linux-kernel - Re: [PATCH v1 0/3] arch_topology: Correct CPU capacity scaling

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: <YjBhH0qlOZ7BykQV@bogus>
Date:   Tue, 15 Mar 2022 09:49:19 +0000
From:   Sudeep Holla <sudeep.holla@....com>
To:     Leo Yan <leo.yan@...aro.org>
Cc:     Ionela Voinescu <ionela.voinescu@....com>,
        Sudeep Holla <sudeep.holla@....com>,
        Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        "Rafael J. Wysocki" <rafael@...nel.org>,
        Vincent Guittot <vincent.guittot@...aro.org>,
        Bryan O'Donoghue <bryan.odonoghue@...aro.org>,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH v1 0/3] arch_topology: Correct CPU capacity scaling

Hi Leo,

On Mon, Mar 14, 2022 at 06:10:58PM +0000, Ionela Voinescu wrote:
> Hi Leo,
> 
> On Sunday 13 Mar 2022 at 13:55:09 (+0800), Leo Yan wrote:
> > This patch set is to address issues for CPU capacity scaling.
> > 
> > "capacity-dmips-mhz" property might be absent in all CPU nodes, and in
> > another situation, DT might have inconsistent binding issue, e.g. some
> > CPU nodes have "capacity-dmips-mhz" property and some nodes miss the
> > property.  Current code mixes these two cases and always rollback to CPU
> > capacity 1024 for these two cases.
> >

Ideally the schema can be made to catch such issues. While I understand
that it is work in progress, we can flag the error in the code to handle that.
Rollback to 1024 seems correct default behaviour to me.

> > Patches 01 and 02 in this set are used to distinguish the two different
> > DT binding cases, and for the inconsistent binding issue, it rolls back
> > to 1024 without CPU capacity scaling.
> >
> > Patch 03 is to handle the case for absenting "capacity-dmips-mhz"
> > property in CPU nodes, the patch proceeds to do CPU capacity scaling based
> > on CPU maximum capacity.  Thus it can reflect the correct CPU capacity for
> > Arm platforms with "fast" and "slow" clusters (CPUs in two clusters have
> > the same raw capacity but with different maximum frequencies).
> >

NACK for the approach. Just fix the DT.

>
> In my opinion it's difficult to handle absent "capacity-dmips-mhz"
> properties, as they can be a result of 3 scenarios: potential..
>  1. bug in DT
>  2. unwillingness to fill this information in DT
>  3. suggestion that we're dealing with CPUs with same u-arch
>     (same capacity-dmips-mhz)
>
> I'm not sure it's up to us to interpret suggestions in the code so I
> believe treating missing information as error is the right choice, which
> is how we're handling this now.
>

+1 for all the points above and are very much valid.

> For 3. (and patch 03), isn't it easier to populate capacity-dmips-mhz to
> the same value (say 1024) in DT? That is a clear message that we're
> dealing with CPUs with the same u-arch.
>

Indeed.

-- 
Regards,
Sudeep