From: Sumit Gupta <sumitg@nvidia.com>
To: Viresh Kumar <viresh.kumar@linaro.org>
Cc: <rjw@rjwysocki.net>, <sudeep.holla@arm.com>,
<thierry.reding@gmail.com>, <jonathanh@nvidia.com>,
<linux-pm@vger.kernel.org>, <linux-tegra@vger.kernel.org>,
<linux-arm-kernel@lists.infradead.org>,
<linux-kernel@vger.kernel.org>, <ksitaraman@nvidia.com>,
<bbasu@nvidia.com>, Sumit Gupta <sumitg@nvidia.com>
Subject: Re: [PATCH v2 1/2] cpufreq: tegra194: get consistent cpuinfo_cur_freq
Date: Mon, 12 Oct 2020 22:04:50 +0530 [thread overview]
Message-ID: <20018a6a-e230-c39c-e801-5a3d0dd4c0ec@nvidia.com> (raw)
In-Reply-To: <20201012052227.x3bigztr7fit4jdz@vireshk-i7>
>> Frequency returned by 'cpuinfo_cur_freq' using counters is not fixed
>> and keeps changing slightly. This change returns a consistent value
>> from freq_table. If the reconstructed frequency has acceptable delta
>> from the last written value, then return the frequency corresponding
>> to the last written ndiv value from freq_table. Otherwise, print a
>> warning and return the reconstructed freq.
>>
>> Signed-off-by: Sumit Gupta <sumitg@nvidia.com>
>> ---
>> drivers/cpufreq/tegra194-cpufreq.c | 71 +++++++++++++++++++++++++++++++++-----
>> 1 file changed, 62 insertions(+), 9 deletions(-)
>>
>> diff --git a/drivers/cpufreq/tegra194-cpufreq.c b/drivers/cpufreq/tegra194-cpufreq.c
>> index e1d931c..d250e49 100644
>> --- a/drivers/cpufreq/tegra194-cpufreq.c
>> +++ b/drivers/cpufreq/tegra194-cpufreq.c
>> @@ -180,9 +180,70 @@ static unsigned int tegra194_get_speed_common(u32 cpu, u32 delay)
>> return (rate_mhz * KHZ); /* in KHz */
>> }
>>
>> +static void get_cpu_ndiv(void *ndiv)
>> +{
>> + u64 ndiv_val;
>> +
>> + asm volatile("mrs %0, s3_0_c15_c0_4" : "=r" (ndiv_val) : );
>> +
>> + *(u64 *)ndiv = ndiv_val;
>> +}
>> +
>> +static void set_cpu_ndiv(void *data)
>
> You weren't required to do this unnecessary change.
>
ya, moved the function up to keep both {get_|set_} calls together.
>> +{
>> + struct cpufreq_frequency_table *tbl = data;
>> + u64 ndiv_val = (u64)tbl->driver_data;
>> +
>> + asm volatile("msr s3_0_c15_c0_4, %0" : : "r" (ndiv_val));
>> +}
>> +
>> static unsigned int tegra194_get_speed(u32 cpu)
>> {
>> - return tegra194_get_speed_common(cpu, US_DELAY);
>> + struct tegra194_cpufreq_data *data = cpufreq_get_driver_data();
>> + struct cpufreq_frequency_table *pos;
>> + unsigned int rate;
>> + u64 ndiv;
>> + int ret;
>> + u32 cl;
>> +
>> + if (!cpu_online(cpu))
>
> This isn't required. The CPU is guaranteed to be online here.
>
OK, will remove this in next version.
>> + return -EINVAL;
>> +
>> + smp_call_function_single(cpu, get_cpu_cluster, &cl, true);
>> +
>> + if (cl >= data->num_clusters)
>
> Is it really possible here ? I meant you must have already checked
> this at cpufreq-init level already. Else mark it unlikely at least.
>
Ya, will remove the check here.
>> + return -EINVAL;
>> +
>> + /* reconstruct actual cpu freq using counters */
>> + rate = tegra194_get_speed_common(cpu, US_DELAY);
>> +
>> + /* get last written ndiv value */
>> + ret = smp_call_function_single(cpu, get_cpu_ndiv, &ndiv, true);
>> + if (ret) {
>
> What exactly can fail here ? get_cpu_ndiv() can't fail. Do we really
> need this check ? What about WARN_ON_ONCE() ?
>
OK.
>> + pr_err("cpufreq: Failed to get ndiv for CPU%d, ret:%d\n",
>> + cpu, ret);
>> + return rate;
>> + }
>> +
>> + /*
>> + * If the reconstructed frequency has acceptable delta from
>> + * the last written value, then return freq corresponding
>> + * to the last written ndiv value from freq_table. This is
>> + * done to return consistent value.
>> + */
>> + cpufreq_for_each_valid_entry(pos, data->tables[cl]) {
>> + if (pos->driver_data != ndiv)
>> + continue;
>> +
>> + if (abs(pos->frequency - rate) > 115200) {
>
> where does this 115200 comes from ? Strange that it matches tty's baud
> rate :)
>The value is equal to one freq step size.
> This is 115 MHz, right ? Isn't that too big of a delta ?
>
The is the acceptable delta used during our testing keeping some margin.
>> + pr_warn("cpufreq: cpu%d,cur:%u,set:%u,set ndiv:%llu\n",
>> + cpu, rate, pos->frequency, ndiv);
>> + } else {
>> + rate = pos->frequency;
>> + }
>> + break;
>> + }
>> + return rate;
>> }
>
> --
> viresh
>
next prev parent reply other threads:[~2020-10-12 16:35 UTC|newest]
Thread overview: 8+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-10-08 13:01 [PATCH v2 0/2] Tegra194 cpufreq driver misc changes Sumit Gupta
2020-10-08 13:01 ` [PATCH v2 1/2] cpufreq: tegra194: get consistent cpuinfo_cur_freq Sumit Gupta
2020-10-12 5:22 ` Viresh Kumar
2020-10-12 16:34 ` Sumit Gupta [this message]
2020-10-08 13:01 ` [PATCH v2 2/2] cpufreq: tegra194: Fix unlisted boot freq warning Sumit Gupta
2020-10-12 6:13 ` Viresh Kumar
2020-10-12 17:06 ` Sumit Gupta
2020-10-13 5:13 ` Viresh Kumar
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20018a6a-e230-c39c-e801-5a3d0dd4c0ec@nvidia.com \
--to=sumitg@nvidia.com \
--cc=bbasu@nvidia.com \
--cc=jonathanh@nvidia.com \
--cc=ksitaraman@nvidia.com \
--cc=linux-arm-kernel@lists.infradead.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-pm@vger.kernel.org \
--cc=linux-tegra@vger.kernel.org \
--cc=rjw@rjwysocki.net \
--cc=sudeep.holla@arm.com \
--cc=thierry.reding@gmail.com \
--cc=viresh.kumar@linaro.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
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).