Re: [GIT PULL] IPv6 Updates

[Hasso Tepper <hasso@estpak.ee>]

YOSHIFUJI Hideaki / 吉藤英明 wrote:
> commit bd292ad1ccbf15c72351f79d4a130dd50da81613
> Author: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
> Date:   Thu Feb 15 02:07:27 2007 +0900
> 
>     [IPV6] ADDRCONF: Register inet6_dev earlier.
>     
>     Allocate inet6_dev earlier to allow users to set up per-interface variables.
>     
>     Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
> 
> diff --git a/net/ipv6/addrconf.c b/net/ipv6/addrconf.c
> index c6683ea..42b5cc5 100644
> --- a/net/ipv6/addrconf.c
> +++ b/net/ipv6/addrconf.c
> @@ -2246,6 +2246,14 @@ static int addrconf_notify(struct notifier_block *this, unsigned long event,
>  	int run_pending = 0;
>  
>  	switch(event) {
> +	case NETDEV_REGISTER:
> +		if (!idev) {
> +			idev = ipv6_add_dev(dev);
> +			if (!idev)
> +				printk(KERN_WARNING "IPv6: add_dev failed for %s\n",
> +					dev->name);
> +		}
> +		break;
>  	case NETDEV_UP:
>  	case NETDEV_CHANGE:
>  		if (event == NETDEV_UP) {
> 
> ---

Many thanks. Now if only IPv4 would be fixed in similar fashion as well.

Also, this reminds me related issue - "ADDRCONF(NETDEV_UP): <ifname>: link
is not ready" and "ADDRCONF(NETDEV_CHANGE): <ifname>: link becomes ready"
messages. While it may look informational, it's extremely annoying in
systems with hundreds and even thousands vlan interfaces.


regards,

-- 
Hasso Tepper