Re: [PATCH 2/3] RDMA/hfi1, rdmavt: open-code rvt_set_ibdev_name()

[Leon Romanovsky]

On Tue, Mar 24, 2026 at 08:27:45AM +0100, Arnd Bergmann wrote:
> On Mon, Mar 23, 2026, at 22:47, Dennis Dalessandro wrote:
> > On 3/23/26 7:01 AM, Leon Romanovsky wrote:
> >> On Mon, Mar 23, 2026 at 09:48:59AM +0100, Arnd Bergmann wrote:
> >>> - The use of INFINIBAND_RDMAVT seems unnecessary: right now
> >>>    this is only used by hfi1, now shared with hfi2 but later to
> >>>    be exclusive to the latter. Since it is unlikely to ever
> >>>    be used by another driver again, this may be a good time
> >>>    to drop the abstraction again and integrate it all into
> >>>    hfi2, with the old version getting dropped along with hfi1.
> >> 
> >> The best approach is to drop rdmavt as well, since hfi2 is expected to
> >> align with the other drivers in drivers/infiniband/hw.
> >> 
> >> Dennis, is this feasible?
> >
> > Feasible yes. I'd like to get hfi2 crossed off the list and in the tree 
> > first though. Then come back and do that. I'd like to do more than just 
> > plop rdmavt inside hfi2 and call it a day. There is a lot of code 
> > cleanup/simplification that we can do.
> 
> Does rdmavt have its own user-space ABI? If there is anything that
> you'd want to change, this might be the one chance to do it, otherwise
> I see nothing wrong with integrating it only after hfi1 is gone.

rdmavt has no user‑space exposure. It is a leftover from an old and never
fully realized idea to provide a shim and common layer for RDMA drivers
that require software handling in their data path.

Thanks

> 
>       Arnd