Date:   Fri, 14 Jan 2022 01:10:03 +0000
From:   Sean Christopherson <seanjc@google.com>
To:     Raghavendra Rao Ananta <rananta@google.com>
Cc:     Jim Mattson <jmattson@google.com>, kvm@vger.kernel.org,
        Will Deacon <will@kernel.org>, Marc Zyngier <maz@kernel.org>,
        Peter Shier <pshier@google.com>, linux-kernel@vger.kernel.org,
        Catalin Marinas <catalin.marinas@arm.com>,
        Paolo Bonzini <pbonzini@redhat.com>,
        kvmarm@lists.cs.columbia.edu,
        Linux ARM <linux-arm-kernel@lists.infradead.org>
Subject: Re: [RFC PATCH v3 01/11] KVM: Capture VM start
Message-ID: <YeDNa+/rF0YEVJAi@google.com>
References: <CAAeT=Fxyct=WLUvfbpROKwB9huyt+QdJnKTaj8c5NKk+UY51WQ@mail.gmail.com>
 <CAJHc60za+E-zEO5v2QeKuifoXznPnt5n--g1dAN5jgsuq+SxrA@mail.gmail.com>
 <CALMp9eQDzqoJMck=_agEZNU9FJY9LB=iW-8hkrRc20NtqN=gDA@mail.gmail.com>
 <CAJHc60xZ9emY9Rs9ZbV+AH-Mjmkyg4JZU7V16TF48C-HJn+n4A@mail.gmail.com>
 <CALMp9eTPJZDtMiHZ5XRiYw2NR9EBKSfcP5CYddzyd2cgWsJ9hw@mail.gmail.com>
 <CAJHc60xD2U36pM4+Dq3yZw6Cokk-16X83JHMPXj4aFnxOJ3BUQ@mail.gmail.com>
 <CALMp9eR+evJ+w9VTSvR2KHciQDgTsnS=bh=1OUL4yy8gG6O51A@mail.gmail.com>
 <CAJHc60zw1o=JdUJ+sNNtv3mc_JTRMKG3kPp=-cchWkHm74hUYA@mail.gmail.com>
 <YeBfj89mIf8SezfD@google.com>
 <CAJHc60wRrgnvwqPWdXdvoqT0V9isXW5xH=btgdjPWQkqVW31Pw@mail.gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <CAJHc60wRrgnvwqPWdXdvoqT0V9isXW5xH=btgdjPWQkqVW31Pw@mail.gmail.com>
Precedence: bulk

On Thu, Jan 13, 2022, Raghavendra Rao Ananta wrote:
> On Thu, Jan 13, 2022 at 9:21 AM Sean Christopherson <seanjc@google.com> wrote:
> > If restricting updates in the arm64 is necessary to ensure KVM provides sane
> > behavior, then it could be justified.  But if it's purely a sanity check on
> > behalf of the guest, then it's not justified.
> Agreed that KVM doesn't really safeguard the guests, but just curious,
> is there really a downside in adding this thin layer of safety check?

It's more stuff that KVM has to maintain, creates an ABI that KVM must adhere to,
potentially creates inconsistencies in KVM, and prevents using KVM to intentionally
do stupid things to test scenarios that are "impossible".  And we also try to avoid
defining arbitrary CPU behavior in KVM (that may not be the case here).

> On the bright side, the guests would be safe, and it could save the
> developers some time in hunting down the bugs in this path, no?

Yes, but that can be said for lots and lots of things.  This is both a slippery
slope argument and the inconsistency argument above, e.g. if KVM actively prevents
userspace from doing X, why doesn't KVM prevent userspace from doing Y?  Having a
decently defined rule for these types of things, e.g. protect KVM/kernel and adhere
to the architecture but otherwise let userspace do whatever, avoids spending too
much time arguing over what KVM should/shouldn't allow, or wondering why on earth
KVM does XYZ, at least in theory :-)

There are certainly times where KVM could have saved userspace some pain, but
overall I do think KVM is better off staying out of the way when possible.