+ /*
+ * In case cpufreq policy has been released, just return 0.
+ */
+ if (cpufreq_get_policy(&policy, cpu))
+ return 0;
+
+ priv = policy.driver_data;
return brcm_avs_get_frequency(priv->base);
}
--
1.9.1
^ permalink raw reply related [flat|nested] 9+ messages in thread
* Re: [PATCH v3] cpufreq: brcmstb-avs: fix imbalance of cpufreq policy refcount
2020-01-19 7:09 [PATCH v3] cpufreq: brcmstb-avs: fix imbalance of cpufreq policy refcount qiwuchen55
@ 2020-01-20 5:32 ` Viresh Kumar
2020-01-20 5:58 ` chenqiwu
0 siblings, 1 reply; 9+ messages in thread
From: Viresh Kumar @ 2020-01-20 5:32 UTC (permalink / raw)
To: qiwuchen55
Cc: mmayer, rjw, f.fainelli, bcm-kernel-feedback-list, linux-pm,
linux-arm-kernel, linux-kernel, chenqiwu
On 19-01-20, 15:09, qiwuchen55@gmail.com wrote:
> From: chenqiwu <chenqiwu@xiaomi.com>
>
> brcm_avs_cpufreq_get() calls cpufreq_cpu_get() to get the cpufreq policy,
> meanwhile, it also increments the kobject reference count to mark it busy.
> However, a corresponding call of cpufreq_cpu_put() is ignored to decrement
> the kobject reference count back, which may lead to a potential stuck risk
> that the cpuhp thread deadly waits for dropping of kobject refcount when
> cpufreq policy free.
>
> For fixing this bug, cpufreq_get_policy() is referenced to do a proper
> cpufreq_cpu_get()/cpufreq_cpu_put() and fill a policy copy for the user.
> If the policy return NULL, we just return 0 to hit the code path of
> cpufreq_driver->get.
>
> Signed-off-by: chenqiwu <chenqiwu@xiaomi.com>
> ---
> drivers/cpufreq/brcmstb-avs-cpufreq.c | 12 ++++++++++--
> 1 file changed, 10 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/cpufreq/brcmstb-avs-cpufreq.c b/drivers/cpufreq/brcmstb-avs-cpufreq.c
> index 77b0e5d..ee0d404 100644
> --- a/drivers/cpufreq/brcmstb-avs-cpufreq.c
> +++ b/drivers/cpufreq/brcmstb-avs-cpufreq.c
> @@ -452,8 +452,16 @@ static bool brcm_avs_is_firmware_loaded(struct private_data *priv)
>
> static unsigned int brcm_avs_cpufreq_get(unsigned int cpu)
> {
> - struct cpufreq_policy *policy = cpufreq_cpu_get(cpu);
Why can't we just add a corresponding cpufreq_cpu_put() instead of all this ?
> - struct private_data *priv = policy->driver_data;
> + struct cpufreq_policy policy;
> + struct private_data *priv;
> +
> + /*
> + * In case cpufreq policy has been released, just return 0.
> + */
> + if (cpufreq_get_policy(&policy, cpu))
> + return 0;
Why did you move away from the previous implementation of cpufreq_cpu_get() ?
> +
> + priv = policy.driver_data;
>
> return brcm_avs_get_frequency(priv->base);
> }
> --
> 1.9.1
--
viresh
^ permalink raw reply [flat|nested] 9+ messages in thread
* Re: [PATCH v3] cpufreq: brcmstb-avs: fix imbalance of cpufreq policy refcount
2020-01-20 5:32 ` Viresh Kumar
@ 2020-01-20 5:58 ` chenqiwu
2020-01-20 6:01 ` Viresh Kumar
0 siblings, 1 reply; 9+ messages in thread
From: chenqiwu @ 2020-01-20 5:58 UTC (permalink / raw)
To: Viresh Kumar
Cc: mmayer, rjw, f.fainelli, bcm-kernel-feedback-list, linux-pm,
linux-arm-kernel, linux-kernel, chenqiwu
On Mon, Jan 20, 2020 at 11:02:50AM +0530, Viresh Kumar wrote:
> On 19-01-20, 15:09, qiwuchen55@gmail.com wrote:
> > From: chenqiwu <chenqiwu@xiaomi.com>
> >
> > brcm_avs_cpufreq_get() calls cpufreq_cpu_get() to get the cpufreq policy,
> > meanwhile, it also increments the kobject reference count to mark it busy.
