Files
napalm-openwrt/CHANGELOG.md
T
Christian Manivong 0552663d53
driver / standard (pull_request) Successful in 24s
driver / lint (pull_request) Successful in 24s
driver / types (pull_request) Successful in 1m28s
driver / test (3.13) (pull_request) Successful in 1m12s
driver / test (3.12) (pull_request) Successful in 1m21s
driver / lowest (pull_request) Successful in 29s
driver / build (pull_request) Successful in 30s
driver / security (pull_request) Successful in 1m0s
CI / driver (pull_request) Successful in 2m38s
driver / publish (pull_request) Skipped
fix(wireless): raise when a uci step of an ACL or channel push is refused
push_mac_acl and push_radio_channel sent each step through the
interactive shell and never looked at the result: a refused uci set,
add_list, commit or wifi reload passed silently, an SSID no wifi-iface
carried did nothing, and wifi was restarted after a refused channel.

Each step now runs on an SSH exec channel of its own, as root, with its
exit status checked (uci_steps.UciStepsMixin). A refused step raises
CommandErrorException naming the command and what the router said; a
refusal before the commit reverts what was staged, so a later commit
does not apply half of it, and wifi is not restarted. push_mac_acl
reads the sections once, raises when none carries the SSID, and deletes
macfilter/maclist only where a section has them instead of hiding each
delete's status behind `|| true`.

Closes #35
2026-11-12 19:01:07 +01:00

9.0 KiB
Raw Blame History

Changelog

All notable changes to napalm-openwrt. The format follows Keep a Changelog, versions follow Semantic Versioning. Versions up to 0.6.0 were not recorded here; their history is the git log.

[Unreleased]

Changed

  • Needs napalm-device-types 4.36.0 or later (TimeSyncDict.dhcp_ntp_servers).
  • A time sync write no longer sets use_dhcp=0, as the first time sync commit (#36) did; a router that commit wrote keeps use_dhcp=0 until it is set back.

Added

  • Time sync (napalm-device-types TimeSyncMixin, NetOrk/netork#2364): get_time_sync() reads the router's NTP servers and its zone from uci -q show system; configure_time_sync() writes both. The servers replace system.ntp.server and enable the client; sysntpd restarts. use_dhcp stays as the router has it: the servers of a DHCP lease (option 42) count (NetOrk/netork#2380). dhcp_ntp_servers names them -- what sysntpd adds while use_dhcp is on, each interface's data.ntpserver in ubus call network.interface dump (only the dhcp_interface list's where it names any), read in the same round trip; None where netifd gives no answer. The zone is written as its name (zonename, in LuCI's spelling) and its POSIX rule (timezone, from napalm-device-types' posix_tz), then /etc/init.d/system reload and a restart of logd apply it -- the rule, not the name, is what libc reads (NetOrk/netork#1849). A value the router already has is not written again and nothing restarts. The zone reads by its name only where the rule in force is that zone's: a router whose rule differs from its zonename, or that has a rule only, reads as the one zone with that rule, or as no zone where several share it. synchronized is None: BusyBox ntpd keeps no state to read.
  • get_host_status() (napalm-device-types HostStatusMixin), over an SSH exec channel: whether the router needs a reboot, how full its filesystems are, and -- with napalm-device-types 4.32 or newer -- whether apk (OpenWrt 25.12) or opkg (before) was left halfway, and which of the two it read (NetOrk/netork#851).
  • Scheduled jobs (NetOrk/netork#1933, #32), through napalm-device-types' PosixScheduledJobsMixin: get_scheduled_jobs() reads every crontab in /etc/crontabs -- BusyBox crond has no /etc/crontab, cron.d, run-parts, anacron or systemd -- with the time zone from UCI and each crontab's package from opkg; manage_scheduled_job() enables, disables or runs one now; put_scheduled_job() and delete_scheduled_job() write and remove an application's own job as marked lines in the user's crontab. Every write changes only the job's own line, is read back (BusyBox crontab exits 0 when it could not write) and leaves cron enabled and restarted. A run is the command as BusyBox crond runs it (% is no stdin), bounded by timeout where the router has it (OpenWrt 25.x); another user's job runs through start-stop-daemon -c, as OpenWrt's BusyBox has no su. @daily and the other aliases are written as five fields; @reboot is refused, as OpenWrt's crond has no special times.
  • configure_lldp_agent(interfaces) (napalm-device-types LldpAgentMixin): installs lldpd where it is missing, writes the interfaces given as lldpd's interface list in UCI, enables lldpd at boot and restarts it. A VLAN device (br-ap.10) is written as the device under it (br-ap). It is the action for the setup the LLDP getters used to do on every poll (#8). Needs napalm-device-types with LldpAgentMixin.

