On 12/1/23 10:57, Akihiko Odaki wrote:
> igb is a family of Intel's gigabit ethernet controllers. This series implements
> 82576 emulation in particular. You can see the last patch for the documentation.
>
> Note that there is another effort to bring 82576 emulation. This series was
> developed independently by Sriram Yagnaraman.
> https://lists.gnu.org/archive/html/qemu-devel/2022-12/msg04670.html
>
> Patch 1 - 16 are general improvements for e1000 and e1000e.
> Patch 17 - 18 are general improvements for e1000e test code.
> Patch 19 - 21 makes necessary modifications to existing files.
> Patch 22 starts off implementing igb emulation by copying e1000e code.
> Patch 23 renames things so that it won't collide with e1000e.
> Patch 24 makes building igb possible.
> Patch 25 actually transforms e1000e emulation code into igb emulation.
> Patch 26 - 27 makes modifications necessary for tests to existing files.
> Patch 28 copies e1000e test code.
> Patch 29 transforms e1000e test code into igb test code.
> Patch 30 adds ethtool test automation.
> Patch 31 adds the documentation.
>
> The main reason why this series is so huge is that the early part of this series
> includes general improvements for e1000e. They are placed before copying e1000e
> code so we won't need to duplicate those changes for both of e1000e and igb code
> later. As their utility do not depend on the igb implementation, they can be
> merged earlier if necessary.
You could post patches 1-23 as "e1000x cleanups (preliminary for IGB)"
then post the rest, using the 'Based-on:' tag referring to the first
series cover letter.
I remember looking at the various e1000x cleanups independently, all
packed in the same series makes review context-switching cheaper for
me, so thanks.
Not really related to IGB but since you touched MDIO/MII/PHY files,
it would be great if we unify the MDIO as a qbus and the various PHYs
as qdevs, so boards could use any/multiple PHYs. We had 2 or 3 attempts
to do that in the past but none got merged.