linux-kernel - Re: [PATCH v3 06/12] x86/amd_nb: Use topology info to get AMD node count

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: <eda226f9-2441-4fe9-8530-8dafc6f58ce9@amd.com>
Date: Thu, 23 Oct 2025 16:43:52 -0500
From: Mario Limonciello <mario.limonciello@....com>
To: Yazen Ghannam <yazen.ghannam@....com>
Cc: Michal Pecio <michal.pecio@...il.com>, Shyam-sundar.S-k@....com,
 bhelgaas@...gle.com, hdegoede@...hat.com, ilpo.jarvinen@...ux.intel.com,
 jdelvare@...e.com, linux-edac@...r.kernel.org, linux-hwmon@...r.kernel.org,
 linux-kernel@...r.kernel.org, linux-pci@...r.kernel.org, linux@...ck-us.net,
 naveenkrishna.chatradhi@....com, platform-driver-x86@...r.kernel.org,
 suma.hegde@....com, tony.luck@...el.com, x86@...nel.org
Subject: Re: [PATCH v3 06/12] x86/amd_nb: Use topology info to get AMD node
 count



On 10/23/2025 1:25 PM, Yazen Ghannam wrote:
> On Thu, Oct 23, 2025 at 12:12:50PM -0500, Mario Limonciello wrote:
>> On 10/23/25 12:06 PM, Michal Pecio wrote:
>>> On Thu, 23 Oct 2025 11:22:29 -0500, Mario Limonciello wrote:
>>>> As this is an ancient BIOS this reminds me of some related commits:
>>>>
>>>> aa06e20f1be6 ("x86/ACPI: Don't add CPUs that are not online capable")
>>>> a74fabfbd1b70 ("x86/ACPI/boot: Use FADT version to check support for
>>>> online capable")
>>>>
>>>> Does reverting that second one help?
>>>
>>> Not sure if it's worth trying? My BIOS predates the ACPI 6.3 spec by
>>> several years and (if I understand correctly) MADT revision is 1.
>>>
>>> It seems Yazen guessed right: they list 6 APICs and mark absent ones
>>> as not enabled. But I don't think we can assume any ACPI 6.3 flags to
>>> be valid here.
>>>
>>> I wonder if some quick check could recognize those consumer CPUs and
>>> simply ignore hotplug there? AFAIK it was never a thing on AM3.
>>>
>>> Michal
>>
>> Oh if the MADT revision is that old, then yeah reverting won't do anything
>> here.
>>
> 
> I think this commit is interesting:
> fed8d8773b8e ("x86/acpi/boot: Correct acpi_is_processor_usable() check")
> 

Yeah; the reason for that appears to be a virtualization use case.  I'd 
be tempted to revert that and add an extra virtualization check instead.