[PATCH v4 00/11] Measured SEV boot with kernel/initrd/cmdline


Dov Murik
 

BZ: https://bugzilla.tianocore.org/show_bug.cgi?id=3457

Booting with SEV prevented the loading of kernel, initrd, and kernel
command-line via QEMU fw_cfg interface because they arrive from the VMM
which is untrusted in SEV.

However, in some cases the kernel, initrd, and cmdline are not secret
but should not be modified by the host. In such a case, we want to
verify inside the trusted VM that the kernel, initrd, and cmdline are
indeed the ones expected by the Guest Owner, and only if that is the
case go on and boot them up (removing the need for grub inside OVMF in
that mode).

This patch series reserves an area in MEMFD (previously the last 1KB of
the launch secret page) which will contain the hashes of these three
blobs (kernel, initrd, cmdline), each under its own GUID entry. This
tables of hashes is populated by QEMU before launch, and encrypted as
part of the initial VM memory; this makes sure these hashes are part of
the SEV measurement (which has to be approved by the Guest Owner for
secret injection, for example). Note that populating the hashes table
requires QEMU support [1].

OVMF parses the table of hashes populated by QEMU (patch 10), and as it
reads the fw_cfg blobs from QEMU, it will verify each one against the
expected hash. This is all done inside the trusted VM context. If all
the hashes are correct, boot of the kernel is allowed to continue.

Any attempt by QEMU to modify the kernel, initrd, cmdline (including
dropping one of them), or to modify the OVMF code that verifies those
hashes, will cause the initial SEV measurement to change and therefore
will be detectable by the Guest Owner during launch before secret
injection.

Relevant part of OVMF serial log during boot with AmdSevX86 build and
QEMU with -kernel/-initrd/-append:

...
BlobVerifierLibSevHashesConstructor: Found injected hashes table in secure location
Select Item: 0x17
Select Item: 0x8
FetchBlob: loading 7379328 bytes for "kernel"
Select Item: 0x18
Select Item: 0x11
VerifyBlob: Found GUID 4DE79437-ABD2-427F-B835-D5B172D2045B in table
VerifyBlob: Hash comparison succeeded for "kernel"
Select Item: 0xB
FetchBlob: loading 12483878 bytes for "initrd"
Select Item: 0x12
VerifyBlob: Found GUID 44BAF731-3A2F-4BD7-9AF1-41E29169781D in table
VerifyBlob: Hash comparison succeeded for "initrd"
Select Item: 0x14
FetchBlob: loading 86 bytes for "cmdline"
Select Item: 0x15
VerifyBlob: Found GUID 97D02DD8-BD20-4C94-AA78-E7714D36AB2A in table
VerifyBlob: Hash comparison succeeded for "cmdline"
...

The patch series is organized as follows:

1: Simple comment fix in adjacent area in the code.
2: Use GenericQemuLoadImageLib to gain one location for fw_cfg blob
fetching.
3: Allow the (previously blocked) usage of -kernel in AmdSevX64.
4-7: Add BlobVerifierLib with null implementation and use it in the correct
location in QemuKernelLoaderFsDxe.
8-9: Reserve memory for hashes table, declare this area in the reset vector.
10-11: Add the secure implementation BlobVerifierLibSevHashes and use it in
AmdSevX64 builds.

[1] https://lore.kernel.org/qemu-devel/20210624102040.2015280-1-dovmurik@linux.ibm.com/

Code is at
https://github.com/confidential-containers-demo/edk2/tree/sev-hashes-v4

v4 changes:
- BlobVerifierSevHashes (patch 10): more comprehensive overflow tests
when parsing the SEV hashes table structure

v3: https://edk2.groups.io/g/devel/message/77955
v3 changes:
- Rename to BlobVerifierLibNull, use decimal INF_VERSION, remove unused
DebugLib reference, fix doxygen comments, add missing IN attribute
- Rename to BlobVerifierLibSevHashes, use decimal INF_VERSION, fix
doxygen comments, add missing IN attribute,
calculate buffer hash only when the guid is found in hashes table
- SecretPei: use ALIGN_VALUE to round the hob size
- Coding style fixes
- Add missing 'Ref:' in patch 1 commit message
- Fix phrasing and typos in commit messages
- Remove Cc: Laszlo from series

v2: https://edk2.groups.io/g/devel/message/77505
v2 changes:
- Use the last 1KB of the existing SEV launch secret page for hashes table
(instead of reserving a whole new MEMFD page).
- Build on top of commit cf203024745f ("OvmfPkg/GenericQemuLoadImageLib: Read
cmdline from QemuKernelLoaderFs", 2021-06-28) to have a single location in
which all of kernel/initrd/cmdline are fetched from QEMU.
- Use static linking of the two BlobVerifierLib implemenatations.
- Reorganize series.

v1: https://edk2.groups.io/g/devel/message/75567

Cc: Ard Biesheuvel <ardb+tianocore@kernel.org>
Cc: Jordan Justen <jordan.l.justen@intel.com>
Cc: Ashish Kalra <ashish.kalra@amd.com>
Cc: Brijesh Singh <brijesh.singh@amd.com>
Cc: Erdem Aktas <erdemaktas@google.com>
Cc: James Bottomley <jejb@linux.ibm.com>
Cc: Jiewen Yao <jiewen.yao@intel.com>
Cc: Min Xu <min.m.xu@intel.com>
Cc: Tom Lendacky <thomas.lendacky@amd.com>
Cc: Leif Lindholm <leif@nuviainc.com>
Cc: Sami Mujawar <sami.mujawar@arm.com>

---

Ard: please review patch 6 (ArmVirtPkg). Thanks.

Tom, Brijesh: I'll also send the diff for patch 10. Thanks.

---

Dov Murik (8):
OvmfPkg/AmdSev: use GenericQemuLoadImageLib in AmdSev builds
OvmfPkg: add library class BlobVerifierLib with null implementation
OvmfPkg: add BlobVerifierLibNull to DSC
ArmVirtPkg: add BlobVerifierLibNull to DSC
OvmfPkg/QemuKernelLoaderFsDxe: call VerifyBlob after fetch from fw_cfg
OvmfPkg/AmdSev/SecretPei: build hob for full page
OvmfPkg: add BlobVerifierLibSevHashes
OvmfPkg/AmdSev: Enforce hash verification of kernel blobs

James Bottomley (3):
OvmfPkg/AmdSev/SecretDxe: fix header comment to generic naming
OvmfPkg: PlatformBootManagerLibGrub: Allow executing kernel via fw_cfg
OvmfPkg/AmdSev: reserve MEMFD space for for firmware config hashes

OvmfPkg/OvmfPkg.dec | 9 +
ArmVirtPkg/ArmVirtQemu.dsc | 5 +-
ArmVirtPkg/ArmVirtQemuKernel.dsc | 5 +-
OvmfPkg/AmdSev/AmdSevX64.dsc | 9 +-
OvmfPkg/OvmfPkgIa32.dsc | 5 +-
OvmfPkg/OvmfPkgIa32X64.dsc | 5 +-
OvmfPkg/OvmfPkgX64.dsc | 5 +-
OvmfPkg/AmdSev/AmdSevX64.fdf | 5 +-
OvmfPkg/Library/BlobVerifierLib/BlobVerifierLibNull.inf | 24 +++
OvmfPkg/Library/BlobVerifierLib/BlobVerifierLibSevHashes.inf | 37 ++++
OvmfPkg/Library/PlatformBootManagerLibGrub/PlatformBootManagerLibGrub.inf | 2 +
OvmfPkg/ResetVector/ResetVector.inf | 2 +
OvmfPkg/Include/Library/BlobVerifierLib.h | 38 ++++
OvmfPkg/Library/PlatformBootManagerLibGrub/BdsPlatform.h | 11 ++
OvmfPkg/AmdSev/SecretDxe/SecretDxe.c | 2 +-
OvmfPkg/AmdSev/SecretPei/SecretPei.c | 3 +-
OvmfPkg/Library/BlobVerifierLib/BlobVerifierNull.c | 33 ++++
OvmfPkg/Library/BlobVerifierLib/BlobVerifierSevHashes.c | 199 ++++++++++++++++++++
OvmfPkg/Library/PlatformBootManagerLibGrub/BdsPlatform.c | 5 +
OvmfPkg/Library/{PlatformBootManagerLib => PlatformBootManagerLibGrub}/QemuKernel.c | 0
OvmfPkg/QemuKernelLoaderFsDxe/QemuKernelLoaderFsDxe.c | 9 +
OvmfPkg/ResetVector/Ia16/ResetVectorVtf0.asm | 20 ++
OvmfPkg/ResetVector/ResetVector.nasmb | 2 +
23 files changed, 425 insertions(+), 10 deletions(-)
create mode 100644 OvmfPkg/Library/BlobVerifierLib/BlobVerifierLibNull.inf
create mode 100644 OvmfPkg/Library/BlobVerifierLib/BlobVerifierLibSevHashes.inf
create mode 100644 OvmfPkg/Include/Library/BlobVerifierLib.h
create mode 100644 OvmfPkg/Library/BlobVerifierLib/BlobVerifierNull.c
create mode 100644 OvmfPkg/Library/BlobVerifierLib/BlobVerifierSevHashes.c
copy OvmfPkg/Library/{PlatformBootManagerLib => PlatformBootManagerLibGrub}/QemuKernel.c (100%)

--
2.25.1


Yao, Jiewen
 

Hi
I am reviewing this patch series. I am OK with most parts.

And I do have one question:
May I know what is criteria to put a SEV module to OvmfPkg\AmdSev or OvmfPkg directly?

My original understanding is:
If a module is required by OvmfPkg{Ia32,Ia32X64,X64}.{dsc,fdf}, then it should be OvmfPkg.
If a module is only required by OvmfPkg\AmdSev\AmdSevX64.{dsc,fdf}, Then it should be in OvmfPkg\AmdSev.

Am I right?

Thank you
Yao Jiewen

-----Original Message-----
From: devel@edk2.groups.io <devel@edk2.groups.io> On Behalf Of Dov Murik
Sent: Thursday, July 22, 2021 4:43 PM
To: devel@edk2.groups.io
Cc: Dov Murik <dovmurik@linux.ibm.com>; Tobin Feldman-Fitzthum
<tobin@linux.ibm.com>; Tobin Feldman-Fitzthum <tobin@ibm.com>; Jim
Cadden <jcadden@ibm.com>; James Bottomley <jejb@linux.ibm.com>;
Hubertus Franke <frankeh@us.ibm.com>; Ard Biesheuvel
<ardb+tianocore@kernel.org>; Justen, Jordan L <jordan.l.justen@intel.com>;
Ashish Kalra <ashish.kalra@amd.com>; Brijesh Singh <brijesh.singh@amd.com>;
Erdem Aktas <erdemaktas@google.com>; Yao, Jiewen <jiewen.yao@intel.com>;
Xu, Min M <min.m.xu@intel.com>; Tom Lendacky
<thomas.lendacky@amd.com>; Leif Lindholm <leif@nuviainc.com>; Sami
Mujawar <sami.mujawar@arm.com>
Subject: [edk2-devel] [PATCH v4 00/11] Measured SEV boot with
kernel/initrd/cmdline

BZ: https://bugzilla.tianocore.org/show_bug.cgi?id=3457

Booting with SEV prevented the loading of kernel, initrd, and kernel
command-line via QEMU fw_cfg interface because they arrive from the VMM
which is untrusted in SEV.

