Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
MIME-Version: 1.0
References: <20230516141531.791492-1-smostafa@google.com> <168665854178.2681974.13725664134174024912.b4-ty@linux.dev>
 <ZIiY8lrLKv4amCu7@google.com>
In-Reply-To: <ZIiY8lrLKv4amCu7@google.com>
From:   Mostafa Saleh <smostafa@google.com>
Date:   Wed, 14 Jun 2023 13:28:18 +0100
Message-ID: <CAFgf54q=hr9OoNdfDnMD3_4OHN6sz4928gYgvmz6bdSEpztUUw@mail.gmail.com>
Subject: Re: [PATCH] KVM: arm64: Use different pointer authentication keys for pKVM
To:     Oliver Upton <oliver.upton@linux.dev>
Cc:     kvmarm@lists.linux.dev, maz@kernel.org,
        linux-kernel@vger.kernel.org, linux-arm-kernel@lists.infradead.org,
        suzuki.poulose@arm.com, kaleshsingh@google.com, tabba@google.com,
        yuzenghui@huawei.com, catalin.marinas@arm.com, james.morse@arm.com,
        will@kernel.org
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

On Tue, Jun 13, 2023 at 5:27=E2=80=AFPM Mostafa Saleh <smostafa@google.com>=
 wrote:
>
> Hi Oliver,
>
> On Tue, Jun 13, 2023 at 12:16:02PM +0000, Oliver Upton wrote:
> > On Tue, 16 May 2023 14:15:31 +0000, Mostafa Saleh wrote:
> > > When the kernel is compiled with CONFIG_ARM64_PTR_AUTH_KERNEL, it
> > > uses Armv8.3-Pauth for return address protection for the kernel code
> > > including nvhe code in EL2.
> > >
> > > Same keys are used in both kernel(EL1) and nvhe code(EL2), this is
> > > fine for nvhe but not when running in protected mode(pKVM) as the hos=
t
> > > can't be trusted.
> > >
> > > [...]
> >
> > Applied to kvmarm/next, thanks!
> >
> > [1/1] KVM: arm64: Use different pointer authentication keys for pKVM
> >       https://git.kernel.org/kvmarm/kvmarm/c/fb737685beee
> >
> > --
>
> Thanks! I did more testing and I found a bug in this patch.
>
> It seems there is another entry point for the kenrel where pauth was
> not handled properly "kvm_host_psci_cpu_entry", I will investigate this
> further and send V2.
>
> Sorry for the inconvenience!
>
> Thanks,
> Mostafa

I investigated the problem further and actually the bug is missing
isb() after updating
Pauth system registers at EL2, I sent a v2 with this fix.

Although when entering from kvm_host_psci_cpu_entry, the hyp keys are not u=
sed,
but this call directly exists to the host, so I believe this should be fine=
.