> > However, a corresponding call of cpufreq_cpu_put() is ignored to decrement
> > the kobject reference count back, which may lead to a potential stuck risk
> > that the cpuhp thread deadly waits for dropping of kobject refcount when
> > cpufreq policy free.
> >
> > For fixing this bug, cpufreq_get_policy() is referenced to do a proper
> > cpufreq_cpu_get()/cpufreq_cpu_put() and fill a policy copy for the user.
> > If the policy return NULL, we just return 0 to hit the code path of
> > cpufreq_driver->get.
> >
> > Signed-off-by: chenqiwu <chenqiwu@xiaomi.com>
> > ---
> > drivers/cpufreq/brcmstb-avs-cpufreq.c | 12 ++++++++++--
> > 1 file changed, 10 insertions(+), 2 deletions(-)
> >
> > diff --git a/drivers/cpufreq/brcmstb-avs-cpufreq.c b/drivers/cpufreq/brcmstb-avs-cpufreq.c
> > index 77b0e5d..ee0d404 100644
> > --- a/drivers/cpufreq/brcmstb-avs-cpufreq.c
> > +++ b/drivers/cpufreq/brcmstb-avs-cpufreq.c
> > @@ -452,8 +452,16 @@ static bool brcm_avs_is_firmware_loaded(struct private_data *priv)
> >
> > static unsigned int brcm_avs_cpufreq_get(unsigned int cpu)
> > {
> > - struct cpufreq_policy *policy = cpufreq_cpu_get(cpu);
>
> Why can't we just add a corresponding cpufreq_cpu_put() instead of all this ?
>
cpufreq_get_policy() does a proper cpufreq_cpu_get()/cpufreq_cpu_put(),
meanwhile fills a policy copy for the user. It equals to using
cpufreq_cpu_get() and a corresponding cpufreq_cpu_put() around access
to the policy pointer. I think both methods are fine here.
What do you think?
> > - struct private_data *priv = policy->driver_data;
> > + struct cpufreq_policy policy;
> > + struct private_data *priv;
> > +
> > + /*
> > + * In case cpufreq policy has been released, just return 0.
> > + */
> > + if (cpufreq_get_policy(&policy, cpu))
> > + return 0;
>
> Why did you move away from the previous implementation of cpufreq_cpu_get() ?
>
> > +
> > + priv = policy.driver_data;
> >
> > return brcm_avs_get_frequency(priv->base);
> > }
> > --
> > 1.9.1
>
> --
> viresh
Qiwu
^ permalink raw reply [flat|nested] 9+ messages in thread
* Re: [PATCH v3] cpufreq: brcmstb-avs: fix imbalance of cpufreq policy refcount
2020-01-20 5:58 ` chenqiwu
@ 2020-01-20 6:01 ` Viresh Kumar
2020-01-20 6:13 ` chenqiwu
0 siblings, 1 reply; 9+ messages in thread
From: Viresh Kumar @ 2020-01-20 6:01 UTC (permalink / raw)
To: chenqiwu
Cc: mmayer, rjw, f.fainelli, bcm-kernel-feedback-list, linux-pm,
linux-arm-kernel, linux-kernel, chenqiwu
On 20-01-20, 13:58, chenqiwu wrote:
> On Mon, Jan 20, 2020 at 11:02:50AM +0530, Viresh Kumar wrote:
> > On 19-01-20, 15:09, qiwuchen55@gmail.com wrote:
> > > From: chenqiwu <chenqiwu@xiaomi.com>
> > >
> > > brcm_avs_cpufreq_get() calls cpufreq_cpu_get() to get the cpufreq policy,
> > > meanwhile, it also increments the kobject reference count to mark it busy.