However, in some cases the kernel, initrd, and cmdline are not secret
but should not be modified by the host. In such a case, we want to
verify inside the trusted VM that the kernel, initrd, and cmdline are
indeed the ones expected by the Guest Owner, and only if that is the
case go on and boot them up (removing the need for grub inside OVMF in
that mode).

This patch series reserves an area in MEMFD (previously the last 1KB of
the launch secret page) which will contain the hashes of these three
blobs (kernel, initrd, cmdline), each under its own GUID entry. This
tables of hashes is populated by QEMU before launch, and encrypted as
part of the initial VM memory; this makes sure these hashes are part of
the SEV measurement (which has to be approved by the Guest Owner for
secret injection, for example). Note that populating the hashes table
requires QEMU support [1].

OVMF parses the table of hashes populated by QEMU (patch 10), and as it
reads the fw_cfg blobs from QEMU, it will verify each one against the
expected hash. This is all done inside the trusted VM context. If all
the hashes are correct, boot of the kernel is allowed to continue.

Any attempt by QEMU to modify the kernel, initrd, cmdline (including
dropping one of them), or to modify the OVMF code that verifies those
hashes, will cause the initial SEV measurement to change and therefore
will be detectable by the Guest Owner during launch before secret
injection.

Relevant part of OVMF serial log during boot with AmdSevX86 build and
QEMU with -kernel/-initrd/-append:

...
BlobVerifierLibSevHashesConstructor: Found injected hashes table in secure
location
Select Item: 0x17
Select Item: 0x8
FetchBlob: loading 7379328 bytes for "kernel"
Select Item: 0x18
Select Item: 0x11
VerifyBlob: Found GUID 4DE79437-ABD2-427F-B835-D5B172D2045B in table
VerifyBlob: Hash comparison succeeded for "kernel"
Select Item: 0xB
FetchBlob: loading 12483878 bytes for "initrd"
Select Item: 0x12
VerifyBlob: Found GUID 44BAF731-3A2F-4BD7-9AF1-41E29169781D in table
VerifyBlob: Hash comparison succeeded for "initrd"
Select Item: 0x14
FetchBlob: loading 86 bytes for "cmdline"
Select Item: 0x15
VerifyBlob: Found GUID 97D02DD8-BD20-4C94-AA78-E7714D36AB2A in table
VerifyBlob: Hash comparison succeeded for "cmdline"
...

The patch series is organized as follows:

1: Simple comment fix in adjacent area in the code.
2: Use GenericQemuLoadImageLib to gain one location for fw_cfg blob
fetching.
3: Allow the (previously blocked) usage of -kernel in AmdSevX64.
4-7: Add BlobVerifierLib with null implementation and use it in the correct
location in QemuKernelLoaderFsDxe.
8-9: Reserve memory for hashes table, declare this area in the reset vector.
10-11: Add the secure implementation BlobVerifierLibSevHashes and use it in
AmdSevX64 builds.

[1] https://lore.kernel.org/qemu-devel/20210624102040.2015280-1-
dovmurik@linux.ibm.com/

Code is at
https://github.com/confidential-containers-demo/edk2/tree/sev-hashes-v4

v4 changes:
- BlobVerifierSevHashes (patch 10): more comprehensive overflow tests
when parsing the SEV hashes table structure

v3: https://edk2.groups.io/g/devel/message/77955
v3 changes:
- Rename to BlobVerifierLibNull, use decimal INF_VERSION, remove unused
DebugLib reference, fix doxygen comments, add missing IN attribute
- Rename to BlobVerifierLibSevHashes, use decimal INF_VERSION, fix
doxygen comments, add missing IN attribute,
calculate buffer hash only when the guid is found in hashes table
- SecretPei: use ALIGN_VALUE to round the hob size
- Coding style fixes
- Add missing 'Ref:' in patch 1 commit message
- Fix phrasing and typos in commit messages
- Remove Cc: Laszlo from series

v2: https://edk2.groups.io/g/devel/message/77505
v2 changes:
- Use the last 1KB of the existing SEV launch secret page for hashes table
(instead of reserving a whole new MEMFD page).
- Build on top of commit cf203024745f ("OvmfPkg/GenericQemuLoadImageLib:
Read
cmdline from QemuKernelLoaderFs", 2021-06-28) to have a single location in
which all of kernel/initrd/cmdline are fetched from QEMU.
- Use static linking of the two BlobVerifierLib implemenatations.
- Reorganize series.

v1: https://edk2.groups.io/g/devel/message/75567

Cc: Ard Biesheuvel <ardb+tianocore@kernel.org>
Cc: Jordan Justen <jordan.l.justen@intel.com>
Cc: Ashish Kalra <ashish.kalra@amd.com>
Cc: Brijesh Singh <brijesh.singh@amd.com>
Cc: Erdem Aktas <erdemaktas@google.com>
Cc: James Bottomley <jejb@linux.ibm.com>
Cc: Jiewen Yao <jiewen.yao@intel.com>
Cc: Min Xu <min.m.xu@intel.com>
Cc: Tom Lendacky <thomas.lendacky@amd.com>
Cc: Leif Lindholm <leif@nuviainc.com>
Cc: Sami Mujawar <sami.mujawar@arm.com>

---

Ard: please review patch 6 (ArmVirtPkg). Thanks.

Tom, Brijesh: I'll also send the diff for patch 10. Thanks.

---

Dov Murik (8):
OvmfPkg/AmdSev: use GenericQemuLoadImageLib in AmdSev builds
OvmfPkg: add library class BlobVerifierLib with null implementation
OvmfPkg: add BlobVerifierLibNull to DSC
ArmVirtPkg: add BlobVerifierLibNull to DSC
OvmfPkg/QemuKernelLoaderFsDxe: call VerifyBlob after fetch from fw_cfg
OvmfPkg/AmdSev/SecretPei: build hob for full page
OvmfPkg: add BlobVerifierLibSevHashes
OvmfPkg/AmdSev: Enforce hash verification of kernel blobs

James Bottomley (3):
OvmfPkg/AmdSev/SecretDxe: fix header comment to generic naming
OvmfPkg: PlatformBootManagerLibGrub: Allow executing kernel via fw_cfg
OvmfPkg/AmdSev: reserve MEMFD space for for firmware config hashes

OvmfPkg/OvmfPkg.dec | 9 +
ArmVirtPkg/ArmVirtQemu.dsc | 5 +-
ArmVirtPkg/ArmVirtQemuKernel.dsc | 5 +-
OvmfPkg/AmdSev/AmdSevX64.dsc | 9 +-
OvmfPkg/OvmfPkgIa32.dsc | 5 +-
OvmfPkg/OvmfPkgIa32X64.dsc | 5 +-
OvmfPkg/OvmfPkgX64.dsc | 5 +-
OvmfPkg/AmdSev/AmdSevX64.fdf | 5 +-
OvmfPkg/Library/BlobVerifierLib/BlobVerifierLibNull.inf | 24
+++
OvmfPkg/Library/BlobVerifierLib/BlobVerifierLibSevHashes.inf | 37
++++

OvmfPkg/Library/PlatformBootManagerLibGrub/PlatformBootManagerLibGrub.
inf | 2 +
OvmfPkg/ResetVector/ResetVector.inf | 2 +
OvmfPkg/Include/Library/BlobVerifierLib.h | 38 ++++
OvmfPkg/Library/PlatformBootManagerLibGrub/BdsPlatform.h |
11 ++
OvmfPkg/AmdSev/SecretDxe/SecretDxe.c | 2 +-
OvmfPkg/AmdSev/SecretPei/SecretPei.c | 3 +-
OvmfPkg/Library/BlobVerifierLib/BlobVerifierNull.c | 33 ++++
OvmfPkg/Library/BlobVerifierLib/BlobVerifierSevHashes.c | 199
++++++++++++++++++++
OvmfPkg/Library/PlatformBootManagerLibGrub/BdsPlatform.c |
5 +
OvmfPkg/Library/{PlatformBootManagerLib =>
PlatformBootManagerLibGrub}/QemuKernel.c | 0
OvmfPkg/QemuKernelLoaderFsDxe/QemuKernelLoaderFsDxe.c
| 9 +
OvmfPkg/ResetVector/Ia16/ResetVectorVtf0.asm | 20 ++
OvmfPkg/ResetVector/ResetVector.nasmb | 2 +
23 files changed, 425 insertions(+), 10 deletions(-)
create mode 100644 OvmfPkg/Library/BlobVerifierLib/BlobVerifierLibNull.inf
create mode 100644
OvmfPkg/Library/BlobVerifierLib/BlobVerifierLibSevHashes.inf
create mode 100644 OvmfPkg/Include/Library/BlobVerifierLib.h
create mode 100644 OvmfPkg/Library/BlobVerifierLib/BlobVerifierNull.c
create mode 100644 OvmfPkg/Library/BlobVerifierLib/BlobVerifierSevHashes.c
copy OvmfPkg/Library/{PlatformBootManagerLib =>
PlatformBootManagerLibGrub}/QemuKernel.c (100%)

--
2.25.1





Dov Murik
 

Hi Jiewen,

On 25/07/2021 5:43, Yao, Jiewen wrote:
Hi
I am reviewing this patch series. I am OK with most parts.
Thank you.


And I do have one question:
May I know what is criteria to put a SEV module to OvmfPkg\AmdSev or OvmfPkg directly?

My original understanding is:
If a module is required by OvmfPkg{Ia32,Ia32X64,X64}.{dsc,fdf}, then it should be OvmfPkg.
If a module is only required by OvmfPkg\AmdSev\AmdSevX64.{dsc,fdf}, Then it should be in OvmfPkg\AmdSev.

Am I right?
I actually don't know the criteria. What you say sounds reasonable.
I'll also let James (who introduced the AmdSevX64 target) say what he
thinks.


If that is indeed the case, then I need to:

1. Modify patch 4: put the code of the null implementation in
OvmfPkf/BlobVerifierLibNull/

2. Modify patches 5+6: use the new path of the BlobVerifierLibNull inf file

3. Modify patches 10+11: put the code of the SevHashes implementation in
OvmfPkg/AmdSev/BlobVerifierLibSevHashes/

Jiewen, is that what you had in mind?



Two other things to consider:

A. The blob verification by hashes just uses initially-measured memory,
and no other features of SEV. It might be useful for other Confidential
Computing implementations. But maybe if that need arises then we'll
extract it from the AmdSev folder.

B. There were talks about reducing the number of targets, and maybe
unifying AmdSevX64 into OvmfPkgX64. If this is indeed the direction
we're going to, then there's no need to separate the code.


