[LTP] [PATCH] netns: Fix duplicate address detection (dad)

Li Wang liwang@redhat.com
Wed Nov 22 04:09:33 CET 2017


On Wed, Nov 22, 2017 at 5:36 AM, Dan Rue <dan.rue@linaro.org> wrote:
> Symptoms (+ command, error):
>     netns_comm_ip_ipv6_ioctl:
>         + ip netns exec tst_net_ns1 ping6 -q -c2 -I veth1 fd00::2
>         connect: Cannot assign requested address
>
>     netns_comm_ip_ipv6_netlink:
>         + ip netns exec tst_net_ns0 ping6 -q -c2 -I veth0 fd00::3
>         connect: Cannot assign requested address
>
>     netns_comm_ns_exec_ipv6_ioctl:
>         + ns_exec 6689 net ping6 -q -c2 -I veth0 fd00::3
>         connect: Cannot assign requested address
>
>     netns_comm_ns_exec_ipv6_netlin:
>         + ns_exec 6891 net ping6 -q -c2 -I veth0 fd00::3
>         connect: Cannot assign requested address
>
> The error is coming from ping6, which is trying to get an IP address for
> veth0 (due to -I veth0), but cannot, because the interface is not
> immediately available due to dad. The existing code to disable dad is
> disables it for the virtual interfaces, but it needs to be disabled for
> all interfaces in the network namespace to be effective.
>
> Also, changed it to use sysctl -w because I think it is a little cleaner
> to read and understand.
>
> Lastly, dropped setting accept_ra. It does not seem to be necessary to
> disable.
>
> Signed-off-by: Dan Rue <dan.rue@linaro.org>
> ---
>  testcases/kernel/containers/netns/netns_helper.sh | 10 ++++------
>  1 file changed, 4 insertions(+), 6 deletions(-)
>
> diff --git a/testcases/kernel/containers/netns/netns_helper.sh b/testcases/kernel/containers/netns/netns_helper.sh
> index a95cdf206..95add0544 100755
> --- a/testcases/kernel/containers/netns/netns_helper.sh
> +++ b/testcases/kernel/containers/netns/netns_helper.sh
> @@ -256,12 +256,10 @@ netns_set_ip()
>         # there is no other host with the same address, the address is
>         # considered to be "tentative" (attempts to bind() to the address fail
>         # with EADDRNOTAVAIL) which may cause problems for tests using ipv6.
> -       echo 0 | $NS_EXEC $NS_HANDLE0 $NS_TYPE \
> -               tee /proc/sys/net/ipv6/conf/veth0/accept_dad \
> -               /proc/sys/net/ipv6/conf/veth0/accept_ra >/dev/null
> -       echo 0 | $NS_EXEC $NS_HANDLE1 $NS_TYPE \
> -               tee /proc/sys/net/ipv6/conf/veth1/accept_dad \
> -               /proc/sys/net/ipv6/conf/veth1/accept_ra >/dev/null
> +       $NS_EXEC $NS_HANDLE0 $NS_TYPE sysctl -w net.ipv6.conf.all.accept_dad=0
> +       $NS_EXEC $NS_HANDLE0 $NS_TYPE sysctl -w net.ipv6.conf.veth0.accept_dad=0

Seems we have already disabled DAD for all interfaces via
'net.ipv6.conf.all.accept_dad=0', I wonder if it's necessary to do the
repeat setting for veth0 device?

> +       $NS_EXEC $NS_HANDLE1 $NS_TYPE sysctl -w net.ipv6.conf.all.accept_dad=0
> +       $NS_EXEC $NS_HANDLE1 $NS_TYPE sysctl -w net.ipv6.conf.veth1.accept_dad=0

Here as well.


>
>         case $USE_IFCONFIG in
>         1)
> --
> 2.14.3

Anyway, this method works to me.



-- 
Li Wang
liwang@redhat.com


More information about the ltp mailing list