RE: [PATCH v8 2/5] x86/tdx: Add TDX Guest event notify interrupt support

[Yao, Jiewen]

Thank you, Jun.

Yes. I confirmed that we will include below change to GHCI.next spec.

================
3.5 TDG.VP.VMCALL<SetupEventNotifyInterrupt>

From: "The host VMM should use SEAMCALL [TDWRVPS] leaf to inject an interrupt at the requested-interrupt vector into the TD via the posted-interrupt descriptor. "

To: "The host VMM should use SEAMCALL [TDWRVPS] leaf to inject an interrupt at the requested-interrupt vector into the TD VCPU that executed TDG.VP.VMCALL <SetupEventNotifyInterrupt> via the posted-interrupt descriptor. "

3.13 TDG.VP.VMCALL<Service>

Table 3-39: TDG.VP.VMCALL< Service >-Input Operands

From: R14: 
"Event notification interrupt vector - (valid values 32~255) selected by TD
0: blocking action. VMM need get response then return.
1~31: Reserved. Should not be used.
32~255: Non-block action. VMM can return immediately and signal the interrupt vector when the response is ready. "

To: R14:
"Event notification interrupt vector - (valid values 32~255) selected by TD
0: blocking action. VMM need get response then return.
1~31: Reserved. Should not be used.
32~255: Non-block action. VMM can return immediately and signal the interrupt vector when the response is ready. VMM should inject interrupt vector into the TD VCPU that executed TDG.VP.VMCALL<Service>."

================

Thank you
Yao Jiewen


> -----Original Message-----
> From: Nakajima, Jun <jun.nakajima@xxxxxxxxx>
> Sent: Saturday, June 25, 2022 7:42 AM
> To: Sathyanarayanan Kuppuswamy
> <sathyanarayanan.kuppuswamy@xxxxxxxxxxxxxxx>
> Cc: Huang, Kai <kai.huang@xxxxxxxxx>; Thomas Gleixner <tglx@xxxxxxxxxxxxx>;
> Ingo Molnar <mingo@xxxxxxxxxx>; Borislav Petkov <bp@xxxxxxxxx>; Dave
> Hansen <dave.hansen@xxxxxxxxxxxxxxx>; x86@xxxxxxxxxx; H . Peter Anvin
> <hpa@xxxxxxxxx>; Kirill A . Shutemov <kirill.shutemov@xxxxxxxxxxxxxxx>; Luck,
> Tony <tony.luck@xxxxxxxxx>; Andi Kleen <ak@xxxxxxxxxxxxxxx>; Wander Lairson
> Costa <wander@xxxxxxxxxx>; Isaku Yamahata <isaku.yamahata@xxxxxxxxx>;
> marcelo.cerri@xxxxxxxxxxxxx; tim.gardner@xxxxxxxxxxxxx;
> khalid.elmously@xxxxxxxxxxxxx; Cox, Philip <philip.cox@xxxxxxxxxxxxx>; LKML
> <linux-kernel@xxxxxxxxxxxxxxx>; Yao, Jiewen <jiewen.yao@xxxxxxxxx>
> Subject: Re: [PATCH v8 2/5] x86/tdx: Add TDX Guest event notify interrupt
> support
> 
> Replying to this (not the latest one) to reduce the quoting levels, adding Jiewen.
> 
> 
> > On Jun 20, 2022, at 8:44 AM, Sathyanarayanan Kuppuswamy
> <sathyanarayanan.kuppuswamy@xxxxxxxxxxxxxxx> wrote:
> >
> > Hi,
> >
> > + Jun
> >
> > On 6/20/22 5:33 AM, Kai Huang wrote:
> >> On Wed, 2022-06-08 at 19:52 -0700, Kuppuswamy Sathyanarayanan wrote:
> >>> Host-guest event notification via configured interrupt vector is useful
> >>> in cases where a guest makes an asynchronous request and needs a
> >>> callback from the host to indicate the completion or to let the host
> >>> notify the guest about events like device removal. One usage example is,
> >>> callback requirement of GetQuote asynchronous hypercall.
> >>
> >> Although this paragraph is from GHCI spec, IMHO it is not very helpful.  In
> >> fact, I think this paragraph is not that right and should be removed from GHCI.
> >> The reason is such event notification from VMM in cases like "device
> removal" is
> >> too vague.  There's no _specification_ in GHCI around which "device
> removal"
> >> should VMM inject such event.  For instance, I _think_ the Qemu enumerated
> ACPI-
> >> based hotplug should continue to work in TD.
> >
> > Yes. It just says that it *can* be used to signal a device removal. It is just
> > an example for where it can be used. But I agree that such a use case is vague.
> > If it makes it better, I am fine with removing it.
> 
> Yes, the “device removal” is just an example, especially, "the TD OS should be
> designed to not use the event notification for trusted operations”, based on the
> context of the spec.
> 
> >>
> >>>
> >>> Reserve 0xec IRQ vector address for TDX guest to receive the event
> >>> completion notification from VMM. Also add related IDT handler to
> >>> process the notification event.
> >>
> >> Here lacks why we need to choose to reserve a system vector.  For instance,
> why
> >> we cannot choose to use device IRQ way which only requires one vector on
> one
> >
> > As you have explained below, as per current spec, it just expects a system
> > vector.
> >
> >> cpu.  As you can see reserving a system vector isn't ideal especially for
> >> attestation as it is not a frequent operation.  It is wasteful of using IRQ
> >
> > I agree that event notification is currently only used for attestation. But I
> > think in future there could be other use cases for it. If the intention is just
> > to use it for attestation, then we can just modify the GetQuote TDVMCALL to
> pass
> > the vector address, and there is no need for new TDVMCALL. I think the
> intention
> > here is to have generic method for VMM to notify TD about some events. I am
> not
> > clear about the possible future use cases, so I cannot comment on frequency
> of
> > its use.
> >
> > Jun, any comments?
> >
> 
> The GHCI spec was not just clear, and we’ll update the spec, for example:
> 
> 3.5 TDG.VP.VMCALL<SetupEventNotifyInterrupt>
> ...
> From:
> 
> “The host VMM should use SEAMCALL [TDWRVPS] leaf to inject an interrupt at
> the requested-interrupt vector into the TD via the posted-interrupt descriptor. “
> 
> To:
> 
> “The host VMM should use SEAMCALL [TDWRVPS] leaf to inject an interrupt at
> the requested-interrupt vector into the TD VCPU that executed TDG.VP.VMCALL
> <SetupEventNotifyInterrupt>, via the posted-interrupt descriptor. “
> 
> ---
> Jun