Invalid subnet masks

[Jim Thompson]

> On Feb 11, 2015, at 4:51 AM, Julian Elischer <julian at freebsd.org> wrote:
> 
>> On 2/11/15 5:55 PM, Matt Churchyard wrote:
>> 
>> I appreciate that it might be 'valid' as a binary mask, but I'm struggling to find any documentation anywhere that actually suggests that it's valid as a network configuration. The entire modern CIDR notation, and all the routing system & hardware built around it (that shows networks in CIDR form and will collapse routes) has no way of dealing with these subnets.
> most can deal with it, just not optimally
>> 
>> Are there actually valid use cases for these types of network?
> yes.
> I've had networks that were the first and last quarter of a /24, and the middle two quarters were separate nets.
> 
> Sure, it made my skin crawl, but I was in a pinch to get more machines onto that /26.
> all four were served by the same router so only one router needed to know..
> 
> I have however at times though we could think about making ifconfig at give a warning.
> (but not an error).

https://lists.freebsd.org/pipermail/freebsd-hackers/2011-April/034997.html
Subject came up on -hackers in 2011

Quoting RFC-1219:

"While RFC-950 allows the "ones" in the subnet mask to be non-contiguous, RFC-950 recommends that 1) they be contiguous, and 2) that they occupy the most significant bits of the "host" part of the internet address."

Jim