Re: [PATCH net-next] devlink: Require devlink lock during device reload

From: Leon Romanovsky <leon@kernel.org>
To: Jakub Kicinski <kuba@kernel.org>
Cc: Ido Schimmel <idosch@idosch.org>, Jiri Pirko <jiri@resnulli.us>,
	"David S . Miller" <davem@davemloft.net>,
	Jiri Pirko <jiri@nvidia.com>,
	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
Date: Mon, 8 Nov 2021 20:24:37 +0200	[thread overview]
Message-ID: <YYlrZZTdJKhha0FF@unreal> (raw)
In-Reply-To: <20211108101646.0a4e5ca4@kicinski-fedora-pc1c0hjn.dhcp.thefacebook.com>

On Mon, Nov 08, 2021 at 10:16:46AM -0800, Jakub Kicinski wrote:
> On Mon, 8 Nov 2021 19:32:19 +0200 Leon Romanovsky wrote:
> > > I think it's common sense. We're just exporting something to make our
> > > lives easier somewhere else in the three. Do you see a way in which
> > > taking refs on devlink can help out-of-tree code?  
> > 
> > I didn't go such far in my thoughts. My main concern is that you ore
> > exposing broken devlink internals in the hope that drivers will do better
> > locking. I wanted to show that internal locking should be fixed first.
> > 
> > https://lore.kernel.org/netdev/cover.1636390483.git.leonro@nvidia.com/T/#m093f067d0cafcbe6c05ed469bcfd708dd1eb7f36
> > 
> > While this series fixes locking and after all my changes devlink started
> > to be more secure, that works correctly for simple drivers.
> 
> I prefer my version. I think I asked you to show how the changes make
> drivers simpler, which you failed to do.

Why did I fail? My version requires **zero** changes to the drivers.
Everything works without them changing anything. You can't ask for more.

> 
> I already told you how this is going to go, don't expect me to comment
> too much.
> 
> > However for net namespace aware drivers it still stays DOA.
> > 
> > As you can see, devlink reload holds pernet_ops_rwsem, which drivers should
> > take too in order to unregister_netdevice_notifier.
> > 
> > So for me, the difference between netdevsim and real device (mlx5) is
> > too huge to really invest time into netdevsim-centric API, because it
> > won't solve any of real world problems.
> 
> Did we not already go over this? Sorry, it feels like you're repeating
> arguments which I replied to before. This is exhausting.

I don't enjoy it either.

> 
> nfp will benefit from the simplified locking as well, and so will bnxt,
> although I'm not sure the maintainers will opt for using devlink framework
> due to the downstream requirements.

Exactly why devlink should be fixed first.

Thanks