Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:18 as permitted sender) client-ip=2620:137:e000::1:18;
Date:   Tue, 12 Apr 2022 17:01:36 +0200 (CEST)
From:   Christoph Lameter <cl@gentwo.de>
To:     Jason Gunthorpe <jgg@nvidia.com>
cc:     Leon Romanovsky <leon@kernel.org>,
        Mark Zhang <markzhang@nvidia.com>,
        linux-kernel@vger.kernel.org, linux-rdma@vger.kernel.org,
        syzbot+8fcbb77276d43cc8b693@syzkaller.appspotmail.com
Subject: Re: [PATCH rdma-rc] RDMA/cma: Limit join multicast to UD QP type
 only
In-Reply-To: <20220412141134.GI2120790@nvidia.com>
Message-ID: <alpine.DEB.2.22.394.2204121653490.374115@gentwo.de>
References: <4132fdbc9fbba5dca834c84ae383d7fe6a917760.1649083917.git.leonro@nvidia.com> <20220408182440.GA3647277@nvidia.com> <YlLHjFlR8BtCc5Hu@unreal> <20220412141134.GI2120790@nvidia.com>
User-Agent: Alpine 2.22 (DEB 394 2020-01-19)
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Precedence: bulk

On Tue, 12 Apr 2022, Jason Gunthorpe wrote:

> The only places setting non-default qkeys are SIDR, maybe nobody uses
> SIDR with multicast.


IP port numbers provided are ignored by the RDMA subsytem when doing
multicast joins. Thus no need to do SIDRs with RDMA multicast.

Some middleware solutions (like LLM by IBM and Confinity) are creating
their own custom MGID because of this problem. The custom MGID will then
contain the port number.

On the subject of this patch: There is a usecase for Multicast with
IBV_QPT_RAW_PACKET too. A multicast join is required to redirect traffic
for a multicast group to the raw socket.