Re: git: bb50bb45f5f6 - main - vtnet: move offload functions to virtio_net.h

From: Harry Schmalzbauer <freebsd_at_omnilan.de>
Date: Mon, 02 Nov 2026 15:26:08 UTC
On 2026-11-02 15:57, Pouria Mousavizadeh Tehrani wrote:
> Add Faraz (kfv@) to the loop.
> 
> On 11/1/26 23:55, Michael Tuexen wrote:
>>> On 1. Nov 2026, at 13:35, Harry Schmalzbauer <freebsd@omnilan.de> wrote:
>>>
>>> On 2026-08-27 14:44, Michael Tuexen wrote:
>>>> The branch main has been updated by tuexen:
>>>>
>>>> URL: https://cgit.FreeBSD.org/src/commit/? 
>>>> id=bb50bb45f5f650873b202e644bfcd9f2992cdace
>>>>
>>>> commit bb50bb45f5f650873b202e644bfcd9f2992cdace
>>>> Author:     Timo Völker <timo.voelker@fh-muenster.de>
>>>> AuthorDate: 2026-08-27 03:09:55 +0000
>>>> Commit:     Michael Tuexen <tuexen@FreeBSD.org>
>>>> CommitDate: 2026-08-27 03:15:17 +0000
>>>>
>>>>      vtnet: move offload functions to virtio_net.h
>>>>           Move the functions vtnet_rxq_csum() and 
>>>> vtnet_txq_offload() and the
>>>>      subfunctions they call from if_vtnet.c to virtio_net.h. This 
>>>> allows
>>>>      us to call these functions from if_tuntap.c and if_ptnet.c.
>>>>      virtio_net.h already contained a copy of these functions, but a 
>>>> copy
>>>>      of an outdated version. The functions evolved in if_vtnet.c.
>>>>      In if_vtnet.c, the copy has never been used because it increments
>>>>      counters in their own functions.
>>>>      This patch removes the outdated copy from virtio_net.h and 
>>>> moves the
>>>>      new version of the functions from if_vtnet.c to virtio_net.h.
>>>>      if_tuntap.c, if_ptnet.c, and if_vtnet.c just call these functions,
>>>>      and if_vtnet.c increments its counters depending on the return 
>>>> value.
>>>>           Reviewed by:            tuexen
>>>>      MFC after:              1 month
>>>>      MFC to:                 stable/15
>>>>      Differential Revision:  https://reviews.freebsd.org/D57299
>>>
>>>
>>> Hello,
>>>
>>> do you still plan to MFC this?
>>> Along with c1fa70c6e9f6 - main - vtnet: offer VIRTIO_NET_F_GUEST_CSUM 
>>> probably?
>> That was originally that plan, but I think trying to do this ended
>> up in several merge conflicts.
>>
>> Do you need these changes? Then I can try harder than I did initially.
No, thanks, I'm currently happy with vtnet as is and I can't test in 
high-load environments currently.

I just wanted to send a reminder in case this was lost.
And maybe to avoid more diverging code base causing future extensions to 
be even harder to merge...

Best,
-harry