BFD early networking support.

[rcgoodfellow]

In

• BFD support #4852

the only way to enable BFD is after a rack install. For some network configurations, it may be necessary to set up BFD during rack installation.

[rcgoodfellow]

Deprioritizing as

• Graceful transition to/from BFD-based networks maghemite#174

has landed and should provide a viable early-networking path for BFD-based networks. This still needs to happen but not for release 7.

[rcgoodfellow]

Partially addressed in

• add bfd to bootstore and early networking #5313

And BFD is present in the RSS/rack-setup data structures

• omicron/common/src/api/internal/shared.rs

  Lines 160 to 175 in 4dda5cb

  	#[derive(Clone, Debug, Deserialize, Serialize, PartialEq, JsonSchema)]
  	pub struct RackNetworkConfigV1 {
  	pub rack_subnet: Ipv6Network,
  	// TODO: #3591 Consider making infra-ip ranges implicit for uplinks
  	/// First ip address to be used for configuring network infrastructure
  	pub infra_ip_first: Ipv4Addr,
  	/// Last ip address to be used for configuring network infrastructure
  	pub infra_ip_last: Ipv4Addr,
  	/// Uplinks for connecting the rack to external networks
  	pub ports: Vec<PortConfigV1>,
  	/// BGP configurations for connecting the rack to external networks
  	pub bgp: Vec<BgpConfig>,
  	/// BFD configuration for connecting the rack to external networks
  	#[serde(default)]
  	pub bfd: Vec<BfdPeerConfig>,
  	}

However, it's missing from wicket

• omicron/wicket-common/src/rack_setup.rs

  Lines 101 to 109 in 4dda5cb

  	pub struct UserSpecifiedRackNetworkConfig {
  	pub infra_ip_first: Ipv4Addr,
  	pub infra_ip_last: Ipv4Addr,
  	// Map of switch -> port -> configuration, under the assumption that
  	// (switch, port) is unique.
  	pub switch0: BTreeMap<String, UserSpecifiedPortConfig>,
  	pub switch1: BTreeMap<String, UserSpecifiedPortConfig>,
  	pub bgp: Vec<BgpConfig>,
  	}

[rcgoodfellow]

Moving this to 10, as not having BFD in wicket is not a blocker for R9.