linux-kernel - Re: [v5, PATCH 4/5] cpufreq: mediatek: add opp notification for SVS support

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>] [day] [month] [year] [list]

Message-ID: <20200313091038.q7q7exiowoah4nk4@vireshk-i7>
Date:   Fri, 13 Mar 2020 14:40:38 +0530
From:   Viresh Kumar <viresh.kumar@...aro.org>
To:     "andrew-sh.cheng" <andrew-sh.cheng@...iatek.com>
Cc:     Mark Rutland <mark.rutland@....com>, Nishanth Menon <nm@...com>,
        srv_heupstream <srv_heupstream@...iatek.com>,
        "linux-pm@...r.kernel.org" <linux-pm@...r.kernel.org>,
        Stephen Boyd <sboyd@...nel.org>,
        "Rafael J. Wysocki" <rjw@...ysocki.net>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        Rob Herring <robh+dt@...nel.org>,
        Chanwoo Choi <cw00.choi@...sung.com>,
        Kyungmin Park <kyungmin.park@...sung.com>,
        MyungJoo Ham <myungjoo.ham@...sung.com>,
        "linux-mediatek@...ts.infradead.org" 
        <linux-mediatek@...ts.infradead.org>,
        "linux-arm-kernel@...ts.infradead.org" 
        <linux-arm-kernel@...ts.infradead.org>,
        Matthias Brugger <matthias.bgg@...il.com>,
        Fan Chen (陳凡) <fan.chen@...iatek.com>,
        "devicetree@...r.kernel.org" <devicetree@...r.kernel.org>
Subject: Re: [v5, PATCH 4/5] cpufreq: mediatek: add opp notification for SVS
 support

On 13-03-20, 15:22, andrew-sh.cheng wrote:
> I have something want to consult you.
> For your previous comment, you suggest use read-write lock to replace
> mutex lock.
> Will it be more efficiently even when all are write lock?
> (all lock region are "setting VProc voltage")

The data to be protected here isn't the VProc voltage but the list of
valid OPPs. My idea was if we can make the target() routine run a bit
faster as it really matters as it is called from scheduler hot path.

It won't be wrong to use the mutex the way you have used it right now,
but I think the read lock is much faster, though the read/write lock
is more beneficial in case where there are multiple readers and fewer
writers. The target() routine gets called multiple times here, not
in parallel, and the OPP change notifier won't be called so often.

-- 
viresh