netdev - Re: [PATCH net v2] net: lwtunnel: handle MTU calculation in forwading

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Sat, 26 Jun 2021 11:18:42 -0600
From:   David Ahern <dsahern@...il.com>
To:     Vadim Fedorenko <vfedorenko@...ek.ru>,
        David Ahern <dsahern@...nel.org>, netdev@...r.kernel.org,
        Roopa Prabhu <roopa@...dia.com>
Cc:     Jakub Kicinski <kuba@...nel.org>,
        Hideaki YOSHIFUJI <yoshfuji@...ux-ipv6.org>,
        "David S. Miller" <davem@...emloft.net>
Subject: Re: [PATCH net v2] net: lwtunnel: handle MTU calculation in forwading

[ added Roopa's nvidia address ]

On 6/25/21 10:21 AM, Vadim Fedorenko wrote:
> Commit 14972cbd34ff ("net: lwtunnel: Handle fragmentation") moved
> fragmentation logic away from lwtunnel by carry encap headroom and
> use it in output MTU calculation. But the forwarding part was not
> covered and created difference in MTU for output and forwarding and
> further to silent drops on ipv4 forwarding path. Fix it by taking
> into account lwtunnel encap headroom.
> 
> The same commit also introduced difference in how to treat RTAX_MTU
> in IPv4 and IPv6 where latter explicitly removes lwtunnel encap
> headroom from route MTU. Make IPv4 version do the same.
> 
> Fixes: 14972cbd34ff ("net: lwtunnel: Handle fragmentation")
> Suggested-by: David Ahern <dsahern@...il.com>
> Signed-off-by: Vadim Fedorenko <vfedorenko@...ek.ru>
> ---
>  include/net/ip.h        | 12 ++++++++----
>  include/net/ip6_route.h | 16 ++++++++++++----
>  net/ipv4/route.c        |  3 ++-
>  3 files changed, 22 insertions(+), 9 deletions(-)
> 


I think this is the right approach - tunnel overhead should always be
considered for the mtu. Did you run the pmtu.sh selftests to make sure
those still work?

Reviewed-by: David Ahern <dsahern@...nel.org>


> diff --git a/include/net/ip.h b/include/net/ip.h
> index e20874059f82..d9683bef8684 100644
> --- a/include/net/ip.h
> +++ b/include/net/ip.h
> @@ -31,6 +31,7 @@
>  #include <net/flow.h>
>  #include <net/flow_dissector.h>
>  #include <net/netns/hash.h>
> +#include <net/lwtunnel.h>
>  
>  #define IPV4_MAX_PMTU		65535U		/* RFC 2675, Section 5.1 */
>  #define IPV4_MIN_MTU		68			/* RFC 791 */
> @@ -445,22 +446,25 @@ static inline unsigned int ip_dst_mtu_maybe_forward(const struct dst_entry *dst,
>  
>  	/* 'forwarding = true' case should always honour route mtu */
>  	mtu = dst_metric_raw(dst, RTAX_MTU);
> -	if (mtu)
> -		return mtu;
> +	if (!mtu)
> +		mtu = min(READ_ONCE(dst->dev->mtu), IP_MAX_MTU);
>  
> -	return min(READ_ONCE(dst->dev->mtu), IP_MAX_MTU);
> +	return mtu - lwtunnel_headroom(dst->lwtstate, mtu);
>  }
>  
>  static inline unsigned int ip_skb_dst_mtu(struct sock *sk,
>  					  const struct sk_buff *skb)
>  {
> +	unsigned int mtu;
> +
>  	if (!sk || !sk_fullsock(sk) || ip_sk_use_pmtu(sk)) {
>  		bool forwarding = IPCB(skb)->flags & IPSKB_FORWARDED;
>  
>  		return ip_dst_mtu_maybe_forward(skb_dst(skb), forwarding);
>  	}
>  
> -	return min(READ_ONCE(skb_dst(skb)->dev->mtu), IP_MAX_MTU);
> +	mtu = min(READ_ONCE(skb_dst(skb)->dev->mtu), IP_MAX_MTU);
> +	return mtu - lwtunnel_headroom(skb_dst(skb)->lwtstate, mtu);
>  }
>  
>  struct dst_metrics *ip_fib_metrics_init(struct net *net, struct nlattr *fc_mx,
> diff --git a/include/net/ip6_route.h b/include/net/ip6_route.h
> index f51a118bfce8..f14149df5a65 100644
> --- a/include/net/ip6_route.h
> +++ b/include/net/ip6_route.h
> @@ -265,11 +265,18 @@ int ip6_fragment(struct net *net, struct sock *sk, struct sk_buff *skb,
>  
>  static inline int ip6_skb_dst_mtu(struct sk_buff *skb)
>  {
> +	int mtu;
> +
>  	struct ipv6_pinfo *np = skb->sk && !dev_recursion_level() ?
>  				inet6_sk(skb->sk) : NULL;
>  
> -	return (np && np->pmtudisc >= IPV6_PMTUDISC_PROBE) ?
> -	       skb_dst(skb)->dev->mtu : dst_mtu(skb_dst(skb));
> +	if (np && np->pmtudisc >= IPV6_PMTUDISC_PROBE) {
> +		mtu = READ_ONCE(skb_dst(skb)->dev->mtu);
> +		mtu -= lwtunnel_headroom(skb_dst(skb)->lwtstate, mtu);
> +	} else
> +		mtu = dst_mtu(skb_dst(skb));
> +
> +	return mtu;
>  }
>  
>  static inline bool ip6_sk_accept_pmtu(const struct sock *sk)
> @@ -317,7 +324,7 @@ static inline unsigned int ip6_dst_mtu_forward(const struct dst_entry *dst)
>  	if (dst_metric_locked(dst, RTAX_MTU)) {
>  		mtu = dst_metric_raw(dst, RTAX_MTU);
>  		if (mtu)
> -			return mtu;
> +			goto out;
>  	}
>  
>  	mtu = IPV6_MIN_MTU;
> @@ -327,7 +334,8 @@ static inline unsigned int ip6_dst_mtu_forward(const struct dst_entry *dst)
>  		mtu = idev->cnf.mtu6;
>  	rcu_read_unlock();
>  
> -	return mtu;
> +out:
> +	return mtu - lwtunnel_headroom(dst->lwtstate, mtu);
>  }
>  
>  u32 ip6_mtu_from_fib6(const struct fib6_result *res,
> diff --git a/net/ipv4/route.c b/net/ipv4/route.c
> index 6a36ac98476f..78d1e5afc452 100644
> --- a/net/ipv4/route.c
> +++ b/net/ipv4/route.c
> @@ -1306,7 +1306,7 @@ INDIRECT_CALLABLE_SCOPE unsigned int ipv4_mtu(const struct dst_entry *dst)
>  		mtu = dst_metric_raw(dst, RTAX_MTU);
>  
>  	if (mtu)
> -		return mtu;
> +		goto out;
>  
>  	mtu = READ_ONCE(dst->dev->mtu);
>  
> @@ -1315,6 +1315,7 @@ INDIRECT_CALLABLE_SCOPE unsigned int ipv4_mtu(const struct dst_entry *dst)
>  			mtu = 576;
>  	}
>  
> +out:
>  	mtu = min_t(unsigned int, mtu, IP_MAX_MTU);
>  
>  	return mtu - lwtunnel_headroom(dst->lwtstate, mtu);
>