Date:   Tue, 16 Nov 2021 08:44:42 -0400
From:   Jason Gunthorpe <jgg@nvidia.com>
To:     Jiri Pirko <jiri@resnulli.us>
Cc:     Leon Romanovsky <leon@kernel.org>,
        Jakub Kicinski <kuba@kernel.org>, Jiri Pirko <jiri@nvidia.com>,
        Ido Schimmel <idosch@idosch.org>,
        "David S . Miller" <davem@davemloft.net>,
        linux-kernel@vger.kernel.org, netdev@vger.kernel.org,
        edwin.peer@broadcom.com
Subject: Re: [PATCH net-next] devlink: Require devlink lock during device
 reload
Message-ID: <20211116124442.GX2105516@nvidia.com>
References: <20211109182427.GJ1740502@nvidia.com>
 <YY0G90fJpu/OtF8L@nanopsycho>
 <YY0J8IOLQBBhok2M@unreal>
 <YY4aEFkVuqR+vauw@nanopsycho>
 <YZCqVig9GQi/o1iz@unreal>
 <YZJCdSy+wzqlwrE2@nanopsycho>
 <20211115125359.GM2105516@nvidia.com>
 <YZJx8raQt+FkKaeY@nanopsycho>
 <20211115150931.GA2386342@nvidia.com>
 <YZNWRXzzRYMNhUEO@nanopsycho>
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <YZNWRXzzRYMNhUEO@nanopsycho>
MIME-Version: 1.0
X-MS-Exchange-AntiSpam-MessageData-ChunkCount: 1
X-MS-Exchange-AntiSpam-MessageData-0: =?us-ascii?Q?wdsloqoW7au7Fyx5HY3rByGYSjYk5vGwYzRR/Zw/zZ0RAXQyx7z5RkSoQvdC?=
 =?us-ascii?Q?qbqeEeJk8pdkdqkzz2HaCcl3cUIVTJbOnK7h3WdJ9HR4G2p+aXPEyK5I9HdZ?=
 =?us-ascii?Q?yp7h82Z0YaDZw2uuVZuk7g7w+zR+wR3g9IisQhGOjIo4I4sZwZzJVaQOgxR4?=
 =?us-ascii?Q?Onh6c8Hu9GEf2D9BRoiOJ1NXJHSPNgm4KQlsvwnThimMysutPBUyTfsFwpN7?=
 =?us-ascii?Q?qu9w7LDMyIlVf7fQk8QBSxVI4FVJrnK3ql7VCA2CaX4CcTuT50TdBL2YdZBB?=
 =?us-ascii?Q?VWPTOp3YxJNXfmkfzxWGFh8vk0T8b4rX/Aac3kw5DAyQ9JflZ+++Lp7UhOk0?=
 =?us-ascii?Q?9t2s4aEIce2M77tZQbE+TBvv/fMe5d9LhNO13/WQ20GUuJnWvbNZuBH66j9h?=
 =?us-ascii?Q?wd15PZxKgfpyNSJtmbgwKsntf5DNTUXNCFSa8uk2oF0+dxxHB0iP7LO/ilBn?=
 =?us-ascii?Q?9swZRK9jmlBPpTxnV+SG3fiNxxva32JFdUazJbDzzHtsOGOTW3ummckI1Eea?=
 =?us-ascii?Q?c+egrb8eUFxNiHa1JrzTPe2TBaPFuCFqUnW9Tn3Ecc1stGpZzmuQ+PpvX1Uw?=
 =?us-ascii?Q?8TrpQYE201SW/FDGHJISA5wRDK8Ct4n/9d91iFQ2Va+YfaCvaNTgkTGiJizH?=
 =?us-ascii?Q?3sXy0Q/iu9iTvweZf+j4JJYmipRPDQVoRTNUUKAR7dm6yfsNplYxX0jkWcaD?=
 =?us-ascii?Q?FVPUmyfbCevAUs2Sv62NyvtUHRJbnO15VkcX2uOKFPQQSfoj33QZwp4+yGCa?=
 =?us-ascii?Q?rTNohX1W42UuTxSHH162fYMxykYUAh4J7WJ6xqk01jBAcRde6UDNZdd9LL3E?=
 =?us-ascii?Q?Tu25morYpQCNBkvYKChjegnFOT+3H8rDV13PdENYVjkI5IIZacPCUqNNc8VO?=
 =?us-ascii?Q?6/+oxvn7afmBkJIfG7b4bagGYxWvfWrlidflBhUNiAtT+TCRO9bvhjrnraP5?=
 =?us-ascii?Q?f3q8BEX7/SgticIBsy4NSk75LPg0sryWNHKFkigl5iktKu/YUHlcO7t5xCxr?=
 =?us-ascii?Q?AcYs3SoiH/VjGgTREaZVY5OmtmD3YrRuUBrmzths3HbeSUemJxFnUV2PdU3I?=
 =?us-ascii?Q?O6YzGO76HypTh/SOAZA1gaiCBRXp1DGqC16U5GUMAiobCLlckwb5iqYsEbKv?=
 =?us-ascii?Q?3+DbUvemIeXurbOWwxnYLKeQy/OJzYD7r0zWwdeLRaHZKKXaIsC6HlwFxmPR?=
