Re: [Xen-devel] vTPM Manager shuts down

[Daniel De Graaf <dgdegra@xxxxxxxxxxxxx>]

On 05/23/2013 10:35 AM, Ross Philipson wrote:
> > -----Original Message-----
> > From: Konrad Rzeszutek Wilk [mailto:konrad.wilk@xxxxxxxxxx]
> > Sent: Wednesday, May 22, 2013 4:54 PM
> > To: Ross Philipson
> > Cc: Daniel De Graaf; Jordi Cucurull Juan; xen-devel@xxxxxxxxxxxxx
> > Subject: Re: [Xen-devel] vTPM Manager shuts down
> >
> > On Wed, May 22, 2013 at 07:14:04PM +0000, Ross Philipson wrote:
> > > > -----Original Message-----
> > > > From: xen-devel-bounces@xxxxxxxxxxxxx [mailto:xen-devel-
> > > > bounces@xxxxxxxxxxxxx] On Behalf Of Ross Philipson
> > > > Sent: Wednesday, May 22, 2013 3:06 PM
> > > > To: Konrad Rzeszutek Wilk; Daniel De Graaf
> > > > Cc: Jordi Cucurull Juan; xen-devel@xxxxxxxxxxxxx
> > > > Subject: Re: [Xen-devel] vTPM Manager shuts down
> > > >
> > > > > -----Original Message-----
> > > > > From: xen-devel-bounces@xxxxxxxxxxxxx [mailto:xen-devel-
> > > > > bounces@xxxxxxxxxxxxx] On Behalf Of Konrad Rzeszutek Wilk
> > > > > Sent: Wednesday, May 22, 2013 2:00 PM
> > > > > To: Daniel De Graaf
> > > > > Cc: Jordi Cucurull Juan; xen-devel@xxxxxxxxxxxxx
> > > > > Subject: Re: [Xen-devel] vTPM Manager shuts down
> > > > >
> > > > > > (XEN) General information for domain 5:
> > > > > > (XEN)     refcnt=3 dying=0 pause_count=0
> > > > > > (XEN)     nr_pages=7168 xenheap_pages=5 shared_pages=0
> > paged_pages=0
> > > > > dirty_cpus={} max_pages=7424
> > > > > > (XEN)     handle=3097d8b9-8d80-4bde-94b6-978c98c37296
> > > > > vm_assist=00000000
> > > > > > (XEN) Rangesets belonging to domain 5:
> > > > > > (XEN)     I/O Ports  { }
> > > > > > (XEN)     Interrupts { }
> > > > > > (XEN)     I/O Memory { fed40 }
> > > > > >
> > > > > > Note: my config contains "iomem=['fed40,1']" not
> > "iomem=['fed40,5']"
> > > > > so your
> > > > > > output will differ there.
> > > > > Is that address (fed40) standard? Are there other TPM devices at
> > other
> > > > > locations?
> > > > Yes that is defined in the TCG's TPM spec. That is the beginning of
> > the
> > > > PCRs
> > > > for locality 0 and they range up to locality 4: FED4_0xxxh -
> > FED4_4xxxh
> > > > If you search for something like TCG_PCClientTPMSpecification_1-
> > 21_1-
> > > > 00_FINAL.pdf
> > > > you can find that information (chapter 9 is on locality).
> > > >
> > > > > How does one find them? lspci?
> > > > The TPM lives on the LPC bus. Some of its features are exported in
> > > > sysfs:
> > > >
> > > > /sys/class/misc/tpm0
> > > Oh and sorry I sent it before I finished what I was going to say. You
> > can
> > > find information about it in ACPI. E.g. this Dell I have has a
> > Broadcom
> > > 1.2 TPM:
> > >
> > >      Device (\_SB.PCI0.LPCB.TPM)
> > >      {
> > >          Name (_HID, EisaId ("BCM0102"))
> > >          Name (_CID, 0x310CD041)
> > >          Name (_STR, Unicode ("TPM 1.2 Device"))
> > >          Name (_CRS, ResourceTemplate ()
> > >          {
> > >              Memory32Fixed (ReadOnly,
> > >                  0xFED40000,         // Address Base
> > >                  0x00005000,         // Address Length
> > >                  )
> > >          })
> > >          ...
> > >      }
> > Would it make sense to have libxl when creating one of the TPM manager
> > domains to automatically get that info from /sys/class/misc/tpm0?
> Do you mean to get that address? It is pretty well fixed (at least on PC
> platforms). In fact the tpm_tis Linux driver has it hard coded in its
> defaults:
>
> enum tis_defaults {
>         TIS_MEM_BASE = 0xFED40000,
>         TIS_MEM_LEN = 0x5000,
>         TIS_SHORT_TIMEOUT = 750,        /* ms */
>         TIS_LONG_TIMEOUT = 2000,        /* 2 sec */
> };
>
> Though the pnp code in tpm_tis can take external values for
> the memory base. I guess the best place to find the actual value(s)
> would be in ACPI firmware (reading the _CRS).
The ACPI firmware entry is meant to be present for ease of use in an OS
that expects devices to have ACPI entries, not as the primary source of
the address. In an environment using TBOOT, for example, this ACPI table
may not be trusted to report the address correctly.

> As to your other question, I guess I never thought about whether there
> could or would be more than one actual TPM on a platform. I am not sure
> what that would be used for and it seems like the answer is "no" on the
> surface but I could be wrong.
The TPM PC Client specification mandates that a TPM 1.2 be located at
0xFED40000 (in section 9.1 of the PC Client Specific TIS 1.21 or 5.2
in version 1.3, among other documents). This prevents multiple TPMs
claiming to be a PC Client TPM.

In general, having multiple TPMs on a single platform is not a useful
configuration because it causes confusion when updating PCRs - which
is a primary reason you would want your keys on TPM.

> > Hm, I guess it wouldn't really work unless you added some new
> > config option called 'vtpm_manager=1' to do this.
The TPM manager already requires a command-line argument to change the
locality (and therefore the address it uses to access the TPM), so it's
not useful to try to auto-detect the address in one location while the
other is still manual.

--
Daniel De Graaf
National Security Agency

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel