Summary: sntp: Allow both IPv4 and IPv6 NTP servers to be
configured simultaneously via DHCP
Project: lwIP - A Lightweight TCP/IP stack
Submitted by: davefiddes
Submitted on: Sun 29 Sep 2019 05:00:29 PM UTC
Priority: 3 - Low
Assigned to: None
Discussion Lock: Any
Planned Release: None
Currently when a DHCP or DHCPv6 updates containing NTP
server addresses is processed all existing NTP server addresses
are wiped. This can lead to problems when there is intermittent
connectivity to an NTP server over a given IP protocol. This
changes DHCP update processing such that it only overwrites
server addresses for the same protocol. It also does the same
for IPv4 NTP server addresses when processing a DHCPv6 update.
The logging for processing DHCP and DHCPv6 updates are now
a bit more useful.
- On a network with DHCP and DHCPv6 configure differing
numbers (0-3) of NTP servers on each protocol. Ensure that
as updates are processed the list of 3 LWIP NTP servers is
updated as expected.
- On a network with intermittent IPv6 connectivity ensure that
with both IPv4 and IPv6 servers configured via DHCP that NTP
updates are requested over IPv4 even if it is not listed first.
[patch #9856] sntp: Allow both IPv4 and IPv6 NTP servers to be configured simultaneously via DHCP
Follow-up Comment #1, patch #9856 (project lwip):
I could imagine a lot of funny things happen when applying this patch as is.
For example, if the number of available slots was smaller than what we get via
DHCPv6, we wouldn't be able to store IPv4 servers?
Up to now, the assumption was that networks either give us IPv4 or IPv6
If that's not the case, maybe separate lists (or fixed offsets in a single
list) would be better than this kind of magic?