linux-kernel - Re: [PATCHv2 00/12] TDX: Enable Dynamic PAMT

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: <CAGtprH-MnG6sxjbWZBLMM83j6mohzRvc-shv3XCJOjJWqwQzXQ@mail.gmail.com>
Date: Wed, 10 Sep 2025 10:18:40 -0700
From: Vishal Annapurve <vannapurve@...gle.com>
To: Dave Hansen <dave.hansen@...el.com>
Cc: Kiryl Shutsemau <kirill@...temov.name>, 
	"Kirill A. Shutemov" <kirill.shutemov@...ux.intel.com>, pbonzini@...hat.com, seanjc@...gle.com, 
	dave.hansen@...ux.intel.com, rick.p.edgecombe@...el.com, 
	isaku.yamahata@...el.com, kai.huang@...el.com, yan.y.zhao@...el.com, 
	chao.gao@...el.com, tglx@...utronix.de, mingo@...hat.com, bp@...en8.de, 
	kvm@...r.kernel.org, x86@...nel.org, linux-coco@...ts.linux.dev, 
	linux-kernel@...r.kernel.org
Subject: Re: [PATCHv2 00/12] TDX: Enable Dynamic PAMT

On Tue, Sep 9, 2025 at 8:24 AM Dave Hansen <dave.hansen@...el.com> wrote:
>
> On 9/9/25 04:16, Kiryl Shutsemau wrote:
>
> > And it is going to be very wasteful. With huge pages, in most cases, you
> > only need dynamic PAMT for control pages. You will have a lot of memory
> > sitting in stash with zero use.
>
> I think it's going to be hard to convince me without actual data on this
> one.
>

* With 1G page backing and with DPAMT entries created only for 4K EPT mappings
- ~5MB of DPAMT memory usage for 704G guest memory size. We expect the
DPAMT memory usage to be in MBs even with 4096G guest memory size.

* With DPAMT entries created for all private memory irrespective of
mapping granularity
- DPAMT memory usage is around ~3GB for 704G guest memory size and
around ~16G for 4096G guest memory size.

For a 4TB guest memory size with 1G page backing, the DPAMT memory usage

> Even then, we're talking about 0.4% of system memory. So how much code
> and complexity are we talking about in order to save a *maximum* of 0.4%
> of system memory?
>