> > > However, a corresponding call of cpufreq_cpu_put() is ignored to decrement
> > > the kobject reference count back, which may lead to a potential stuck risk
> > > that the cpuhp thread deadly waits for dropping of kobject refcount when
> > > cpufreq policy free.
> > >
> > > For fixing this bug, cpufreq_get_policy() is referenced to do a proper
> > > cpufreq_cpu_get()/cpufreq_cpu_put() and fill a policy copy for the user.
> > > If the policy return NULL, we just return 0 to hit the code path of
> > > cpufreq_driver->get.
> > >
> > > Signed-off-by: chenqiwu <chenqiwu@xiaomi.com>
> > > ---
> > > drivers/cpufreq/brcmstb-avs-cpufreq.c | 12 ++++++++++--
> > > 1 file changed, 10 insertions(+), 2 deletions(-)
> > >
> > > diff --git a/drivers/cpufreq/brcmstb-avs-cpufreq.c b/drivers/cpufreq/brcmstb-avs-cpufreq.c
> > > index 77b0e5d..ee0d404 100644
> > > --- a/drivers/cpufreq/brcmstb-avs-cpufreq.c
> > > +++ b/drivers/cpufreq/brcmstb-avs-cpufreq.c
> > > @@ -452,8 +452,16 @@ static bool brcm_avs_is_firmware_loaded(struct private_data *priv)
> > >
> > > static unsigned int brcm_avs_cpufreq_get(unsigned int cpu)
> > > {
> > > - struct cpufreq_policy *policy = cpufreq_cpu_get(cpu);
> >
> > Why can't we just add a corresponding cpufreq_cpu_put() instead of all this ?
> >
>
> cpufreq_get_policy() does a proper cpufreq_cpu_get()/cpufreq_cpu_put(),
> meanwhile fills a policy copy for the user. It equals to using
> cpufreq_cpu_get() and a corresponding cpufreq_cpu_put() around access
> to the policy pointer. I think both methods are fine here.
> What do you think?
cpufreq_get_policy() does an extra memcpy as well, which isn't required at all
in your case.
--
viresh
^ permalink raw reply [flat|nested] 9+ messages in thread
* Re: [PATCH v3] cpufreq: brcmstb-avs: fix imbalance of cpufreq policy refcount
2020-01-20 6:01 ` Viresh Kumar
@ 2020-01-20 6:13 ` chenqiwu
2020-01-20 6:21 ` Viresh Kumar
0 siblings, 1 reply; 9+ messages in thread
From: chenqiwu @ 2020-01-20 6:13 UTC (permalink / raw)
To: Viresh Kumar
Cc: mmayer, rjw, f.fainelli, bcm-kernel-feedback-list, linux-pm,
linux-arm-kernel, linux-kernel, chenqiwu
On Mon, Jan 20, 2020 at 11:31:34AM +0530, Viresh Kumar wrote:
> On 20-01-20, 13:58, chenqiwu wrote:
> > On Mon, Jan 20, 2020 at 11:02:50AM +0530, Viresh Kumar wrote:
> > > On 19-01-20, 15:09, qiwuchen55@gmail.com wrote:
> > > > From: chenqiwu <chenqiwu@xiaomi.com>
> > > >
> > > > brcm_avs_cpufreq_get() calls cpufreq_cpu_get() to get the cpufreq policy,
> > > > meanwhile, it also increments the kobject reference count to mark it busy.
> > > > However, a corresponding call of cpufreq_cpu_put() is ignored to decrement
> > > > the kobject reference count back, which may lead to a potential stuck risk
> > > > that the cpuhp thread deadly waits for dropping of kobject refcount when
> > > > cpufreq policy free.
> > > >
> > > > For fixing this bug, cpufreq_get_policy() is referenced to do a proper
> > > > cpufreq_cpu_get()/cpufreq_cpu_put() and fill a policy copy for the user.
> > > > If the policy return NULL, we just return 0 to hit the code path of
> > > > cpufreq_driver->get.
