[libvirt] TPM support in libvirt

[Daniel P. Berrange]

On Thu, Mar 31, 2011 at 09:41:44AM -0400, Stefan Berger wrote:
> Hello!
> 
>   I'd like to start a discussion and get some feedback on support
> for TPM (Trusted Platform Module) management in libvirt. I have been
> posting patches to the Qemu and SeaBIOS mailing lists that will
> provide the ground work for TPM support in Qemu, though I anticipate
> that it will take a while until the code goes into those
> repositories. So everything I am saying related to command lines
> etc. can still change. Nevertheless, the next obvious layer for
> support is libvirt. Well, and I do have patches...
> 
>   First some background: the TPM is a crypto device built into many
> modern laptops and desktops. It's a chip connected to the
> motherboard and has firmware for running the TPM specifications. I
> would say it has a full processor on its own to run this firmware.
> It also has some very limited space for persistent storage to
> permanently store for example the owner's password (that he can
> establish with the TPM), store private keys and it also has some
> area for NVRAM spaces where the user can put any data into it he
> wants to.
> 
>   So now, on the qemu command line the TPM support (currently) looks
> as follows:
> 
> ... -tpm ?
> 
> displays:
> 
> Supported TPM types (choose only one):
> builtin   Qemu's built-in TPM; requires 63kb of block storage
> 
> I use this for reading the size of the necessary (QCoW2)
> blockstorage from it, 63kb.
> 
> To start a VM with TPM support:
> 
> ... -tpm type=builtin,path=/tmp/tpmstate.bin
> 
>   The above indicates the type of TPM to use. Currently there is
> only one type available, which is the 'builtin' one. It also
> provides a path to persistent storage, here /tmp/tpmstate.bin. The
> persistent storage file is actually a Qcow2 file so we can properly
> support snapshotting of the image and the TPM and go back to
> different states.
> 
>   Obviously we are now adding another 'disk' to the VM, so that the
> TPM can store its own persistent data. Now a user has to make sure
> to always keep his image file and this TPM 'disk' together,
> particularly once he stores vital keys in the TPM.
> 
>   So now for modeling this in libvirt I thought of this XML here for
> a user-provided Qcow2 image:
> 
> <tpm type='built-in'>
> <storage file='/tmp/tpmstate.bin'/>
> </tpm>
> 
>  In this case the user would create the above QCoW2
> /tmp/tmpstate.bin at the appropriate size (63kb).
> 
>   This XML here is for a libvirt-created QCoW2 image:
> 
> <tpm type='built-in'>
> <storage/>
> </tpm>
> 
>   This then causes libvirt to create a QCoW2 image of 63kb size
> (reads the 63kb from .. tpm ?) once the VM is about to start and
> produces for example this command line:
> 
> -tpm type=builtin,path=/var/lib/libvirt/tpm/a4d7cd22-da89-3094-6212-079a48a309a1.bin
> 
> The UUID then corresponds to the UUID of the VM. Since only one TPM
> per VM is supported, this should work just fine.
> 
> 
>  Now to support an encrypted QCoW2 image, the following XML could be used:
> 
> <tpm type='built-in'>
> <storage>
> <encryption format='qcow'>
> <secret type='passphrase' uuid='13ea49f7-2553-7308-5168-e337ade36232'/>
> </encryption>
> </storage>
> </tpm>
> 
>  Here the user has created a 'secret' and passes that secret's UUID
> into the XML above. So this is similar to how storage is dealt with
> with the difference that no path to the storage is provided and also
> internally no comparisons for storage paths are being made. I did
> also not want to force the user to create storage spaces (volumes)
> on his own, but let libvirt handle this implicitly for the TPM (use
> qemu-img to create the QCoW2). Also, multiple VMs' TPMs could share
> the same secret.

I'm not sure I'm a fan of the idea of libvirt creating the TPM
qcow2 file itself at VM startup, since that's at odds with the
way we deal with other host OS resources where the app is required
to create everything the VM needs upfront.

Also, why QCow2 ? It seems like rather a large hammer, if all we
really need is to ensure the data is encrypted on disk.

Might it be better to just use a raw 64kb file and just give QEMU
a decryption key to use for it. It should be trivial for QEMU to
have a variant on the raw block driver which passes all data
writes/reads through an encrypt/decrypt API. QCow2 is not exactly
everyone's favourite bit of QEMU code, being rather complex and
unreliable at times.

>  Please let me know of any comments regarding this.

My other main thought is how does the Xen TPM model fit in ? That has been
around for quite a while, but not supported in libvirt. If we add TPM to
libvirt we need to make sure the way we model it works for Xen too.
Not sure if anyone else like VMWare/VirtualBox does any TPM stuff yet.

Regards,
Daniel
-- 
|: http://berrange.com      -o-    http://www.flickr.com/photos/dberrange/ :|
|: http://libvirt.org              -o-             http://virt-manager.org :|
|: http://autobuild.org       -o-         http://search.cpan.org/~danberr/ :|
|: http://entangle-photo.org       -o-       http://live.gnome.org/gtk-vnc :|