| Message ID | 20170404154957.19678-1-bigeasy@linutronix.de (mailing list archive) |
|---|---|
| State | Not Applicable, archived |
| Headers | show |
On 04-04-17, 17:49, Sebastian Andrzej Siewior wrote: > It is pure mystery to me why we need to be on a specific CPU while > looking up a value in an array. > My best shot at this is that before commit d4019f0a92ab ("cpufreq: move > freq change notifications to cpufreq core") it was required to invoke > cpufreq_notify_transition() on a special CPU. > > Since it looks like a waste, remove it. > > Cc: "Rafael J. Wysocki" <rjw@rjwysocki.net> > Cc: Viresh Kumar <viresh.kumar@linaro.org> > Cc: linux-pm@vger.kernel.org > Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> > --- > drivers/cpufreq/loongson2_cpufreq.c | 7 ------- > 1 file changed, 7 deletions(-) > > diff --git a/drivers/cpufreq/loongson2_cpufreq.c b/drivers/cpufreq/loongson2_cpufreq.c > index 6bbdac1065ff..9ac27b22476c 100644 > --- a/drivers/cpufreq/loongson2_cpufreq.c > +++ b/drivers/cpufreq/loongson2_cpufreq.c > @@ -51,19 +51,12 @@ static int loongson2_cpu_freq_notifier(struct notifier_block *nb, > static int loongson2_cpufreq_target(struct cpufreq_policy *policy, > unsigned int index) > { > - unsigned int cpu = policy->cpu; > - cpumask_t cpus_allowed; > unsigned int freq; > > - cpus_allowed = current->cpus_allowed; > - set_cpus_allowed_ptr(current, cpumask_of(cpu)); > - > freq = > ((cpu_clock_freq / 1000) * > loongson2_clockmod_table[index].driver_data) / 8; > > - set_cpus_allowed_ptr(current, &cpus_allowed); > - > /* setting the cpu frequency */ > clk_set_rate(policy->clk, freq * 1000); It was Zhangjin who wrote the first version and he may be able to answer the questions we have. Anyway, it is safe to apply this patch right now. Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
On Mon, Apr 10, 2017 at 04:01:53PM +0530, Viresh Kumar wrote: > On 04-04-17, 17:49, Sebastian Andrzej Siewior wrote: > > It is pure mystery to me why we need to be on a specific CPU while > > looking up a value in an array. > > My best shot at this is that before commit d4019f0a92ab ("cpufreq: move > > freq change notifications to cpufreq core") it was required to invoke > > cpufreq_notify_transition() on a special CPU. > > > > Since it looks like a waste, remove it. > > > > Cc: "Rafael J. Wysocki" <rjw@rjwysocki.net> > > Cc: Viresh Kumar <viresh.kumar@linaro.org> > > Cc: linux-pm@vger.kernel.org > > Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> > > --- > > drivers/cpufreq/loongson2_cpufreq.c | 7 ------- > > 1 file changed, 7 deletions(-) > > > > diff --git a/drivers/cpufreq/loongson2_cpufreq.c b/drivers/cpufreq/loongson2_cpufreq.c > > index 6bbdac1065ff..9ac27b22476c 100644 > > --- a/drivers/cpufreq/loongson2_cpufreq.c > > +++ b/drivers/cpufreq/loongson2_cpufreq.c > > @@ -51,19 +51,12 @@ static int loongson2_cpu_freq_notifier(struct notifier_block *nb, > > static int loongson2_cpufreq_target(struct cpufreq_policy *policy, > > unsigned int index) > > { > > - unsigned int cpu = policy->cpu; > > - cpumask_t cpus_allowed; > > unsigned int freq; > > > > - cpus_allowed = current->cpus_allowed; > > - set_cpus_allowed_ptr(current, cpumask_of(cpu)); > > - > > freq = > > ((cpu_clock_freq / 1000) * > > loongson2_clockmod_table[index].driver_data) / 8; > > > > - set_cpus_allowed_ptr(current, &cpus_allowed); > > - > > /* setting the cpu frequency */ > > clk_set_rate(policy->clk, freq * 1000); > > It was Zhangjin who wrote the first version and he may be able to > answer the questions we have. Anyway, it is safe to apply this patch > right now. > > Acked-by: Viresh Kumar <viresh.kumar@linaro.org> Thanks for the ack, patch applied. I noticed some of the other cpufreq drivers seem to be using similar constructs. Ralf
On 2017-04-10 13:05:11 [+0200], Ralf Baechle wrote: > I noticed some of the other cpufreq drivers seem to be using similar > constructs. Don't worry, we are working on that. > Ralf Sebastian
diff --git a/drivers/cpufreq/loongson2_cpufreq.c b/drivers/cpufreq/loongson2_cpufreq.c index 6bbdac1065ff..9ac27b22476c 100644 --- a/drivers/cpufreq/loongson2_cpufreq.c +++ b/drivers/cpufreq/loongson2_cpufreq.c @@ -51,19 +51,12 @@ static int loongson2_cpu_freq_notifier(struct notifier_block *nb, static int loongson2_cpufreq_target(struct cpufreq_policy *policy, unsigned int index) { - unsigned int cpu = policy->cpu; - cpumask_t cpus_allowed; unsigned int freq; - cpus_allowed = current->cpus_allowed; - set_cpus_allowed_ptr(current, cpumask_of(cpu)); - freq = ((cpu_clock_freq / 1000) * loongson2_clockmod_table[index].driver_data) / 8; - set_cpus_allowed_ptr(current, &cpus_allowed); - /* setting the cpu frequency */ clk_set_rate(policy->clk, freq * 1000);
It is pure mystery to me why we need to be on a specific CPU while looking up a value in an array. My best shot at this is that before commit d4019f0a92ab ("cpufreq: move freq change notifications to cpufreq core") it was required to invoke cpufreq_notify_transition() on a special CPU. Since it looks like a waste, remove it. Cc: "Rafael J. Wysocki" <rjw@rjwysocki.net> Cc: Viresh Kumar <viresh.kumar@linaro.org> Cc: linux-pm@vger.kernel.org Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> --- drivers/cpufreq/loongson2_cpufreq.c | 7 ------- 1 file changed, 7 deletions(-)