linux-kernel - Re: [PATCH V3 0/4] cpufreq: cppc: Add support for frequency invariance

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20210622065259.nw3e7ajwgzgnlm5e@vireshk-i7>
Date:   Tue, 22 Jun 2021 12:22:59 +0530
From:   Viresh Kumar <viresh.kumar@...aro.org>
To:     Qian Cai <quic_qiancai@...cinc.com>
Cc:     Rafael Wysocki <rjw@...ysocki.net>,
        Ionela Voinescu <ionela.voinescu@....com>,
        Ben Segall <bsegall@...gle.com>,
        Daniel Bristot de Oliveira <bristot@...hat.com>,
        Dietmar Eggemann <dietmar.eggemann@....com>,
        Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        Ingo Molnar <mingo@...hat.com>,
        Juri Lelli <juri.lelli@...hat.com>,
        Mel Gorman <mgorman@...e.de>,
        Peter Zijlstra <peterz@...radead.org>,
        "Rafael J. Wysocki" <rafael@...nel.org>,
        Steven Rostedt <rostedt@...dmis.org>,
        Sudeep Holla <sudeep.holla@....com>,
        Vincent Guittot <vincent.guittot@...aro.org>,
        Will Deacon <will@...nel.org>, linux-pm@...r.kernel.org,
        linux-acpi@...r.kernel.org, linux-kernel@...r.kernel.org,
        "Paul E. McKenney" <paulmck@...nel.org>,
        "Rafael J. Wysocki" <rafael.j.wysocki@...el.com>
Subject: Re: [PATCH V3 0/4] cpufreq: cppc: Add support for frequency
 invariance

On 21-06-21, 16:48, Qian Cai wrote:
> Viresh, this series works fine on my quick tests so far.

Thanks for testing.

> BTW, I
> noticed some strange things even with the series applied mentioned
> below when reading acpi_cppc vs cpufreq sysfs. Do you happen to know
> are those just hardware/firmware issues because Linux just
> faithfully exported the register values?

The values are exported by drivers/acpi/cppc_acpi.c I believe and they
look to be based on simple register reads.

> == Arm64 server Foo ==
> CPU max MHz:                     3000.0000
> CPU min MHz:                     1000.0000
> 
> /sys/devices/system/cpu/cpu0/acpi_cppc/highest_perf
> 300
> /sys/devices/system/cpu/cpu0/acpi_cppc/lowest_freq
> 1000
> /sys/devices/system/cpu/cpu0/acpi_cppc/lowest_nonlinear_perf
> 200
> /sys/devices/system/cpu/cpu0/acpi_cppc/lowest_perf
> 100
> /sys/devices/system/cpu/cpu0/acpi_cppc/nominal_freq <--- should be 3000?
> 2800
> /sys/devices/system/cpu/cpu0/acpi_cppc/nominal_perf <--- should be 300?
> 280

nominal-perf is max perf, and highest-perf is boost-perf. Same goes
with nominal-freq (i.e. policy->max).

So 280 and 2800 look to be the correct values, 300 and 3000 come with
boost enabled. Look at the first entry, highest_perf.

> /sys/devices/system/cpu/cpu0/acpi_cppc/reference_perf
> 100
> 
> == Arm64 server Bar ==
> CPU max MHz:                     3000.0000
> CPU min MHz:                     375.0000
> 
> /sys/devices/system/cpu/cpu0/acpi_cppc/highest_perf <--- should be 3000? There is no cpufreq boost.
> 3300

This isn't exported by cpufreq driver but acpi, and it just exports
hardware values of highest_perf (with boost i.e.). cpufreq may or
may not use this to support boost.

> /sys/devices/system/cpu/cpu0/acpi_cppc/lowest_freq <--- don't understand why 0.
> 0

Because corresponding hardware registers aren't implemented for your
platform, this is the function that reads these registers:

int cppc_get_perf_caps(int cpunum, struct cppc_perf_caps *perf_caps)
{
        ...

	/* Read optional lowest and nominal frequencies if present */
	if (CPC_SUPPORTED(low_freq_reg))
		cpc_read(cpunum, low_freq_reg, &low_f);

	if (CPC_SUPPORTED(nom_freq_reg))
		cpc_read(cpunum, nom_freq_reg, &nom_f);

	perf_caps->lowest_freq = low_f;
	perf_caps->nominal_freq = nom_f;
}

> /sys/devices/system/cpu/cpu0/acpi_cppc/lowest_nonlinear_perf
> 375
> /sys/devices/system/cpu/cpu0/acpi_cppc/lowest_perf
> 375
> /sys/devices/system/cpu/cpu0/acpi_cppc/nominal_freq <--- ditto
> 0
> /sys/devices/system/cpu/cpu0/acpi_cppc/nominal_perf
> 3000
> /sys/devices/system/cpu/cpu0/acpi_cppc/reference_perf
> 100

-- 
viresh