Message ID | 1461048206-23567-1-git-send-email-hongtao.jia@nxp.com (mailing list archive) |
---|---|
State | Not Applicable |
Delegated to: | Scott Wood |
Headers | show |
On 19-04-16, 14:43, Jia Hongtao wrote: > Cooling device is registered by ready callback. It's also invoked while > system resuming from sleep (Enabling non-boot cpus). Thus cooling device > may be multiple registered. Matchable unregistration is added to exit > callback to fix this issue. > > Signed-off-by: Jia Hongtao <hongtao.jia@nxp.com> > --- > Changes for V2: > * Using qoriq_cpufreq_cpu_exit() callback instead of adding stop_cpu(). > > drivers/cpufreq/qoriq-cpufreq.c | 5 +++-- > 1 file changed, 3 insertions(+), 2 deletions(-) > > diff --git a/drivers/cpufreq/qoriq-cpufreq.c b/drivers/cpufreq/qoriq-cpufreq.c > index b23e525..0b85f90 100644 > --- a/drivers/cpufreq/qoriq-cpufreq.c > +++ b/drivers/cpufreq/qoriq-cpufreq.c > @@ -301,10 +301,11 @@ err_np: > return -ENODEV; > } > > -static int __exit qoriq_cpufreq_cpu_exit(struct cpufreq_policy *policy) > +static int qoriq_cpufreq_cpu_exit(struct cpufreq_policy *policy) Yes, __exit should have been removed, but not in this patch to create confusion. You should have submitted two separate patches for these changes. > { > struct cpu_data *data = policy->driver_data; > > + cpufreq_cooling_unregister(data->cdev); > kfree(data->pclk); > kfree(data->table); > kfree(data); > @@ -348,7 +349,7 @@ static struct cpufreq_driver qoriq_cpufreq_driver = { > .name = "qoriq_cpufreq", > .flags = CPUFREQ_CONST_LOOPS, > .init = qoriq_cpufreq_cpu_init, > - .exit = __exit_p(qoriq_cpufreq_cpu_exit), > + .exit = qoriq_cpufreq_cpu_exit, > .verify = cpufreq_generic_frequency_table_verify, > .target_index = qoriq_cpufreq_target, > .get = cpufreq_generic_get, > -- > 2.1.0.27.g96db324
> -----邮件原件----- > 发件人: Viresh Kumar [mailto:viresh.kumar@linaro.org] > 发送时间: Tuesday, April 19, 2016 3:54 PM > 收件人: Hongtao Jia <hongtao.jia@nxp.com> > 抄送: linux-pm@vger.kernel.org; linuxppc-dev@lists.ozlabs.org; Scott Wood > <scott.wood@nxp.com>; Yuantian Tang <yuantian.tang@nxp.com> > 主题: Re: [PATCH V2] cpufreq: qoriq: Fix cooling device registration issue during > suspend > > On 19-04-16, 14:43, Jia Hongtao wrote: > > Cooling device is registered by ready callback. It's also invoked > > while system resuming from sleep (Enabling non-boot cpus). Thus > > cooling device may be multiple registered. Matchable unregistration is > > added to exit callback to fix this issue. > > > > Signed-off-by: Jia Hongtao <hongtao.jia@nxp.com> > > --- > > Changes for V2: > > * Using qoriq_cpufreq_cpu_exit() callback instead of adding stop_cpu(). > > > > drivers/cpufreq/qoriq-cpufreq.c | 5 +++-- > > 1 file changed, 3 insertions(+), 2 deletions(-) > > > > diff --git a/drivers/cpufreq/qoriq-cpufreq.c > > b/drivers/cpufreq/qoriq-cpufreq.c index b23e525..0b85f90 100644 > > --- a/drivers/cpufreq/qoriq-cpufreq.c > > +++ b/drivers/cpufreq/qoriq-cpufreq.c > > @@ -301,10 +301,11 @@ err_np: > > return -ENODEV; > > } > > > > -static int __exit qoriq_cpufreq_cpu_exit(struct cpufreq_policy > > *policy) > > +static int qoriq_cpufreq_cpu_exit(struct cpufreq_policy *policy) > > Yes, __exit should have been removed, but not in this patch to create > confusion. > You should have submitted two separate patches for these changes. Thanks. I will send a new patch set soon. > > > { > > struct cpu_data *data = policy->driver_data; > > > > + cpufreq_cooling_unregister(data->cdev); > > kfree(data->pclk); > > kfree(data->table); > > kfree(data); > > @@ -348,7 +349,7 @@ static struct cpufreq_driver qoriq_cpufreq_driver = { > > .name = "qoriq_cpufreq", > > .flags = CPUFREQ_CONST_LOOPS, > > .init = qoriq_cpufreq_cpu_init, > > - .exit = __exit_p(qoriq_cpufreq_cpu_exit), > > + .exit = qoriq_cpufreq_cpu_exit, > > .verify = cpufreq_generic_frequency_table_verify, > > .target_index = qoriq_cpufreq_target, > > .get = cpufreq_generic_get, > > -- > > 2.1.0.27.g96db324 > > -- > viresh
diff --git a/drivers/cpufreq/qoriq-cpufreq.c b/drivers/cpufreq/qoriq-cpufreq.c index b23e525..0b85f90 100644 --- a/drivers/cpufreq/qoriq-cpufreq.c +++ b/drivers/cpufreq/qoriq-cpufreq.c @@ -301,10 +301,11 @@ err_np: return -ENODEV; } -static int __exit qoriq_cpufreq_cpu_exit(struct cpufreq_policy *policy) +static int qoriq_cpufreq_cpu_exit(struct cpufreq_policy *policy) { struct cpu_data *data = policy->driver_data; + cpufreq_cooling_unregister(data->cdev); kfree(data->pclk); kfree(data->table); kfree(data); @@ -348,7 +349,7 @@ static struct cpufreq_driver qoriq_cpufreq_driver = { .name = "qoriq_cpufreq", .flags = CPUFREQ_CONST_LOOPS, .init = qoriq_cpufreq_cpu_init, - .exit = __exit_p(qoriq_cpufreq_cpu_exit), + .exit = qoriq_cpufreq_cpu_exit, .verify = cpufreq_generic_frequency_table_verify, .target_index = qoriq_cpufreq_target, .get = cpufreq_generic_get,
Cooling device is registered by ready callback. It's also invoked while system resuming from sleep (Enabling non-boot cpus). Thus cooling device may be multiple registered. Matchable unregistration is added to exit callback to fix this issue. Signed-off-by: Jia Hongtao <hongtao.jia@nxp.com> --- Changes for V2: * Using qoriq_cpufreq_cpu_exit() callback instead of adding stop_cpu(). drivers/cpufreq/qoriq-cpufreq.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-)