[edk2-devel] [RFC PATCH 0/7] OVMF: Disable the TPM2 platform hierarchy

Yao, Jiewen jiewen.yao at intel.com
Sun Aug 8 00:31:58 UTC 2021


Hi Marvin
There are many factors impacting where to put the feature code - EDKII or EDKII-platform/Feature.
I can list some of them:
If it is a basic feature v.s. an advanced feature.
If it is a mature production feature v.s. immature feature.
If it is a generic core feature v.s. a platform specific feature.
......


The position can be changed. For example, we can put a feature to EDKII-platform at first. Later, if everyone wants this feature in a production, we can move it to EDKII.

For Ext4, I think it had better be in EDKII-platform at first, because it seems to be an advanced feature. I am not sure how many production need it. And I am not sure the maturity level.

For TPM Hierarchy, we put it to EDKII-platform at first, because we think the platform hierarchy management is platform specific behavior. Each platform may choose their own way to do that. But if all people think the example in EDKII-platform can be used in their production. I think it is OK to move to SecurityPkg.

Thank you
Yao Jiewen

> -----Original Message-----
> From: devel at edk2.groups.io <devel at edk2.groups.io> On Behalf Of Marvin
> Häuser
> Sent: Saturday, August 7, 2021 6:22 PM
> To: devel at edk2.groups.io; Yao, Jiewen <jiewen.yao at intel.com>; Stefan Berger
> <stefanb at linux.vnet.ibm.com>
> Cc: marcandre.lureau at redhat.com; lersek at redhat.com;
> dick_wilkins at phoenix.com
> Subject: Re: [edk2-devel] [RFC PATCH 0/7] OVMF: Disable the TPM2 platform
> hierarchy
> 
> Hey Jiewen,
> Hey Stefan,
> 
> On 07.08.21 08:00, Yao, Jiewen wrote:
> > Hi Stefan
> > It seems this patch series is not a production fix. It is more like a prototype, my
> personally feeling.
> >
> > A common issue in patch 2, 3, 4, 5, is that using "comment" to remove the
> code. Please remove the unnecessary code directly without // or /**/ in C, and #
> in INF.
> >
> > For patch 1, if you want to move the code to SecurityPkg, that is fine. Please
> move the whole driver their and you should not remove and code by comment.
> Please fix the issue to make it pass build, instead of commenting the code like
> work-around.
> > Otherwise, you may copy the module to OvmfPkg. Then you can modify it as
> you need.
> 
> Maybe there should be stricter policies about what code goes where, and
> duplication should be outright banned? My PE/COFF loader proposal merged
> at least 4 copies of the Authenticode hashing and 3 copies of a record
> construction algorithm together. There have been other code
> centralisations, such as certificate iteration. If code exists twice, it
> needs to me maintained twice, and in reality this does not happen.
> 
> Regarding what goes where, e.g. FatPkg is in edk2, but the new EXT4
> driver was recommended to be located in edk2-platforms. I know there are
> a bunch of TPM-related things (e.g. Image measuring) in SecurityPkg, and
> as someone with no expertise around TPM or the edk2-platforms design, I
> would never even have thought to look in edk2-platforms for such code.
> And especially for library code edk2-platforms seems to be an
> unfortunate location, as edk2 packages can never depend on them.
> 
> > Please also merge 2, 3, 4 into 1. I don’t think we want a broken patch in 1,
> then add fix in 2, 3, 4.
> 
> I think for this initial inspection this was actually a good thing. The
> separation and the patches after 1 signal to me that there have been no
> functional changes to the library at all. Probably the best idea is to
> just move it to SecurityPkg entirely, including the PCDs, and update
> MinPlatformPkg to consume it from there instead?
