dhcpcd-discuss

On 21/04/2017 14:42, Stuart Henderson wrote:
> On 2017/04/21 14:25, Roy Marples wrote:
>> Hi Stuart
>>
>> On 21/04/2017 13:48, Stuart Henderson wrote:
>>> Unless I misunderstand the problem I think this may have been fixed
>>> in -current (as of sys/netinet6/in6.c r1.201, committed this week).
>>
>> It's easy to test.
>> Here is the issue on OpenBSD dhcpcd tries to work-around (present in
>> OpenBSD-6.1)
>>
>> for x in 1 2 3 4 5; do \
>> 	ifconfig xnf0 inet6 alias fd73:79::1/64 vltime 128 pltime 120;\
>> 	ifconfig xnf0 | grep inet6; \
>> 	sleep 5; \
>> done
>>
>>         inet6 fd73:79::1 prefixlen 64 tentative pltime 120 vltime 128
>>         inet6 fd73:79::1 prefixlen 64 pltime 115 vltime 123
>>         inet6 fd73:79::1 prefixlen 64 pltime 110 vltime 118
>>         inet6 fd73:79::1 prefixlen 64 pltime 105 vltime 113
>>         inet6 fd73:79::1 prefixlen 64 pltime 100 vltime 108
>>
>> As you can see, the vltime and pltime are not updated, so no matter what
>> userland does the address will always expire as will any associated route.
>>
>> Can you test this on OpenBSD -current and report back?
> 
> Yes, that's exactly what Florian's commit fixes.
> 
>         inet6 fd73:79::1 prefixlen 64 tentative pltime 120 vltime 128
>         inet6 fd73:79::1 prefixlen 64 pltime 120 vltime 128
>         inet6 fd73:79::1 prefixlen 64 pltime 120 vltime 128
>         inet6 fd73:79::1 prefixlen 64 pltime 120 vltime 128
>         inet6 fd73:79::1 prefixlen 64 pltime 120 vltime 128
> 
> 

Nice!

Is there an amd64 build of -current available? ideally an install.iso
for an easy upgrade to test.

Roy