TPM ACPI HID creation

TPM ACPI HID creation

[jason.spottswood]

[-- Attachment #1: Type: text/plain, Size: 1780 bytes --]

Tcg2Smm.c has a function "UpdateHID to create the ACPI HID for the TPM.  This function uses the TPM vendor ID combined with the firmware version number to create the ACPI HID.  The use of the TPM firmware version is not specified in any spec from the TCG or otherwise that I have been able to find.  I believe this was a design choice specific to EDK2.  However, using the TPM firmware version does not match the intended use case from the TCG PC Client spec, where the HID should be comprised of the vendor ID and device ID.  See below.  One problem that arises from this design is that the ACPI tables will change when the TPM FW has been updated.  Since the ACPI tables are hashed into PCR[0], it consequently means that a TPM FW change will cause a change to PCR[0].  It is not intuitive nor spec'd that TPM FW be included in PCR[0] measurements.  In fact, PCR[0] is used only for system FW/UEFI code measurements.  If a user does not update UEFI, then there is no expectation of PCR[0] measurement changes.  I propose that EDK2 change the UpdateHID function to use the vendor ID and device ID read from register TPM_DID_VID_0 (locality 0 offset 0xF00) in the creation of the ACPI HID.

The following was taken from the TCG PC Client Platform Firmware spec:

According to the ACPI Specification (version 5, Errata A, Section 6.1.5 and 6.1.3) a

hardware ID or compatibility ID is either a PNP ID with format “AAA####” or ACPI ID

with format “NNNN####”. The manufacturer ID returned by a TPM2_GetCapability

command can be used to set the “AAA” or “NNNN” portion of the ID. The remaining four

hexadecimal digits should be set to a value that allows software to differentiate different

device classes built by the same manufacturer.

[-- Attachment #2: Type: text/html, Size: 2924 bytes --]

Re: [edk2-devel] TPM ACPI HID creation

[Yao, Jiewen]

[-- Attachment #1: Type: text/plain, Size: 2185 bytes --]

Thanks Jason.

I think we should NOT measure TPM2 table *after* ACPI table patch.
The measurement should happen *before* ACPI table patch.

Hi Chao
Do you agree on that?

Thank you
Yao Jiewen


From: devel@edk2.groups.io [mailto:devel@edk2.groups.io] On Behalf Of jason.spottswood@hpe.com
Sent: Tuesday, June 25, 2019 11:58 PM
To: devel@edk2.groups.io
Subject: [edk2-devel] TPM ACPI HID creation

Tcg2Smm.c has a function "UpdateHID to create the ACPI HID for the TPM.  This function uses the TPM vendor ID combined with the firmware version number to create the ACPI HID.  The use of the TPM firmware version is not specified in any spec from the TCG or otherwise that I have been able to find.  I believe this was a design choice specific to EDK2.  However, using the TPM firmware version does not match the intended use case from the TCG PC Client spec, where the HID should be comprised of the vendor ID and device ID.  See below.  One problem that arises from this design is that the ACPI tables will change when the TPM FW has been updated.  Since the ACPI tables are hashed into PCR[0], it consequently means that a TPM FW change will cause a change to PCR[0].  It is not intuitive nor spec'd that TPM FW be included in PCR[0] measurements.  In fact, PCR[0] is used only for system FW/UEFI code measurements.  If a user does not update UEFI, then there is no expectation of PCR[0] measurement changes.  I propose that EDK2 change the UpdateHID function to use the vendor ID and device ID read from register TPM_DID_VID_0 (locality 0 offset 0xF00) in the creation of the ACPI HID.

The following was taken from the TCG PC Client Platform Firmware spec:

According to the ACPI Specification (version 5, Errata A, Section 6.1.5 and 6.1.3) a

hardware ID or compatibility ID is either a PNP ID with format “AAA####” or ACPI ID

with format “NNNN####”. The manufacturer ID returned by a TPM2_GetCapability

command can be used to set the “AAA” or “NNNN” portion of the ID. The remaining four

hexadecimal digits should be set to a value that allows software to differentiate different

device classes built by the same manufacturer.


[-- Attachment #2: Type: text/html, Size: 8488 bytes --]

Re: [edk2-devel] TPM ACPI HID creation

[Yao, Jiewen]

[-- Attachment #1: Type: text/plain, Size: 2726 bytes --]

Hi Jason
Would you mind to help us to file a Bugzilla - https://bugzilla.tianocore.org/ for tracking purpose?


Thank you
Yao Jiewen


From: devel@edk2.groups.io [mailto:devel@edk2.groups.io] On Behalf Of Yao, Jiewen
Sent: Wednesday, June 26, 2019 12:08 AM
To: devel@edk2.groups.io; jason.spottswood@hpe.com
Cc: Zhang, Chao B <chao.b.zhang@intel.com>; Yao, Jiewen <jiewen.yao@intel.com>
Subject: Re: [edk2-devel] TPM ACPI HID creation

Thanks Jason.

I think we should NOT measure TPM2 table *after* ACPI table patch.
The measurement should happen *before* ACPI table patch.

Hi Chao
Do you agree on that?

