linux-kernel - Re: [PATCH v2 3/3] arm64: disable ACPI PPTT support temporarily

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:   Tue, 5 Jun 2018 11:38:28 -0500
From:   Jeremy Linton <jeremy.linton@....com>
To:     Sudeep Holla <sudeep.holla@....com>,
        "Rafael J. Wysocki" <rafael@...nel.org>
Cc:     ACPI Devel Maling List <linux-acpi@...r.kernel.org>,
        "Rafael J . Wysocki" <rjw@...ysocki.net>,
        Catalin Marinas <catalin.marinas@....com>,
        Morten Rasmussen <morten.rasmussen@....com>,
        Linux ARM <linux-arm-kernel@...ts.infradead.org>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
        Hanjun Guo <hanjun.guo@...aro.org>,
        Geert Uytterhoeven <geert@...ux-m68k.org>,
        Will Deacon <will.deacon@....com>
Subject: Re: [PATCH v2 3/3] arm64: disable ACPI PPTT support temporarily

Hi,

On 06/05/2018 10:27 AM, Sudeep Holla wrote:
> 
> 
> On 05/06/18 16:02, Rafael J. Wysocki wrote:
>> On Tue, Jun 5, 2018 at 4:35 PM, Sudeep Holla <sudeep.holla@....com> wrote:
>>> Currently, ARM64 doesn't support updating the CPU topology masks on
>>> CPU hotplug operations. ACPI PPTT support rely on that missing feature
>>> which is technically not incorrect. Instead of reverting all the PPTT
>>> support, let's keep it simple and disable ACPI PPTT support on ARM64
>>> for time-being until the topology updates are added for CPU hotplug
>>> operations.
>>
>> When is that going to happen?
>>> Nobody else uses PPTT now.
>>
> 
> No we need it on ARM64. I have even the fixes for the original issue
> but it's bit invasive and too late as we are already in the merge
> window. It needs to be tested on NUMA platforms, I have asked for that
> explicitly.
> 

So, there is a lot simpler fix for this, and its to remove the numa mask 
dependency from the scheduler domain building. That is just a couple 
line fix, and it reverts !ACPI to its original behavior, while the ACPI 
LLC building will work around the NUMA in package issue that caused 
patch 12/12 as the machine in question doesn't have a LLC crossing numa 
domains.