ip: a /32 mask is refused on a port, in the vendor's own sentence #1921

Merged
jaylouisw merged 1 commit from rev/1118-ip32-solo into main 2026-11-18 01:54:59 +00:00
Owner

A 32-bit mask names a subnet with one address and no host space, so a box will not put one on a link that has to reach another box. Cisco's addressing guide: 255.255.255.255 "describes a subnet with only one IPv4 host address ... strictly reserved for use on links that can have only one address. The example for Cisco routers is the loopback interface".

Three public transcripts, three platforms, print the same refusal for the same command - a GRE tunnel (Bad mask /32 for address 10.0.0.1), a 2811 serial port on 15.0(1)M10, and a 4331 routed port on IOS XE (Bad mask /32 for address 12.0.0.1).

This console accepted the address and stored it, so show ip interface brief and show running-config carried a /32 no device would have. _ipHostRefusal() now answers with the one line the network- and broadcast-address cases already use, before anything is stored or the subnet compared, and stands down for an interface whose type is loopback (the vendor's exception, which no catalogue device currently declares).

  • research/ip-address-not-a-host.md: the /32 moves out of "unresolved"; four sources added; what is still open recorded (an SVI or port-channel, and the loopback exception the model cannot demonstrate).
  • test/ip-address-not-a-host.test.js: 27 checks (was 23) - the /32 line on a port, the bare-prefix spelling, the held address and the running-config untouched, a /31 still taken at the boundary, a dot1q subinterface refused like its parent.
  • docs/CLI.md, CHANGELOG.md say the same.

Gates: node scripts/syntax-check.js (401 files), npm test 328 suites / 11407 assertions / 0 failed.

A 32-bit mask names a subnet with one address and no host space, so a box will not put one on a link that has to reach another box. Cisco's addressing guide: `255.255.255.255` "describes a subnet with only one IPv4 host address ... strictly reserved for use on links that can have only one address. The example for Cisco routers is the loopback interface". Three public transcripts, three platforms, print the same refusal for the same command - a GRE tunnel (`Bad mask /32 for address 10.0.0.1`), a 2811 serial port on 15.0(1)M10, and a 4331 routed port on IOS XE (`Bad mask /32 for address 12.0.0.1`). This console accepted the address and stored it, so `show ip interface brief` and `show running-config` carried a /32 no device would have. `_ipHostRefusal()` now answers with the one line the network- and broadcast-address cases already use, before anything is stored or the subnet compared, and stands down for an interface whose type is `loopback` (the vendor's exception, which no catalogue device currently declares). - `research/ip-address-not-a-host.md`: the /32 moves out of "unresolved"; four sources added; what is still open recorded (an SVI or port-channel, and the loopback exception the model cannot demonstrate). - `test/ip-address-not-a-host.test.js`: 27 checks (was 23) - the /32 line on a port, the bare-prefix spelling, the held address and the running-config untouched, a `/31` still taken at the boundary, a dot1q subinterface refused like its parent. - `docs/CLI.md`, `CHANGELOG.md` say the same. Gates: `node scripts/syntax-check.js` (401 files), `npm test` 328 suites / 11407 assertions / 0 failed.
ip: a /32 mask is refused on a port, in the vendor's own sentence
All checks were successful
CI / Classify the change (pull_request) Successful in 6s
PR rebase / Rebase onto main and arm auto-merge (pull_request) Successful in 9s
CI / Server smoke test (pull_request) Successful in 5s
CI / Docker build (pull_request) Successful in 20s
CI / Test (Node 20) (pull_request) Successful in 47s
CI / Test (Node 22) (pull_request) Successful in 51s
CI / Test (Node 18) (pull_request) Successful in 54s
CI / Arm auto-merge (pull_request) Successful in 5s
9f23996404
A 32-bit mask names a subnet with one address and no host space, so a box
will not put one on a link that has to reach another box. Cisco's guide:
"255.255.255.255 ... describes a subnet with only one IPv4 host address.
These subnets cannot be used to assign address to network links ... strictly
reserved for use on links that can have only one address. The example for
Cisco routers is the loopback interface."

Three transcripts, three platforms, print the refusal for the same command:
a GRE tunnel (`Bad mask /32 for address 10.0.0.1`), a 2811 serial port on
15.0(1)M10 (`Bad mask /32 for address 123.123.123.123`) and a 4331 routed
port on IOS XE (`Bad mask /32 for address 12.0.0.1`).

This console accepted the address and stored it, so `show ip interface brief`
and `show running-config` carried a /32 no device would have. `_ipHostRefusal()`
now answers with the same one line the network- and broadcast-address cases
use, before anything is stored or the subnet compared, and stands down for an
interface whose type is `loopback` - the vendor's exception, which no catalogue
device declares.

research/ip-address-not-a-host.md moves the /32 out of "unresolved", adds the
four sources, and records what is still open (an SVI or port-channel, and the
loopback exception the model cannot demonstrate). docs/CLI.md and CHANGELOG.md
say the same.

test/ip-address-not-a-host.test.js: 27 checks (was 23) - the /32 line on a
port, the bare-prefix spelling, the held address and the running-config
untouched, a /31 still taken at the boundary, and a dot1q subinterface refused
like its parent. Full suite: 329 suites, 11435 assertions, 0 failed.
jaylouisw force-pushed rev/1118-ip32-solo from 9f23996404
All checks were successful
CI / Classify the change (pull_request) Successful in 6s
PR rebase / Rebase onto main and arm auto-merge (pull_request) Successful in 9s
CI / Server smoke test (pull_request) Successful in 5s
CI / Docker build (pull_request) Successful in 20s
CI / Test (Node 20) (pull_request) Successful in 47s
CI / Test (Node 22) (pull_request) Successful in 51s
CI / Test (Node 18) (pull_request) Successful in 54s
CI / Arm auto-merge (pull_request) Successful in 5s
to 3a0c10888f
All checks were successful
CI / Classify the change (pull_request) Successful in 7s
PR rebase / Rebase onto main and arm auto-merge (pull_request) Successful in 5s
CI / Server smoke test (pull_request) Successful in 6s
CI / Docker build (pull_request) Successful in 20s
CI / Test (Node 20) (pull_request) Successful in 52s
CI / Test (Node 18) (pull_request) Successful in 59s
CI / Test (Node 22) (pull_request) Successful in 54s
CI / Arm auto-merge (pull_request) Successful in 7s
2026-11-18 01:53:03 +00:00
Compare
Sign in to join this conversation.
No reviewers
No milestone
No project
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference
jaylouisw/packetforge!1921
No description provided.