Re: [PATCH 4/4] ipv6: symbol_get to access a sit symbol

[David Miller]

From: Christoph Hellwig <hch@xxxxxx>
Date: Thu, 14 May 2020 16:51:01 +0200

> Instead of going through the ioctl handler from kernel space, use
> symbol_get to the newly factored out ipip6_set_dstaddr helper, bypassing
> addrconf.c entirely.
> 
> Signed-off-by: Christoph Hellwig <hch@xxxxxx>
 ...
> -		memset(&p, 0, sizeof(p));
> -		p.iph.daddr = ireq.ifr6_addr.s6_addr32[3];
> -		p.iph.saddr = 0;
> -		p.iph.version = 4;
> -		p.iph.ihl = 5;
> -		p.iph.protocol = IPPROTO_IPV6;
> -		p.iph.ttl = 64;
> -		ifr.ifr_ifru.ifru_data = (__force void __user *)&p;
> -
> -		if (ops->ndo_do_ioctl) {
> -			mm_segment_t oldfs = get_fs();
> -
> -			set_fs(KERNEL_DS);
> -			err = ops->ndo_do_ioctl(dev, &ifr, SIOCADDTUNNEL);
> -			set_fs(oldfs);
> -		} else
> -			err = -EOPNOTSUPP;
 ...
> +	p.iph.daddr = ireq.ifr6_addr.s6_addr32[3];
> +	p.iph.version = 4;
> +	p.iph.ihl = 5;
> +	p.iph.protocol = IPPROTO_IPV6;
> +	p.iph.ttl = 64;
> +	p.iph.frag_off |= htons(IP_DF);
> +
> +	err = -ENOBUFS;
> +	if (!ipip6_tunnel_locate(dev_net(tunnel_dev), &p, true))
> +		goto out_unlock;

You're not undoing one, but two levels of abstraction here.

Is this "ipip6_tunnel_locate()" call part of the SIT ioctl implementation?
Where did it come from?   Why are ->ndo_do_ioctl() implementations no longer
allowed from here?

Honestly, this feels like a bit much.