> > > >
> > > > Signed-off-by: chenqiwu <chenqiwu@xiaomi.com>
> > > > ---
> > > > drivers/cpufreq/brcmstb-avs-cpufreq.c | 12 ++++++++++--
> > > > 1 file changed, 10 insertions(+), 2 deletions(-)
> > > >
> > > > diff --git a/drivers/cpufreq/brcmstb-avs-cpufreq.c b/drivers/cpufreq/brcmstb-avs-cpufreq.c
> > > > index 77b0e5d..ee0d404 100644
> > > > --- a/drivers/cpufreq/brcmstb-avs-cpufreq.c
> > > > +++ b/drivers/cpufreq/brcmstb-avs-cpufreq.c
> > > > @@ -452,8 +452,16 @@ static bool brcm_avs_is_firmware_loaded(struct private_data *priv)
> > > >
> > > > static unsigned int brcm_avs_cpufreq_get(unsigned int cpu)
> > > > {
> > > > - struct cpufreq_policy *policy = cpufreq_cpu_get(cpu);
> > >
> > > Why can't we just add a corresponding cpufreq_cpu_put() instead of all this ?
> > >
> >
> > cpufreq_get_policy() does a proper cpufreq_cpu_get()/cpufreq_cpu_put(),
> > meanwhile fills a policy copy for the user. It equals to using
> > cpufreq_cpu_get() and a corresponding cpufreq_cpu_put() around access
> > to the policy pointer. I think both methods are fine here.
> > What do you think?
>
> cpufreq_get_policy() does an extra memcpy as well, which isn't required at all
> in your case.
>
> --
> viresh
Huha..Do you worry about the race conditon with cpufreq policy free path?
If the policy has been released, cpufreq_get_policy() will return -EINVAL,
it won't do an extra memcpy.
Qiwu
^ permalink raw reply [flat|nested] 9+ messages in thread
* Re: [PATCH v3] cpufreq: brcmstb-avs: fix imbalance of cpufreq policy refcount
2020-01-20 6:13 ` chenqiwu
@ 2020-01-20 6:21 ` Viresh Kumar
2020-01-20 6:27 ` chenqiwu
0 siblings, 1 reply; 9+ messages in thread
From: Viresh Kumar @ 2020-01-20 6:21 UTC (permalink / raw)
To: chenqiwu
Cc: mmayer, rjw, f.fainelli, bcm-kernel-feedback-list, linux-pm,
linux-arm-kernel, linux-kernel, chenqiwu
On 20-01-20, 14:13, chenqiwu wrote:
> On Mon, Jan 20, 2020 at 11:31:34AM +0530, Viresh Kumar wrote:
> > On 20-01-20, 13:58, chenqiwu wrote:
> > > On Mon, Jan 20, 2020 at 11:02:50AM +0530, Viresh Kumar wrote:
> > > > On 19-01-20, 15:09, qiwuchen55@gmail.com wrote:
> > > > > From: chenqiwu <chenqiwu@xiaomi.com>
> > > > >
> > > > > brcm_avs_cpufreq_get() calls cpufreq_cpu_get() to get the cpufreq policy,
> > > > > meanwhile, it also increments the kobject reference count to mark it busy.
> > > > > However, a corresponding call of cpufreq_cpu_put() is ignored to decrement
> > > > > the kobject reference count back, which may lead to a potential stuck risk
> > > > > that the cpuhp thread deadly waits for dropping of kobject refcount when
> > > > > cpufreq policy free.
> > > > >
> > > > > For fixing this bug, cpufreq_get_policy() is referenced to do a proper
> > > > > cpufreq_cpu_get()/cpufreq_cpu_put() and fill a policy copy for the user.
> > > > > If the policy return NULL, we just return 0 to hit the code path of
> > > > > cpufreq_driver->get.