Thanks,
-Dov


Thank you
Yao Jiewen


-----Original Message-----
From: devel@edk2.groups.io <devel@edk2.groups.io> On Behalf Of Dov Murik
Sent: Thursday, July 22, 2021 4:43 PM
To: devel@edk2.groups.io
Cc: Dov Murik <dovmurik@linux.ibm.com>; Tobin Feldman-Fitzthum
<tobin@linux.ibm.com>; Tobin Feldman-Fitzthum <tobin@ibm.com>; Jim
Cadden <jcadden@ibm.com>; James Bottomley <jejb@linux.ibm.com>;
Hubertus Franke <frankeh@us.ibm.com>; Ard Biesheuvel
<ardb+tianocore@kernel.org>; Justen, Jordan L <jordan.l.justen@intel.com>;
Ashish Kalra <ashish.kalra@amd.com>; Brijesh Singh <brijesh.singh@amd.com>;
Erdem Aktas <erdemaktas@google.com>; Yao, Jiewen <jiewen.yao@intel.com>;
Xu, Min M <min.m.xu@intel.com>; Tom Lendacky
<thomas.lendacky@amd.com>; Leif Lindholm <leif@nuviainc.com>; Sami
Mujawar <sami.mujawar@arm.com>
Subject: [edk2-devel] [PATCH v4 00/11] Measured SEV boot with
kernel/initrd/cmdline

BZ: https://bugzilla.tianocore.org/show_bug.cgi?id=3457

Booting with SEV prevented the loading of kernel, initrd, and kernel
command-line via QEMU fw_cfg interface because they arrive from the VMM
which is untrusted in SEV.

However, in some cases the kernel, initrd, and cmdline are not secret
but should not be modified by the host. In such a case, we want to
verify inside the trusted VM that the kernel, initrd, and cmdline are
indeed the ones expected by the Guest Owner, and only if that is the
case go on and boot them up (removing the need for grub inside OVMF in
that mode).

This patch series reserves an area in MEMFD (previously the last 1KB of
the launch secret page) which will contain the hashes of these three
blobs (kernel, initrd, cmdline), each under its own GUID entry. This
tables of hashes is populated by QEMU before launch, and encrypted as
part of the initial VM memory; this makes sure these hashes are part of
the SEV measurement (which has to be approved by the Guest Owner for
secret injection, for example). Note that populating the hashes table
requires QEMU support [1].

OVMF parses the table of hashes populated by QEMU (patch 10), and as it
reads the fw_cfg blobs from QEMU, it will verify each one against the
expected hash. This is all done inside the trusted VM context. If all
the hashes are correct, boot of the kernel is allowed to continue.

Any attempt by QEMU to modify the kernel, initrd, cmdline (including
dropping one of them), or to modify the OVMF code that verifies those
hashes, will cause the initial SEV measurement to change and therefore
will be detectable by the Guest Owner during launch before secret
injection.

Relevant part of OVMF serial log during boot with AmdSevX86 build and
QEMU with -kernel/-initrd/-append:

...
BlobVerifierLibSevHashesConstructor: Found injected hashes table in secure
location
Select Item: 0x17
Select Item: 0x8
FetchBlob: loading 7379328 bytes for "kernel"
Select Item: 0x18
Select Item: 0x11
VerifyBlob: Found GUID 4DE79437-ABD2-427F-B835-D5B172D2045B in table
VerifyBlob: Hash comparison succeeded for "kernel"
Select Item: 0xB
FetchBlob: loading 12483878 bytes for "initrd"
Select Item: 0x12
VerifyBlob: Found GUID 44BAF731-3A2F-4BD7-9AF1-41E29169781D in table
VerifyBlob: Hash comparison succeeded for "initrd"
Select Item: 0x14
FetchBlob: loading 86 bytes for "cmdline"
Select Item: 0x15
VerifyBlob: Found GUID 97D02DD8-BD20-4C94-AA78-E7714D36AB2A in table
VerifyBlob: Hash comparison succeeded for "cmdline"
...

The patch series is organized as follows:

1: Simple comment fix in adjacent area in the code.
2: Use GenericQemuLoadImageLib to gain one location for fw_cfg blob
fetching.
3: Allow the (previously blocked) usage of -kernel in AmdSevX64.
4-7: Add BlobVerifierLib with null implementation and use it in the correct
location in QemuKernelLoaderFsDxe.
8-9: Reserve memory for hashes table, declare this area in the reset vector.
10-11: Add the secure implementation BlobVerifierLibSevHashes and use it in
AmdSevX64 builds.

[1] https://lore.kernel.org/qemu-devel/20210624102040.2015280-1-
dovmurik@linux.ibm.com/

Code is at
https://github.com/confidential-containers-demo/edk2/tree/sev-hashes-v4

v4 changes:
- BlobVerifierSevHashes (patch 10): more comprehensive overflow tests
when parsing the SEV hashes table structure

v3: https://edk2.groups.io/g/devel/message/77955
v3 changes:
- Rename to BlobVerifierLibNull, use decimal INF_VERSION, remove unused
DebugLib reference, fix doxygen comments, add missing IN attribute
- Rename to BlobVerifierLibSevHashes, use decimal INF_VERSION, fix
doxygen comments, add missing IN attribute,
calculate buffer hash only when the guid is found in hashes table
- SecretPei: use ALIGN_VALUE to round the hob size
- Coding style fixes
- Add missing 'Ref:' in patch 1 commit message
- Fix phrasing and typos in commit messages
- Remove Cc: Laszlo from series

v2: https://edk2.groups.io/g/devel/message/77505
v2 changes:
- Use the last 1KB of the existing SEV launch secret page for hashes table
(instead of reserving a whole new MEMFD page).
- Build on top of commit cf203024745f ("OvmfPkg/GenericQemuLoadImageLib:
Read
cmdline from QemuKernelLoaderFs", 2021-06-28) to have a single location in
which all of kernel/initrd/cmdline are fetched from QEMU.
- Use static linking of the two BlobVerifierLib implemenatations.
- Reorganize series.

v1: https://edk2.groups.io/g/devel/message/75567

Cc: Ard Biesheuvel <ardb+tianocore@kernel.org>
Cc: Jordan Justen <jordan.l.justen@intel.com>
Cc: Ashish Kalra <ashish.kalra@amd.com>
Cc: Brijesh Singh <brijesh.singh@amd.com>
Cc: Erdem Aktas <erdemaktas@google.com>
Cc: James Bottomley <jejb@linux.ibm.com>
Cc: Jiewen Yao <jiewen.yao@intel.com>
Cc: Min Xu <min.m.xu@intel.com>
Cc: Tom Lendacky <thomas.lendacky@amd.com>
Cc: Leif Lindholm <leif@nuviainc.com>
Cc: Sami Mujawar <sami.mujawar@arm.com>

---

Ard: please review patch 6 (ArmVirtPkg). Thanks.

Tom, Brijesh: I'll also send the diff for patch 10. Thanks.

---

Dov Murik (8):
OvmfPkg/AmdSev: use GenericQemuLoadImageLib in AmdSev builds
OvmfPkg: add library class BlobVerifierLib with null implementation
OvmfPkg: add BlobVerifierLibNull to DSC
ArmVirtPkg: add BlobVerifierLibNull to DSC
OvmfPkg/QemuKernelLoaderFsDxe: call VerifyBlob after fetch from fw_cfg
OvmfPkg/AmdSev/SecretPei: build hob for full page
OvmfPkg: add BlobVerifierLibSevHashes
OvmfPkg/AmdSev: Enforce hash verification of kernel blobs

James Bottomley (3):
OvmfPkg/AmdSev/SecretDxe: fix header comment to generic naming
OvmfPkg: PlatformBootManagerLibGrub: Allow executing kernel via fw_cfg
OvmfPkg/AmdSev: reserve MEMFD space for for firmware config hashes

OvmfPkg/OvmfPkg.dec | 9 +
ArmVirtPkg/ArmVirtQemu.dsc | 5 +-
ArmVirtPkg/ArmVirtQemuKernel.dsc | 5 +-
OvmfPkg/AmdSev/AmdSevX64.dsc | 9 +-
OvmfPkg/OvmfPkgIa32.dsc | 5 +-
OvmfPkg/OvmfPkgIa32X64.dsc | 5 +-
OvmfPkg/OvmfPkgX64.dsc | 5 +-
OvmfPkg/AmdSev/AmdSevX64.fdf | 5 +-
OvmfPkg/Library/BlobVerifierLib/BlobVerifierLibNull.inf | 24
+++
OvmfPkg/Library/BlobVerifierLib/BlobVerifierLibSevHashes.inf | 37
++++

OvmfPkg/Library/PlatformBootManagerLibGrub/PlatformBootManagerLibGrub.
inf | 2 +
OvmfPkg/ResetVector/ResetVector.inf | 2 +
OvmfPkg/Include/Library/BlobVerifierLib.h | 38 ++++
OvmfPkg/Library/PlatformBootManagerLibGrub/BdsPlatform.h |
11 ++
OvmfPkg/AmdSev/SecretDxe/SecretDxe.c | 2 +-
OvmfPkg/AmdSev/SecretPei/SecretPei.c | 3 +-
OvmfPkg/Library/BlobVerifierLib/BlobVerifierNull.c | 33 ++++
OvmfPkg/Library/BlobVerifierLib/BlobVerifierSevHashes.c | 199
++++++++++++++++++++
OvmfPkg/Library/PlatformBootManagerLibGrub/BdsPlatform.c |
5 +
OvmfPkg/Library/{PlatformBootManagerLib =>
PlatformBootManagerLibGrub}/QemuKernel.c | 0
OvmfPkg/QemuKernelLoaderFsDxe/QemuKernelLoaderFsDxe.c
| 9 +
OvmfPkg/ResetVector/Ia16/ResetVectorVtf0.asm | 20 ++
OvmfPkg/ResetVector/ResetVector.nasmb | 2 +
23 files changed, 425 insertions(+), 10 deletions(-)
create mode 100644 OvmfPkg/Library/BlobVerifierLib/BlobVerifierLibNull.inf
create mode 100644
OvmfPkg/Library/BlobVerifierLib/BlobVerifierLibSevHashes.inf
create mode 100644 OvmfPkg/Include/Library/BlobVerifierLib.h
create mode 100644 OvmfPkg/Library/BlobVerifierLib/BlobVerifierNull.c
create mode 100644 OvmfPkg/Library/BlobVerifierLib/BlobVerifierSevHashes.c
copy OvmfPkg/Library/{PlatformBootManagerLib =>
PlatformBootManagerLibGrub}/QemuKernel.c (100%)

--
2.25.1





Yao, Jiewen
 

Thank you Dov.
Comment below:

-----Original Message-----
From: Dov Murik <dovmurik@linux.ibm.com>
Sent: Sunday, July 25, 2021 3:53 PM
To: Yao, Jiewen <jiewen.yao@intel.com>; devel@edk2.groups.io
Cc: Tobin Feldman-Fitzthum <tobin@linux.ibm.com>; Tobin Feldman-Fitzthum
<tobin@ibm.com>; Jim Cadden <jcadden@ibm.com>; James Bottomley
<jejb@linux.ibm.com>; Hubertus Franke <frankeh@us.ibm.com>; Ard Biesheuvel
<ardb+tianocore@kernel.org>; Justen, Jordan L <jordan.l.justen@intel.com>;
Ashish Kalra <ashish.kalra@amd.com>; Brijesh Singh <brijesh.singh@amd.com>;
Erdem Aktas <erdemaktas@google.com>; Xu, Min M <min.m.xu@intel.com>;
Tom Lendacky <thomas.lendacky@amd.com>; Leif Lindholm
<leif@nuviainc.com>; Sami Mujawar <sami.mujawar@arm.com>; Dov Murik
<dovmurik@linux.ibm.com>
Subject: Re: [edk2-devel] [PATCH v4 00/11] Measured SEV boot with
kernel/initrd/cmdline

Hi Jiewen,

On 25/07/2021 5:43, Yao, Jiewen wrote:
Hi
I am reviewing this patch series. I am OK with most parts.
Thank you.


