The live suite asserted the datapath over IPv4 only. Give every guest a
ULA IPv6 address next to its v4 one and twin every v4 assertion point,
sharing the impairment window (one filter application, both families
asserted) so the runtime grows by minutes, not double.
- harness: configure_ios(eth_ipv6=), ping() dispatches to `ping ipv6`,
docker_bring_up_ipv6 (the Docker daemon leaves eth0 with
disable_ipv6=1 on v6-less bridge networks, so the interfaces-file
route cannot work), family-aware docker_wait_address, pcap_records
scans every record (a leading DAD/MLD frame must not mask the
sampled ethertype), and pcap_icmp_rtt measures the wire round trip
from capture timestamps;
- scenarios: connectivity/delay/suspend/isolation/churn checks in both
families, captures assert 86dd next to 0800, the libpcap bpf drops
are asserted family-selective (icmp vs icmp6, each killing its own
family and leaving the other alive), the eBPF classifier modes on
both families, marker v6 twins with an echo-specific expression
(bare icmp6 also matches DAD/MLD and would break the exact counts),
the v6 sharkd replay, and the link-local matrix gains the 86dd
all-nodes row; serial (HDLC) legs stay v4-only;
- the dynamips-emulated classic IOS under-reports its `ping ipv6` RTT
by ~40 % of a delayed round trip (pcap-verified: the wire carries the
v6 echo exchange in the same ~230 ms as v4 while the console reports
~130 ms), so the dynamips-family scenarios assert the v6 delay from
the capture's own timestamps — recorded in tests/e2e/README.md.
Full suite: 17 passed in 58m50s.
The 'impairment cleared' asserts used avg < 50 on three-packet IOS pings;
the unimpaired baseline is 20-45 ms with 4 ms quantization, and host
scheduling jitter pushed averages to 54/56 in two full-suite runs
(esw, dynamips) — a margin, not an impairment (the qdisc asserts pass,
success is 100 %). The contrasted delayed regime is >= 150 ms, so a
100 ms bound still separates the two cleanly. Docker's sub-millisecond
veth asserts keep their 50 ms bound (three orders of magnitude of
headroom, never observed to flake).
The esw kernel fast-path scenario deletes and re-creates a link, which
re-enslaves the peer anchor — and on NetworkManager hosts NM's netdev
state machine can release an NM-managed TAP from its bridge some 60-70 ms
after enslavement, ~50% of re-enslavements, while the tap is in NM's
not-yet-assumed state. The freed port stays administratively up and no
GNS3 process logs anything: the membership assert sees the join, then the
ping measures a silent blackhole (this was the line-277 e2e failure).
Root-caused with a server-free harness (one uBridge, one fresh tap, one
bridge, an addif/delif loop): the same tap frees 20/40 while NM-managed,
0/40 after `nmcli device set <tap> managed no`, 20/40 again when
re-managed. Docker's veth host ends are not NM-managed, which is why only
TAP anchors ever showed it — including the "anchor lost its bridge
membership after a ports_mapping update (TAP peers only)" incident from
the switch's development, whose conclusion is corrected in the docs.
Unmanage the anchors once the routers have booted, before the scenario
re-enslaves them; document the failure signature and the host-side remedy
(per-device nmcli call, or NM unmanaged-devices configuration) in the
L2-anchor spec and the switch feature doc.
The peer end of an Ethernet switch fast-path link is a NIOBridge with no
bridge of its own (the switch absorbed its anchor into the switch bridge).
Its update reconcile still ran the impairment pass with the empty filter
set the controller routes to that end — and an empty set is a plain
`tc reset` on the shared anchor, silently wiping the netem/cls_bpf/eBPF
the switch had just applied. Any link update did it (filters, markers,
suspend, capture start/stop), and the outcome depended on the link's
endpoint order: UdpLink.update() PUTs node1 then node2, so with the
switch first its apply ran first and the peer's reset ran last (API
green, data plane unimpaired), while with the peer first the reset ran
first and the filter survived. That order dependence is also why the
shipped e2e never caught it — it only created peer-first links.
Ownership is now explicit: only the end that owns the anchor's tc state
applies the impairment reconcile (_owns_anchor_impairments — a NIOBridge
with `bridge is None` is a passive carrier, the same idiom the
bridge-delete path already used). Markers, capture and the suspend
carrier still apply on both ends: markers reconcile per process, so
neither end can strip the other's, and the carrier converges by
agreement.
Red/green: two new docker unit tests fail on the unfixed tree with
`tc reset "<anchor>"` and pass after; the switch e2e also creates one
link switch-first and asserts the netem qdisc and the RTT hold under
that order (failed with the wipe, passes with the fix). Full unit suite
2110 passed; the switch e2e file passed end to end (378s).
The L2-anchor spec's behavioral guarantee — an idle link stays silent for
5 s once the enslavement burst settles — only ran in the Dynamips and
Docker suites. It now also runs in QEMU, IOU, the IOL container, the
switch fast path and the switch cascade, which covers every host-device
creation path:
* tap create (Dynamips/QEMU/IOU anchors), docker create_veth (Docker host
ends; the cascade's race-loser end is hardened after create by the
server), bridge add_nio_tap (the IOL container anchor, the fifth
creator), and the brctl switch bridges, whose multicast snooping is off
precisely so the bridge role can reach silence;
* each scenario silences its guests first (IOS interfaces shut through
the console; Docker guests' links down) and re-verifies the ping after;
* the cascade measures 8 devices at once (four absorbed router anchors,
both cascade ends, both switch bridges).