> > > > >
> > > > > Signed-off-by: chenqiwu <chenqiwu@xiaomi.com>
> > > > > ---
> > > > > drivers/cpufreq/brcmstb-avs-cpufreq.c | 12 ++++++++++--
> > > > > 1 file changed, 10 insertions(+), 2 deletions(-)
> > > > >
> > > > > diff --git a/drivers/cpufreq/brcmstb-avs-cpufreq.c b/drivers/cpufreq/brcmstb-avs-cpufreq.c
> > > > > index 77b0e5d..ee0d404 100644
> > > > > --- a/drivers/cpufreq/brcmstb-avs-cpufreq.c
> > > > > +++ b/drivers/cpufreq/brcmstb-avs-cpufreq.c
> > > > > @@ -452,8 +452,16 @@ static bool brcm_avs_is_firmware_loaded(struct private_data *priv)
> > > > >
> > > > > static unsigned int brcm_avs_cpufreq_get(unsigned int cpu)
> > > > > {
> > > > > - struct cpufreq_policy *policy = cpufreq_cpu_get(cpu);
> > > >
> > > > Why can't we just add a corresponding cpufreq_cpu_put() instead of all this ?
> > > >
> > >
> > > cpufreq_get_policy() does a proper cpufreq_cpu_get()/cpufreq_cpu_put(),
> > > meanwhile fills a policy copy for the user. It equals to using
> > > cpufreq_cpu_get() and a corresponding cpufreq_cpu_put() around access
> > > to the policy pointer. I think both methods are fine here.
> > > What do you think?
> >
> > cpufreq_get_policy() does an extra memcpy as well, which isn't required at all
> > in your case.
> >
> > --
> > viresh
>
> Huha..Do you worry about the race conditon with cpufreq policy free path?
No. I just worry about an unnecessary memcpy, nothing else.
--
viresh
^ permalink raw reply [flat|nested] 9+ messages in thread
* Re: [PATCH v3] cpufreq: brcmstb-avs: fix imbalance of cpufreq policy refcount
2020-01-20 6:21 ` Viresh Kumar
@ 2020-01-20 6:27 ` chenqiwu
2020-01-20 6:30 ` Viresh Kumar
0 siblings, 1 reply; 9+ messages in thread
From: chenqiwu @ 2020-01-20 6:27 UTC (permalink / raw)
To: Viresh Kumar
Cc: mmayer, rjw, f.fainelli, bcm-kernel-feedback-list, linux-pm,
linux-arm-kernel, linux-kernel, chenqiwu
On Mon, Jan 20, 2020 at 11:51:26AM +0530, Viresh Kumar wrote:
> On 20-01-20, 14:13, chenqiwu wrote:
> > On Mon, Jan 20, 2020 at 11:31:34AM +0530, Viresh Kumar wrote:
> > > On 20-01-20, 13:58, chenqiwu wrote:
> > > > On Mon, Jan 20, 2020 at 11:02:50AM +0530, Viresh Kumar wrote:
> > > > > On 19-01-20, 15:09, qiwuchen55@gmail.com wrote:
> > > > > > From: chenqiwu <chenqiwu@xiaomi.com>
> > > > > >
> > > > > > brcm_avs_cpufreq_get() calls cpufreq_cpu_get() to get the cpufreq policy,
> > > > > > meanwhile, it also increments the kobject reference count to mark it busy.
> > > > > > However, a corresponding call of cpufreq_cpu_put() is ignored to decrement
> > > > > > the kobject reference count back, which may lead to a potential stuck risk
> > > > > > that the cpuhp thread deadly waits for dropping of kobject refcount when
> > > > > > cpufreq policy free.
> > > > > >
> > > > > > For fixing this bug, cpufreq_get_policy() is referenced to do a proper
> > > > > > cpufreq_cpu_get()/cpufreq_cpu_put() and fill a policy copy for the user.
> > > > > > If the policy return NULL, we just return 0 to hit the code path of
> > > > > > cpufreq_driver->get.