And I do have one question:
May I know what is criteria to put a SEV module to OvmfPkg\AmdSev or
OvmfPkg directly?

My original understanding is:
If a module is required by OvmfPkg{Ia32,Ia32X64,X64}.{dsc,fdf}, then it should
be OvmfPkg.
If a module is only required by OvmfPkg\AmdSev\AmdSevX64.{dsc,fdf}, Then it
should be in OvmfPkg\AmdSev.

Am I right?
I actually don't know the criteria. What you say sounds reasonable.
I'll also let James (who introduced the AmdSevX64 target) say what he
thinks.
[Jiewen] OK.




If that is indeed the case, then I need to:

1. Modify patch 4: put the code of the null implementation in
OvmfPkf/BlobVerifierLibNull/
[Jiewen] Since this is a library, I expect to be OvmfPkf/Library/BlobVerifierLibNull/


2. Modify patches 5+6: use the new path of the BlobVerifierLibNull inf file

3. Modify patches 10+11: put the code of the SevHashes implementation in
OvmfPkg/AmdSev/BlobVerifierLibSevHashes/

Jiewen, is that what you had in mind?
[Jiewen] Let's comply with the exiting rule.
1) A library in a packet should be in XXXPkg/Library/AAALib in general. (For example, OvmfPkg/Library)
2) If a library in a packet belongs to a feature, then it should be XXXPkg/FeatureYYY/AAALib. (For example, OvmfPkg/Csm/CsmSupportLib)





Two other things to consider:

A. The blob verification by hashes just uses initially-measured memory,
and no other features of SEV. It might be useful for other Confidential
Computing implementations. But maybe if that need arises then we'll
extract it from the AmdSev folder.
[Jiewen] I think so. Because it is generic, that is why I agree that the *library class* name does not include SEV keyword.
The *library instance* should add *Sev* keyword, because the implementation does include SEV specific, such as SEV_HASH_TABLE_GUID, SEV_KERNEL_HASH_GUID.

If you want to make it for generic confidential computing, then more work should be done. For example:
1) The instance name should be BlobVerificationLibTeeLinuxModuleHash. (TEE to replace SEV. We need Linux keyword, because I don’t think it is applicable for Windows)
2) We need consider crypto agile. The current instance only consider SHA256, which is not allowed in TDX.
3) The GUID definition should be in OvmfPkg.dec, as the interface.

Since we don’t have a TEE folder yet, I prefer we defer that work.
Later, when we finish all Sev/Tdx work, we can consider create a common Tee dir or event a package. But that may happen in next year.




B. There were talks about reducing the number of targets, and maybe
unifying AmdSevX64 into OvmfPkgX64. If this is indeed the direction
we're going to, then there's no need to separate the code.
[Jiewen] I think we are following what Laslo suggested.
1) The OvmfPkg includes the *basic* feature at first.
2) The advanced feature is checked into SEV folder or TDX folder, at first.
3) We can revisit those advanced feature once we think they are mature.

We agree that direction, and we should follow that.
Let's keep #1 and #2 at first to finish the feature at first (in this year). Then we can see how to enhance in #3 (maybe next year).
The more we know each other, the better we can create an architecture to support TEE.

Thank you
Yao Jiewen



Thanks,
-Dov


Thank you
Yao Jiewen


-----Original Message-----
From: devel@edk2.groups.io <devel@edk2.groups.io> On Behalf Of Dov
Murik
Sent: Thursday, July 22, 2021 4:43 PM
To: devel@edk2.groups.io
Cc: Dov Murik <dovmurik@linux.ibm.com>; Tobin Feldman-Fitzthum
<tobin@linux.ibm.com>; Tobin Feldman-Fitzthum <tobin@ibm.com>; Jim
Cadden <jcadden@ibm.com>; James Bottomley <jejb@linux.ibm.com>;
Hubertus Franke <frankeh@us.ibm.com>; Ard Biesheuvel
<ardb+tianocore@kernel.org>; Justen, Jordan L <jordan.l.justen@intel.com>;
Ashish Kalra <ashish.kalra@amd.com>; Brijesh Singh
<brijesh.singh@amd.com>;
Erdem Aktas <erdemaktas@google.com>; Yao, Jiewen
<jiewen.yao@intel.com>;
Xu, Min M <min.m.xu@intel.com>; Tom Lendacky
<thomas.lendacky@amd.com>; Leif Lindholm <leif@nuviainc.com>; Sami
Mujawar <sami.mujawar@arm.com>
Subject: [edk2-devel] [PATCH v4 00/11] Measured SEV boot with
kernel/initrd/cmdline

BZ: https://bugzilla.tianocore.org/show_bug.cgi?id=3457

Booting with SEV prevented the loading of kernel, initrd, and kernel
command-line via QEMU fw_cfg interface because they arrive from the VMM
which is untrusted in SEV.

However, in some cases the kernel, initrd, and cmdline are not secret
but should not be modified by the host. In such a case, we want to
verify inside the trusted VM that the kernel, initrd, and cmdline are
indeed the ones expected by the Guest Owner, and only if that is the
case go on and boot them up (removing the need for grub inside OVMF in
that mode).

This patch series reserves an area in MEMFD (previously the last 1KB of
the launch secret page) which will contain the hashes of these three
blobs (kernel, initrd, cmdline), each under its own GUID entry. This
tables of hashes is populated by QEMU before launch, and encrypted as
part of the initial VM memory; this makes sure these hashes are part of
the SEV measurement (which has to be approved by the Guest Owner for
secret injection, for example). Note that populating the hashes table
requires QEMU support [1].

OVMF parses the table of hashes populated by QEMU (patch 10), and as it
reads the fw_cfg blobs from QEMU, it will verify each one against the
expected hash. This is all done inside the trusted VM context. If all
the hashes are correct, boot of the kernel is allowed to continue.

Any attempt by QEMU to modify the kernel, initrd, cmdline (including
dropping one of them), or to modify the OVMF code that verifies those
hashes, will cause the initial SEV measurement to change and therefore
will be detectable by the Guest Owner during launch before secret
injection.

Relevant part of OVMF serial log during boot with AmdSevX86 build and
QEMU with -kernel/-initrd/-append:

...
BlobVerifierLibSevHashesConstructor: Found injected hashes table in secure
location
Select Item: 0x17
Select Item: 0x8
FetchBlob: loading 7379328 bytes for "kernel"
Select Item: 0x18
Select Item: 0x11
VerifyBlob: Found GUID 4DE79437-ABD2-427F-B835-D5B172D2045B in
table
VerifyBlob: Hash comparison succeeded for "kernel"
Select Item: 0xB
FetchBlob: loading 12483878 bytes for "initrd"
Select Item: 0x12
VerifyBlob: Found GUID 44BAF731-3A2F-4BD7-9AF1-41E29169781D in table
VerifyBlob: Hash comparison succeeded for "initrd"
Select Item: 0x14
FetchBlob: loading 86 bytes for "cmdline"
Select Item: 0x15
VerifyBlob: Found GUID 97D02DD8-BD20-4C94-AA78-E7714D36AB2A in
table
VerifyBlob: Hash comparison succeeded for "cmdline"
...

The patch series is organized as follows:

1: Simple comment fix in adjacent area in the code.
2: Use GenericQemuLoadImageLib to gain one location for fw_cfg blob
fetching.
3: Allow the (previously blocked) usage of -kernel in AmdSevX64.
4-7: Add BlobVerifierLib with null implementation and use it in the correct
location in QemuKernelLoaderFsDxe.
8-9: Reserve memory for hashes table, declare this area in the reset vector.
10-11: Add the secure implementation BlobVerifierLibSevHashes and use it in
AmdSevX64 builds.

[1] https://lore.kernel.org/qemu-devel/20210624102040.2015280-1-
dovmurik@linux.ibm.com/

Code is at
https://github.com/confidential-containers-demo/edk2/tree/sev-hashes-v4

v4 changes:
- BlobVerifierSevHashes (patch 10): more comprehensive overflow tests
when parsing the SEV hashes table structure

v3: https://edk2.groups.io/g/devel/message/77955
v3 changes:
- Rename to BlobVerifierLibNull, use decimal INF_VERSION, remove unused
DebugLib reference, fix doxygen comments, add missing IN attribute
- Rename to BlobVerifierLibSevHashes, use decimal INF_VERSION, fix
doxygen comments, add missing IN attribute,
calculate buffer hash only when the guid is found in hashes table
- SecretPei: use ALIGN_VALUE to round the hob size
- Coding style fixes
- Add missing 'Ref:' in patch 1 commit message
- Fix phrasing and typos in commit messages
- Remove Cc: Laszlo from series

v2: https://edk2.groups.io/g/devel/message/77505
v2 changes:
- Use the last 1KB of the existing SEV launch secret page for hashes table
(instead of reserving a whole new MEMFD page).
- Build on top of commit cf203024745f
("OvmfPkg/GenericQemuLoadImageLib:
Read
cmdline from QemuKernelLoaderFs", 2021-06-28) to have a single location
in
which all of kernel/initrd/cmdline are fetched from QEMU.
- Use static linking of the two BlobVerifierLib implemenatations.
- Reorganize series.

v1: https://edk2.groups.io/g/devel/message/75567

Cc: Ard Biesheuvel <ardb+tianocore@kernel.org>
Cc: Jordan Justen <jordan.l.justen@intel.com>
Cc: Ashish Kalra <ashish.kalra@amd.com>
Cc: Brijesh Singh <brijesh.singh@amd.com>
Cc: Erdem Aktas <erdemaktas@google.com>
Cc: James Bottomley <jejb@linux.ibm.com>
Cc: Jiewen Yao <jiewen.yao@intel.com>
Cc: Min Xu <min.m.xu@intel.com>
Cc: Tom Lendacky <thomas.lendacky@amd.com>
Cc: Leif Lindholm <leif@nuviainc.com>
Cc: Sami Mujawar <sami.mujawar@arm.com>

---

Ard: please review patch 6 (ArmVirtPkg). Thanks.

Tom, Brijesh: I'll also send the diff for patch 10. Thanks.

