[ofa-general] Re: RFC: "mlx4" drivers for Mellanox ConnectX HCAs

[Michael S. Tsirkin]

> Quoting Roland Dreier <rdreier at cisco.com>:
> Subject: Re: RFC: "mlx4" drivers for Mellanox ConnectX HCAs
> 
>  > >     git://git.kernel.org/pub/scm/linux/kernel/git/roland/infiniband.git connectx
> 
>  > I cloned your git and checked out the connectx branch.  When I do "git log" on the
>  > connectx branch, I see that the last commit was done on March 4. Is that so?
> 
> Yes, I have been using stgit to update the two patches in that branch
> rather than adding new commits.  So it seems that the patches keep the
> date they were created, although they have been updated quite a bit
> since then.

So, how can this git branch be used?
I guess doing git pull's won't work too well since
you are rewriting the history all the time ...

Maybe using plain git would be a better idea,
you can always smash the history before you submit
code upstream ...

> By the way, I've tried to put "FIXME" comments in the places that
> still need work.  I think the most important things that still need to
> be done are:
> 
>   - Fix CQ locking on destroy QP; not completely trivial since I want
>     to keep a somewhat clean division between mlx4_core and mlx4_ib

It seems core could just export a "cleanup" functions then, and
mlx4_ib would call that under the appropriate lock?

>   - clean stale CQEs on destroy QP or modify QP to RESET

BTW, I went over that code in mthca and it seems that it does
not handle CQ resize correctly. Right?

>   - inline send support

Inline send from userspace, or from kernel as well?
If from kernel - note that we never had inline
in kernel for older HW, so ULPs don't use it.
So I guess this is a low priority feature?

> And of course we still need to get write combining support in the core
> kernel to make blueflame work well.

You haven't gootten anhy feedback from your last request for
comments on this, did you?

> Please let me know if you're starting to work on something so we don't
> duplicate effort.


-- 
MST