Changed

  • get_ntp_servers() and get_ntp_peers() read over an SSH exec channel, with the same reading as get_time_sync(), and name no servers while the NTP client is disabled (system.ntp.enabled='0'): sysntpd then starts no ntpd.
  • The floor on napalm-device-types is 4.35.0, the first release with TimeSyncMixin (it was 4.26.0 for PosixScheduledJobsMixin, and 4.24.0 for the shared lldpctl parser before that).
  • Python 3.12 is the floor (requires-python = ">=3.12"): netOrk runs every driver on 3.12, so 3.10 and 3.11 are no longer tested (NAPALM/driver-standard#6).
  • The floor on napalm is 5.1: older releases import pkg_resources, which a Python 3.12 venv no longer has.
  • get_lldp_neighbors() and get_lldp_neighbors_detail() only read. They no longer enable and start a stopped lldpd, and no longer rewrite lldpd's interface list and restart it, which could happen on every poll. A device whose lldpd is stopped now reports no neighbours until lldpd is started, for example with manage_service("lldpd", "enable") and manage_service("lldpd", "start") (#8).

Fixed

  • push_mac_acl() and push_radio_channel() raise when the router refuses a step (CommandErrorException, naming the uci command or the wifi restart and what the router said). They used to return normally after a refused uci set, add_list, commit or wifi reload, so an ACL that never landed looked applied. Each step runs on an SSH exec channel of its own and as root, with its exit status checked; a refused change reverts what was staged, so nothing of it is committed by a later write, and wifi is not restarted for it. push_mac_acl() raises when no wifi-iface carries the SSID, where it did nothing, and deletes macfilter/maclist only where a section has them instead of hiding every delete's status behind || true (#35).
  • get_ntp_servers() names every server of the system.ntp.server list; it kept only the first, as uci prints a list as 'a' 'b' 'c' on one line.
  • get_lldp_neighbors() and get_lldp_neighbors_detail() keep an interface name with a dot whole: the neighbours of a VLAN interface such as eth0.1 used to come back under eth0 with every field empty (#26). A neighbour that names its port by MAC address or a local id is reported with that port instead of an empty one, and a chassis named by a local id or an address with that id. The detail now also carries the neighbour's system description and capabilities, which were always empty. lldpd's output is parsed by napalm-device-types' shared parser.
  • get_lldp_neighbors() reports a neighbour's chassis MAC as mac (lower case) when the neighbour names its chassis by one, as napalm-linux does, so netOrk can match the neighbour by MAC rather than by name only (#31).
  • get_lldp_neighbors() and get_lldp_neighbors_detail() raise when lldpctl fails (not installed, lldpd not running or unreachable), with lldpctl's message; they used to return no neighbours, the same answer as a device without any (#27). An empty result now means lldpctl ran and listed none. get_environment() raises when /proc/stat or /proc/meminfo cannot be read or parsed, instead of reporting 0 % CPU and 0 bytes of memory.
  • get_arp_table() reports an unknown age as -1.0 (napalm-device-types' ARP_AGE_UNKNOWN) instead of 0.0, which means "confirmed just now": neither ip neigh nor /proc/net/arp has an age (NetOrk/netork#1327).
  • The driver no longer removes port from the optional_args dict it is given: the caller's dict stays as it was (#19).
  • manage_service() reports a failed action or a missing init script as failed: only the init script's exit status decides, where any action used to come back as a success (#3).
  • apply_updates() reports a failed upgrade as failed, with the package manager's exit status in error; it used to report every upgrade as done (#11).
  • get_environment() reports the CPU usage of the last second, from two reads of /proc/stat one second apart; it used to report the average since boot, so a box busy for the last minute could show a low figure (#10).
  • get_lldp_neighbors() and get_lldp_neighbors_detail() report every neighbour on a port (a phone with a PC behind it, a switch without LLDP of its own); only the last one came back, mixed with fields of the others (#9).
  • install_package() and uninstall_package() report a failed install or removal as failed: only the package manager's exit status decides. They used to search the output for words like "error:", so an opkg dependency failure ("Collected errors:", "Cannot install package …") came back as a success, and a successful install of a package whose name contains "failed" as a failure (#28).

[0.7.0] - 2026-10-23

The first release on PyPI.

Changed

  • The driver follows the NAPALM driver standard (https://git.netork.io/NAPALM/driver-standard): CI on Gitea runs ruff, mypy, the tests on Python 3.10–3.13 and at the lowest declared versions, bandit, pip-audit and a build check, and a release tag publishes to PyPI. The package now ships py.typed and a LICENSE file. Python 3.8 and 3.9 are no longer supported; napalm-device-types never ran on them.
  • napalm-device-types 4.21.1 or newer is required.
  • Exceptions the driver deliberately ignores (the snmpd community in get_snmp_config(), the network VLANs in get_ssids()) are now logged at debug level.
  • The project links point at git.netork.io instead of GitHub.