IOL 158 s, cascade 293 s, ESW fast path 378 s, QEMU 428 s, IOU 308 s —
all green.
Two real IOU routers on a real L3 image (the image's own IOL fabric, the
real IOS CLI on the server's telnet console), driving the anchor TAPs the
kernel datapath builds on: anchors born with the node (one per Ethernet
bay/unit), real ICMP through the [fabric <-> TAP] port bridge, netem
delay, suspend, capture, link delete/re-create, node stop/start rewiring,
the serial link staying on the relay (serial bays never anchor) and a
relay negative control on an isolated instance. Every scenario ends with
no anchors, bridges or tmpdirs left behind.
harness: create_iou_node/iou_images/iou_anchor_name, and the IOS console
helper is now ios_console (dynamips, IOU and the IOL containers share it).
Without one every c7200 busy-waits a full CPU core for the whole scenario
-- the tests create raw nodes and no template supplies a value (measured:
constant ~100 % per router, up to ~6 minutes per dynamips scenario).
harness.dynamips_idlepc() detects a value through the controller's
auto_idlepc endpoint (the GUI idle-PC finder path: a throwaway project
boots the image, candidate values are validated by measuring actual CPU
usage) and caches it in ~/.cache/gns3-e2e/idlepc.json keyed by the image
checksum, so the cost (~100 s here) is paid once per image, not per run.
GNS3_E2E_IDLEPC overrides detection; any failure returns None and the
scenarios run without one rather than fail on a CPU optimisation.
All four scenarios that boot c7200 routers (dynamips kernel/relay,
ethernet switch kernel/relay) pass the value at creation. Live: detected
0x60189234 for the M11 image (238 s first run including detection); the
routers drop from ~100 % during boot to 3-8 % steady (sampled per
process), and a re-run reports the cached value with no detection.
The spec's gns3-server alignment section promised three assertions the
suites never carried; they land in harness.py and run in every kernel
scenario (plus the relay controls' anchor checks):
- E.1 assert_pure_l2: anchors and bridges carry no L3 identity (addrgenmode
none, no address in /proc/net/if_inet6, no ip -o -4 addr) -- the kernel
scenarios' anchors and per-link/switch bridges, relay-mode anchors and
the switch's relay TAPs alike (hardening is a creation-time property,
not a datapath one);
- E.2 assert_idle_silence (Dynamips suite): guest interfaces shut to
silence IOS's own CDP/keepalives, settle 2.5 s, then a 5 s window of
zero traffic over both anchors and the per-link bridge;
- assert_forwarding: every attached port reports brport/state == 3 -- a
bridge left DOWN keeps its ports disabled and forwards nothing, with no
error anywhere; the assertion is the only signal.
harness.l2only_supported() probes a throwaway uBridge once (link l2only
against a nonexistent device: 208 = command present, 202-Unknown command
= old build); without the command the assertions skip -- the spec's
degradation contract, mirroring the tc-capability skip.
The first live run caught a fifth-creator gap: the switch's relay port
TAPs (minted implicitly by bridge add_nio_tap) carried a live fe80::.
Delivered uBridge-side per the spec (fb72758: the transient TAP is
hardened at creation, attaches untouched); the server needed no change,
and the relay-control E.1 check is that creator's end-to-end coverage.
The Ethernet switch's kernel fast path: instead of a link crossing the
switch's per-port uBridge relay, the switch absorbs the peer's anchor into
its own kernel bridge. The peer's interface (Docker veth host end, QEMU
TAP, IOU port TAP, Dynamips port TAP) becomes the switch port, and frames
cross anchor -> switch bridge -> switch ports entirely in the kernel.
Wire format: the switch end receives nio_anchor ('join this interface to
my own bridge' — the mirror of nio_bridge), the peer end nio_bridge with a
null bridge ('your anchor is bridged elsewhere': carrier, markers and
impairments only — never bridge membership, and never a bridge delete, on
an interface this node does not own). Anchor names come from the shared
naming contract (utils.kernel_anchor), a pure function of node type, id,
adapter and port — the controller computes the peer's anchor while the
node may be stopped, which is what makes links to stopped nodes wireable.
Stopped peers defer the join; after any node of a link starts
(UDPLink.node_started) the controller re-pushes the switch NIO, and the
switch completes — or repairs, when a peer restart replaced its interface
under the same name — the join against the kernel's actual membership
instead of assuming it.
VLAN changes now edit membership in place: only the ports whose mode/VLAN
actually changed are touched, each losing exactly the previous mode's
entries the new mode does not want (the dot1q admits-all entry is one
range operation in both directions), and no port ever leaves the bridge to
change its VLANs. This replaces the old 'reset by re-enslaving' shortcut
(delif+addif on every wired port, added when the ports were the switch's
own TAPs): on absorbed anchors, re-enslaving an interface owned by another
node is where live testing watched links die — anchors occasionally lost
their bridge membership right after a ports_mapping update (later updates
then failed with delif EINVAL), reproducibly with qemu, IOU and Dynamips
peers and never with Docker's veths.
Eligibility: the switch side needs no capability of its own (its bridge
exists by construction); the peer must be able to anchor. Switch-to-switch
cascades stay on the relay — a kernel interface belongs to one bridge, and
joining two bridges needs a veth pair, deferred.
Verified live on a real server: qemu peers through four VLAN cycles with
both anchors enslaved throughout; a mixed switch (2 IOU + 2 Dynamips, four
kernel ports); Docker peers with real ICMP isolating and recovering (with
both anchors still enslaved); VPCS relay ports likewise in place.
