hi Sami,

On Fri, 19 Feb 2021 at 07:36, Sughosh Ganu <sughosh.ganu@linaro.org> wrote:
>
> The following patch series adds support for using the Firmware
> Framework(FF-A) as a transport mechanism for requesting services from
> the Secure Partition Manager(SPM). This is done through a Pcd which
> can be used to enable the FF-A mechanism or to use the earlier used
> SVC calls.
>
> The patches have been pushed to my github repository[1]
>
> Ran the CI tests through the github draft pull request, and all the CI
> test pass. Ran the PatchCheck script, with no errors.
>
> Changes since V4:
> * Define all variable Pcd's under the [Pcd] section instead of
>   [FixedPcd], as suggested by Liming Gao
>
> Changes since V3:
> * Put the PcdFfaEnable under the PcdsFeatureFlag.AARCH64 section to
>   avoid build breakage for the X64 StandaloneMm builds.
> * Put the macro definitions for the SPM major and minor versions in a
>   separate patch, as suggested by Sami.
> * Separated out the declaration of the SPM major and minor version
>   macros in the earlier patch as was suggested by Sami.
> * Put the macro definitions for the SPM major and minor versions with
>   FF-A support in a separate patch, as suggested by Sami.
> * Declare the PcdFfaEnable Pcd Feature flag under FeaturePcd.AARCH64
>   to avoid build break for the X64 build of StandaloneMm.
> * Change the patch header to have the ArmPkg prefix instead of
>   StandaloneMmMmuLib as suggested by Sami.
>
> Changes since V2:
> * Added a STATIC storage class specifier for mSpmMajorVer and
>   mSpmMinorVer variables
> * Added a STATIC storage class specifier for mSpmMajorVerFfa and
>   mSpmMinorVerFfa variables
> * Add braces for if/else statements
> * Add a check for EFI_NOT_FOUND as a possible return value from
>   LocateStandaloneMmCorePeCoffData in _ModuleEntryPoint function
> * Check for the return value in Arg0 after the Direct Request call to
>   handle errors returned
> * Put a comment to reflect the fact that FFA_INTERRUPT and FFA_SUCCESS
>   response won't be expected in return to a Direct Request call to get
>   the memory attributes
> * Check for the return value in Arg0 after the Direct Request call to
>   handle errors returned
> * Put a comment to reflect the fact that FFA_INTERRUPT and FFA_SUCCESS
>   response won't be expected in return to a Direct Request call to set
>   the memory attributes
>
> Changes since V1:
> Handled review comments from Sami Mujawar
>
> [1] - https://github.com/sughoshg/edk2/tree/implement_ffa_svc_optional_v5
>
>
> Achin Gupta (7):
>   ArmPkg/IndustryStandard: Add barebones FF-A header
>   ArmPkg/ArmSvcLib: Return x4-x7 in output parameters
>   StandaloneMmPkg: Use FF-A header file in Standalone MM Core entry
>     point
>   StandaloneMmPkg: Add option to use FF-A calls for communication with
>     SPM
>   ArmPkg: Use FF-A header file in Standalone MM Arm MMU library
>   ArmPkg: Allow FF-A calls to get memory region's attributes
>   ArmPkg: Allow FF-A calls to set memory region's attributes
>
> Ilias Apalodimas (2):
>   MdeModulePkg/VariableStandaloneMm: Set PcdFlashNvStorageVariableBase
>     to Pcd
>   StandaloneMmPkg: Allow sending FFA Direct Request message to
>     StandaloneMm
>
> Sughosh Ganu (5):
>   ArmPkg: Introduce support for PcdFfaEnable
>   ArmPkg: Add macros for SPM version
>   StandaloneMmPkg: Use macros for SPM version check
>   ArmPkg: Add macros for SPM version with FF-A support enabled
>   StandaloneMmPkg: Add option to use FF-A calls for getting SPM version
>

Unfortunately, I won't be able to do a detailed review of this series,
but I have discussed these patches before with Sami off-list, and
based on that and on my earlier review of v3:

Acked-by: Ard Biesheuvel <ardb@kernel.org>

for the series, where necessary.