fix(charts): filter default-route discovery helpers to IPv4 family (#145)

* test(charts): pin IPv4-only contract on default-route discovery helpers

Add two failing tests covering the dual-stack regression:

- TestDefaultGatewayIsIPv4OnDualStack: helper-level assertion that
  talm.discovered.default_gateway returns the IPv4 default-route gateway
  on a node with both IPv4 and IPv6 default routes (IPv6 ordered first
  in discovery, mirroring real Hetzner-style nodes).

- TestCozystackChartRendersIPv4GatewayOnDualStack: end-to-end assertion
  on the cozystack v1.12 multi-doc chart output. Both LinkConfig.routes
  gateway and LinkConfig.addresses must use the IPv4 default-route data
  even when an IPv6 default route also exists.

Both fail today: default_gateway and default_addresses_by_gateway both
iterate routes and break at the first default route without filtering
by family, so an IPv6-first discovery yields:

  - gateway: fe80::1 paired with an implicit IPv4 destination, which
    Talos rejects as a malformed route (the headline VIP-not-addressable
    symptom users hit).
  - addresses: empty, because the IPv6 family selected by the first
    default route filters out the node's IPv4 addresses, leaving the
    primary NIC unconfigured.

Add a dualStackNicLookup() fixture next to simpleNicLookup() so the
chart-rendering test path can exercise dual-stack discovery without
disturbing the existing IPv4-only fixtures the rest of the suite
depends on.

Assisted-By: Claude <noreply@anthropic.com>
Signed-off-by: Aleksei Sviridkin <f@lex.la>

* fix(charts): filter default-route discovery helpers to IPv4 family

talm.discovered.default_gateway and talm.discovered.default_addresses_by_gateway
now skip routes whose family is not inet4. The chart consumers (cozystack
and generic) pair the returned gateway with a hardcoded IPv4 destination
(network: 0.0.0.0/0 on the legacy schema, or no network field on the typed
RouteConfig schema where Talos defaults to IPv4), and the addresses helper
filters per-link addresses by the route's family.

Without the filter, a node with both IPv4 and IPv6 default routes (the
typical Hetzner / dual-stack setup) ends up with:

  - gateway: <ipv6 link-local> in a route block whose destination is
    IPv4 — Talos rejects the malformed entry and dependent features
    (Layer2 VIP, default routing through the rendered chart config)
    silently break.
  - addresses: empty in the rendered LinkConfig because the helper
    inherits the IPv6 family from the first iterated default route and
    the address filter then drops every IPv4 entry on the link.

The IPv4-only filter is symmetric with gateway_by_link, which already
documents "IPv4-only by convention to avoid family/address mismatch on
dual-stack nodes". Doc comments updated to spell out the chain so a
future maintainer who adds an IPv6-aware variant for new typed routes
does not silently widen the existing helpers and re-introduce the
regression.

Assisted-By: Claude <noreply@anthropic.com>
Signed-off-by: Aleksei Sviridkin <f@lex.la>

---------

Signed-off-by: Aleksei Sviridkin <f@lex.la>

Author: lexfrei

charts/cozystack/values.yaml

@@ -31,8 +31,8 @@ clusterDomain: cozy.local
 floatingIP: ""
 
 # Optional override for the link Layer2VIPConfig is pinned to. When
-# left empty the chart picks the default-gateway-bearing link the
-# node already has (the VLAN sub-interface if one carries the
+# left empty the chart picks the IPv4-default-gateway-bearing link
+# the node already has (the VLAN sub-interface if one carries the
 # default route, otherwise the physical NIC). Set this when the
 # target link does not yet exist on the live system at first apply
 # -- typically a VLAN sub-interface that the same template is about
@@ -57,12 +57,14 @@ serviceSubnets:
 
