[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <086f08d257d5d50f5dca5e85b474831358d37c22.camel@nvidia.com>
Date: Thu, 22 Aug 2024 02:07:45 +0000
From: Jianbo Liu <jianbol@...dia.com>
To: "jv@...sburgh.net" <jv@...sburgh.net>, "kuba@...nel.org" <kuba@...nel.org>
CC: "liuhangbin@...il.com" <liuhangbin@...il.com>, "davem@...emloft.net"
<davem@...emloft.net>, Tariq Toukan <tariqt@...dia.com>, Gal Pressman
<gal@...dia.com>, "andy@...yhouse.net" <andy@...yhouse.net>,
"netdev@...r.kernel.org" <netdev@...r.kernel.org>, "pabeni@...hat.com"
<pabeni@...hat.com>, "edumazet@...gle.com" <edumazet@...gle.com>, Saeed
Mahameed <saeedm@...dia.com>, Leon Romanovsky <leonro@...dia.com>
Subject: Re: [PATCH net V5 3/3] bonding: change ipsec_lock from spin lock to
mutex
On Wed, 2024-08-21 at 17:11 -0700, Jakub Kicinski wrote:
> On Wed, 21 Aug 2024 09:00:30 -0700 Jay Vosburgh wrote:
> > Is it really safe to access real_dev once we've left the
> > rcu
> > critical section? What prevents the device referenced by real_dev
> > from
> > being deleted as soon as rcu_read_unlock() completes?
>
> Hah, I asked them this question at least 2 times.
> Let's see if your superior communication skills help :)
Sorry, I maybe misunderstood. I really need your suggestion about how
to change it. Always waiting for you, hangbin or anyother's confirm
before do any update. V5 is the rebase. If you and Jay want to move
rcu_read_unlock after xfrm callbacks, I will send new version soon.
Thanks!
Jianbo
Powered by blists - more mailing lists