> On Oct 24, 2019, at 4:01 PM, Lauri Tirkkonen <[email protected]> wrote:
>
> On Thu, Oct 24 2019 11:15:08 -0400, [email protected] wrote:
>>>> On Oct 24, 2019, at 8:25 AM, Lauri Tirkkonen <[email protected]> wrote:
>>>
>>> On Thu, Oct 24 2019 11:06:04 +0200, Martin Pieuchot wrote:
>>>> Could you repeat the steps to make that issue arise while capturing
>>>> routing socket messages with "# route monitor" on another terminal?
>>>>
>>>> If you do that before and after the commit it might help figuring out
>>>> what's different.
>>>
>>> Certainly. Though this is bsd.rd, so there is no another terminal :) I
>>> did the following instead on the same terminal:
>>>
>>> # route monitor >/tmp/route-monitor.good &
>>> # dhclient vio0
>>> # ping 1.1.1.1
>>
>> Can you add a "ifconfig vio0" at this point,
>
> Using the same bsd.rd's:
>
> $ cat ifconfig.bad.dhclient
> vio0: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> mtu 1500
> lladdr 96:00:00:34:06:8e
> llprio 3
> groups: egress
> media: Ethernet autoselect
> status: active
> inet 95.217.7.62 netmask 0xffffffff
> $ cat ifconfig.good.dhclient
> vio0: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> mtu 1500
> lladdr 96:00:00:34:06:8e
> llprio 3
> groups: egress
> media: Ethernet autoselect
> status: active
> inet 95.217.7.62 netmask 0xffffffff
>
>>> ^C
>>> # ifconfig vio0 -soii
>>> # ping 1.1.1.1
>>
>> and here.
>
> $ cat ifconfig.bad.-soii
> vio0: flags=408843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST,INET6_NOSOII> mtu
> 1500
> lladdr 96:00:00:34:06:8e
> llprio 3
> groups: egress
> media: Ethernet autoselect
> status: active
> inet 95.217.7.62 netmask 0xffffffff
> $ cat ifconfig.good.-soii
> vio0: flags=408843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST,INET6_NOSOII> mtu
> 1500
> lladdr 96:00:00:34:06:8e
> llprio 3
> groups: egress
> media: Ethernet autoselect
> status: active
> inet 95.217.7.62 netmask 0xffffffff
>
>> I am puzzled that the AUTOCONF4 flag appears to be missing from IFINFO
>> messages after dhclient has configured the interface. Just doing -soii
>> doesn't remove the flag here.
>>
>> The fact that AUTOCONF4 has been 'lost' triggers getting a new IPv4 lease.
>> No idea why this would cause the problem being reported.
>
> I think AUTOCONF4 is missing because it wasn't added until July (from my
> reading of the commit logs), so the images under test, being built from
> commits in May, just don't contain that code yet :)
Aha. :-)
>
> I can of course repeat any of these tests on a -CURRENT system vs. a
> -CURRENT system that has the commit that introduced the problem
> reverted, if you like.
Yes, please.
.... Ken
>
> --
> Lauri Tirkkonen | lotheac @ IRCnet
