Re: [PATCH 2/2] OvmfPkg/PlatformPei: DENY_EXECUTE_ON_SECURITY_VIOLATION when SEV is active

[Laszlo Ersek <lersek@redhat.com>]

Hi Brijesh,

I've got three comments:

On 10/05/17 20:48, Brijesh Singh wrote:
> The following commit:
> 
> 1fea9ddb4e3f OvmfPkg: execute option ROM images regardless of Secure Boot
> 
> sets the OptionRomImageVerificationPolicy to ALWAYS_EXECUTE the expansion
> ROMs attached to the emulated PCI devices. A expansion ROM constitute
> another channel through which a cloud provider (i.e hypervisor) can
> inject a code in guest boot flow to compromise it.
> 
> When SEV is enabled, the bios code has been verified by the guest owner
> via the SEV guest launch sequence before its executed. When secure boot,
> is enabled, lets make sure that we do not allow guest bios to execute a
> code which is not signed by the guest owner.
> 
> Fixes: https://bugzilla.tianocore.org/show_bug.cgi?id=728
> Cc: Chao Zhang <chao.b.zhang@intel.com>
> Cc: Jordan Justen <jordan.l.justen@intel.com>
> Cc: Laszlo Ersek <lersek@redhat.com>
> Cc: Tom Lendacky <thomas.lendacky@amd.com>
> Signed-off-by: Brijesh Singh <brijesh.singh@amd.com>
> ---
>  OvmfPkg/OvmfPkgIa32X64.dsc          | 9 +++++----
>  OvmfPkg/OvmfPkgX64.dsc              | 9 +++++----
>  OvmfPkg/PlatformPei/PlatformPei.inf | 2 ++
>  OvmfPkg/PlatformPei/AmdSev.c        | 7 +++++++
>  4 files changed, 19 insertions(+), 8 deletions(-)
> 
> diff --git a/OvmfPkg/OvmfPkgIa32X64.dsc b/OvmfPkg/OvmfPkgIa32X64.dsc
> index 7f9220ccb90a..4bcbddb95768 100644
> --- a/OvmfPkg/OvmfPkgIa32X64.dsc
> +++ b/OvmfPkg/OvmfPkgIa32X64.dsc
> @@ -489,10 +489,6 @@
>    gUefiCpuPkgTokenSpaceGuid.PcdCpuSmmStackSize|0x4000
>  !endif
>  
> -!if $(SECURE_BOOT_ENABLE) == TRUE
> -  gEfiSecurityPkgTokenSpaceGuid.PcdOptionRomImageVerificationPolicy|0x00
> -!endif
> -
>    # IRQs 5, 9, 10, 11 are level-triggered
>    gPcAtChipsetPkgTokenSpaceGuid.Pcd8259LegacyModeEdgeLevel|0x0E20
>  
> @@ -552,6 +548,11 @@
>    gUefiCpuPkgTokenSpaceGuid.PcdCpuSmmApSyncTimeout|100000
>  !endif
>  
> +!if $(SECURE_BOOT_ENABLE) == TRUE
> +  gEfiSecurityPkgTokenSpaceGuid.PcdOptionRomImageVerificationPolicy|0x00
> +!endif
> +
> +
>  ################################################################################
>  #
>  # Components Section - list of all EDK II Modules needed by this Platform.

(1) It's hard to see what section of the DSC file the lines are moved
from, and to what section.

Git can be configured so that it display the section names in the diff
hunk header (following the "@@" marks). Please refer to the following:

https://github.com/tianocore/tianocore.github.io/wiki/Laszlo's-unkempt-git-guide-for-edk2-contributors-and-maintainers#contrib-05

https://github.com/tianocore/tianocore.github.io/wiki/Laszlo's-unkempt-git-guide-for-edk2-contributors-and-maintainers#contrib-09

In particular, see the "diff.ini.xfuncname" config option, and the
matching "diff=ini" attributes.

(This remark does not affect the patch, only how it is formatted.
Showing the DSC section names in the diff hunk headers helps quite a bit
with review.)