> > > > > >
> > > > > > Signed-off-by: chenqiwu <chenqiwu@xiaomi.com>
> > > > > > ---
> > > > > > drivers/cpufreq/brcmstb-avs-cpufreq.c | 12 ++++++++++--
> > > > > > 1 file changed, 10 insertions(+), 2 deletions(-)
> > > > > >
> > > > > > diff --git a/drivers/cpufreq/brcmstb-avs-cpufreq.c b/drivers/cpufreq/brcmstb-avs-cpufreq.c
> > > > > > index 77b0e5d..ee0d404 100644
> > > > > > --- a/drivers/cpufreq/brcmstb-avs-cpufreq.c
> > > > > > +++ b/drivers/cpufreq/brcmstb-avs-cpufreq.c
> > > > > > @@ -452,8 +452,16 @@ static bool brcm_avs_is_firmware_loaded(struct private_data *priv)
> > > > > >
> > > > > > static unsigned int brcm_avs_cpufreq_get(unsigned int cpu)
> > > > > > {
> > > > > > - struct cpufreq_policy *policy = cpufreq_cpu_get(cpu);
> > > > >
> > > > > Why can't we just add a corresponding cpufreq_cpu_put() instead of all this ?
> > > > >
> > > >
> > > > cpufreq_get_policy() does a proper cpufreq_cpu_get()/cpufreq_cpu_put(),
> > > > meanwhile fills a policy copy for the user. It equals to using
> > > > cpufreq_cpu_get() and a corresponding cpufreq_cpu_put() around access
> > > > to the policy pointer. I think both methods are fine here.
> > > > What do you think?
> > >
> > > cpufreq_get_policy() does an extra memcpy as well, which isn't required at all
> > > in your case.
> > >
> > > --
> > > viresh
> >
> > Huha..Do you worry about the race conditon with cpufreq policy free path?
>
> No. I just worry about an unnecessary memcpy, nothing else.
>
Is there any question about this extra memcpy?
Qiwu
^ permalink raw reply [flat|nested] 9+ messages in thread
* Re: [PATCH v3] cpufreq: brcmstb-avs: fix imbalance of cpufreq policy refcount
2020-01-20 6:27 ` chenqiwu
@ 2020-01-20 6:30 ` Viresh Kumar
2020-01-20 6:50 ` chenqiwu
0 siblings, 1 reply; 9+ messages in thread
From: Viresh Kumar @ 2020-01-20 6:30 UTC (permalink / raw)
To: chenqiwu
Cc: mmayer, rjw, f.fainelli, bcm-kernel-feedback-list, linux-pm,
linux-arm-kernel, linux-kernel, chenqiwu
On 20-01-20, 14:27, chenqiwu wrote:
> On Mon, Jan 20, 2020 at 11:51:26AM +0530, Viresh Kumar wrote:
> > On 20-01-20, 14:13, chenqiwu wrote:
> > > On Mon, Jan 20, 2020 at 11:31:34AM +0530, Viresh Kumar wrote:
> > > > On 20-01-20, 13:58, chenqiwu wrote:
> > > > > On Mon, Jan 20, 2020 at 11:02:50AM +0530, Viresh Kumar wrote:
> > > > > > On 19-01-20, 15:09, qiwuchen55@gmail.com wrote:
> > > > > > > From: chenqiwu <chenqiwu@xiaomi.com>
> > > > > > >
> > > > > > > brcm_avs_cpufreq_get() calls cpufreq_cpu_get() to get the cpufreq policy,
> > > > > > > meanwhile, it also increments the kobject reference count to mark it busy.
> > > > > > > However, a corresponding call of cpufreq_cpu_put() is ignored to decrement
> > > > > > > the kobject reference count back, which may lead to a potential stuck risk
> > > > > > > that the cpuhp thread deadly waits for dropping of kobject refcount when
> > > > > > > cpufreq policy free.
> > > > > > >
> > > > > > > For fixing this bug, cpufreq_get_policy() is referenced to do a proper
> > > > > > > cpufreq_cpu_get()/cpufreq_cpu_put() and fill a policy copy for the user.
> > > > > > > If the policy return NULL, we just return 0 to hit the code path of
> > > > > > > cpufreq_driver->get.
