Add configurable IPv6 prefix length to address pools (Implements #901)

docs/addressing.md

@@ -45,7 +45,7 @@ You can override or augment them in topology **addressing** element. You can als
 
 ## Dual-Stack Support
 
-Every address pool could have an IPv4 and an IPv6 prefix, supporting IPv4-only, dual-stack or IPv6-only deployments. Pool address space is specified in **ipv4** or **ipv6** CIDR prefix. Size of individual IPv4 address allocations is specified with the **prefix** parameter, IPv6 prefixes are currently fixed  (/64).
+Every address pool could have an IPv4 and an IPv6 prefix, supporting IPv4-only, dual-stack or IPv6-only deployments. Pool address space is specified in **ipv4** or **ipv6** CIDR prefix. Size of individual IPv4 address allocations is specified with the **prefix** parameter. *netlab* uses /64 IPv6 prefixes unless you change the prefix size with **prefix6** attribute.
 
 (address-pool-specs)=
 ## Specifying Address Pools
@@ -65,13 +65,15 @@ Each address pool specification is a dictionary of address pools. Individual add
 * **ipv4** -- IPv4 CIDR prefix or **true** for unnumbered IPv4 links
 * **ipv6** -- IPv6 CIDR prefix or **true** for LLA-only IPv6 links.
 * **unnumbered** -- unnumbered address pool. Interfaces attached to nodes based on this address pool will have IPv4 and/or IPv6 enabled based on the protocols enabled on node's loopback interface.
-* **prefix** -- IPv4 subnet allocation size. IPv6 subnets use /64 prefixes that cannot be changed.
+* **prefix** -- IPv4 subnet allocation size.
+* **prefix6** -- Optional IPv6 subnet allocation size. *netlab* creates /64 IPv6 subnets when the **prefix6** address pool parameter is not specified.
 * **start** -- first subnet or first IP address offset. Used primarily with **loopback** pools to ensure the first address assignment gets the x.x.x.1/32 IP address, and with **mgmt** pool to specify the first management IP address.
 * **allocation** -- [address allocation policy](addressing-tutorial-lan-links) (`id_based`, `sequential`, `p2p` or `loopback`).
 
 **Notes:**
 
 * Default IPv4 subnet allocation size is 32 for `loopback` pool, 24 for `lan` pool and 30 for `p2p` pool. All other pools must specify the **prefix** parameter whenever the **ipv4** parameter is specified.
+* Default IPv6 subnet allocation is /64. You can change this behavior with the **prefix6** pool parameter.
 * IPv4 loopback pool starts at .1 (**start** parameter is assumed to be 1)
 * IPv6 loopback pool starts at the second subnet to make loopback IPv6 address similar to its IPv4 counterpart.
 

docs/example/addr-builtin.txt

@@ -257,7 +257,7 @@ Point to point links get their prefixes from the **p2p** pool -- unless you chan
 
 IP address assignment on a point-to-point link does not use node identifiers. Nodes attached to a P2P link are sorted alphabetically, then the .1 address is assigned to the first node and .2 address to the second node. You can force *netlab* to use this allocation method by setting **allocation: p2p** in a [pool](address-pool-specs) or a **prefix**.
 
-The same process is used for IPv6 addresses even though you get a /64 prefix assigned to every P2P link. You know the reason: it was easier to do consistent address allocation across address families.
+The same process is used for IPv6 addresses even though you usually get a /64 prefix assigned to every P2P link unless you change the size of IPv6 prefixes allocated from the **p2p** pool with the **prefix6** parameter.
 
 How about a quick example: two nodes, one link:
 

docs/example/addr-ipv6.txt

@@ -62,7 +62,7 @@ nodes:
 ```
 
 ```{tip}
-The IPAM logic creates /64 IPv6 prefixes from address pools. You cannot specify target IPv6 prefix size like you could with IPv4. To build a network with non-/64 IPv6 prefixes you'll have to use [static link prefixes](#static-link-prefixes).
+The IPAM logic creates /64 IPv6 prefixes from address pools. To change the target IPv6 prefix size, use the **‌prefix6** pool attribute. For example, you could set **prefix6: 128** on the loopback pool to have /128 loopback IPv6 addresses.
 ```
 
 ### IPv6-Only Networks

docs/example/addr-static.txt

@@ -31,10 +31,6 @@ links:
     ipv6: 2001:db8:0:42:0:1::/96
 ```
 
-```{tip}
-Static link **prefix** is the only way to create IPv6 subnets with non-/64 subnet mask. We're a bit opinionated about the right way of running IPv6.
-```
-
 Finally, you can set the link **prefix** to *False* to create a link without IP addresses (you could also use **pool: l2only**, but we're talking about static prefixes here):
 
 ```

netsim/augment/addressing.py

@@ -35,7 +35,8 @@
 * _prefix_ length - prefix length of individual subnets in the pool
 
 IPv6 component of an addressing pool is defined with an _ipv6_ prefix,
-prefix length of individual subnets is assumed to be /64.
+prefix length of individual subnets is assumed to be /64 unless the
+pool has _prefix6_ attribute.
 
 Legacy setup migration
 ----------------------
@@ -219,9 +220,9 @@ def create_pool_generators(addrs: Box, no_copy_list: list) -> Box:
     for key,data in pfx.items():
       if "_pfx" in key:
         af   = key.replace('_pfx','')
-        plen = pfx['prefix'] if af == 'ipv4' else 64
+        plen = pfx['prefix'] if af == 'ipv4' else pfx.get('prefix6',64)
         gen[pool][af] = data.subnet(plen)
-        if (af == 'ipv4' and plen == 32) or (pool == 'loopback'):
+        if (af == 'ipv4' and plen == 32) or (af == 'ipv6' and plen >= 127) or (pool == 'loopback'):
           next(gen[pool][af])
       elif not key in no_copy_list:
         gen[pool][key] = data

netsim/augment/nodes.py

@@ -357,7 +357,10 @@ def transform(topology: Box, defaults: Box, pools: Box) -> None:
             log.fatal( f"Loopback addresses must be valid IP prefixes, not 'True': {prefix_list}" )
         elif not n.loopback[af]:
           if af == 'ipv6':
-            n.loopback[af] = addressing.get_addr_mask(prefix_list[af],1)
+            if prefix_list[af].prefixlen == 128:
+              n.loopback[af] = str(prefix_list[af])
+            else:
+              n.loopback[af] = addressing.get_addr_mask(prefix_list[af],1)
           else:
             n.loopback[af] = str(prefix_list[af])
           n.af[af] = True

netsim/defaults/attributes.yml

@@ -83,6 +83,7 @@ pool:
   ipv6: { type: ipv6, use: prefix }
   start: { type: int, min_value: 1 }
   prefix: { type: int, min_value: 1, max_value: 32 }
+  prefix6: { type: int, min_value: 1, max_value: 128 }
   allocation: { type: str, valid_values: [ p2p, sequential, id_based ] }
   mac: mac
   unnumbered: bool