On Thu, May 16, 2019 at 03:51:47PM +0000, Bernard Metzler wrote:
>
>
> >To: "Bernard Metzler" <BMT@xxxxxxxxxxxxxx>
> >From: "Jason Gunthorpe" <jgg@xxxxxxxx>
> >Date: 05/16/2019 05:47PM
> >Cc: "Leon Romanovsky" <leon@xxxxxxxxxx>, "Doug Ledford"
> ><dledford@xxxxxxxxxx>, "linux-rdma" <linux-rdma@xxxxxxxxxxxxxxx>
> >Subject: Re: iWARP and soft-iWARP interop testing
> >
> >On Thu, May 16, 2019 at 03:39:10PM +0000, Bernard Metzler wrote:
> >>
> >> >To: "Doug Ledford" <dledford@xxxxxxxxxx>
> >> >From: "Jason Gunthorpe" <jgg@xxxxxxxx>
> >> >Date: 05/07/2019 06:13PM
> >> >Cc: "linux-rdma" <linux-rdma@xxxxxxxxxxxxxxx>, "Bernard Metzler"
> >> ><BMT@xxxxxxxxxxxxxx>
> >> >Subject: Re: iWARP and soft-iWARP interop testing
> >> >
> >> >On Mon, May 06, 2019 at 04:38:27PM -0400, Doug Ledford wrote:
> >> >> So, Jason and I were discussing the soft-iWARP driver
> >submission,
> >> >and he
> >> >> thought it would be good to know if it even works with the
> >various
> >> >iWARP
> >> >> hardware devices. I happen to have most of them on hand in one
> >> >form or
> >> >> another, so I set down to test it. In the process, I ran across
> >> >some
> >> >> issues just with the hardware versions themselves, let alone
> >with
> >> >soft-
> >> >> iWARP. So, here's the results of my matrix of tests. These
> >aren't
> >> >> performance tests, just basic "does it work" smoke tests...
> >> >
> >> >Well, lets imagine to merge this at 5.2-rc1?
> >> >
> >> >Bernard you'll need to rebase and resend when it comes out in two
> >> >weeks.
> >> >
> >> >Thanks,
> >> >Jason
> >> >
> >> >
> >> I think I addressed all major issues of the current siw RFC.
> >>
> >> Probably most important, it's now guaranteed that the remaining
> >> two objects (QP and MR) are kfree'd after return from the
> >> ib_devices free call. This makes it easier for future development
> >> of mid layers resource management, as Leon was pointing out.
> >>
> >> All IDR usage is gone as well.
> >>
> >> I removed the siw protection domain, since there is nothing
> >> siw specific to be stored within. I just keep a structure
> >> definition of 'struct siw_pd {struct ib_pd *base_pd}' to
> >> keep INIT_RDMA_OBJ_SIZE() happy.
> >
> >? Really? I iwarp doesn't use a protection domain?
>
> Aehm no, siw does not need any siw specific additions to the ib_pd.
> So there is no need to define any extra siw_pd besides of ib_pd.
> Except for the INIT_RDMA_OBJ_SIZE thing.
>
> siw can just use struct ib_pd. It has all it needs...
So you are using the pointer value of the ib_pd to identify if a
MR/QP/etc are in the same PD?
Jason
