ip6_tunnel: Correct tos value in collect_md mode

Same as ip_gre, geneve and vxlan, use key->tos as tos value.

Signed-off-by: Haishuang Yan <yanhaishuang@cmss.chinamobile.com>
---
 net/ipv6/ip6_tunnel.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

On 06/13/2017 09:55 AM, Haishuang Yan wrote:
> Same as ip_gre, geneve and vxlan, use key->tos as tos value.
>

Please also add Fixes tag and Cc original authors.

Fixes: 0e9a709560db ("ip6_tunnel, ip6_gre: fix setting of DSCP on encapsulated packets")

> Signed-off-by: Haishuang Yan <yanhaishuang@cmss.chinamobile.com>
> ---
>   net/ipv6/ip6_tunnel.c | 4 ++--
>   1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/net/ipv6/ip6_tunnel.c b/net/ipv6/ip6_tunnel.c
> index ef99d59..5f4aff5 100644
> --- a/net/ipv6/ip6_tunnel.c
> +++ b/net/ipv6/ip6_tunnel.c
> @@ -1249,7 +1249,7 @@ int ip6_tnl_xmit(struct sk_buff *skb, struct net_device *dev, __u8 dsfield,
>   		fl6.flowi6_proto = IPPROTO_IPIP;
>   		fl6.daddr = key->u.ipv6.dst;
>   		fl6.flowlabel = key->label;
> -		dsfield = ip6_tclass(key->label);
> +		dsfield = key->tos;

Not exactly the same as vxlan / geneve. They do use key->tos
indeed, but they also set fl6.flowlabel differently:

   [...]
   prio = info->key.tos;
   [...]
   fl6->flowlabel = ip6_make_flowinfo(RT_TOS(prio),
                                      info->key.label);
   [...]

But looks like the latter is done later in ip6_tnl_xmit() eventually
before the route lookup as per 5f733ee68f9a ("ip6_tunnel: fix traffic
class routing for tunnels") fix. However, you still might want to
mask key->tos with RT_TOS().

>   	} else {
>   		if (!(t->parms.flags & IP6_TNL_F_IGN_ENCAP_LIMIT))
>   			encap_limit = t->parms.encap_limit;
> @@ -1320,7 +1320,7 @@ int ip6_tnl_xmit(struct sk_buff *skb, struct net_device *dev, __u8 dsfield,
>   		fl6.flowi6_proto = IPPROTO_IPV6;
>   		fl6.daddr = key->u.ipv6.dst;
>   		fl6.flowlabel = key->label;
> -		dsfield = ip6_tclass(key->label);
> +		dsfield = key->tos;
>   	} else {
>   		offset = ip6_tnl_parse_tlv_enc_lim(skb, skb_network_header(skb));
>   		/* ip6_tnl_parse_tlv_enc_lim() might have reallocated skb->head */
>

On 06/13/2017 03:31 PM, 严海双 wrote:
>> On 13 Jun 2017, at 5:57 PM, Daniel Borkmann <daniel@iogearbox.net> wrote:
>> On 06/13/2017 09:55 AM, Haishuang Yan wrote:
>>> Same as ip_gre, geneve and vxlan, use key->tos as tos value.
>>
>> Please also add Fixes tag and Cc original authors.
>>
>> Fixes: 0e9a709560db ("ip6_tunnel, ip6_gre: fix setting of DSCP on encapsulated packets”)
>
> Ok, I will add it, thanks
>
>>> Signed-off-by: Haishuang Yan <yanhaishuang@cmss.chinamobile.com>
>>> ---
>>>   net/ipv6/ip6_tunnel.c | 4 ++--
>>>   1 file changed, 2 insertions(+), 2 deletions(-)
>>>
>>> diff --git a/net/ipv6/ip6_tunnel.c b/net/ipv6/ip6_tunnel.c
>>> index ef99d59..5f4aff5 100644
>>> --- a/net/ipv6/ip6_tunnel.c
>>> +++ b/net/ipv6/ip6_tunnel.c
>>> @@ -1249,7 +1249,7 @@ int ip6_tnl_xmit(struct sk_buff *skb, struct net_device *dev, __u8 dsfield,
>>>   		fl6.flowi6_proto = IPPROTO_IPIP;
>>>   		fl6.daddr = key->u.ipv6.dst;
>>>   		fl6.flowlabel = key->label;
>>> -		dsfield = ip6_tclass(key->label);
>>> +		dsfield = key->tos;
>>
>> Not exactly the same as vxlan / geneve. They do use key->tos
>> indeed, but they also set fl6.flowlabel differently:
>>
>>   [...]
>>   prio = info->key.tos;
>>   [...]
>>   fl6->flowlabel = ip6_make_flowinfo(RT_TOS(prio),
>>                                      info->key.label);
>>   [...]
>>
>> But looks like the latter is done later in ip6_tnl_xmit() eventually
>> before the route lookup as per 5f733ee68f9a ("ip6_tunnel: fix traffic
>> class routing for tunnels") fix. However, you still might want to
>> mask key->tos with RT_TOS().
>
> Do you mean this way?

This one from my last comment: dsfield = RT_TOS(key->tos)