9 Commits
Author SHA1 Message Date
Zakk 6e09c172db fix(transparent): restore the redirect ingress wildcard, trim whitelist IPs, keep DNS readiness advisory
Three regressions from the previous transparent-proxy commit:

- the redirect ingress rules lost the "*" to "+" translation, so an excluded
  interface pattern such as docker* stopped matching and Docker traffic was
  redirected; the egress rules and both tproxy directions still translated it;
- the whitelist now parses its entries, and the GUI submits the textarea lines
  as typed, so a pasted CIDR with surrounding whitespace aborted the whole
  setup where the shell used to swallow it;
- the DNS readiness probe resolves a name, so a dead or slow upstream failed it
  even with the listener up, and the core could no longer start at all. The
  probe is now advisory: it still waits when DNS is healthy, but a failure is
  logged instead of aborting the start.
2026-09-18 05:07:03 +10:00
Zakk a4c0a5a37c fix(transparent): validate rule inputs, real DNS readiness probe, rule cleanup
Interface names and whitelist IPs from the settings were interpolated
into iptables/nft command text run through sh -c; they are now validated
(interface names may still use the "*" wildcard, translated to "+" for
iptables) and rejected with a message naming the value. nftTproxy emitted
"elements = { }" for an empty whitelist, an nft syntax error that failed
the whole setup. waitForDnsPort dialed UDP, which succeeds without a
listener; it now sends a DNS query and waits for an answer. The resolv
hijacker's ticker goroutine never exited because Stop does not close the
channel. A failing transparent pre-stop hook skipped rule deletion and
left the tproxy chains and resolv hijack behind. The DNS REDIRECT rules
ran through raw ExecCommands instead of the iptables backend setter, so
nft-only hosts silently skipped them. On darwin/BSD sysctl was invoked
with the whole command line as the program name.
2026-09-18 02:53:32 +10:00
xuweizhi 22856c189a fix(iptables): add protocol to DNS_REDIRECT redirect rules
iptables requires -p tcp or -p udp when using the REDIRECT target with
--to-port. The generated rule lacked the protocol match, so on legacy
iptables the command failed with "Need TCP, UDP, SCTP or DCCP with port
specification", aborting transparent proxy startup entirely.
2026-08-26 19:39:17 +08:00
zhangthree78 525654ec9b fix: prevent DNS redirect loop causing unbounded memory growth
Fixes defects in the v2.4.8+ transparent-proxy DNS path that cause the
core to exhaust memory within seconds on low-RAM hosts (see discussion
#1915: '2.4.9 内存占用过高' / '有严重的内存泄漏,别用了。本地DNS完全废掉了'):

1. core/dns: upstream queries from NewUpstreamManager /
   GetOrCreateUpstream / getClientForProtocol used bare &dns.Client{}
   with no Dialer configured, so their sockets never carried
   SO_MARK=0x80. The unmarked queries were hijacked by the module's own
   transparent-proxy DNS redirect rules and looped back into :52353 —
   an infinite self-query loop and memory snowball.

2. service/kernel/v2ray/transparent.go: the unconditional DNS hijack
   rules (REDIRECT --to-port 52353) had no fwmark-0x80 exemption, and
   being inserted with -I they sat at the head of nat OUTPUT/PREROUTING,
   so marked traffic was looped back as well.

3. core/dns: per-query info logging (query/route/upstream) produced
   thousands of lines/second under load, amplifying CPU/IO/memory
   pressure.

Fix:
- Route every upstream client through newMarkedDnsClient, whose
  net.Dialer sets SO_MARK=0x80 on all sockets. miekg/dns dials UDP
  through net.Dialer too, so Dialer.Control covers UDP sockets — no
  manual socket path is needed. UDPSize=4096 keeps a sane receive
  buffer for queries without an EDNS0 OPT record.
- Append the DNS redirects with -A and insert the mark exemption with
  -I so the exemption is matched first; also delete the exemption
  rules in the residual-rule cleanup so they don't accumulate across
  restarts.
- redirect mode: insert the DNS_REDIRECT jumps after the TP_PRE/TP_OUT
  jumps so they actually take precedence (-I inserts at chain head),
  keeping local and LAN TCP DNS flowing to the DNS module (:52353)
  instead of the proxy inbound (:52345).
- Gate per-query info logs behind V2RAYA_DNS_DEBUG=1; error logs stay
  always visible.

