]> git.proxmox.com Git - mirror_edk2.git/commit
ArmVirtPkg/PlatformHasAcpiDtDxe: don't expose DT if QEMU provides ACPI
authorLaszlo Ersek <lersek@redhat.com>
Fri, 17 Mar 2017 16:50:24 +0000 (17:50 +0100)
committerLaszlo Ersek <lersek@redhat.com>
Tue, 28 Mar 2017 12:18:47 +0000 (14:18 +0200)
commit110316a995ed09078028044acce49d3db32245d2
tree75bd570f0501708010cac60df1d49dfd85ad27c7
parent51b09a2c50c60eb5dd3996b97f375eb596c5cc1d
ArmVirtPkg/PlatformHasAcpiDtDxe: don't expose DT if QEMU provides ACPI

This will let QEMU's "-no-acpi" option exclusively expose DT vs. ACPI to
the guest. Showing both is never needed (it is actually detrimental to the
adoption of standards, such as SBSA / SBBR).

* Without "-no-acpi", the firmware logs (from PlatformHasAcpiDtDxe)

> Found FwCfg @ 0x9020008/0x9020000
> Found FwCfg DMA @ 0x9020010
> InstallProtocolInterface: [EdkiiPlatformHasAcpi] 0

plus the usual messages. Later the guest kernel logs

> [    0.000000] efi:  SMBIOS 3.0=0x13bdb0000  ACPI 2.0=0x138440000
>                      MEMATTR=0x13a675018

before it lists the ACPI tables one by one.

In addition, in the guest, the "/sys/firmware/devicetree/*" shell pattern
matches no files, while the "/sys/firmware/acpi/tables/*" pattern matches
the ACPI tables.

* With "-no-acpi", the firmware logs:

> PlatformHasAcpiDtDxe | Found FwCfg @ 0x9020008/0x9020000
> PlatformHasAcpiDtDxe | Found FwCfg DMA @ 0x9020010
> PlatformHasAcpiDtDxe | InstallProtocolInterface:
> PlatformHasAcpiDtDxe | [EdkiiPlatformHasDeviceTree] 0
> FdtClientDxe         | OnPlatformHasDeviceTree: exposing DTB @
> FdtClientDxe         | 0x13FFBF000 to OS
> ...
> DXE_CORE             | Driver [AcpiTableDxe] was discovered but not
> DXE_CORE             | loaded!!
> DXE_CORE             | Driver [QemuFwCfgAcpiPlatform] was discovered but
> DXE_CORE             | not loaded!!
> ...
> RamDiskDxe           | RamDiskAcpiCheck: Cannot locate the EFI ACPI
> RamDiskDxe           | Table Protocol, unable to publish RAM disks to
> RamDiskDxe           | NFIT.

(BootGraphicsResourceTableDxe's ReadyToBoot callback --
InstallBootGraphicsResourceTable() -- handles the lack of
EFI_ACPI_TABLE_PROTOCOL silently.) Later the guest kernel logs

> [    0.000000] efi:  SMBIOS 3.0=0x13bdb0000  MEMATTR=0x138caa018

In addition, in the guest, the "/sys/firmware/devicetree/*" shell pattern
matches the directory "/sys/firmware/devicetree/base", which contains a
large number of DT nodes, while the "/sys/firmware/acpi/tables/*" pattern
matches no files.

Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Leif Lindholm <leif.lindholm@linaro.org>
Ref: https://bugzilla.redhat.com/show_bug.cgi?id=1430262
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
ArmVirtPkg/PlatformHasAcpiDtDxe/PlatformHasAcpiDtDxe.c
ArmVirtPkg/PlatformHasAcpiDtDxe/PlatformHasAcpiDtDxe.inf