---

Dov Murik (8):
OvmfPkg/AmdSev: use GenericQemuLoadImageLib in AmdSev builds
OvmfPkg: add library class BlobVerifierLib with null implementation
OvmfPkg: add BlobVerifierLibNull to DSC
ArmVirtPkg: add BlobVerifierLibNull to DSC
OvmfPkg/QemuKernelLoaderFsDxe: call VerifyBlob after fetch from fw_cfg
OvmfPkg/AmdSev/SecretPei: build hob for full page
OvmfPkg: add BlobVerifierLibSevHashes
OvmfPkg/AmdSev: Enforce hash verification of kernel blobs

James Bottomley (3):
OvmfPkg/AmdSev/SecretDxe: fix header comment to generic naming
OvmfPkg: PlatformBootManagerLibGrub: Allow executing kernel via fw_cfg
OvmfPkg/AmdSev: reserve MEMFD space for for firmware config hashes

OvmfPkg/OvmfPkg.dec | 9 +
ArmVirtPkg/ArmVirtQemu.dsc | 5 +-
ArmVirtPkg/ArmVirtQemuKernel.dsc | 5 +-
OvmfPkg/AmdSev/AmdSevX64.dsc | 9 +-
OvmfPkg/OvmfPkgIa32.dsc | 5 +-
OvmfPkg/OvmfPkgIa32X64.dsc | 5 +-
OvmfPkg/OvmfPkgX64.dsc | 5 +-
OvmfPkg/AmdSev/AmdSevX64.fdf | 5 +-
OvmfPkg/Library/BlobVerifierLib/BlobVerifierLibNull.inf | 24
+++
OvmfPkg/Library/BlobVerifierLib/BlobVerifierLibSevHashes.inf |
37
++++

OvmfPkg/Library/PlatformBootManagerLibGrub/PlatformBootManagerLibGrub.
inf | 2 +
OvmfPkg/ResetVector/ResetVector.inf | 2 +
OvmfPkg/Include/Library/BlobVerifierLib.h | 38 ++++
OvmfPkg/Library/PlatformBootManagerLibGrub/BdsPlatform.h
|
11 ++
OvmfPkg/AmdSev/SecretDxe/SecretDxe.c | 2 +-
OvmfPkg/AmdSev/SecretPei/SecretPei.c | 3 +-
OvmfPkg/Library/BlobVerifierLib/BlobVerifierNull.c | 33
++++
OvmfPkg/Library/BlobVerifierLib/BlobVerifierSevHashes.c |
199
++++++++++++++++++++
OvmfPkg/Library/PlatformBootManagerLibGrub/BdsPlatform.c
|
5 +
OvmfPkg/Library/{PlatformBootManagerLib =>
PlatformBootManagerLibGrub}/QemuKernel.c | 0
OvmfPkg/QemuKernelLoaderFsDxe/QemuKernelLoaderFsDxe.c
| 9 +
OvmfPkg/ResetVector/Ia16/ResetVectorVtf0.asm | 20
++
OvmfPkg/ResetVector/ResetVector.nasmb | 2 +
23 files changed, 425 insertions(+), 10 deletions(-)
create mode 100644 OvmfPkg/Library/BlobVerifierLib/BlobVerifierLibNull.inf
create mode 100644
OvmfPkg/Library/BlobVerifierLib/BlobVerifierLibSevHashes.inf
create mode 100644 OvmfPkg/Include/Library/BlobVerifierLib.h
create mode 100644 OvmfPkg/Library/BlobVerifierLib/BlobVerifierNull.c
create mode 100644
OvmfPkg/Library/BlobVerifierLib/BlobVerifierSevHashes.c
copy OvmfPkg/Library/{PlatformBootManagerLib =>
PlatformBootManagerLibGrub}/QemuKernel.c (100%)

--
2.25.1





Dov Murik
 

Thanks for the explanations. Comments below:


On 25/07/2021 11:17, Yao, Jiewen wrote:
Thank you Dov.
Comment below:

-----Original Message-----
From: Dov Murik <dovmurik@linux.ibm.com>
Sent: Sunday, July 25, 2021 3:53 PM
To: Yao, Jiewen <jiewen.yao@intel.com>; devel@edk2.groups.io
Cc: Tobin Feldman-Fitzthum <tobin@linux.ibm.com>; Tobin Feldman-Fitzthum
<tobin@ibm.com>; Jim Cadden <jcadden@ibm.com>; James Bottomley
<jejb@linux.ibm.com>; Hubertus Franke <frankeh@us.ibm.com>; Ard Biesheuvel
<ardb+tianocore@kernel.org>; Justen, Jordan L <jordan.l.justen@intel.com>;
Ashish Kalra <ashish.kalra@amd.com>; Brijesh Singh <brijesh.singh@amd.com>;
Erdem Aktas <erdemaktas@google.com>; Xu, Min M <min.m.xu@intel.com>;
Tom Lendacky <thomas.lendacky@amd.com>; Leif Lindholm
<leif@nuviainc.com>; Sami Mujawar <sami.mujawar@arm.com>; Dov Murik
<dovmurik@linux.ibm.com>
Subject: Re: [edk2-devel] [PATCH v4 00/11] Measured SEV boot with
kernel/initrd/cmdline

Hi Jiewen,

On 25/07/2021 5:43, Yao, Jiewen wrote:
Hi
I am reviewing this patch series. I am OK with most parts.
Thank you.


And I do have one question:
May I know what is criteria to put a SEV module to OvmfPkg\AmdSev or
OvmfPkg directly?

My original understanding is:
If a module is required by OvmfPkg{Ia32,Ia32X64,X64}.{dsc,fdf}, then it should
be OvmfPkg.
If a module is only required by OvmfPkg\AmdSev\AmdSevX64.{dsc,fdf}, Then it
should be in OvmfPkg\AmdSev.

Am I right?
I actually don't know the criteria. What you say sounds reasonable.
I'll also let James (who introduced the AmdSevX64 target) say what he
thinks.
[Jiewen] OK.




If that is indeed the case, then I need to:

1. Modify patch 4: put the code of the null implementation in
OvmfPkf/BlobVerifierLibNull/
[Jiewen] Since this is a library, I expect to be OvmfPkf/Library/BlobVerifierLibNull/
Yes, you're right, I missed that "/Library/" in my description.



2. Modify patches 5+6: use the new path of the BlobVerifierLibNull inf file

3. Modify patches 10+11: put the code of the SevHashes implementation in
OvmfPkg/AmdSev/BlobVerifierLibSevHashes/

Jiewen, is that what you had in mind?
[Jiewen] Let's comply with the exiting rule.
1) A library in a packet should be in XXXPkg/Library/AAALib in general. (For example, OvmfPkg/Library)
2) If a library in a packet belongs to a feature, then it should be XXXPkg/FeatureYYY/AAALib. (For example, OvmfPkg/Csm/CsmSupportLib)
OK. I'll send a new version with the paths corrected.






Two other things to consider:

A. The blob verification by hashes just uses initially-measured memory,
and no other features of SEV. It might be useful for other Confidential
Computing implementations. But maybe if that need arises then we'll
extract it from the AmdSev folder.
[Jiewen] I think so. Because it is generic, that is why I agree that the *library class* name does not include SEV keyword.
The *library instance* should add *Sev* keyword, because the implementation does include SEV specific, such as SEV_HASH_TABLE_GUID, SEV_KERNEL_HASH_GUID.

If you want to make it for generic confidential computing, then more work should be done. For example:
1) The instance name should be BlobVerificationLibTeeLinuxModuleHash. (TEE to replace SEV. We need Linux keyword, because I don’t think it is applicable for Windows)
2) We need consider crypto agile. The current instance only consider SHA256, which is not allowed in TDX.
3) The GUID definition should be in OvmfPkg.dec, as the interface.

Since we don’t have a TEE folder yet, I prefer we defer that work.
Later, when we finish all Sev/Tdx work, we can consider create a common Tee dir or event a package. But that may happen in next year.
OK I'll keep that in mind for later. Also note that these require
corresponding changes in QEMU.





B. There were talks about reducing the number of targets, and maybe
unifying AmdSevX64 into OvmfPkgX64. If this is indeed the direction
we're going to, then there's no need to separate the code.
[Jiewen] I think we are following what Laslo suggested.
1) The OvmfPkg includes the *basic* feature at first.
2) The advanced feature is checked into SEV folder or TDX folder, at first.
3) We can revisit those advanced feature once we think they are mature.

We agree that direction, and we should follow that.
Let's keep #1 and #2 at first to finish the feature at first (in this year). Then we can see how to enhance in #3 (maybe next year).
The more we know each other, the better we can create an architecture to support TEE.
Sounds good.

Thanks,
-Dov



Thank you
Yao Jiewen



Thanks,
-Dov


Thank you
Yao Jiewen


-----Original Message-----
From: devel@edk2.groups.io <devel@edk2.groups.io> On Behalf Of Dov
Murik
Sent: Thursday, July 22, 2021 4:43 PM
To: devel@edk2.groups.io
Cc: Dov Murik <dovmurik@linux.ibm.com>; Tobin Feldman-Fitzthum
<tobin@linux.ibm.com>; Tobin Feldman-Fitzthum <tobin@ibm.com>; Jim
Cadden <jcadden@ibm.com>; James Bottomley <jejb@linux.ibm.com>;
Hubertus Franke <frankeh@us.ibm.com>; Ard Biesheuvel
<ardb+tianocore@kernel.org>; Justen, Jordan L <jordan.l.justen@intel.com>;
Ashish Kalra <ashish.kalra@amd.com>; Brijesh Singh
<brijesh.singh@amd.com>;
Erdem Aktas <erdemaktas@google.com>; Yao, Jiewen
<jiewen.yao@intel.com>;
Xu, Min M <min.m.xu@intel.com>; Tom Lendacky
<thomas.lendacky@amd.com>; Leif Lindholm <leif@nuviainc.com>; Sami
Mujawar <sami.mujawar@arm.com>
Subject: [edk2-devel] [PATCH v4 00/11] Measured SEV boot with
kernel/initrd/cmdline

BZ: https://bugzilla.tianocore.org/show_bug.cgi?id=3457

Booting with SEV prevented the loading of kernel, initrd, and kernel
command-line via QEMU fw_cfg interface because they arrive from the VMM
which is untrusted in SEV.

However, in some cases the kernel, initrd, and cmdline are not secret
but should not be modified by the host. In such a case, we want to
verify inside the trusted VM that the kernel, initrd, and cmdline are
indeed the ones expected by the Guest Owner, and only if that is the
case go on and boot them up (removing the need for grub inside OVMF in
that mode).

This patch series reserves an area in MEMFD (previously the last 1KB of
the launch secret page) which will contain the hashes of these three
blobs (kernel, initrd, cmdline), each under its own GUID entry. This
tables of hashes is populated by QEMU before launch, and encrypted as
part of the initial VM memory; this makes sure these hashes are part of
the SEV measurement (which has to be approved by the Guest Owner for
secret injection, for example). Note that populating the hashes table
requires QEMU support [1].

OVMF parses the table of hashes populated by QEMU (patch 10), and as it
reads the fw_cfg blobs from QEMU, it will verify each one against the
expected hash. This is all done inside the trusted VM context. If all
the hashes are correct, boot of the kernel is allowed to continue.

Any attempt by QEMU to modify the kernel, initrd, cmdline (including
dropping one of them), or to modify the OVMF code that verifies those
hashes, will cause the initial SEV measurement to change and therefore
will be detectable by the Guest Owner during launch before secret
injection.

