linux-kernel - Re: [PATCH] arm64: topology: Stop using MPIDR for topology information

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

Date:   Wed, 09 Sep 2020 12:21:03 +0100
From:   Valentin Schneider <valentin.schneider@....com>
To:     "Zengtao \(B\)" <prime.zeng@...ilicon.com>
Cc:     "linux-kernel\@vger.kernel.org" <linux-kernel@...r.kernel.org>,
        "linux-arm-kernel\@lists.infradead.org" 
        <linux-arm-kernel@...ts.infradead.org>,
        Catalin Marinas <catalin.marinas@....com>,
        Will Deacon <will@...nel.org>,
        Sudeep Holla <sudeep.holla@....com>,
        Robin Murphy <robin.murphy@....com>,
        Jeremy Linton <Jeremy.Linton@....com>,
        Dietmar Eggemann <dietmar.eggemann@....com>,
        Morten Rasmussen <morten.rasmussen@....com>
Subject: Re: [PATCH] arm64: topology: Stop using MPIDR for topology information


On 03/09/20 02:44, B wrote:
>> -----Original Message-----
>> From: Valentin Schneider [mailto:valentin.schneider@....com]
>> On 02/09/20 04:24, B wrote:
>> > I agree with your idea to remove the topology functionality of MPIDR ,
>> > but I think we need also consider ARM32 and GIC.
>> >
>>
>> Could you please elaborate? This change doesn't impact arch_topology, so
>> only arm64 is affected.
>
> Yes, this change only affects arm64, my question is that do we need to
>  leverage it to arm32 since arm32 got the same issue.
>
> And for GIC we are also using MPIDR for the topology info, but I am sure
> It's got the same issue or not, just a suggestion to have a look.

So technically yes, we can be bothered by this on arm32 - Sudeep pointed
out a list of DT files that shows platforms with non-zero values in Aff1 or
above.

However, the bigger issue is that artificial separation in clusters of 16
CPUs due to extra limitations on Aff0 (mainly due to GICv3 AIUI). Given
that GICv2 can support at most 8 CPU interfaces, I don't think we have it
as bad on arm32.