[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <4581ac35-b879-4d32-91fd-ab486ec4b63c@gmail.com>
Date: Tue, 22 Jul 2025 00:02:11 +0800
From: Zihuan Zhang <zzhwaxy.kernel@...il.com>
To: "Rafael J. Wysocki" <rafael@...nel.org>
Cc: Viresh Kumar <viresh.kumar@...aro.org>, linux-pm@...r.kernel.org,
linux-kernel@...r.kernel.org
Subject: Re: [PATCH v1] cpufreq: Avoid creating sysfs link for offline CPUs
在 2025/7/21 23:51, Rafael J. Wysocki 写道:
> On Mon, Jul 21, 2025 at 5:37 PM Zihuan Zhang <zzhwaxy.kernel@...il.com> wrote:
>> Currently, cpufreq_add_dev() attempts to create a sysfs link
>> even when the target CPU is offline. Although guarded by a NULL
>> check on cpufreq_policy, this behavior is not strictly correct,
>> since offline CPUs shouldn't have their sysfs interface created.
> Why?
Thanks for the comment.
The idea was to avoid unnecessary sysfs operations for CPUs that are
offline, even though the existing checks make it safe. But I understand
it's not strictly needed — I'm fine with dropping it.
>> This patch cleans up the logic to only add the sysfs link
>> when the CPU is online and cpufreq policy is properly initialized.
> I don't really think that this change is necessary. Thanks!
>
>> Signed-off-by: Zihuan Zhang <zzhwaxy.kernel@...il.com>
>> ---
>> drivers/cpufreq/cpufreq.c | 10 +++++-----
>> 1 file changed, 5 insertions(+), 5 deletions(-)
>>
>> diff --git a/drivers/cpufreq/cpufreq.c b/drivers/cpufreq/cpufreq.c
>> index d7426e1d8bdd..0a77892d366c 100644
>> --- a/drivers/cpufreq/cpufreq.c
>> +++ b/drivers/cpufreq/cpufreq.c
>> @@ -1660,12 +1660,12 @@ static int cpufreq_add_dev(struct device *dev, struct subsys_interface *sif)
>> ret = cpufreq_online(cpu);
>> if (ret)
>> return ret;
>> - }
>>
>> - /* Create sysfs link on CPU registration */
>> - policy = per_cpu(cpufreq_cpu_data, cpu);
>> - if (policy)
>> - add_cpu_dev_symlink(policy, cpu, dev);
>> + /* Create sysfs link on CPU registration */
>> + policy = per_cpu(cpufreq_cpu_data, cpu);
>> + if (policy)
>> + add_cpu_dev_symlink(policy, cpu, dev);
>> + }
>>
>> return 0;
>> }
>> --
Powered by blists - more mailing lists