> 
> Best regards,
> Marvin
> 
> >
> > Thank you
> > Yao Jiewen
> >
> >> -----Original Message-----
> >> From: Stefan Berger <stefanb at linux.vnet.ibm.com>
> >> Sent: Friday, August 6, 2021 11:33 PM
> >> To: devel at edk2.groups.io; Yao, Jiewen <jiewen.yao at intel.com>
> >> Cc: marcandre.lureau at redhat.com; lersek at redhat.com;
> >> dick_wilkins at phoenix.com; Stefan Berger <stefanb at linux.vnet.ibm.com>
> >> Subject: [RFC PATCH 0/7] OVMF: Disable the TPM2 platform hierarchy
> >>
> >> This series imports code from the edk2-platforms project related to
> >> changing the password of the TPM2 platform hierarchy and uses it to
> >> disable the TPM2 platform hierarchy in OVMF. It addresses the OVMF
> >> aspects of the following bugs:
> >>
> >> https://bugzilla.tianocore.org/show_bug.cgi?id=3510
> >> https://bugzilla.tianocore.org/show_bug.cgi?id=3499
> >>
> >> There's no doubt that my struggles with the build system and handling
> >> of dependencies are visible in this series. Quite a few aspects of
> >> getting things right are more or less guesswork and I am often not sure
> >> what the correct way of doing things are. If 'you' wanted to fix
> >> things up and repost it, please go ahead...
> >>
> >> Stefan
> >>
> >> Stefan Berger (7):
> >>    SecurityPkg/TPM: Import PeiDxeTpmPlatformHierarchyLib.c from
> >>      edk2-platforms
> >>    SecruityPkg/TPM: Disable dependency on MinPlatformPkg
> >>    SecurityPkg/TPM: Disable PcdGetBool (PcdRandomizePlatformHierarchy)
> >>    SecurityPkg/TPM: Disable a Pcd
> >>    SecurityPkg/TPM: Add a NULL implementation of
> >>      PeiDxeTpmPlatformHierarchyLib
> >>    OVMF: Reference new classes in the build system for compilation
> >>    OVMF: Disable the TPM2 platform hierarchy
> >>
> >>   OvmfPkg/AmdSev/AmdSevX64.dsc                  |   3 +
> >>   .../PlatformBootManagerLib/BdsPlatform.c      |   6 +
> >>   .../PlatformBootManagerLib.inf                |   1 +
> >>   .../PlatformBootManagerLibBhyve/BdsPlatform.c |   6 +
> >>   .../PlatformBootManagerLibGrub/BdsPlatform.c  |   6 +
> >>   OvmfPkg/OvmfPkgIa32.dsc                       |   3 +
> >>   OvmfPkg/OvmfPkgIa32X64.dsc                    |   3 +
> >>   OvmfPkg/OvmfPkgX64.dsc                        |   3 +
> >>   .../Include/Library/TpmPlatformHierarchyLib.h |  27 ++
> >>   .../PeiDxeTpmPlatformHierarchyLib.c           | 266 ++++++++++++++++++
> >>   .../PeiDxeTpmPlatformHierarchyLib.inf         |  46 +++
> >>   .../PeiDxeTpmPlatformHierarchyLib.c           |  23 ++
> >>   .../PeiDxeTpmPlatformHierarchyLib.inf         |  39 +++
> >>   13 files changed, 432 insertions(+)
> >>   create mode 100644
> SecurityPkg/Include/Library/TpmPlatformHierarchyLib.h
> >>   create mode 100644
> >>
> SecurityPkg/Library/PeiDxeTpmPlatformHierarchyLib/PeiDxeTpmPlatformHierar
> >> chyLib.c
> >>   create mode 100644
> >>
> SecurityPkg/Library/PeiDxeTpmPlatformHierarchyLib/PeiDxeTpmPlatformHierar
> >> chyLib.inf
> >>   create mode 100644
> >>
> SecurityPkg/Library/PeiDxeTpmPlatformHierarchyLibNull/PeiDxeTpmPlatformHi
> >> erarchyLib.c
> >>   create mode 100644
> >>
> SecurityPkg/Library/PeiDxeTpmPlatformHierarchyLibNull/PeiDxeTpmPlatformHi
> >> erarchyLib.inf
> >>
> >> --
> >> 2.31.1
> >
> >
> >
> >
> >
> 
> 
> 
> 
> 



-=-=-=-=-=-=-=-=-=-=-=-
Groups.io Links: You receive all messages sent to this group.
View/Reply Online (#78864): https://edk2.groups.io/g/devel/message/78864
Mute This Topic: https://groups.io/mt/84712022/1813853
Group Owner: devel+owner at edk2.groups.io
Unsubscribe: https://edk2.groups.io/g/devel/unsub [edk2-devel-archive at redhat.com]
-=-=-=-=-=-=-=-=-=-=-=-






More information about the edk2-devel-archive mailing list