> > > > > > >
> > > > > > > Signed-off-by: chenqiwu <chenqiwu@xiaomi.com>
> > > > > > > ---
> > > > > > > drivers/cpufreq/brcmstb-avs-cpufreq.c | 12 ++++++++++--
> > > > > > > 1 file changed, 10 insertions(+), 2 deletions(-)
> > > > > > >
> > > > > > > diff --git a/drivers/cpufreq/brcmstb-avs-cpufreq.c b/drivers/cpufreq/brcmstb-avs-cpufreq.c
> > > > > > > index 77b0e5d..ee0d404 100644
> > > > > > > --- a/drivers/cpufreq/brcmstb-avs-cpufreq.c
> > > > > > > +++ b/drivers/cpufreq/brcmstb-avs-cpufreq.c
> > > > > > > @@ -452,8 +452,16 @@ static bool brcm_avs_is_firmware_loaded(struct private_data *priv)
> > > > > > >
> > > > > > > static unsigned int brcm_avs_cpufreq_get(unsigned int cpu)
> > > > > > > {
> > > > > > > - struct cpufreq_policy *policy = cpufreq_cpu_get(cpu);
> > > > > >
> > > > > > Why can't we just add a corresponding cpufreq_cpu_put() instead of all this ?
> > > > > >
> > > > >
> > > > > cpufreq_get_policy() does a proper cpufreq_cpu_get()/cpufreq_cpu_put(),
> > > > > meanwhile fills a policy copy for the user. It equals to using
> > > > > cpufreq_cpu_get() and a corresponding cpufreq_cpu_put() around access
> > > > > to the policy pointer. I think both methods are fine here.
> > > > > What do you think?
> > > >
> > > > cpufreq_get_policy() does an extra memcpy as well, which isn't required at all
> > > > in your case.
> > > >
> > > > --
> > > > viresh
> > >
> > > Huha..Do you worry about the race conditon with cpufreq policy free path?
> >
> > No. I just worry about an unnecessary memcpy, nothing else.
> >
> Is there any question about this extra memcpy?
What do you mean by that?
The whole point I am trying to make is that for your specific case, doing an
explicit cpufreq_cpu_get() and cpufreq_cpu_put() is far more efficient than
calling cpufreq_get_policy() which has a different purpose and usecase.
--
viresh
^ permalink raw reply [flat|nested] 9+ messages in thread
* Re: [PATCH v3] cpufreq: brcmstb-avs: fix imbalance of cpufreq policy refcount
2020-01-20 6:30 ` Viresh Kumar
@ 2020-01-20 6:50 ` chenqiwu
0 siblings, 0 replies; 9+ messages in thread
From: chenqiwu @ 2020-01-20 6:50 UTC (permalink / raw)
To: Viresh Kumar
Cc: mmayer, rjw, f.fainelli, bcm-kernel-feedback-list, linux-pm,
linux-arm-kernel, linux-kernel, chenqiwu
On Mon, Jan 20, 2020 at 12:00:04PM +0530, Viresh Kumar wrote:
> On 20-01-20, 14:27, chenqiwu wrote:
> > On Mon, Jan 20, 2020 at 11:51:26AM +0530, Viresh Kumar wrote:
> > > On 20-01-20, 14:13, chenqiwu wrote:
> > > > On Mon, Jan 20, 2020 at 11:31:34AM +0530, Viresh Kumar wrote:
> > > > > On 20-01-20, 13:58, chenqiwu wrote:
> > > > > > On Mon, Jan 20, 2020 at 11:02:50AM +0530, Viresh Kumar wrote:
> > > > > > > On 19-01-20, 15:09, qiwuchen55@gmail.com wrote:
> > > > > > > > From: chenqiwu <chenqiwu@xiaomi.com>
> > > > > > > >
> > > > > > > > brcm_avs_cpufreq_get() calls cpufreq_cpu_get() to get the cpufreq policy,
> > > > > > > > meanwhile, it also increments the kobject reference count to mark it busy.
> > > > > > > > However, a corresponding call of cpufreq_cpu_put() is ignored to decrement
> > > > > > > > the kobject reference count back, which may lead to a potential stuck risk
> > > > > > > > that the cpuhp thread deadly waits for dropping of kobject refcount when
> > > > > > > > cpufreq policy free.
> > > > > > > >
> > > > > > > > For fixing this bug, cpufreq_get_policy() is referenced to do a proper
> > > > > > > > cpufreq_cpu_get()/cpufreq_cpu_put() and fill a policy copy for the user.