Relevant part of OVMF serial log during boot with AmdSevX86 build and
QEMU with -kernel/-initrd/-append:

...
BlobVerifierLibSevHashesConstructor: Found injected hashes table in secure
location
Select Item: 0x17
Select Item: 0x8
FetchBlob: loading 7379328 bytes for "kernel"
Select Item: 0x18
Select Item: 0x11
VerifyBlob: Found GUID 4DE79437-ABD2-427F-B835-D5B172D2045B in
table
VerifyBlob: Hash comparison succeeded for "kernel"
Select Item: 0xB
FetchBlob: loading 12483878 bytes for "initrd"
Select Item: 0x12
VerifyBlob: Found GUID 44BAF731-3A2F-4BD7-9AF1-41E29169781D in table
VerifyBlob: Hash comparison succeeded for "initrd"
Select Item: 0x14
FetchBlob: loading 86 bytes for "cmdline"
Select Item: 0x15
VerifyBlob: Found GUID 97D02DD8-BD20-4C94-AA78-E7714D36AB2A in
table
VerifyBlob: Hash comparison succeeded for "cmdline"
...

The patch series is organized as follows:

1: Simple comment fix in adjacent area in the code.
2: Use GenericQemuLoadImageLib to gain one location for fw_cfg blob
fetching.
3: Allow the (previously blocked) usage of -kernel in AmdSevX64.
4-7: Add BlobVerifierLib with null implementation and use it in the correct
location in QemuKernelLoaderFsDxe.
8-9: Reserve memory for hashes table, declare this area in the reset vector.
10-11: Add the secure implementation BlobVerifierLibSevHashes and use it in
AmdSevX64 builds.

[1] https://lore.kernel.org/qemu-devel/20210624102040.2015280-1-
dovmurik@linux.ibm.com/

Code is at
https://github.com/confidential-containers-demo/edk2/tree/sev-hashes-v4

v4 changes:
- BlobVerifierSevHashes (patch 10): more comprehensive overflow tests
when parsing the SEV hashes table structure

v3: https://edk2.groups.io/g/devel/message/77955
v3 changes:
- Rename to BlobVerifierLibNull, use decimal INF_VERSION, remove unused
DebugLib reference, fix doxygen comments, add missing IN attribute
- Rename to BlobVerifierLibSevHashes, use decimal INF_VERSION, fix
doxygen comments, add missing IN attribute,
calculate buffer hash only when the guid is found in hashes table
- SecretPei: use ALIGN_VALUE to round the hob size
- Coding style fixes
- Add missing 'Ref:' in patch 1 commit message
- Fix phrasing and typos in commit messages
- Remove Cc: Laszlo from series

v2: https://edk2.groups.io/g/devel/message/77505
v2 changes:
- Use the last 1KB of the existing SEV launch secret page for hashes table
(instead of reserving a whole new MEMFD page).
- Build on top of commit cf203024745f
("OvmfPkg/GenericQemuLoadImageLib:
Read
cmdline from QemuKernelLoaderFs", 2021-06-28) to have a single location
in
which all of kernel/initrd/cmdline are fetched from QEMU.
- Use static linking of the two BlobVerifierLib implemenatations.
- Reorganize series.

v1: https://edk2.groups.io/g/devel/message/75567

Cc: Ard Biesheuvel <ardb+tianocore@kernel.org>
Cc: Jordan Justen <jordan.l.justen@intel.com>
Cc: Ashish Kalra <ashish.kalra@amd.com>
Cc: Brijesh Singh <brijesh.singh@amd.com>
Cc: Erdem Aktas <erdemaktas@google.com>
Cc: James Bottomley <jejb@linux.ibm.com>
Cc: Jiewen Yao <jiewen.yao@intel.com>
Cc: Min Xu <min.m.xu@intel.com>
Cc: Tom Lendacky <thomas.lendacky@amd.com>
Cc: Leif Lindholm <leif@nuviainc.com>
Cc: Sami Mujawar <sami.mujawar@arm.com>

---

Ard: please review patch 6 (ArmVirtPkg). Thanks.

Tom, Brijesh: I'll also send the diff for patch 10. Thanks.

---

Dov Murik (8):
OvmfPkg/AmdSev: use GenericQemuLoadImageLib in AmdSev builds
OvmfPkg: add library class BlobVerifierLib with null implementation
OvmfPkg: add BlobVerifierLibNull to DSC
ArmVirtPkg: add BlobVerifierLibNull to DSC
OvmfPkg/QemuKernelLoaderFsDxe: call VerifyBlob after fetch from fw_cfg
OvmfPkg/AmdSev/SecretPei: build hob for full page
OvmfPkg: add BlobVerifierLibSevHashes
OvmfPkg/AmdSev: Enforce hash verification of kernel blobs

James Bottomley (3):
OvmfPkg/AmdSev/SecretDxe: fix header comment to generic naming
OvmfPkg: PlatformBootManagerLibGrub: Allow executing kernel via fw_cfg
OvmfPkg/AmdSev: reserve MEMFD space for for firmware config hashes

OvmfPkg/OvmfPkg.dec | 9 +
ArmVirtPkg/ArmVirtQemu.dsc | 5 +-
ArmVirtPkg/ArmVirtQemuKernel.dsc | 5 +-
OvmfPkg/AmdSev/AmdSevX64.dsc | 9 +-
OvmfPkg/OvmfPkgIa32.dsc | 5 +-
OvmfPkg/OvmfPkgIa32X64.dsc | 5 +-
OvmfPkg/OvmfPkgX64.dsc | 5 +-
OvmfPkg/AmdSev/AmdSevX64.fdf | 5 +-
OvmfPkg/Library/BlobVerifierLib/BlobVerifierLibNull.inf | 24
+++
OvmfPkg/Library/BlobVerifierLib/BlobVerifierLibSevHashes.inf |
37
++++

OvmfPkg/Library/PlatformBootManagerLibGrub/PlatformBootManagerLibGrub.
inf | 2 +
OvmfPkg/ResetVector/ResetVector.inf | 2 +
OvmfPkg/Include/Library/BlobVerifierLib.h | 38 ++++
OvmfPkg/Library/PlatformBootManagerLibGrub/BdsPlatform.h
|
11 ++
OvmfPkg/AmdSev/SecretDxe/SecretDxe.c | 2 +-
OvmfPkg/AmdSev/SecretPei/SecretPei.c | 3 +-
OvmfPkg/Library/BlobVerifierLib/BlobVerifierNull.c | 33
++++
OvmfPkg/Library/BlobVerifierLib/BlobVerifierSevHashes.c |
199
++++++++++++++++++++
OvmfPkg/Library/PlatformBootManagerLibGrub/BdsPlatform.c
|
5 +
OvmfPkg/Library/{PlatformBootManagerLib =>
PlatformBootManagerLibGrub}/QemuKernel.c | 0
OvmfPkg/QemuKernelLoaderFsDxe/QemuKernelLoaderFsDxe.c
| 9 +
OvmfPkg/ResetVector/Ia16/ResetVectorVtf0.asm | 20
++
OvmfPkg/ResetVector/ResetVector.nasmb | 2 +
23 files changed, 425 insertions(+), 10 deletions(-)
create mode 100644 OvmfPkg/Library/BlobVerifierLib/BlobVerifierLibNull.inf
create mode 100644
OvmfPkg/Library/BlobVerifierLib/BlobVerifierLibSevHashes.inf
create mode 100644 OvmfPkg/Include/Library/BlobVerifierLib.h
create mode 100644 OvmfPkg/Library/BlobVerifierLib/BlobVerifierNull.c
create mode 100644
OvmfPkg/Library/BlobVerifierLib/BlobVerifierSevHashes.c
copy OvmfPkg/Library/{PlatformBootManagerLib =>
PlatformBootManagerLibGrub}/QemuKernel.c (100%)

--
2.25.1





James Bottomley
 

On Sun, 2021-07-25 at 10:52 +0300, Dov Murik wrote:
And I do have one question:
May I know what is criteria to put a SEV module to OvmfPkg\AmdSev
or OvmfPkg directly?

My original understanding is:
If a module is required by OvmfPkg{Ia32,Ia32X64,X64}.{dsc,fdf},
then it should be OvmfPkg.
If a module is only required by OvmfPkg\AmdSev\AmdSevX64.{dsc,fdf},
Then it should be in OvmfPkg\AmdSev.

Am I right?
I actually don't know the criteria. What you say sounds reasonable.
I'll also let James (who introduced the AmdSevX64 target) say what he
thinks.
The original reason for the AmdSev package was actually for
attestation: The only way to get attested boot using a standard VM
image for SEV and SEV-ES was to pull grub inside the measurement
envelope and have a stripped down hard failing boot path, so if the key
didn't decode the encrypted boot volume for some reason, the whole
thing would fail without revealing the injected secret. This stripped
down hard failing boot path is much easier to construct as a separate
target.

Essentially that means that lots of SEV exists outside the AmdSev
directory and things should only be in it if they're either modified to
support the encrypted volume boot path or are only required by it.
However, this ran into problems when it was decided AmdSev shouldn't
have it's own Library, so the modified boot path now lives in
OvmfPkg/Library/PlatformBootManagerLibGrub, so now it's unclear even to
me what the criteria are.

James


Yao, Jiewen
 

Hi James
"However, this ran into problems when it was decided AmdSev shouldn't have it's own Library."

I am not clear on the history. Would you please clarify why AmdSev should not have its own library?

It looks not reasonable to me. AmdSev is just a feature. A feature may have its own library. We have enough examples.

Also, the instance name "Grub" is very confusing. I compared PlatformBootManagerLib and PlatformBootManagerLibGrub. This is just a customized PlatformBootManagerLib.

For example, XEN feature removing and PIIX4 difference has nothing to do with Grub...
=================
PciWrite8 (PCI_LIB_ADDRESS (0, 1, 0, 0x60), 0x0b); // A
PciWrite8 (PCI_LIB_ADDRESS (0, 1, 0, 0x61), 0x0b); // B
PciWrite8 (PCI_LIB_ADDRESS (0, 1, 0, 0x62), 0x0a); // C
PciWrite8 (PCI_LIB_ADDRESS (0, 1, 0, 0x63), 0x0a); // D
=================

It is a big misleading. Can we move the PlatformBootManagerLibGrub To AmdSev now?

-----Original Message-----
From: James Bottomley <jejb@linux.ibm.com>
Sent: Monday, July 26, 2021 5:10 AM
To: devel@edk2.groups.io; dovmurik@linux.ibm.com; Yao, Jiewen
<jiewen.yao@intel.com>
Cc: Tobin Feldman-Fitzthum <tobin@linux.ibm.com>; Tobin Feldman-Fitzthum
<tobin@ibm.com>; Jim Cadden <jcadden@ibm.com>; Hubertus Franke
<frankeh@us.ibm.com>; Ard Biesheuvel <ardb+tianocore@kernel.org>; Justen,
Jordan L <jordan.l.justen@intel.com>; Ashish Kalra <ashish.kalra@amd.com>;
Brijesh Singh <brijesh.singh@amd.com>; Erdem Aktas
<erdemaktas@google.com>; Xu, Min M <min.m.xu@intel.com>; Tom Lendacky
<thomas.lendacky@amd.com>; Leif Lindholm <leif@nuviainc.com>; Sami
Mujawar <sami.mujawar@arm.com>
Subject: Re: [edk2-devel] [PATCH v4 00/11] Measured SEV boot with
kernel/initrd/cmdline

On Sun, 2021-07-25 at 10:52 +0300, Dov Murik wrote:
And I do have one question:
May I know what is criteria to put a SEV module to OvmfPkg\AmdSev
or OvmfPkg directly?

