linux-kernel - Re: [ANNOUNCE] PUCK Agenda - 2025.04.02

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]

Message-ID: <diqz1pt8xl01.fsf@ackerleytng-ctop.c.googlers.com>
Date: Thu, 01 May 2025 11:07:42 -0700
From: Ackerley Tng <ackerleytng@...gle.com>
To: Sean Christopherson <seanjc@...gle.com>
Cc: kvm@...r.kernel.org, linux-kernel@...r.kernel.org, roypat@...zon.co.uk, 
	kalyazin@...zon.com, Fuad Tabba <tabba@...gle.com>, 
	Vishal Annapurve <vannapurve@...gle.com>, david@...hat.com
Subject: Re: [ANNOUNCE] PUCK Agenda - 2025.04.02 - No Topic

Sean Christopherson <seanjc@...gle.com> writes:

> On Tue, Apr 29, 2025, Ackerley Tng wrote:
>> 
>> Would like to add an agenda item for 2025-04-30's PUCK meeting: KVM
>> memory attributes vs guest_memfd shareability.
>
> Does next week work for you?  I.e. May 7th.  I won't be able to make tomorrow's
> PUCK (about to send a cancelation mail).
>
>> guest_memfd tracks shareability to determine whether a page can be
>> faulted by the host into userspace.
>> 
>> pKVM does not use kvm->mem_attr_array for tracking private/shared status
>> of a page, and for Coco VMs like TDX, there seems to be duplicate
>> tracking of private/shared status in guest_memfd's shareability and in
>> KVM's memory attributes.
>> 
>> I would like to discuss a proposal for shared/private conversions to be
>> performed through a guest_memfd (not KVM) ioctl instead of using
>> KVM_SET_MEMORY_ATTRIBUTES, where Coco VMs using guest_memfd for both
>> shared and private memory can be able to (with some other changes around
>> KVM memory attributes) skip tracking private/shared in KVM's memory
>> attributes.
>
> Has the proposal been posted on-list anywhere?  I haven't been following the
> guest_memfd threads very closely (understatement).

We managed to get it discussed at the guest_memfd upstream call. Here are
the slides, updated with discussion notes:
https://lpc.events/event/18/contributions/1764/attachments/1409/3708/2025-05-01-kvm-memory-attributes-vs-guest_memfd-shareability.pdf

Please remove the topic from the next PUCK! Thanks.