 =?us-ascii?Q?E38Dy9c1T/74LC4FBNDY1A5c1+OHL8ZaDuxAR+cWKhmYfpKuZBZiMrqPTSn/?=
 =?us-ascii?Q?zdwcCZiaiRNOaFXmHvOyTCEHUvbdfA8mKF0sUmTsgT/t6o7F/9zJFZAWVt56?=
 =?us-ascii?Q?74sx6+4FK3q5cvQEFJJ0CXbItZhfo+CuA2bKdOLml2PpMC2fQeKyvT+Gitme?=
 =?us-ascii?Q?SOu1Uee5pT+mnJUsULOduPr1msIMrQ8AfOFJRvqYyJIOvsZNUF6ABU6/pdPi?=
 =?us-ascii?Q?7plE4Cm56uKMHAvH594S4Ko70g4Adks51ZFazHkDLJS+akwLY0Gj0gWB3IM9?=
 =?us-ascii?Q?romTPQ9xnAnZIbsfOzNNE2PNzwYrw5hhtR7AsxEtVzmzPoFhcObsiksLK+0v?=
 =?us-ascii?Q?40dEnEbVi+XcprD6SXM=3D?=
X-OriginatorOrg: Nvidia.com
X-MS-Exchange-CrossTenant-Network-Message-Id: cd5d21dd-4f77-405b-2eb9-08d9a8fedda1
X-MS-Exchange-CrossTenant-AuthSource: BL0PR12MB5506.namprd12.prod.outlook.com
X-MS-Exchange-CrossTenant-AuthAs: Internal
X-MS-Exchange-CrossTenant-OriginalArrivalTime: 16 Nov 2021 12:44:44.5224
 (UTC)
X-MS-Exchange-CrossTenant-FromEntityHeader: Hosted
X-MS-Exchange-CrossTenant-Id: 43083d15-7273-40c1-b7db-39efd9ccc17a
X-MS-Exchange-CrossTenant-MailboxType: HOSTED
X-MS-Exchange-CrossTenant-UserPrincipalName: NE9qY5qOIH622hNmRKdKBv2opqsq8atJB6301GFwpTftePX9R4L+NHLlzb4tR6ae
X-MS-Exchange-Transport-CrossTenantHeadersStamped: BL1PR12MB5141
Precedence: bulk
List-ID: <linux-kernel.vger.kernel.org>
X-Mailing-List: linux-kernel@vger.kernel.org

On Tue, Nov 16, 2021 at 07:57:09AM +0100, Jiri Pirko wrote:

> >There is only one place in the entire kernel calling the per-ns
> >register_netdevice_notifier_dev_net() and it is burred inside another
> >part of mlx5 for some reason..
> 
> Yep. I added it there to solve this deadlock.

I wonder how it can work safely inside a driver, since when are
drivers NS aware?

        uplink_priv->bond->nb.notifier_call = mlx5e_rep_esw_bond_netevent;
        ret = register_netdevice_notifier_dev_net(netdev,
                                                  &uplink_priv->bond->nb,
                                                  &uplink_priv->bond->nn);

Doesn't that just loose events when the user moves netdev to another
namespace?

> >I believe Parav already looked at using that in rdma and it didn't
> >work for some reason I've forgotten. 
> >
> >It is not that we care about events in different namespaces, it is
> >that rdma, like everything else, doesn't care about namespaces and
> >wants events from the netdev no matter where it is located.
> 
> Wait, so there is no notion of netnamespaces in rdma? I was under
> impression rdma supports netnamespaces...

It does, but that doesn't change things, when it is attached to a
netdev it needs events, without any loss, no matter what NS that
netdev is in.

Jason