Verified: unpatched v2.4.11 core RSS 230MB->1GB in ~6s then OOM/hang;
patched build stable at ~25MB idle / ~243MB loaded, DNS and SOCKS proxy
working.
2026-08-16 23:59:43 +08:00
Xing Shi CaiandClaude Fable 5 cfe0e877a0 🐛 (iptables): Exempt DNS module port from generic TPROXY hijack
Locally originated DNS queries are marked 0x40 by DNS_MARK in mangle OUTPUT, then nat OUTPUT REDIRECT rewrites their destination port from 53 to 52353. When the packet re-enters PREROUTING via loopback, the dport-53 TPROXY rule no longer matches, so the generic mark-0x40 rule diverts it to the transparent proxy port 52345 where it is dropped. All local UDP DNS times out and transparent proxy appears completely broken (TCP DNS survives because the proxied connection still reaches 52353).

Add TP_PRE RETURN rules for dport 52353 ahead of the TPROXY rules in the legacy IPv4, legacy IPv6, and nftables variants so already-redirected DNS traffic is delivered to the DNS module untouched.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-23 10:48:41 +08:00
Markson Joe Hon 1bcd63eb1f fix(iptables): resolve firewall rule bugs and improve cleanup reliability
- Fix nftTproxy.GetCleanCommands using wrong table name ('ip' instead of
  'inet') when IPv6 is disabled, causing nftables table to persist after
  teardown
- Fix legacyRedirect.RemoveIPWhitelist not handling IPv6 CIDRs (always
  used iptables even for IPv6 addresses, causing silent delete failure)
- Suppress errors in ip rule/route del cleanup commands with
  2>/dev/null || true (applies to both legacy and nft TProxy modes)
- Normalize mixed tab/space indentation in nftRedirect's tp_rule chain
  for consistency
2026-07-22 13:09:01 +08:00
Markson Pigeonzilla Plus 1fdeb62bbe fix(windows): fix type mismatch in readRegistryProxyState
k.GetIntegerValue() returns uint64 on Windows, but proxyEnable was
declared as uint32. Add explicit uint32() conversion to fix the
cross-compilation error:

  cannot use k.GetIntegerValue("ProxyEnable") (value of type uint64)
  as uint32 value in assignment
2026-07-02 21:24:31 +08:00
Markson Pigeonzilla Plus 54efcae3ac fix(iptables): fix interface exclusion for OUTPUT path and rename nftables sets
Two fixes for the redirect/tproxy interface exclusion feature:

1. Add missing -o/oifname rules for the OUTPUT path
   - The TP_RULE chain is jumped from both PREROUTING (tp_pre) and
     OUTPUT (tp_out). The existing -i/iifname rules only match the
     PREROUTING path correctly (packets arriving on a given interface).
   - For the OUTPUT path (locally-originated traffic), the input
     interface is always 'lo', so -i never matches the excluded
     interface. Add corresponding -o/oifname rules so that locally-
     originated traffic going out through an excluded interface is
     also bypassed (RETURN before the final REDIRECT/TPROXY action).
   - Applies to both redirect and tproxy modes, legacy iptables and
     nftables implementations, for both IPv4 and IPv6.

2. Rename nftables set 'interface'/'interface6' to 'local_ips'/'local_ips6'
   - These sets store IPv4/IPv6 addresses of local interfaces,
     dynamically maintained by the LocalIPWatcher to prevent proxy
     loops (traffic destined to local IPs returns early).
   - The old name 'interface' was misleading: it suggested it stores
     interface names, but the type is ipv4_addr/ipv6_addr.
   - The rename affects nftRedirect and nftTproxy, including the
     set definitions, chain references (ip daddr @local_ips), and
     AddIPWhitelist/RemoveIPWhitelist nft command strings.
2026-07-02 21:21:18 +08:00
Markson Joe Hon 47f1462baf feat: init a new for v2raya core and fix a firewall mistake 2026-07-02 13:10:12 +08:00