AccessPointDriver used to declare get_services, manage_service,
get_available_updates, apply_updates, get_packages, install_package and
remove_package as NotImplementedError stubs. Those stubs preceded
OpenWrtSystemMixin and OpenWrtPackageMixin in the MRO and shadowed their working
implementations, so this driver carried a forwarder for each one purely to
delegate past the base.
napalm-device-types v1.0 makes role bases declaration-only, so nothing shadows
anything and all seven forwarders are dead weight. Every method now resolves
straight to the mixin that implements it.
remove_package went with them: the base now calls the method by the name the
package mixin already used, uninstall_package, so the name adapter is gone too.
fix_snmp reported success on APs where the rule never reached nftables.
Five defects stacked up:
1. Zone detection required ".src=" and "ssh" in the same `uci show` line.
UCI prints one option per line, so anonymous rules never matched and
every device fell through to the hardcoded "lan" fallback.
2. That fallback was never checked against the zones that actually exist.
On an AP whose zone section has no `option name`, fw4 skips the section,
so `src='lan'` referenced a zone that was not there and the rule was
dropped with it.
3. The "already present" guard was a substring test, so a rule written by
an earlier broken run was skipped forever instead of repaired.
4. Stale-rule deletion never committed — the only `uci commit firewall`
sat in the add branch that the guard had just skipped.
5. `fw4 reload` errors were swallowed by `|| true`, and with no local
snmpget the action hardcoded success = True.
Now: the management address comes from $SSH_CONNECTION and is mapped to
its network section (via ipaddr, or via `ip -o -4 addr` -> device when the
interface is DHCP-addressed) and from there to the owning zone. A zone
section without a name aborts the action with the repair command rather
than writing a dead rule — naming it is left to the operator, since an
inert zone becoming active changes what the AP filters. Rules are written
in full every run, stale ones are deleted highest anonymous index first
(uci renumbers @rule[n] on delete) and committed, reload output is no
longer truncated or ignored, and success is verified on the device via
`ss -lun` and a udp/161 lookup in the live ruleset.
NetworkDriver (parent of AccessPointDriver) raises NotImplementedError for all
standard NAPALM methods. With AccessPointDriver listed first, get_interfaces()
and get_vlans() from the mixins were shadowed and always raised NotImplementedError
(empty message) — causing all AP polls to report 0 interfaces and 0 VLANs.
Fix: reorder inheritance so mixins come before AccessPointDriver. The standard
NAPALM method stubs in NetworkDriver are now shadowed by the mixin implementations.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
board_name format is "vendor,model" (e.g. "sophos,ap100"). The prefix
before the comma is title-cased to give the manufacturer name (e.g.
"Sophos"). Falls back to "OpenWrt" when board_name is unavailable.
The model is already read from /tmp/sysinfo/model which gives the full
human-readable name (e.g. "Sophos AP100") — no change needed there.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
The previous attempt added a rule without src= which lands in the global
nftables input chain. Traffic from the management interface (br-ap.10)
jumps immediately to input_mgmt, so the global rule was never reached.
Now detects the management zone name by finding the zone whose allow-SSH
rule already exists (the named rule pattern allow_ssh_from_<zone>).
Removes any mis-scoped previous SNMP rule, then adds a named UCI rule
allow_snmp_from_<zone> with src=<zone> so fw4 places it in the correct
chain (input_mgmt). Persists across reboots via uci commit + fw4 reload.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
The method was called but missing — always raised AttributeError, so
Fix SNMP did nothing on OpenWRT devices.
OpenWRT's default firewall (fw4) policy-drops everything except the
ports explicitly listed in the management zone (SSH/HTTP/HTTPS/ICMP).
SNMP (UDP/161) is not in that list, which is why snmpd runs but is
unreachable from outside the device.
Fix adds a persistent UCI firewall rule for UDP/161, reloads fw4
immediately, verifies snmpd is running, and probes locally if SNMP
client tools are available. Returns success so callers can clear the
snmp_no_data warning.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Added _lldpd_fix_interface() helper that detects the management interface
via the default route and strips any .VID suffix (e.g. br-ap.10 → br-ap).
LLDP is L2 and must run on the bridge itself — sending on a VLAN subinterface
produces tagged frames the switch won't recognize as LLDP.
The helper runs every poll so existing wrong configs (e.g. eth0 from the
original install action) are corrected automatically on the next poll.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
opkg list-upgradable and apk version output contains version strings and
comparison operators; meta.packages now stores only the bare package name
so the schedule-updates API validation passes.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Install snmpd-nossl + luci-app-snmpd via opkg, configure via UCI with correct
field names (group/viewname/context='none'), bare port 161, stop+pkill before
start to break crash loops. get_snmp_config() reads current UCI snmpd state.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Christian Manivong