Re: [edk2-devel] [PATCH v2 2/5] MdeMoudlePkg: introduce new PCD for Acpi/rsdp

["Laszlo Ersek" <lersek@redhat.com>]

On 05/17/21 08:50, Jianyong Wu wrote:
> As there is lack of a machnism in Cloud Hypervisor to pass the base
> address of Rsdp in Acpi, so a PCD varialbe is introduced here to
> feed it.
> 
> Cc: Hao A Wu <hao.a.wu@intel.com>
> Cc: Jian J Wang <jian.j.wang@intel.com>
> Signed-off-by: Jianyong Wu <jianyong.wu@arm.com>
> ---
>  MdeModulePkg/MdeModulePkg.dec | 7 +++++++
>  1 file changed, 7 insertions(+)
> 
> diff --git a/MdeModulePkg/MdeModulePkg.dec b/MdeModulePkg/MdeModulePkg.dec
> index 148395511034..4c8baac35a9e 100644
> --- a/MdeModulePkg/MdeModulePkg.dec
> +++ b/MdeModulePkg/MdeModulePkg.dec
> @@ -910,6 +910,13 @@ [PcdsFixedAtBuild]
>    # @Expression 0x80000001 | (gEfiMdeModulePkgTokenSpaceGuid.PcdLoadModuleAtFixAddressEnable == 0xFFFFFFFFFFFFFFFF || gEfiMdeModulePkgTokenSpaceGuid.PcdLoadModuleAtFixAddressEnable <= 0x0FFFFFFFFFFFFFFF)
>    gEfiMdeModulePkgTokenSpaceGuid.PcdLoadModuleAtFixAddressEnable|0|UINT64|0x30001015
>  
> +  ##
> +  # This is the physical address of rsdp which is the core struct of Acpi.
> +  # Some hypervisor may has no way to pass rsdp address to the guest, so a PCD
> +  # is worth for those.
> +  #
> +  gEfiMdeModulePkgTokenSpaceGuid.PcdAcpiRsdpBaseAddress|0x0|UINT64|0x30001056
> +
>    ## Progress Code for OS Loader LoadImage start.<BR><BR>
>    #  PROGRESS_CODE_OS_LOADER_LOAD   = (EFI_SOFTWARE_DXE_BS_DRIVER | (EFI_OEM_SPECIFIC | 0x00000000)) = 0x03058000<BR>
>    # @Prompt Progress Code for OS Loader LoadImage start.
> 

This PCD is not useful enough to be placed in MdeModulePkg -- it is only
used in the next two patches, which are for ArmVirtPkg.

(1) Therefore, please add the PCD to the "ArmVirtPkg.dec" file.

(2) The PCD should arguably refer to "CloudHv" in the name.

(3) In my opinion, this patch (once reimplemented for ArmVirtPkg.dec)
should be squashed into the CloudHvAcpiPlatformDxe patch. The PCD is
being introduced *for* CloudHvAcpiPlatformDxe, and *only* for that
driver. In such cases, we usually keep the DEC modifications in the same
patch as the driver addition, assuming the PCD goes in the same package
as the driver.

(4) "some hypervisor" in the DEC comment is bogus. Please be as explicit
about the use case as possible.

Thanks
Laszlo