Commit Graph

3 Commits

Author SHA1 Message Date
Kuzz007 d1b77b2aa4 feat(amneziawg): Phase 2c — RouteViaXray (TPROXY into Xray)
Per-client toggle (RouteThroughXray + RouteOutboundTag) that TPROXYs a
peer's traffic into Xray instead of NAT'ing it straight out the host's
network interface, so it can egress through any configured Xray
outbound (or balancer) — a VLESS/proxy chain, WARP, etc.

Discovered mid-design that internal/mtproto already solved the "let a
native sidecar's traffic egress through Xray" problem once, via
routeThroughXray/routeXrayPort/outboundTag + injectMtprotoEgress: a
loopback bridge inbound plus a routing rule. AmneziaWG can't reuse it
directly — mtg is a userspace process that dials *out* through a local
SOCKS proxy, while AmneziaWG is a kernel tunnel interface with no
process of its own to redirect. The Xray-side shape carries over
almost exactly, the kernel-side plumbing is new:

- internal/amneziawg/route_egress.go: EgressPort/EgressTag/EgressFwmark/
  EgressTable are one shared constant set, not one bridge per peer.
  Every routed peer, across every AmneziaWG instance, TPROXYs into the
  *same* loopback dokodemo-door bridge; the per-peer distinction happens
  downstream, in Xray's own router, matched against each peer's
  TPROXY-preserved source IP (Xray's field-rule `source` matcher — a
  capability the router already had). This avoids two independent
  reconcile loops (the AWG manager and the Xray-config generator) ever
  having to agree on a dynamically-picked port for each peer.
- manager.go's defaultPostUpDown emits a per-peer mangle-table TPROXY
  rule (matched by tunnel source IP) for each opted-in peer, plus the
  fwmark->table->local-everywhere policy route TPROXY needs to deliver
  those packets to the bridge. That policy route is system-wide, not
  interface-specific, so — like the existing IPv6-forwarding sysctl —
  it's added idempotently and never torn down in PostDown; a second
  AmneziaWG instance with its own routed peers must find it already in
  place, not race to remove what the first still needs.
- The existing portForwardFingerprint became hostRulesFingerprint,
  covering both ForwardedPorts and RouteThroughXray/RouteOutboundTag:
  both only ever take effect through PostUp/PostDown, which `awg
  syncconf` never re-runs, so either one changing must force the same
  full interface bounce.
- internal/web/service/xray.go's new injectAmneziawgEgress mirrors
  injectMtprotoEgress/injectPanelEgress's safety rules, adapted for one
  bridge serving many peers: an invalid or missing outbound target
  skips only that one peer's rule (not the whole bridge, since other
  peers may still need it), while the bridge itself is skipped
  entirely when nothing needs it or its tag is already taken by a real
  inbound.

Frontend: a Switch + conditional outbound Select on the client form
(showAmneziawg only), mirroring mtproto's own routeThroughXray UI and
reusing its useOutboundTags hook. install.sh now modprobes the
mainline TPROXY modules (xt_TPROXY, nf_tproxy_ipv4/ipv6) alongside the
existing AmneziaWG setup — ordinary upstream kernel modules, no
DKMS/PPA needed unlike the AmneziaWG module itself.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
2026-07-25 18:34:37 +03:00
Kuzz007 69de904bf6 feat(amneziawg): Phase 2b — per-client port-forwarding
Admins can now set a per-client ForwardedPorts string (e.g. "80, 443,
8000-8100") that gets DNAT'd + FORWARD'd to that peer's tunnel address
via iptables rules in PostUp/PostDown, ported and simplified from
coinman-dev/3ax-ui's shared/portfwd.

Two decisions worth flagging for future readers:

- The iptables --comment tag on each rule is awg-fwd-<fnv32a(email)>,
  not the raw client email. Email is admin/API-supplied free text that
  ends up embedded in a shell-executed PostUp/PostDown line; a hash
  can never carry a shell metacharacter through where raw
  interpolation could.
- The reconcile manager gained a third fingerprint (portFwdFP, next to
  the existing structural/peers ones). `awg syncconf` only touches the
  WireGuard peer table — it never re-applies PostUp/PostDown iptables
  rules — so a port-forward-only change has to force a full
  awg-quick down+up bounce, same as a structural change, rather than
  the lighter sync a plain peer add/remove can use.

Also fixes a real pre-existing bug found while wiring up IPv6 client
allocation in the previous commit's spirit: allocateWireguardAddress
always suffixed "/32" regardless of address family, which produced
invalid host bits for IPv6 (needs "/128").

ForwardedPorts flows through model.Client -> model.ClientRecord
(gorm column wg_forwarded_ports, auto-migrated) -> ToRecord/ToClient/
MergeClientRecord, mirroring the awgServer field's earlier lesson
that new fields need checking against a second, hand-maintained
persistence-layer struct.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
2026-07-25 17:54:05 +03:00
Kuzz007 83cc545953 feat(amneziawg): add native AmneziaWG protocol backend
AmneziaWG (WireGuard plus DPI-resistant obfuscation) needs no Docker
here — it runs as a genuine kernel interface via awg-quick/awg, managed
the same way internal/mtproto manages mtg: one Inbound row is one
desired Instance, and a Manager reconciles running interfaces toward
the database every 10s (internal/web/job/amneziawg_job.go) plus
immediately after a client edit (applyLocalAmneziaWG).

Clients reuse model.Client verbatim (the same PrivateKey/PublicKey/
PreSharedKey/AllowedIPs fields WireGuard already uses), so bulk
operations, the QR/share-link modal and subscriptions come from the
shared inbound infrastructure instead of a parallel implementation.
internal/amneziawg owns the obfuscation param generator/validator
(ported from coinman-dev/3ax-ui, upgraded to AmneziaWG 2.0's S3/S4
padding and I1 signature packet) and the exec wrapper around
awg-quick/awg, with fingerprint-based reconcile (noop / reload-via-
syncconf / full restart) mirroring mtproto.Manager so a same-protocol
edit doesn't force an unnecessary interface bounce that would drop
every peer's connection.

Frontend and install.sh's DKMS/awg-tools setup are tracked separately;
this is backend-only.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
2026-07-25 00:39:41 +03:00