linux-kernel - Re: [PATCH 0/4] CPUFreq statistics retrieved by drivers

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]

Date:   Wed, 5 Aug 2020 13:36:43 +0100
From:   Sudeep Holla <sudeep.holla@....com>
To:     Florian Fainelli <f.fainelli@...il.com>
Cc:     Lukasz Luba <lukasz.luba@....com>,
        Viresh Kumar <viresh.kumar@...aro.org>,
        linux-kernel@...r.kernel.org, linux-arm-kernel@...ts.infradead.org,
        linux-pm@...r.kernel.org, cristian.marussi@....com,
        Sudeep Holla <sudeep.holla@....com>, rjw@...ysocki.net
Subject: Re: [PATCH 0/4] CPUFreq statistics retrieved by drivers

On Tue, Aug 04, 2020 at 10:19:23AM -0700, Florian Fainelli wrote:
>
>
> On 7/31/2020 8:56 AM, Sudeep Holla wrote:
> > On Thu, Jul 30, 2020 at 10:36:51AM +0100, Lukasz Luba wrote:
> >>
> >> In this case I think we would have to create debugfs.
> >> Sudeep do you think these debugfs should be exposed from the protocol
> >> layer:
> >> drivers/firmware/arm_scmi/perf.c
> >
> > I prefer above over cpufreq as we can support for all the devices not
> > just cpus which avoids adding similar support elsewhere(mostly devfreq)
> >
> >> or maybe from the cpufreq scmi driver? I would probably be safer to have
> >> it in the cpufreq driver because we have scmi_handle there.
> >>
> >
> > Cristian was thinking if we can consolidate all such debugfs under one
> > device may be and that should eliminate your handle restriction. I would
> > like to see how that works out in implementation but I don't have any
> > better suggestion ATM.
>
> debugfs is not enabled in production kernels, and especially not with
> Android kernels, so sticking those in sysfs like the existing cpufreq
> subsystem statistics may be a better choice.

Fair enough. I was suggesting that only if we can't push this into existing
sysfs support. If we can, then we need not worry about it. If not, I don't
want a user ABI just for SCMI for this firmware stats, I would rather keep
it in debugfs for debug purposes. This will be useless once we start seeing
AMU in the hardware and hence I was pushing for debugfs.

--
Regards,
Sudeep