linux-kernel - Re: [PATCH V2 4/5] virtio-pci: implement synchronize

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: <20220412095519.245cf9f7.pasic@linux.ibm.com>
Date:   Tue, 12 Apr 2022 09:55:19 +0200
From:   Halil Pasic <pasic@...ux.ibm.com>
To:     Jason Wang <jasowang@...hat.com>
Cc:     Cornelia Huck <cohuck@...hat.com>,
        "Michael S. Tsirkin" <mst@...hat.com>,
        "Paul E. McKenney" <paulmck@...nel.org>,
        Peter Zijlstra <peterz@...radead.org>,
        Marc Zyngier <maz@...nel.org>,
        linux-kernel <linux-kernel@...r.kernel.org>,
        virtualization <virtualization@...ts.linux-foundation.org>,
        Thomas Gleixner <tglx@...utronix.de>,
        Halil Pasic <pasic@...ux.ibm.com>
Subject: Re: [PATCH V2 4/5] virtio-pci: implement synchronize_vqs()

On Tue, 12 Apr 2022 10:24:35 +0800
Jason Wang <jasowang@...hat.com> wrote:

> > Regarding the question "are we safe against notifications before
> > indicators have been registered" I think we really need to think about
> > something like Secure Execution. We don't have, and we are unlikely
> > to have in hardware virtio-ccw implementations, and for a malicious hypervisor
> > that has full access to the guest memory hardening makes no sense.  
> 
> Does s390 have something like memory encryption? (I guess yes). In the
> case of x86 VM encryption, the I/O buffers were now done via software
> IOTLB, that's why hardening of the virtio driver is needed to prevent
> the hypervisor to poke the swiotlb etc.

Yep! Secure Execution is a confidential computing solution which is much
like encrypted guest memory, except for one gets exceptions when trying
to access private memory instead of ending up with garbage  because of
the encryption. These improvements are IMHO relevant to us!

Regards,
Halil