My original understanding is:
If a module is required by OvmfPkg{Ia32,Ia32X64,X64}.{dsc,fdf},
then it should be OvmfPkg.
If a module is only required by OvmfPkg\AmdSev\AmdSevX64.{dsc,fdf},
Then it should be in OvmfPkg\AmdSev.

Am I right?
I actually don't know the criteria. What you say sounds reasonable.
I'll also let James (who introduced the AmdSevX64 target) say what he
thinks.
The original reason for the AmdSev package was actually for
attestation: The only way to get attested boot using a standard VM
image for SEV and SEV-ES was to pull grub inside the measurement
envelope and have a stripped down hard failing boot path, so if the key
didn't decode the encrypted boot volume for some reason, the whole
thing would fail without revealing the injected secret. This stripped
down hard failing boot path is much easier to construct as a separate
target.

Essentially that means that lots of SEV exists outside the AmdSev
directory and things should only be in it if they're either modified to
support the encrypted volume boot path or are only required by it.
However, this ran into problems when it was decided AmdSev shouldn't
have it's own Library, so the modified boot path now lives in
OvmfPkg/Library/PlatformBootManagerLibGrub, so now it's unclear even to
me what the criteria are.

James


James Bottomley
 

On Mon, 2021-07-26 at 00:55 +0000, Yao, Jiewen wrote:
Hi James
"However, this ran into problems when it was decided AmdSev shouldn't
have it's own Library."

I am not clear on the history. Would you please clarify why AmdSev
should not have its own library?
The history predates me. It was already done for the Bhyve package
which also has a modified PlatformBootManagerLib when I came along with
this. However, only having Library in the top level package seems to
be a common edk2 pattern if you run a find.

It looks not reasonable to me. AmdSev is just a feature. A feature
may have its own library. We have enough examples.
We do? Running

find . -name Library -print

only turns up

./FmpDevicePkg/Test/UnitTest/Library

As not following the top level package only pattern.

Also, the instance name "Grub" is very confusing. I compared
PlatformBootManagerLib and PlatformBootManagerLibGrub. This is just a
customized PlatformBootManagerLib.
It's called Grub because it places Grub in the Fv for combined pre-
attestation. Either SEV or TDX could use this (Although TDX looks
likely not to want to).

For example, XEN feature removing and PIIX4 difference has nothing to
do with Grub...
=================
PciWrite8 (PCI_LIB_ADDRESS (0, 1, 0, 0x60), 0x0b); // A
PciWrite8 (PCI_LIB_ADDRESS (0, 1, 0, 0x61), 0x0b); // B
PciWrite8 (PCI_LIB_ADDRESS (0, 1, 0, 0x62), 0x0a); // C
PciWrite8 (PCI_LIB_ADDRESS (0, 1, 0, 0x63), 0x0a); // D
=================
It's part of the boot path stripping to make sure there's a hard
failure if Grub fails to execute. There's a Bugzilla requiring more of
this because a grub only booting platform library needs fewer
extraneous things which could constitute an attack surface for the
injected secret.

It is a big misleading. Can we move the PlatformBootManagerLibGrub To
AmdSev now?
I think you probably want to ask around older edk2 package maintainers
and see if there's any reason for this pattern, which seems to be
strongly enforced. If no-one can remember, then likely it can be
broken.

James


Yao, Jiewen
 

Hi James
You are right that initially EDKII did recommend to put a lib under Library dir.
But with more and more feature, people start realizing that it is NOT efficient way to identify a *feature*.
We changed the direction later - if we can define a feature dir, we can lib to feature dir.


I used " find . -name *Lib*.inf -print", because I cannot assume the Lib is under Library dir for feature.

I can find lots of stuff. Most of them are under Library, below is NOT in *Pkg/Library/
=============
./edk2/ArmPkg/Drivers/ArmGic/ArmGicLib.inf
./edk2/ArmPlatformPkg/MemoryInitPei/MemoryInitPeiLib.inf
./edk2/ArmPlatformPkg/PlatformPei/PlatformPeiLib.inf
./edk2/CryptoPkg/Test/UnitTest/Library/BaseCryptLib/TestBaseCryptLibHost.inf
./edk2/CryptoPkg/Test/UnitTest/Library/BaseCryptLib/TestBaseCryptLibShell.inf
./edk2/FmpDevicePkg/FmpDxe/FmpDxeLib.inf
./edk2/FmpDevicePkg/Test/UnitTest/Library/FmpDependencyLib/FmpDependencyLibUnitTestsHost.inf
./edk2/FmpDevicePkg/Test/UnitTest/Library/FmpDependencyLib/FmpDependencyLibUnitTestsUefi.inf
./edk2/MdePkg/Test/UnitTest/Library/BaseLib/BaseLibUnitTestsHost.inf
./edk2/MdePkg/Test/UnitTest/Library/BaseLib/BaseLibUnitTestsUefi.inf
./edk2/MdePkg/Test/UnitTest/Library/BaseSafeIntLib/TestBaseSafeIntLibDxe.inf
./edk2/MdePkg/Test/UnitTest/Library/BaseSafeIntLib/TestBaseSafeIntLibHost.inf
./edk2/MdePkg/Test/UnitTest/Library/BaseSafeIntLib/TestBaseSafeIntLibPei.inf
./edk2/MdePkg/Test/UnitTest/Library/BaseSafeIntLib/TestBaseSafeIntLibSmm.inf
./edk2/MdePkg/Test/UnitTest/Library/BaseSafeIntLib/TestBaseSafeIntLibUefiShell.inf
./edk2/OvmfPkg/Csm/CsmSupportLib/CsmSupportLib.inf
./edk2/OvmfPkg/Csm/LegacyBootMaintUiLib/LegacyBootMaintUiLib.inf
./edk2/OvmfPkg/Csm/LegacyBootManagerLib/LegacyBootManagerLib.inf
============