Thank you
Yao Jiewen


From: devel@edk2.groups.io<mailto:devel@edk2.groups.io> [mailto:devel@edk2.groups.io] On Behalf Of jason.spottswood@hpe.com<mailto:jason.spottswood@hpe.com>
Sent: Tuesday, June 25, 2019 11:58 PM
To: devel@edk2.groups.io<mailto:devel@edk2.groups.io>
Subject: [edk2-devel] TPM ACPI HID creation

Tcg2Smm.c has a function "UpdateHID to create the ACPI HID for the TPM.  This function uses the TPM vendor ID combined with the firmware version number to create the ACPI HID.  The use of the TPM firmware version is not specified in any spec from the TCG or otherwise that I have been able to find.  I believe this was a design choice specific to EDK2.  However, using the TPM firmware version does not match the intended use case from the TCG PC Client spec, where the HID should be comprised of the vendor ID and device ID.  See below.  One problem that arises from this design is that the ACPI tables will change when the TPM FW has been updated.  Since the ACPI tables are hashed into PCR[0], it consequently means that a TPM FW change will cause a change to PCR[0].  It is not intuitive nor spec'd that TPM FW be included in PCR[0] measurements.  In fact, PCR[0] is used only for system FW/UEFI code measurements.  If a user does not update UEFI, then there is no expectation of PCR[0] measurement changes.  I propose that EDK2 change the UpdateHID function to use the vendor ID and device ID read from register TPM_DID_VID_0 (locality 0 offset 0xF00) in the creation of the ACPI HID.

The following was taken from the TCG PC Client Platform Firmware spec:

According to the ACPI Specification (version 5, Errata A, Section 6.1.5 and 6.1.3) a

hardware ID or compatibility ID is either a PNP ID with format “AAA####” or ACPI ID

with format “NNNN####”. The manufacturer ID returned by a TPM2_GetCapability

command can be used to set the “AAA” or “NNNN” portion of the ID. The remaining four

hexadecimal digits should be set to a value that allows software to differentiate different

device classes built by the same manufacturer.


[-- Attachment #2: Type: text/html, Size: 10625 bytes --]

Re: [edk2-devel] TPM ACPI HID creation

[Zhang, Chao B]

[-- Attachment #1: Type: text/plain, Size: 2730 bytes --]

Hi Jason:
   The behavior is defined in PFP spec v0.51 Chapter 8.   2 ACPI measuring points don’t  comply with PFP spec. All ACPI DATA for EV_POST_CODE must be measured before fixup.
Tks for pointing it.

From: Yao, Jiewen
Sent: Wednesday, June 26, 2019 12:08 AM
To: devel@edk2.groups.io; jason.spottswood@hpe.com
Cc: Zhang, Chao B <chao.b.zhang@intel.com>; Yao, Jiewen <jiewen.yao@intel.com>
Subject: RE: [edk2-devel] TPM ACPI HID creation

Thanks Jason.

I think we should NOT measure TPM2 table *after* ACPI table patch.
The measurement should happen *before* ACPI table patch.

Hi Chao
Do you agree on that?

Thank you
Yao Jiewen


From: devel@edk2.groups.io<mailto:devel@edk2.groups.io> [mailto:devel@edk2.groups.io] On Behalf Of jason.spottswood@hpe.com<mailto:jason.spottswood@hpe.com>
Sent: Tuesday, June 25, 2019 11:58 PM
To: devel@edk2.groups.io<mailto:devel@edk2.groups.io>
Subject: [edk2-devel] TPM ACPI HID creation

Tcg2Smm.c has a function "UpdateHID to create the ACPI HID for the TPM.  This function uses the TPM vendor ID combined with the firmware version number to create the ACPI HID.  The use of the TPM firmware version is not specified in any spec from the TCG or otherwise that I have been able to find.  I believe this was a design choice specific to EDK2.  However, using the TPM firmware version does not match the intended use case from the TCG PC Client spec, where the HID should be comprised of the vendor ID and device ID.  See below.  One problem that arises from this design is that the ACPI tables will change when the TPM FW has been updated.  Since the ACPI tables are hashed into PCR[0], it consequently means that a TPM FW change will cause a change to PCR[0].  It is not intuitive nor spec'd that TPM FW be included in PCR[0] measurements.  In fact, PCR[0] is used only for system FW/UEFI code measurements.  If a user does not update UEFI, then there is no expectation of PCR[0] measurement changes.  I propose that EDK2 change the UpdateHID function to use the vendor ID and device ID read from register TPM_DID_VID_0 (locality 0 offset 0xF00) in the creation of the ACPI HID.

The following was taken from the TCG PC Client Platform Firmware spec:

According to the ACPI Specification (version 5, Errata A, Section 6.1.5 and 6.1.3) a

hardware ID or compatibility ID is either a PNP ID with format “AAA####” or ACPI ID

with format “NNNN####”. The manufacturer ID returned by a TPM2_GetCapability

command can be used to set the “AAA” or “NNNN” portion of the ID. The remaining four

hexadecimal digits should be set to a value that allows software to differentiate different

device classes built by the same manufacturer.


[-- Attachment #2: Type: text/html, Size: 9527 bytes --]