[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <Z7Vjd7Xx1wplacKC@mini-arch>
Date: Tue, 18 Feb 2025 20:52:07 -0800
From: Stanislav Fomichev <stfomichev@...il.com>
To: Jakub Kicinski <kuba@...nel.org>
Cc: Stanislav Fomichev <sdf@...ichev.me>, netdev@...r.kernel.org,
davem@...emloft.net, edumazet@...gle.com, pabeni@...hat.com,
Saeed Mahameed <saeed@...nel.org>
Subject: Re: [PATCH net-next v4 11/12] docs: net: document new locking reality
On 02/18, Jakub Kicinski wrote:
> On Mon, 17 Feb 2025 18:09:47 -0800 Stanislav Fomichev wrote:
> > +RTNL and netdev instance lock
> > +=============================
> > +
> > +Historically, all networking control operations were protected by a single
> > +global lock known as RTNL. There is an ongoing effort to replace this global
>
> I think RTNL stands for RouTeNetLink. RTNL -> rtnl_lock here?
SG. Will do s/RTNL/rtnl_lock/ in a bunch of other (new) places.
> > +lock with separate locks for each network namespace. The netdev instance lock
> > +represents another step towards making the locking mechanism more granular.
>
> Reads a bit like the per-netns and instance locks are related.
> Maybe rephrase as:
>
> lock with separate locks for each network namespace. Additionally, properties
> of individual netdev are increasingly protected by per-netdev locks.
Sure.
> > +For device drivers that implement shaping or queue management APIs, all control
> > +operations will be performed under the netdev instance lock. Currently, this
> > +instance lock is acquired within the context of RTNL. In the future, there will
> > +be an option for individual drivers to opt out of using RTNL and instead
> > +perform their control operations directly under the netdev instance lock.
> > +
> > +Devices drivers are encouraged to rely on the instance lock where possible.
Powered by blists - more mailing lists