> diff --git a/OvmfPkg/OvmfPkgX64.dsc b/OvmfPkg/OvmfPkgX64.dsc
> index 36c60fc19c40..e52a3bd4db9b 100644
> --- a/OvmfPkg/OvmfPkgX64.dsc
> +++ b/OvmfPkg/OvmfPkgX64.dsc
> @@ -488,10 +488,6 @@
>    gUefiCpuPkgTokenSpaceGuid.PcdCpuSmmStackSize|0x4000
>  !endif
>  
> -!if $(SECURE_BOOT_ENABLE) == TRUE
> -  gEfiSecurityPkgTokenSpaceGuid.PcdOptionRomImageVerificationPolicy|0x00
> -!endif
> -
>    # IRQs 5, 9, 10, 11 are level-triggered
>    gPcAtChipsetPkgTokenSpaceGuid.Pcd8259LegacyModeEdgeLevel|0x0E20
>  
> @@ -551,6 +547,11 @@
>    gUefiCpuPkgTokenSpaceGuid.PcdCpuSmmApSyncTimeout|100000
>  !endif
>  
> +!if $(SECURE_BOOT_ENABLE) == TRUE
> +  gEfiSecurityPkgTokenSpaceGuid.PcdOptionRomImageVerificationPolicy|0x00
> +!endif
> +
> +
>  ################################################################################
>  #
>  # Components Section - list of all EDK II Modules needed by this Platform.

(2) Please apply the same code movement to "OvmfPkg/OvmfPkgIa32.dsc".

As written, the PcdSet32S() macro invocation in PlatformPei below will
not even compile for OvmfPkgIa32. Namely, the PCD will remain
fixed-at-build, and because of that, the replacement text of the
PcdSet32S() macro, i.e.

  _PCD_SET_MODE_32_S_##TokenName    ((Value))

from

  MdePkg/Include/Library/PcdLib.h

will not have any *further* replacement text, from "AutoGen.h".


> diff --git a/OvmfPkg/PlatformPei/PlatformPei.inf b/OvmfPkg/PlatformPei/PlatformPei.inf
> index 16a8db7b0bd2..de7434d93dc0 100644
> --- a/OvmfPkg/PlatformPei/PlatformPei.inf
> +++ b/OvmfPkg/PlatformPei/PlatformPei.inf
> @@ -41,6 +41,7 @@
>    IntelFrameworkModulePkg/IntelFrameworkModulePkg.dec
>    MdePkg/MdePkg.dec
>    MdeModulePkg/MdeModulePkg.dec
> +  SecurityPkg/SecurityPkg.dec
>    UefiCpuPkg/UefiCpuPkg.dec
>    OvmfPkg/OvmfPkg.dec
>  
> @@ -96,6 +97,7 @@
>    gEfiMdeModulePkgTokenSpaceGuid.PcdPropertiesTableEnable
>    gEfiMdeModulePkgTokenSpaceGuid.PcdAcpiS3Enable
>    gEfiMdeModulePkgTokenSpaceGuid.PcdPteMemoryEncryptionAddressOrMask
> +  gEfiSecurityPkgTokenSpaceGuid.PcdOptionRomImageVerificationPolicy
>    gUefiCpuPkgTokenSpaceGuid.PcdCpuLocalApicBaseAddress
>    gUefiCpuPkgTokenSpaceGuid.PcdCpuMaxLogicalProcessorNumber
>    gUefiCpuPkgTokenSpaceGuid.PcdCpuApInitTimeOutInMicroSeconds
> diff --git a/OvmfPkg/PlatformPei/AmdSev.c b/OvmfPkg/PlatformPei/AmdSev.c
> index 26f7c3fdbb13..1539e5b5cdce 100644
> --- a/OvmfPkg/PlatformPei/AmdSev.c
> +++ b/OvmfPkg/PlatformPei/AmdSev.c
> @@ -59,4 +59,11 @@ AmdSevInitialize (
>    ASSERT_RETURN_ERROR (PcdStatus);
>  
>    DEBUG ((DEBUG_INFO, "SEV is enabled (mask 0x%lx)\n", EncryptionMask));
> +
> +  //
> +  // Set Pcd to Deny the execution of option ROM when security
> +  // violation.
> +  //
> +  PcdStatus = PcdSet32S (PcdOptionRomImageVerificationPolicy, 0x4);
> +  ASSERT_RETURN_ERROR (PcdStatus);
>  }
> 

(3) I find it sub-optimal that the

  DENY_EXECUTE_ON_SECURITY_VIOLATION

macro, from

  SecurityPkg/Library/DxeImageVerificationLib/DxeImageVerificationLib.h

is internal to that library instance, and isn't exposed in some library
class header.

Anyway, I guess we can't do much about it here, and at least
SecurityPkg.dec spells out the meanings of the values. So this hunk
looks fine to me.

Thanks!
Laszlo