On 12/08/2017 10:02 AM, Steffen Klassert wrote:
> On Wed, Dec 06, 2017 at 07:38:19PM +0300, Alexey Kodanev wrote:
>> LTP/udp6_ipsec_vti tests fail when sending large UDP datagrams
>> that require fragmentation and underlying device MTU <= 1500.
>> This happens because ip6_vti sets mtu to ETH_DATA_LEN and not
>> updating it depending on a destiantion address.
>>
>> Futhure attempts to send UDP packets may succeed because pmtu
>> get updated on ICMPV6_PKT_TOOBIG in vti6_err().
>>
>> Here is the example when output device MTU set to 9000:
>>
>> # ip a sh ltp_ns_veth2
>> ltp_ns_veth2@if7: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 9000 ...
>> inet 10.0.0.2/24 scope global ltp_ns_veth2
>> inet6 fd00::2/64 scope global
>> ...
>> # ip li add vti6 type vti6 local fd00::2 remote fd00::1
>> # ip li show vti6
>> vti6@NONE: <POINTOPOINT,NOARP> mtu 1500 ...
>> link/tunnel6 fd00::2 peer fd00::1
>>
>> After the patch:
>>
>> # ip li add vti6 type vti6 local fd00::2 remote fd00::1
>> # ip li show vti6
>> vti6@NONE: <POINTOPOINT,NOARP> mtu 8832 ...
>> link/tunnel6 fd00::2 peer fd00::1
>>
>> Regarding ip_vti, it already tunes mtu with ip_tunnel_bind_dev():
>>
>> # ip li add vti4 type vti local 10.0.0.2 remote 10.0.0.1
>> # ip li sh vti4
>> vti4@NONE: <POINTOPOINT,NOARP> mtu 8832 ...
>> link/ipip 10.0.0.2 peer 10.0.0.1
>>
>> Reported-by: Petr Vorel <pvo...@suse.cz>
>> Signed-off-by: Alexey Kodanev <alexey.koda...@oracle.com>
>> ---
>>
>> ip6_vti mtu offset is the same (168) as in ip_vti because ip_vti
>> offset includes two sizes of struct iphdr: in dev->hard_header_len
>> and in t_hlen in ip_tunnel_bind_dev(). I'm not sure if it's correct.
>>
>> net/ipv6/ip6_vti.c | 18 ++++++++++++++++++
>> 1 files changed, 18 insertions(+), 0 deletions(-)
>>
>> diff --git a/net/ipv6/ip6_vti.c b/net/ipv6/ip6_vti.c
>> index dbb74f3..47e6464 100644
>> --- a/net/ipv6/ip6_vti.c
>> +++ b/net/ipv6/ip6_vti.c
>> @@ -638,6 +638,24 @@ static void vti6_link_config(struct ip6_tnl *t)
>> dev->flags |= IFF_POINTOPOINT;
>> else
>> dev->flags &= ~IFF_POINTOPOINT;
>> +
>> + if (p->flags & IP6_TNL_F_CAP_XMIT) {
>> + int strict = (ipv6_addr_type(&p->raddr) &
>> + (IPV6_ADDR_MULTICAST | IPV6_ADDR_LINKLOCAL));
>> +
>> + struct rt6_info *rt = rt6_lookup(t->net,
>> + &p->raddr, &p->laddr,
>> + p->link, strict);
>> +
>> + if (!rt)
>> + return;
>> +
>> + if (rt->dst.dev) {
>> + dev->mtu = max(rt->dst.dev->mtu - dev->hard_header_len,
>> + IPV6_MIN_MTU);
>
> Hm, I'm gettting this when compiling with your patch:
>
> In file included from /home/klassert/git/ipsec-next/include/linux/list.h:9:0,
> from /home/klassert/git/ipsec-next/include/linux/module.h:9,
> from /home/klassert/git/ipsec-next/net/ipv6/ip6_vti.c:18:
> /home/klassert/git/ipsec-next/net/ipv6/ip6_vti.c: In function
> ‘vti6_link_config’:
> /home/klassert/git/ipsec-next/include/linux/kernel.h:808:16: warning:
> comparison of distinct pointer types lacks a cast
> (void) (&max1 == &max2); \
> ^
> /home/klassert/git/ipsec-next/include/linux/kernel.h:817:2: note: in
> expansion of macro ‘__max’
> __max(typeof(x), typeof(y), \
> ^~~~~
> /home/klassert/git/ipsec-next/net/ipv6/ip6_vti.c:654:15: note: in expansion
> of macro ‘max’
> dev->mtu = max(rt->dst.dev->mtu - dev->hard_header_len,
>
rt->dst.dev->mtu and dev->hard_header_len are both unsigned and
IPV6_MIN_MTU considered as int, I guess IPV6_MIN_MTU can be changed
to dev->min_mtu as it is set to the same value in setup, but checking
in the way it is done in ip6_tnl_link_config() looks better.
I'll send 2nd version.
Thanks,
Alexey
