linux-kernel - Re: [RFC PATCH 0/3] ACPI, arm64: Backport for ACPI PPTT 6.3 thread flag for stable 4.19.x

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 for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Date:   Fri, 11 Oct 2019 08:53:32 +0800
From:   Hanjun Guo <guohanjun@...wei.com>
To:     John Garry <john.garry@...wei.com>, <catalin.marinas@....com>,
        <will@...nel.org>, <rjw@...ysocki.net>, <lenb@...nel.org>,
        <robert.moore@...el.com>, <erik.schmauss@...el.com>,
        <sudeep.holla@....com>, <rrichter@...vell.com>,
        <jeremy.linton@....com>
CC:     <linux-arm-kernel@...ts.infradead.org>,
        <linux-kernel@...r.kernel.org>, <linux-acpi@...r.kernel.org>,
        <linuxarm@...wei.com>, <gregkh@...uxfoundation.org>,
        <wanghuiqiang@...wei.com>
Subject: Re: [RFC PATCH 0/3] ACPI, arm64: Backport for ACPI PPTT 6.3 thread
 flag for stable 4.19.x

Hi John,

On 2019/10/10 21:29, John Garry wrote:
> This series is a backport of the ACPI PPTT 6.3 thread flag feature for
> supporting arm64 systems.
> 
> The background is that some arm64 implementations are broken, in that they
> incorrectly advertise that a CPU is mutli-threaded, when it is not - the
> HiSilicon Taishanv110 rev 2, aka tsv110, being an example.
> 
> This leads to the system topology being incorrect. The reason being that
> arm64 topology code uses a combination of ACPI PPTT (Processor Properties
> Topology Table) and the system MPIDR (Multiprocessor Affinity Register) MT
> bit to determine the topology.
> 
> Until ACPI 6.3, the PPTT did not have any method to determine whether
> a CPU was multi-threaded, so only the MT bit is used - hence the
> broken topology for some systems.
> 
> In ACPI 6.3, a PPTT thread flag was introduced, which - when supported -
> would be used by the kernel to determine really if a CPU is multi-threaded
> or not, so that we don't get incorrect topology.

Thanks for doing this, and this patch set fix my CPU topology issue
in 4.19 kernel with updated firmware.

> 
> Note: I'm sending this as an RFC before sending to stable proper. I also
> have a 5.2 and 5.3 backport which are almost the same, and only
> significant change being that the ACPICA patch is not required.

5.2 stable was end of life, so only 4.19 and 5.3 are needed I think.

Thanks
Hanjun