Re: [PATCH 1/2] can: bcm: registration process optimization in bcm_module_init()

[Oliver Hartkopp]

Just another reference which make it clear that the reordering of 
function calls in your patch is likely not correct:

https://elixir.bootlin.com/linux/v5.19.7/source/net/packet/af_packet.c#L4734

static int __init packet_init(void)
{
        int rc;

        rc = proto_register(&packet_proto, 0);
        if (rc)
                goto out;
        rc = sock_register(&packet_family_ops);
        if (rc)
                goto out_proto;
        rc = register_pernet_subsys(&packet_net_ops);
        if (rc)
                goto out_sock;
        rc = register_netdevice_notifier(&packet_netdev_notifier);
        if (rc)
                goto out_pernet;

        return 0;

out_pernet:
        unregister_pernet_subsys(&packet_net_ops);
out_sock:
        sock_unregister(PF_PACKET);
out_proto:
        proto_unregister(&packet_proto);
out:
        return rc;
}



On 08.09.22 09:10, Oliver Hartkopp wrote:
> On 08.09.22 05:04, Ziyang Xuan wrote:
> > Now, register_netdevice_notifier() and register_pernet_subsys() are both
> > after can_proto_register(). It can create CAN_BCM socket and process 
> > socket
> > once can_proto_register() successfully, so it is possible missing 
> > notifier
> > event or proc node creation because notifier or bcm proc directory is not
> > registered or created yet. Although this is a low probability 
> > scenario, it
> > is not impossible.
> >
> > Move register_pernet_subsys() and register_netdevice_notifier() to the
> > front of can_proto_register(). In addition, register_pernet_subsys() and
> > register_netdevice_notifier() may fail, check their results are 
> > necessary.
> >
> > Signed-off-by: Ziyang Xuan <william.xuanziyang@xxxxxxxxxx>
> > ---
> >   net/can/bcm.c | 18 +++++++++++++++---
> >   1 file changed, 15 insertions(+), 3 deletions(-)
> >
> > diff --git a/net/can/bcm.c b/net/can/bcm.c
> > index e60161bec850..e2783156bfd1 100644
> > --- a/net/can/bcm.c
> > +++ b/net/can/bcm.c
> > @@ -1744,15 +1744,27 @@ static int __init bcm_module_init(void)
> >       pr_info("can: broadcast manager protocol\n");
> > +    err = register_pernet_subsys(&canbcm_pernet_ops);
> > +    if (err)
> > +        return err;
>
> Analogue to your patch for the CAN_RAW socket here (which has been 
> applied to can-next right now) ...
>
> https://lore.kernel.org/linux-can/7af9401f0d2d9fed36c1667b5ac9b8df8f8b87ee.1661584485.git.william.xuanziyang@xxxxxxxxxx/T/#u 
>
>
> ... I'm not sure whether this is the right sequence to acquire the 
> different resources here.
>
> E.g. in ipsec_pfkey_init() in af_key.c
>
> https://elixir.bootlin.com/linux/v5.19.7/source/net/key/af_key.c#L3887
>
> proto_register() is executed before register_pernet_subsys()
>
> Which seems to be more natural to me.
>
> Best regards,
> Oliver
>
> > +
> > +    err = register_netdevice_notifier(&canbcm_notifier);
> > +    if (err)
> > +        goto register_notifier_failed;
> > +
> >       err = can_proto_register(&bcm_can_proto);
> >       if (err < 0) {
> >           printk(KERN_ERR "can: registration of bcm protocol failed\n");
> > -        return err;
> > +        goto register_proto_failed;
> >       }
> > -    register_pernet_subsys(&canbcm_pernet_ops);
> > -    register_netdevice_notifier(&canbcm_notifier);
> >       return 0;
> > +
> > +register_proto_failed:
> > +    unregister_netdevice_notifier(&canbcm_notifier);
> > +register_notifier_failed:
> > +    unregister_pernet_subsys(&canbcm_pernet_ops);
> > +    return err;
> >   }
> >   static void __exit bcm_module_exit(void)