> > > > > > > > If the policy return NULL, we just return 0 to hit the code path of
> > > > > > > > cpufreq_driver->get.
> > > > > > > >
> > > > > > > > Signed-off-by: chenqiwu <chenqiwu@xiaomi.com>
> > > > > > > > ---
> > > > > > > > drivers/cpufreq/brcmstb-avs-cpufreq.c | 12 ++++++++++--
> > > > > > > > 1 file changed, 10 insertions(+), 2 deletions(-)
> > > > > > > >
> > > > > > > > diff --git a/drivers/cpufreq/brcmstb-avs-cpufreq.c b/drivers/cpufreq/brcmstb-avs-cpufreq.c
> > > > > > > > index 77b0e5d..ee0d404 100644
> > > > > > > > --- a/drivers/cpufreq/brcmstb-avs-cpufreq.c
> > > > > > > > +++ b/drivers/cpufreq/brcmstb-avs-cpufreq.c
> > > > > > > > @@ -452,8 +452,16 @@ static bool brcm_avs_is_firmware_loaded(struct private_data *priv)
> > > > > > > >
> > > > > > > > static unsigned int brcm_avs_cpufreq_get(unsigned int cpu)
> > > > > > > > {
> > > > > > > > - struct cpufreq_policy *policy = cpufreq_cpu_get(cpu);
> > > > > > >
> > > > > > > Why can't we just add a corresponding cpufreq_cpu_put() instead of all this ?
> > > > > > >
> > > > > >
> > > > > > cpufreq_get_policy() does a proper cpufreq_cpu_get()/cpufreq_cpu_put(),
> > > > > > meanwhile fills a policy copy for the user. It equals to using
> > > > > > cpufreq_cpu_get() and a corresponding cpufreq_cpu_put() around access
> > > > > > to the policy pointer. I think both methods are fine here.
> > > > > > What do you think?
> > > > >
> > > > > cpufreq_get_policy() does an extra memcpy as well, which isn't required at all
> > > > > in your case.
> > > > >
> > > > > --
> > > > > viresh
> > > >
> > > > Huha..Do you worry about the race conditon with cpufreq policy free path?
> > >
> > > No. I just worry about an unnecessary memcpy, nothing else.
> > >
> > Is there any question about this extra memcpy?
>
> What do you mean by that?
>
> The whole point I am trying to make is that for your specific case, doing an
> explicit cpufreq_cpu_get() and cpufreq_cpu_put() is far more efficient than
> calling cpufreq_get_policy() which has a different purpose and usecase.
>
For efficiency, I agree your idea.
So we have change as follows:
diff --git a/drivers/cpufreq/brcmstb-avs-cpufreq.c b/drivers/cpufreq/brcmstb-avs-cpufreq.c
index 77b0e5d..b2ddde3 100644
--- a/drivers/cpufreq/brcmstb-avs-cpufreq.c
+++ b/drivers/cpufreq/brcmstb-avs-cpufreq.c
@@ -455,6 +455,11 @@ static unsigned int brcm_avs_cpufreq_get(unsigned int cpu)
struct cpufreq_policy *policy = cpufreq_cpu_get(cpu);
struct private_data *priv = policy->driver_data;
+ if (!policy)
+ return 0;
+
+ cpufreq_cpu_put(policy);
+
return brcm_avs_get_frequency(priv->base);
}
Qiwu
^ permalink raw reply related [flat|nested] 9+ messages in thread
end of thread, other threads:[~2020-01-20 6:50 UTC | newest]
Thread overview: 9+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-01-19 7:09 [PATCH v3] cpufreq: brcmstb-avs: fix imbalance of cpufreq policy refcount qiwuchen55
2020-01-20 5:32 ` Viresh Kumar
2020-01-20 5:58 ` chenqiwu
2020-01-20 6:01 ` Viresh Kumar
2020-01-20 6:13 ` chenqiwu
2020-01-20 6:21 ` Viresh Kumar
2020-01-20 6:27 ` chenqiwu
2020-01-20 6:30 ` Viresh Kumar
2020-01-20 6:50 ` chenqiwu
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).