 # Optional override for machine.kubelet.nodeIP.validSubnets and
 # cluster.etcd.advertisedSubnets. When left empty the chart derives
-# the value from the node's default-gateway-bearing link at render
-# time (via talm.discovered.default_addresses_by_gateway), so the
-# generated machine config matches the node's actual network without
-# any values.yaml edit. Set this only when you want to pin a specific
-# subnet — typically for multi-homed nodes where the default-gateway
-# link is not the subnet you want kubelet/etcd to use.
+# the value from the node's IPv4-default-gateway-bearing link at
+# render time (via talm.discovered.default_addresses_by_gateway), so
+# the generated machine config matches the node's actual network
+# without any values.yaml edit. Set this only when you want to pin a
+# specific subnet — typically for multi-homed nodes where the
+# IPv4-default-gateway link is not the subnet you want kubelet/etcd
+# to use, or for IPv6-only setups where the chart's IPv4 default
+# would otherwise leave validSubnets empty.
 # Example:
 #   advertisedSubnets:
 #     - "10.0.0.0/8"

charts/generic/values.yaml

@@ -19,8 +19,8 @@ endpoint: ""
 floatingIP: ""
 
 # Optional override for the link Layer2VIPConfig is pinned to. When
-# left empty the chart picks the default-gateway-bearing link the
-# node already has (the VLAN sub-interface if one carries the
+# left empty the chart picks the IPv4-default-gateway-bearing link
+# the node already has (the VLAN sub-interface if one carries the
 # default route, otherwise the physical NIC). Set this when the
 # target link does not yet exist on the live system at first apply
 # — typically a VLAN sub-interface that the same template is about
@@ -45,12 +45,14 @@ serviceSubnets:
 
 # Optional override for machine.kubelet.nodeIP.validSubnets and
 # cluster.etcd.advertisedSubnets. When left empty the chart derives
-# the value from the node's default-gateway-bearing link at render
-# time (via talm.discovered.default_addresses_by_gateway), so the
-# generated machine config matches the node's actual network without
-# any values.yaml edit. Set this only when you want to pin a specific
-# subnet — typically for multi-homed nodes where the default-gateway
-# link is not the subnet you want kubelet/etcd to use.
+# the value from the node's IPv4-default-gateway-bearing link at
+# render time (via talm.discovered.default_addresses_by_gateway), so
+# the generated machine config matches the node's actual network
+# without any values.yaml edit. Set this only when you want to pin a
+# specific subnet — typically for multi-homed nodes where the
+# IPv4-default-gateway link is not the subnet you want kubelet/etcd
+# to use, or for IPv6-only setups where the chart's IPv4 default
+# would otherwise leave validSubnets empty.
 # Example:
 #   advertisedSubnets:
 #     - "10.0.0.0/8"

charts/talm/templates/_helpers.tpl

@@ -91,11 +91,19 @@
 {{- end }}
 {{- end }}
 
+{{- /* JSON list of IPv4 addresses on the link carrying the IPv4
+       default route. Filters routes to family=inet4 because consumers
+       (cozystack and generic chart helpers) pair the returned addresses
+       with a hardcoded IPv4 destination route — selecting the first
+       default route regardless of family on a dual-stack node would
+       cascade into the IPv6 family being applied to the address filter,
+       and the node's IPv4 addresses would silently disappear from the
+       rendered config. Mirrors gateway_by_link's IPv4-only convention. */ -}}
 {{- define "talm.discovered.default_addresses_by_gateway" }}
 {{- $linkName := "" }}
 {{- $family := "" }}
 {{- range (lookup "routes" "" "").items }}
-{{- if and (eq .spec.dst "") (not (eq .spec.gateway "")) (eq .spec.table "main") }}
+{{- if and (eq .spec.dst "") (not (eq .spec.gateway "")) (eq .spec.table "main") (eq (.spec.family | toString) "inet4") }}
 {{- $linkName = .spec.outLinkName }}
 {{- $family = .spec.family }}
 {{- break }}
@@ -134,9 +142,20 @@
 {{- end }}
 {{- end }}
 
