your netlink permission checks backport to v3.4

your netlink permission checks backport to v3.4

[David Miller]

I don't think you properly compile tested this code:

diff --git a/net/decnet/dn_dev.c b/net/decnet/dn_dev.c
index c00e307..b79ce1e 100644
--- a/net/decnet/dn_dev.c
+++ b/net/decnet/dn_dev.c
@@ -440,7 +440,7 @@ int dn_dev_ioctl(unsigned int cmd, void __user *arg)
 	case SIOCGIFADDR:
 		break;
 	case SIOCSIFADDR:
-		if (!capable(CAP_NET_ADMIN))
+		if (!netlink_capable(skb, CAP_NET_ADMIN))
 			return -EACCES;
 		if (sdn->sdn_family != AF_DECnet)
 			return -EINVAL;

There is no "skb" variable in this function.

I think we're at the point where this backport makes no sense to submit
to the -stable tree.

Multiple rounds of attempting to backport these changes has been tries
and all such attempts have had problems.  Compile problems, authorship
errors, lack of follow-up bug fixes, etc.

This is why I did not try to backport these changes to v3.4 and
earlier.  The possiblity for introducing more regressions is larger
than the benefit from including these changes.

Please do not ask me to include this backport into my v3.4 -stable
series submissions any longer, thank you.

Sorry.

Re: your netlink permission checks backport to v3.4

[wangyufen]

On 2014/8/8 4:53, David Miller wrote:
> 
> I don't think you properly compile tested this code:
> 
> diff --git a/net/decnet/dn_dev.c b/net/decnet/dn_dev.c
> index c00e307..b79ce1e 100644
> --- a/net/decnet/dn_dev.c
> +++ b/net/decnet/dn_dev.c
> @@ -440,7 +440,7 @@ int dn_dev_ioctl(unsigned int cmd, void __user *arg)
>  	case SIOCGIFADDR:
>  		break;
>  	case SIOCSIFADDR:
> -		if (!capable(CAP_NET_ADMIN))
> +		if (!netlink_capable(skb, CAP_NET_ADMIN))
>  			return -EACCES;
>  		if (sdn->sdn_family != AF_DECnet)
>  			return -EINVAL;
> 
> There is no "skb" variable in this function.
> 
> I think we're at the point where this backport makes no sense to submit
> to the -stable tree.
> 
> Multiple rounds of attempting to backport these changes has been tries
> and all such attempts have had problems.  Compile problems, authorship
> errors, lack of follow-up bug fixes, etc.
> 
> This is why I did not try to backport these changes to v3.4 and
> earlier.  The possiblity for introducing more regressions is larger
> than the benefit from including these changes.
> 
> Please do not ask me to include this backport into my v3.4 -stable
> series submissions any longer, thank you.
> 
> Sorry.
> --
> To unsubscribe from this list: send the line "unsubscribe netdev" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> 
> 

I did not open the CONFIG_DECNET option and the code did not build-in.
It's my first time to do backport. I'm ashamed that I made so many mistakes.

Sincerely sorry to trouble you so much.