Am 2019-12-11 14:04, schrieb Alexandru Marginean:
On 12/10/2019 11:47 PM, Michael Walle wrote:
Am 2019-12-10 15:55, schrieb Alex Marginean:
Passes on the primary address used by u-boot to Linux. The code does
a DT
fix-up for ENETC PFs and sets the primary MAC address in IERB. The
address
in IERB is restored on ENETC PCI functions at FLR.
I don't get the reason why this is done in a proprietary way. What is
the
difference between any other network interface whose hardware address
is
set up in the generic u-boot code.
By proprietary do you mean IERB?
I meant the fdt fixups for mac-address/local-mac-address which is not
handled in the generic u-boot code, see my previous mail.
Because if it would, then there would be no need for the board specific
code below (just the one function call actually).
Network cards normally have a ROM which comes preset from the factory
with default MAC addresses. At run-time drivers can use the default
address or replace it. The MAC address in IERB is the default MAC
address for ENETC interfaces at run-time, this address is available to
the driver/stack after issuing an FLR and in the absence of a DT node
for the PCI function.
We can pass MAC addresses to Linux through DT alone, but that's more
of a hassle if Linux decides to assign the PCI function to a guest or
user-space app. Using the IERB values doesn't require any fix-up for
guest DTs.
Understood.
IERB is not actually a ROM though and we need U-Boot to set factory
MAC addresses into IERB some time before Linux boots.
Also, shouldn't write_hwaddr callback be implemented instead of the
enetc_set_ierb_primary_mac()?
OK, makes sense, I will move the code there.
-michael