+{{- /* All four default_link_*_by_gateway helpers below are IPv4-only.
+       They identify the link that carries the chart's primary uplink, and
+       that uplink is the same one default_gateway / default_addresses_by_gateway
+       configure — selecting a different link here would produce a config
+       where LinkConfig.name attaches to one NIC while LinkConfig.routes /
+       addresses describe a different NIC, leaving both unconfigured.
+       Picking the IPv4 default route's outLinkName keeps the whole
+       chain pointed at the same NIC on multi-NIC dual-stack nodes
+       (typical Hetzner shape: management NIC on IPv4, public NIC on
+       IPv6, default routes on different links). Symmetric with
+       gateway_by_link / default_gateway / default_addresses_by_gateway. */ -}}
 {{- define "talm.discovered.default_link_name_by_gateway" }}
 {{- range (lookup "routes" "" "").items }}
-{{- if and (eq .spec.dst "") (not (eq .spec.gateway "")) (eq .spec.table "main") }}
+{{- if and (eq .spec.dst "") (not (eq .spec.gateway "")) (eq .spec.table "main") (eq (.spec.family | toString) "inet4") }}
 {{- .spec.outLinkName }}
 {{- break }}
 {{- end }}
@@ -145,7 +164,7 @@
 
 {{- define "talm.discovered.default_link_address_by_gateway" }}
 {{- range (lookup "routes" "" "").items }}
-{{- if and (eq .spec.dst "") (not (eq .spec.gateway "")) (eq .spec.table "main") }}
+{{- if and (eq .spec.dst "") (not (eq .spec.gateway "")) (eq .spec.table "main") (eq (.spec.family | toString) "inet4") }}
 {{- (lookup "links" "" .spec.outLinkName).spec.hardwareAddr }}
 {{- break }}
 {{- end }}
@@ -154,7 +173,7 @@
 
 {{- define "talm.discovered.default_link_bus_by_gateway" }}
 {{- range (lookup "routes" "" "").items }}
-{{- if and (eq .spec.dst "") (not (eq .spec.gateway "")) (eq .spec.table "main") }}
+{{- if and (eq .spec.dst "") (not (eq .spec.gateway "")) (eq .spec.table "main") (eq (.spec.family | toString) "inet4") }}
 {{- (lookup "links" "" .spec.outLinkName).spec.busPath }}
 {{- break }}
 {{- end }}
@@ -163,7 +182,7 @@
 
 {{- define "talm.discovered.default_link_selector_by_gateway" }}
 {{- range (lookup "routes" "" "").items }}
-{{- if and (eq .spec.dst "") (not (eq .spec.gateway "")) (eq .spec.table "main") }}
+{{- if and (eq .spec.dst "") (not (eq .spec.gateway "")) (eq .spec.table "main") (eq (.spec.family | toString) "inet4") }}
 {{- with (lookup "links" "" .spec.outLinkName) }}
 busPath: {{ .spec.busPath }}
 {{- break }}
@@ -176,9 +195,20 @@ busPath: {{ .spec.busPath }}
 {{ printf "enx%s" (lookup "links" "" . | dig "spec" "hardwareAddr" . | replace ":" "") }}
 {{- end }}
 
+{{- /* Scalar IPv4 default-route gateway (dst="", main table, family=inet4).
+       Empty if the node has no IPv4 default route. IPv4-only by convention
+       so the helper stays symmetric with gateway_by_link and so consumers
+       that pair the returned gateway with an IPv4 destination
+       (`network: 0.0.0.0/0` on the legacy schema, or no `network:` field
+       on the typed RouteConfig schema) never end up with a malformed
+       IPv4-dst + IPv6-gateway route on a dual-stack node. Talos derives
+       the route family from the gateway literal at validation time, so a
+       RouteConfig with gateway `fe80::1` and no `network:` field is
+       rejected outright — silently breaking Layer2 VIP and any other
+       feature that depends on the chart-emitted route entry. */ -}}
 {{- define "talm.discovered.default_gateway" }}
 {{- range (lookup "routes" "" "").items }}
-{{- if and (eq .spec.dst "") (not (eq .spec.gateway "")) (eq .spec.table "main") }}
+{{- if and (eq .spec.dst "") (not (eq .spec.gateway "")) (eq .spec.table "main") (eq (.spec.family | toString) "inet4") }}
 {{- .spec.gateway }}
 {{- break }}
 {{- end }}