linux-kernel - Re: [PATCH 1/2] KVM: x86/mmu: Add RET_PF_RETRY_INVALID

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] [thread-next>] [day] [month] [year] [list]

Message-ID: <aCtT9zsGmPiH2S6L@google.com>
Date: Mon, 19 May 2025 08:53:27 -0700
From: Sean Christopherson <seanjc@...gle.com>
To: Rick P Edgecombe <rick.p.edgecombe@...el.com>
Cc: Reinette Chatre <reinette.chatre@...el.com>, Yan Y Zhao <yan.y.zhao@...el.com>, 
	"kvm@...r.kernel.org" <kvm@...r.kernel.org>, "pbonzini@...hat.com" <pbonzini@...hat.com>, 
	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH 1/2] KVM: x86/mmu: Add RET_PF_RETRY_INVALID_SLOT for fault
 retry on invalid slot

On Mon, May 19, 2025, Rick P Edgecombe wrote:
> On Mon, 2025-05-19 at 08:05 -0700, Reinette Chatre wrote:
> > > Was this hit by a real VMM?  If so, why is a TDX VMM removing a memslot
> > > without kicking vCPUs out of KVM?
> > 
> > No, this was not hit by a real VMM. This was hit by a TDX MMU stress test
> > (built on top of [1]) that is still under development.
> 
> Yea, the context is that this TDX MMU stress test has grown more and more
> stressful. Mostly it has found TDX module issues. But recently it added this
> case which turned out to be a general issue. The TDX specific MMU stress test is
> not ready yet, so Yan added the case to the general test and fixed it for both
> VM types.
> 
> For TDX, since it's an pretty edge case and nothing catastrophic happens, I'd
> prefer to not rush a fix into the TDX PR.

Yeah, and I'd prefer not to bleed these details into userspace (or into KVM in
general), hence my question about whether or not a "real" VMM hit this.

Thanks!