[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <CANn89iJB_QYWxr5tjR+4uRpO=b5jhe21pAX=c0sFmEHH4+OG6w@mail.gmail.com>
Date: Wed, 31 Mar 2021 07:53:50 +0200
From: Eric Dumazet <edumazet@...gle.com>
To: "Wong, Vee Khee" <vee.khee.wong@...el.com>
Cc: "eric.dumazet@...il.com" <eric.dumazet@...il.com>,
"davem@...emloft.net" <davem@...emloft.net>,
"kuba@...nel.org" <kuba@...nel.org>,
"netdev@...r.kernel.org" <netdev@...r.kernel.org>,
"syzkaller@...glegroups.com" <syzkaller@...glegroups.com>,
"Ismail, Mohammad Athari" <mohammad.athari.ismail@...el.com>
Subject: Re: [PATCH net-next] sit: proper dev_{hold|put} in ndo_[un]init methods
On Wed, Mar 31, 2021 at 2:05 AM Wong, Vee Khee <vee.khee.wong@...el.com> wrote:
>
> Hi all,
>
> This patch introduced the following massive warnings printouts on a
> Intel x86 Alderlake platform with STMMAC MAC and Marvell 88E2110 PHY.
>
> [ 149.674232] unregister_netdevice: waiting for sit0 to become free. Usage count = 2
> [ 159.930310] unregister_netdevice: waiting for sit0 to become free. Usage count = 2
> [ 170.186205] unregister_netdevice: waiting for sit0 to become free. Usage count = 2
> [ 180.434311] unregister_netdevice: waiting for sit0 to become free. Usage count = 2
> [ 190.682309] unregister_netdevice: waiting for sit0 to become free. Usage count = 2
> [ 200.690176] unregister_netdevice: waiting for sit0 to become free. Usage count = 2
> [ 210.938310] unregister_netdevice: waiting for sit0 to become free. Usage count = 2
> [ 221.186311] unregister_netdevice: waiting for sit0 to become free. Usage count = 2
> [ 231.442311] unregister_netdevice: waiting for sit0 to become free. Usage count = 2
> [ 241.690186] unregister_netdevice: waiting for sit0 to become free. Usage count = 2
> [ 251.698288] unregister_netdevice: waiting for sit0 to become free. Usage count = 2
> [ 261.946311] unregister_netdevice: waiting for sit0 to become free. Usage count = 2
> [ 272.194181] unregister_netdevice: waiting for sit0 to become free. Usage count = 2
> [ 282.442311] unregister_netdevice: waiting for sit0 to become free. Usage count = 2
> [ 292.690310] unregister_netdevice: waiting for sit0 to become free. Usage count = 2
> [ 302.938313] unregister_netdevice: waiting for sit0 to become free. Usage count = 2
> [ 313.186255] unregister_netdevice: waiting for sit0 to become free. Usage count = 2
> [ 323.442329] unregister_netdevice: waiting for sit0 to become free. Usage count = 2
> [ 333.698309] unregister_netdevice: waiting for sit0 to become free. Usage count = 2
> [ 343.946310] unregister_netdevice: waiting for sit0 to become free. Usage count = 2
> [ 354.202166] unregister_netdevice: waiting for sit0 to become free. Usage count = 2
> [ 364.450190] unregister_netdevice: waiting for sit0 to become free. Usage count = 2
> [ 374.706314] unregister_netdevice: waiting for sit0 to become free. Usage count = 2
>
> Is this an expected behavior?
>
Nope, I already have a fix, but it depends on a pending patch.
https://patchwork.kernel.org/project/netdevbpf/patch/20210330064551.545964-1-eric.dumazet@gmail.com/
(I need the patch being merged to add a corresponding Fixes: tag)
You can try the attached patch :
View attachment "0001-ipv6-remove-extra-dev_hold-for-fallback-tunnels.patch" of type "text/x-patch" (3126 bytes)
Powered by blists - more mailing lists