On Fri, Dec 06, 2019 at 11:02:17 +0000, Ard Biesheuvel wrote:
> The platform DXE driver of the Juno platform enumerates the PCIe explicitly
> by calling ConnectController on the PCIe root bridge, in order to ensure
> that all PCI I/O protocols have been instantiated when we try to locate the
> one describing the Marvel Yukon NIC, so that we can program its MAC address.
>
> Taking control of core firmware behavior like this has already caused
> confusion when reasoning about why and when PCIe option ROMs get dispatched,
> and is likely to cause more confusion down the road.
>
> So let's fix this, by triggering the MAC programming via a protocol notify
> on the PCI I/O protocol.
>
> This has been build tested only at the moment.
>
> Cc: Sami Mujawar <sami.muja...@arm.com>
> Cc: Laszlo Ersek <ler...@redhat.com>
> Cc: Leif Lindholm <leif.lindh...@linaro.org>
This looks sensible to me (and way more can-of-wormish than I would
have expected) - so:
Acked-by: Leif Lindholm <leif.lindh...@linaro.org>
*but* we need tested-by from somewhere before merging 2/2.
1/2 could go in whenever.
/
Leif
> Ard Biesheuvel (2):
> Platform/ARM/ArmJunoDxe: make mAcpiRegistration STATIC
> Platform/ARM/ArmJunoDxe: use PciIo protocol notify to program the MAC
>
> .../JunoPkg/Drivers/ArmJunoDxe/ArmJunoDxe.c | 163 ++++--------------
> .../JunoPkg/Drivers/ArmJunoDxe/ArmJunoDxe.inf | 1 -
> 2 files changed, 31 insertions(+), 133 deletions(-)
>
> --
> 2.17.1
>
-=-=-=-=-=-=-=-=-=-=-=-
Groups.io Links: You receive all messages sent to this group.
View/Reply Online (#51833): https://edk2.groups.io/g/devel/message/51833
Mute This Topic: https://groups.io/mt/67467391/21656
Group Owner: devel+ow...@edk2.groups.io
Unsubscribe: https://edk2.groups.io/g/devel/unsub [arch...@mail-archive.com]
-=-=-=-=-=-=-=-=-=-=-=-
