Re: [PATCH v8 06/11] x86/traps: Add #VE support for TDX guest

From: Sean Christopherson <seanjc@google.com>
To: Lai Jiangshan <jiangshanlai+lkml@gmail.com>
Cc: Kuppuswamy Sathyanarayanan 
	<sathyanarayanan.kuppuswamy@linux.intel.com>,
	Thomas Gleixner <tglx@linutronix.de>,
	Ingo Molnar <mingo@redhat.com>, Borislav Petkov <bp@alien8.de>,
	X86 ML <x86@kernel.org>, Paolo Bonzini <pbonzini@redhat.com>,
	David Hildenbrand <david@redhat.com>,
	Andrea Arcangeli <aarcange@redhat.com>,
	Josh Poimboeuf <jpoimboe@redhat.com>,
	Juergen Gross <jgross@suse.com>, Deep Shah <sdeep@vmware.com>,
	VMware Inc <pv-drivers@vmware.com>,
	Vitaly Kuznetsov <vkuznets@redhat.com>,
	Wanpeng Li <wanpengli@tencent.com>,
	Jim Mattson <jmattson@google.com>, Joerg Roedel <joro@8bytes.org>,
	Peter H Anvin <hpa@zytor.com>,
	Dave Hansen <dave.hansen@intel.com>,
	Tony Luck <tony.luck@intel.com>,
	Dan Williams <dan.j.williams@intel.com>,
	Andi Kleen <ak@linux.intel.com>,
	Kirill Shutemov <kirill.shutemov@linux.intel.com>,
	Kuppuswamy Sathyanarayanan <knsathya@kernel.org>,
	LKML <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH v8 06/11] x86/traps: Add #VE support for TDX guest
Date: Mon, 11 Oct 2021 15:06:11 +0000	[thread overview]
Message-ID: <YWRS41aZH9A6fekt@google.com> (raw)
In-Reply-To: <CAJhGHyAqapG2MHfANeHG+LFHYr3a8AcHtbxcL3xUR_rmEOTqiQ@mail.gmail.com>

On Sat, Oct 09, 2021, Lai Jiangshan wrote:
> On Tue, Oct 5, 2021 at 10:54 AM Kuppuswamy Sathyanarayanan
> <sathyanarayanan.kuppuswamy@linux.intel.com> wrote:
> 
> >
> > The entry paths do not access TD-shared memory, MMIO regions or use
> > those specific MSRs, instructions, CPUID leaves that might generate #VE.
> > In addition, all interrupts including NMIs are blocked by the hardware
> > starting with #VE delivery until TDGETVEINFO is called.  This eliminates
> > the chance of a #VE during the syscall gap or paranoid entry paths and
> > simplifies #VE handling.

Minor clarification: it eliminates the chance of a #VE during the syscall gap
_if the VMM is benign_.  If the VMM is malicious, it can unmap and remap the
syscall page to induce an EPT Violation #VE due to the page not being accepted.

> Hello
> 
> If the reason is applied to #VE, I think it can be applied to SVM-ES's
> #VC too.  (I wish the entry code for #VC to be simplified since I'm
> moving some the asm entry code to C code)
>
> And I'm sorry I haven't read all the emails.
> Has the question asked by Andy Lutomirski been answered in any emails?
> 
> https://lore.kernel.org/lkml/CALCETrU9XypKbj-TrXLB3CPW6=MZ__5ifLz0ckbB=c=Myegn9Q@mail.gmail.com/

This question?

  Can the hypervisor cause an already-accepted secure-EPT page to transition to
  the unaccepted state?

Yep.  I wrote the above before following the link, I should have guessed which
question it was :-)

IIRC, the proposed middle ground was to add a TDCALL and/or TDPARAMS setting that
would allow the guest to opt-out of EPT Violation #VE due to page not accepted,
and instead terminate the VM on such a condition.  The caveat is that that would
require the kernel to never take an "page not accepted #VE" when doing lazy page
acceptance, but that was deemed doable.

That also raises the question of whether Andy's NAK applies to SEV-SNP without
support for "Enhanced SYSCALL Behavior"[*], otherwise SEV-SNP has the same "#VC
in syscall gap" attack.

[*] https://www.amd.com/system/files/TechDocs/57115.pdf