Re: [RFC PATCH 1/2] KVM: arm64: Add PSCI SYSTEM_OFF2 function for hibernation

[David Woodhouse <dwmw2@infradead.org>]

[-- Attachment #1: Type: text/plain, Size: 1572 bytes --]

On Tue, 2024-03-12 at 08:47 -0700, Oliver Upton wrote:
> Hi,
> 
> On Tue, Mar 12, 2024 at 01:51:28PM +0000, David Woodhouse wrote:
> > From: David Woodhouse <dwmw@amazon.co.uk>
> > 
> > The PSCI v1.3 specification (alpha) adds support for a SYSTEM_OFF2 function
> > which is analogous to ACPI S4 state. This will allow hosting environments
> > to determine that a guest is hibernated rather than just powered off, and
> > ensure that they preserve the virtual environment appropriately to allow
> > the guest to resume safely (or bump the hardware_signature in the FACS to
> > trigger a clean reboot instead).
> > 
> > The beta version will be changed to say that PSCI_FEATURES returns a bit
> > mask of the supported hibernate types, which is implemented here.
> 
> Have you considered doing the PSCI implementation in userspace? The
> SMCCC filter [*] was added for this exact purpose. 
> 

For the purpose of deprecating the in-KVM PSCI implementation and
reimplementing it in VMMs? So we're never going to add new features and
versions to the kernel PSCI?

If that's the case then I suppose I can send the patch to clearly
document the in-KVM PSCI as deprecated, and do it that way.

But to answer your question directly, no I hadn't considered that. I
was just following the existing precedent of adding new optional PSCI
features like SYSTEM_RESET2.

I didn't think that the addition of SYSTEM_OFF2 in precisely the same
fashion would be the straw that broke the camel's back, and pushed us
to reimplement it in userspace instead.

[-- Attachment #2: smime.p7s --]
[-- Type: application/pkcs7-signature, Size: 5965 bytes --]