On 7/31/26 19:22, Bjoern A. Zeeb wrote:
On Wed, 29 Jul 2026, John Baldwin wrote:
The branch main has been updated by jhb:
URL:
https://cgit.FreeBSD.org/src/commit/?id=ea44c37d0bd15e796c85f84b57c211ac220c550c
commit ea44c37d0bd15e796c85f84b57c211ac220c550c
Author: John Baldwin <[email protected]>
AuthorDate: 2026-07-29 18:42:19 +0000
Commit: John Baldwin <[email protected]>
CommitDate: 2026-07-29 18:42:19 +0000
OFED: Add generic function to extract IB speed from netdev
This is really confusing to me and I point this out given one of the many
"this is not wireles" LinuxKPI parts I had to do in the past was to
"divorce" netdev being an ifnet as one was aliased to the other in the
past not allowing us to have the Linux native fields we nowhere used and
had edited away (all before my time).
And I was like "we have no support for any of this in LinuxKPI -- how
can this work" until I saw the if_t.
We should really start calling ifnet ifnet and no longer netdev unless
we make netdev work on top of ifnet and it is a net_device again.
The title was borrowed from the upstream Linux commit. It's true that
s/struct netdev/if_t/ is one of the changes I have to make every time I
backport an RDMA/OFED commit which is pretty annoying, but this commit
continues that trend. OTOH, renaming the local 'netdev_speed' variable
in the function would just be a gratuitous diff relative to upstream that
would make future merges harder.
--
John Baldwin
