smoke: decouple nexthop assertions from route add events

[rjarry]

Starting with FRR 10.8, routes are always installed via nexthop groups regardless of the number of nexthops. When a route has a single nexthop, grout resolves the group and stores the individual nexthop directly, but the initial "route add" event now carries the group reference instead of the resolved nexthop details.

The smoke tests were matching the full nexthop description in the route add event pattern (e.g. "via type=L3 .*addr=172.16.0.2"). This broke with FRR 10.8 because the event now reads "via type=group ...".

Split the route verification in two steps: first, wait for the route add event matching only on prefix and origin (which is stable across FRR versions). Then, look up the route's resolved nexthop via grcli and assert its attributes separately. Two new helpers (route_nh_id and assert_nexthop) handle the second step and work whether the route was installed with a nexthop group or a direct nexthop.

Link: FRRouting/frr@99b1afa202d0d28273

Summary

• Added helpers to resolve FRR nexthop IDs and validate resolved nexthop attributes via grcli.
• Updated route and SRv6 smoke-test checks to support nexthop groups and direct nexthops.
• Updated BGP tests to match route events first, then validate resolved nexthop addresses, VRFs, behaviors, and SRv6 segment lists.

[coderabbitai]

📝 Walkthrough

Walkthrough

FRR smoke helpers now resolve route nexthop IDs and validate nexthop attributes through grcli. IP route checks validate addresses or interfaces, while SRv6 checks validate VRFs, behaviors, and segment lists. BGP tests wait for route additions before asserting the corresponding nexthop properties.

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[coderabbitai]

Actionable comments posted: 1

🤖 Prompt for all review comments with AI agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Inline comments:
In `@smoke/_init_frr.sh`:
- Around line 158-164: Update the next-hop classification in the case statement
around assert_nexthop so dotted interface names such as p0.100 are treated as
interfaces, not addresses. Restrict the IPv4 branch to values matching a numeric
dotted-quad pattern, while preserving the existing IPv6 detection and .addr
validation for actual addresses and the .iface validation for interface names.

🪄 Autofix (Beta)

Fix all unresolved CodeRabbit comments on this PR:

• Push a commit to this branch (recommended)
• Create a new PR with the fixes

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro Plus

Run ID: 7dda97e8-4a3e-42d7-9ce9-973c0a2397dc

📥 Commits

Reviewing files that changed from the base of the PR and between 57d92e3 and 385429d.

📒 Files selected for processing (3)

• smoke/_init_frr.sh
• smoke/bgp6_srv6_frr_test.sh
• smoke/bgp_frr_test.sh

[coderabbitai]

🎯 Functional Correctness | 🟡 Minor | ⚡ Quick win

Handle dotted interface names correctly.

Line 159 treats every value containing . as an address. A route via an interface such as p0.100 will be configured successfully but then fail validation against .addr. Restrict IPv4 detection to numeric dotted-quad values.

Proposed fix

-	case "$nh" in
-	*.*|*:*)
+	if [[ "$nh" == *:* || "$nh" =~ ^([0-9]{1,3}\.){3}[0-9]{1,3}$ ]]; then
 		assert_nexthop "$nh_id" ".vrf == \"${nexthop_vrf_name:-$gr_vrf_name}\" and .addr ==\"$nh\""
-		;;
-	*)
+	else
 		assert_nexthop "$nh_id" ".vrf == \"${nexthop_vrf_name:-$gr_vrf_name}\" and .iface ==\"$nh\""
-		;;
-	esac
+	fi

📝 Committable suggestion

‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.

Suggested change

	case "$nh" in
	*.*|*:*)
	assert_nexthop "$nh_id" ".vrf == \"${nexthop_vrf_name:-$gr_vrf_name}\" and .addr ==\"$nh\""
	;;
	*)
	assert_nexthop "$nh_id" ".vrf == \"${nexthop_vrf_name:-$gr_vrf_name}\" and .iface ==\"$nh\""
	;;
	if [[ "$nh" == *:* || "$nh" =~ ^([0-9]{1,3}\.){3}[0-9]{1,3}$ ]]; then
	assert_nexthop "$nh_id" ".vrf == \"${nexthop_vrf_name:-$gr_vrf_name}\" and .addr ==\"$nh\""
	else
	assert_nexthop "$nh_id" ".vrf == \"${nexthop_vrf_name:-$gr_vrf_name}\" and .iface ==\"$nh\""
	fi

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@smoke/_init_frr.sh` around lines 158 - 164, Update the next-hop
classification in the case statement around assert_nexthop so dotted interface
names such as p0.100 are treated as interfaces, not addresses. Restrict the IPv4
branch to values matching a numeric dotted-quad pattern, while preserving the
existing IPv6 detection and .addr validation for actual addresses and the .iface
validation for interface names.