If I search on edk2-platform (https://github.com/tianocore/edk2-platforms), I can find more:
============
./Platform/Intel/CometlakeOpenBoardPkg/CometlakeURvp/Library/BaseFuncLib/BaseFuncLib.inf
./Platform/Intel/CometlakeOpenBoardPkg/CometlakeURvp/Library/BaseGpioCheckConflictLib/BaseGpioCheckConflictLib.inf
./Platform/Intel/CometlakeOpenBoardPkg/CometlakeURvp/Library/BaseGpioCheckConflictLibNull/BaseGpioCheckConflictLibNull.inf
./Platform/Intel/CometlakeOpenBoardPkg/CometlakeURvp/Library/BasePlatformHookLib/BasePlatformHookLib.inf
./Platform/Intel/CometlakeOpenBoardPkg/CometlakeURvp/Library/BoardAcpiLib/SmmBoardAcpiEnableLib.inf
./Platform/Intel/CometlakeOpenBoardPkg/CometlakeURvp/Library/BoardAcpiLib/SmmMultiBoardAcpiSupportLib.inf
./Platform/Intel/CometlakeOpenBoardPkg/CometlakeURvp/Library/BoardInitLib/PeiBoardInitPostMemLib.inf
./Platform/Intel/CometlakeOpenBoardPkg/CometlakeURvp/Library/BoardInitLib/PeiBoardInitPreMemLib.inf
./Platform/Intel/CometlakeOpenBoardPkg/CometlakeURvp/Library/BoardInitLib/PeiMultiBoardInitPostMemLib.inf
./Platform/Intel/CometlakeOpenBoardPkg/CometlakeURvp/Library/BoardInitLib/PeiMultiBoardInitPreMemLib.inf
./Platform/Intel/CometlakeOpenBoardPkg/CometlakeURvp/Library/DxePolicyBoardConfigLib/DxePolicyBoardConfigLib.inf
./Platform/Intel/CometlakeOpenBoardPkg/CometlakeURvp/Library/PeiPolicyBoardConfigLib/PeiPolicyBoardConfigLib.inf
./Platform/Intel/CometlakeOpenBoardPkg/Features/Tbt/Library/DxeTbtPolicyLib/DxeTbtPolicyLib.inf
./Platform/Intel/CometlakeOpenBoardPkg/Features/Tbt/Library/PeiDxeSmmTbtCommonLib/TbtCommonLib.inf
./Platform/Intel/CometlakeOpenBoardPkg/Features/Tbt/Library/PeiTbtPolicyLib/PeiTbtPolicyLib.inf
./Platform/Intel/CometlakeOpenBoardPkg/Features/Tbt/Library/Private/PeiDTbtInitLib/PeiDTbtInitLib.inf
./Platform/Intel/CometlakeOpenBoardPkg/FspWrapper/Library/PeiFspPolicyInitLib/PeiFspPolicyInitLib.inf
./Platform/Intel/CometlakeOpenBoardPkg/FspWrapper/Library/PeiSiliconPolicyUpdateLibFsp/PeiSiliconPolicyUpdateLibFsp.inf
./Platform/Intel/CometlakeOpenBoardPkg/Policy/Library/DxePolicyUpdateLib/DxePolicyUpdateLib.inf
./Platform/Intel/CometlakeOpenBoardPkg/Policy/Library/PeiPolicyInitLib/PeiPolicyInitLib.inf
./Platform/Intel/CometlakeOpenBoardPkg/Policy/Library/PeiPolicyUpdateLib/PeiPolicyUpdateLib.inf
./Platform/Intel/KabylakeOpenBoardPkg/Features/Tbt/Library/DxeTbtPolicyLib/DxeTbtPolicyLib.inf
./Platform/Intel/KabylakeOpenBoardPkg/Features/Tbt/Library/PeiDxeSmmTbtCommonLib/TbtCommonLib.inf
./Platform/Intel/KabylakeOpenBoardPkg/Features/Tbt/Library/PeiTbtPolicyLib/PeiTbtPolicyLib.inf
./Platform/Intel/KabylakeOpenBoardPkg/Features/Tbt/Library/Private/PeiDTbtInitLib/PeiDTbtInitLib.inf
./Platform/Intel/KabylakeOpenBoardPkg/FspWrapper/Library/PeiSiliconPolicyNotifyLib/PeiPreMemSiliconPolicyNotifyLib.inf
./Platform/Intel/KabylakeOpenBoardPkg/GalagoPro3/FspWrapper/Library/PeiSiliconPolicyNotifyLib/PeiPreMemSiliconPolicyNotifyLib.inf
./Platform/Intel/KabylakeOpenBoardPkg/GalagoPro3/FspWrapper/Library/PeiSiliconPolicyUpdateLibFsp/PeiSiliconPolicyUpdateLibFsp.inf
./Platform/Intel/KabylakeOpenBoardPkg/GalagoPro3/Library/BasePlatformHookLib/BasePlatformHookLib.inf
./Platform/Intel/KabylakeOpenBoardPkg/GalagoPro3/Library/BoardAcpiLib/DxeBoardAcpiTableLib.inf
./Platform/Intel/KabylakeOpenBoardPkg/GalagoPro3/Library/BoardAcpiLib/DxeMultiBoardAcpiSupportLib.inf
./Platform/Intel/KabylakeOpenBoardPkg/GalagoPro3/Library/BoardAcpiLib/SmmBoardAcpiEnableLib.inf
./Platform/Intel/KabylakeOpenBoardPkg/GalagoPro3/Library/BoardAcpiLib/SmmMultiBoardAcpiSupportLib.inf
./Platform/Intel/KabylakeOpenBoardPkg/GalagoPro3/Library/BoardInitLib/PeiBoardInitPostMemLib.inf
./Platform/Intel/KabylakeOpenBoardPkg/GalagoPro3/Library/BoardInitLib/PeiBoardInitPreMemLib.inf
./Platform/Intel/KabylakeOpenBoardPkg/GalagoPro3/Library/BoardInitLib/PeiMultiBoardInitPostMemLib.inf
./Platform/Intel/KabylakeOpenBoardPkg/GalagoPro3/Library/BoardInitLib/PeiMultiBoardInitPreMemLib.inf
./Platform/Intel/KabylakeOpenBoardPkg/GalagoPro3/Policy/Library/DxeSiliconPolicyUpdateLib/DxeSiliconPolicyUpdateLib.inf
./Platform/Intel/KabylakeOpenBoardPkg/KabylakeRvp3/FspWrapper/Library/PeiSiliconPolicyUpdateLibFsp/PeiSiliconPolicyUpdateLibFsp.inf
./Platform/Intel/KabylakeOpenBoardPkg/KabylakeRvp3/Library/BasePlatformHookLib/BasePlatformHookLib.inf
./Platform/Intel/KabylakeOpenBoardPkg/KabylakeRvp3/Library/BoardAcpiLib/DxeBoardAcpiTableLib.inf
./Platform/Intel/KabylakeOpenBoardPkg/KabylakeRvp3/Library/BoardAcpiLib/DxeMultiBoardAcpiSupportLib.inf
./Platform/Intel/KabylakeOpenBoardPkg/KabylakeRvp3/Library/BoardAcpiLib/SmmBoardAcpiEnableLib.inf
./Platform/Intel/KabylakeOpenBoardPkg/KabylakeRvp3/Library/BoardAcpiLib/SmmMultiBoardAcpiSupportLib.inf
./Platform/Intel/KabylakeOpenBoardPkg/KabylakeRvp3/Library/BoardInitLib/PeiBoardInitPostMemLib.inf
./Platform/Intel/KabylakeOpenBoardPkg/KabylakeRvp3/Library/BoardInitLib/PeiBoardInitPreMemLib.inf
./Platform/Intel/KabylakeOpenBoardPkg/KabylakeRvp3/Library/BoardInitLib/PeiMultiBoardInitPostMemLib.inf
./Platform/Intel/KabylakeOpenBoardPkg/KabylakeRvp3/Library/BoardInitLib/PeiMultiBoardInitPreMemLib.inf
./Platform/Intel/KabylakeOpenBoardPkg/KabylakeRvp3/Policy/Library/DxeSiliconPolicyUpdateLib/DxeSiliconPolicyUpdateLib.inf
./Platform/Intel/KabylakeOpenBoardPkg/KabylakeRvp3/Policy/Library/PeiSiliconPolicyUpdateLib/PeiSiliconPolicyUpdateLib.inf
./Platform/Intel/MinPlatformPkg/Acpi/Library/BoardAcpiEnableLibNull/BoardAcpiEnableLibNull.inf
./Platform/Intel/MinPlatformPkg/Acpi/Library/BoardAcpiTableLibNull/BoardAcpiTableLibNull.inf
./Platform/Intel/MinPlatformPkg/Acpi/Library/DxeAslUpdateLib/DxeAslUpdateLib.inf
./Platform/Intel/MinPlatformPkg/Acpi/Library/MultiBoardAcpiSupportLib/DxeMultiBoardAcpiSupportLib.inf
./Platform/Intel/MinPlatformPkg/Acpi/Library/MultiBoardAcpiSupportLib/SmmMultiBoardAcpiSupportLib.inf
./Platform/Intel/MinPlatformPkg/Bds/Library/BoardBootManagerLibNull/BoardBootManagerLibNull.inf
./Platform/Intel/MinPlatformPkg/Bds/Library/DxePlatformBootManagerLib/DxePlatformBootManagerLib.inf
./Platform/Intel/MinPlatformPkg/Flash/Library/SpiFlashCommonLibNull/SpiFlashCommonLibNull.inf
./Platform/Intel/MinPlatformPkg/FspWrapper/Library/DxeFspWrapperPlatformLib/DxeFspWrapperPlatformLib.inf
./Platform/Intel/MinPlatformPkg/FspWrapper/Library/PeiFspWrapperHobProcessLib/PeiFspWrapperHobProcessLib.inf
./Platform/Intel/MinPlatformPkg/FspWrapper/Library/PeiFspWrapperPlatformLib/PeiFspWrapperPlatformLib.inf
./Platform/Intel/MinPlatformPkg/FspWrapper/Library/SecFspWrapperPlatformSecLib/SecFspWrapperPlatformSecLib.inf
./Platform/Intel/MinPlatformPkg/Pci/Library/PciHostBridgeLibSimple/PciHostBridgeLibSimple.inf
./Platform/Intel/MinPlatformPkg/Pci/Library/PciSegmentInfoLibSimple/PciSegmentInfoLibSimple.inf
./Platform/Intel/MinPlatformPkg/PlatformInit/Library/BoardInitLibNull/BoardInitLibNull.inf
./Platform/Intel/MinPlatformPkg/PlatformInit/Library/MultiBoardInitSupportLib/DxeMultiBoardInitSupportLib.inf
./Platform/Intel/MinPlatformPkg/PlatformInit/Library/MultiBoardInitSupportLib/PeiMultiBoardInitSupportLib.inf
./Platform/Intel/MinPlatformPkg/PlatformInit/Library/PeiReportFvLib/PeiReportFvLib.inf
./Platform/Intel/MinPlatformPkg/PlatformInit/Library/ReportCpuHobLib/ReportCpuHobLib.inf
./Platform/Intel/MinPlatformPkg/PlatformInit/Library/SecBoardInitLibNull/SecBoardInitLibNull.inf
./Platform/Intel/MinPlatformPkg/PlatformInit/Library/SiliconPolicyInitLibNull/SiliconPolicyInitLibNull.inf
./Platform/Intel/MinPlatformPkg/PlatformInit/Library/SiliconPolicyUpdateLibNull/SiliconPolicyUpdateLibNull.inf
./Platform/Intel/MinPlatformPkg/Tcg/Library/PeiDxeTpmPlatformHierarchyLib/PeiDxeTpmPlatformHierarchyLib.inf
./Platform/Intel/MinPlatformPkg/Test/Library/TestPointCheckLib/DxeTestPointCheckLib.inf
./Platform/Intel/MinPlatformPkg/Test/Library/TestPointCheckLib/PeiTestPointCheckLib.inf
./Platform/Intel/MinPlatformPkg/Test/Library/TestPointCheckLib/SecTestPointCheckLib.inf
./Platform/Intel/MinPlatformPkg/Test/Library/TestPointCheckLib/SmmTestPointCheckLib.inf
./Platform/Intel/MinPlatformPkg/Test/Library/TestPointCheckLibNull/TestPointCheckLibNull.inf
./Platform/Intel/MinPlatformPkg/Test/Library/TestPointLib/DxeTestPointLib.inf
./Platform/Intel/MinPlatformPkg/Test/Library/TestPointLib/PeiTestPointLib.inf
./Platform/Intel/MinPlatformPkg/Test/Library/TestPointLib/SmmTestPointLib.inf
......
============

-----Original Message-----
From: James Bottomley <jejb@linux.ibm.com>
Sent: Monday, July 26, 2021 10:50 PM
To: Yao, Jiewen <jiewen.yao@intel.com>; devel@edk2.groups.io;
dovmurik@linux.ibm.com
Cc: Tobin Feldman-Fitzthum <tobin@linux.ibm.com>; Tobin Feldman-Fitzthum
<tobin@ibm.com>; Jim Cadden <jcadden@ibm.com>; Hubertus Franke
<frankeh@us.ibm.com>; Ard Biesheuvel <ardb+tianocore@kernel.org>; Justen,
Jordan L <jordan.l.justen@intel.com>; Ashish Kalra <ashish.kalra@amd.com>;
Brijesh Singh <brijesh.singh@amd.com>; Erdem Aktas
<erdemaktas@google.com>; Xu, Min M <min.m.xu@intel.com>; Tom Lendacky
<thomas.lendacky@amd.com>; Leif Lindholm <leif@nuviainc.com>; Sami
Mujawar <sami.mujawar@arm.com>
Subject: Re: [edk2-devel] [PATCH v4 00/11] Measured SEV boot with
kernel/initrd/cmdline

On Mon, 2021-07-26 at 00:55 +0000, Yao, Jiewen wrote:
Hi James
"However, this ran into problems when it was decided AmdSev shouldn't
have it's own Library."

I am not clear on the history. Would you please clarify why AmdSev
should not have its own library?
The history predates me. It was already done for the Bhyve package
which also has a modified PlatformBootManagerLib when I came along with
this. However, only having Library in the top level package seems to
be a common edk2 pattern if you run a find.

It looks not reasonable to me. AmdSev is just a feature. A feature
may have its own library. We have enough examples.
We do? Running

find . -name Library -print

only turns up

./FmpDevicePkg/Test/UnitTest/Library

As not following the top level package only pattern.

Also, the instance name "Grub" is very confusing. I compared
PlatformBootManagerLib and PlatformBootManagerLibGrub. This is just a
customized PlatformBootManagerLib.
It's called Grub because it places Grub in the Fv for combined pre-
attestation. Either SEV or TDX could use this (Although TDX looks
likely not to want to).

For example, XEN feature removing and PIIX4 difference has nothing to
do with Grub...
=================
PciWrite8 (PCI_LIB_ADDRESS (0, 1, 0, 0x60), 0x0b); // A
PciWrite8 (PCI_LIB_ADDRESS (0, 1, 0, 0x61), 0x0b); // B
PciWrite8 (PCI_LIB_ADDRESS (0, 1, 0, 0x62), 0x0a); // C
PciWrite8 (PCI_LIB_ADDRESS (0, 1, 0, 0x63), 0x0a); // D
=================
It's part of the boot path stripping to make sure there's a hard
failure if Grub fails to execute. There's a Bugzilla requiring more of
this because a grub only booting platform library needs fewer
extraneous things which could constitute an attack surface for the
injected secret.

It is a big misleading. Can we move the PlatformBootManagerLibGrub To
AmdSev now?
I think you probably want to ask around older edk2 package maintainers
and see if there's any reason for this pattern, which seems to be
strongly enforced. If no-one can remember, then likely it can be
broken.

James