82d8a049cc
stats: add rx_errors, tx_errors, rx_dropped, tx_dropped counters
...
Stats struct gains 4 new u64 fields per RFC 1213 (MIB-II):
- rx_errors: incoming packets with errors
- tx_errors: outgoing packets with errors
- rx_dropped: incoming packets dropped (buffer full, etc.)
- tx_dropped: outgoing packets dropped (qdisc dropped, buffer full)
EthernetLink tx path now increments tx_dropped when qdisc returns None
(qdisc decided to drop the packet, e.g., TokenBucket rate exceeded).
netcfg stats output now exposes all 8 counters:
rx_bytes=N rx_packets=N tx_bytes=N tx_packets=N
rx_errors=N tx_errors=N rx_dropped=N tx_dropped=N
mtu=N link=up
Aggregate fields work for bridge/bond via existing sum logic.
2026-07-08 18:35:22 +03:00
f3c18f410c
qdisc: configure traffic shaping type via netcfg
...
LinkDevice trait gains set_qdisc(kind) — Result<(), String>.
EthernetLink supports three kinds:
'none' / unset → QdiscConfig::None
'token_bucket' / 'tbf' → 1 Mbps rate, 1500 byte burst
'priority_queue' / 'pfifo_fast' → 1000 packet max
netcfg/ifaces/eth0/qdisc becomes rw:
cat /scheme/netcfg/ifaces/eth0/qdisc # current config
echo token_bucket > .../qdisc
echo priority_queue > .../qdisc
echo none > .../qdisc (default)
Mirrors Linux 'tc qdisc add/replace' for basic shaping.
TokenBucket rate/burst defaults applied on creation;
set via tb.set_rate() / tb.set_burst() (already in qdisc.rs).
2026-07-08 18:26:47 +03:00
9bab466704
netcfg: interface enable/disable control (ip link set eth0 down/up)
...
LinkDevice trait gains is_enabled() and set_enabled() with defaults.
EthernetLink stores enabled flag (default: true). When disabled,
link_state() reports 'down' regardless of hardware_address presence.
netcfg/ifaces/eth0/enabled is rw:
cat /scheme/netcfg/ifaces/eth0/enabled → 'up' or 'down'
echo up > .../enabled → bring interface up
echo down > .../enabled → bring interface down
echo on/off/1/0/yes/no → also accepted
Mirrors Linux 'ip link set dev <state>'.
Pre-existing link state (hardware_address Some/None) unchanged;
the enabled flag adds a software override layer that can be
toggled without restarting the netstack.
2026-07-08 18:24:02 +03:00
89083a25f9
netcfg: UDP socket listing + routes/count node
...
- UDP sockets now enumerated in /scheme/netcfg/sockets/list:
udp: local=0.0.0.0:53
udp: local=127.0.0.1:8080
- New RouteTable::len() method + /scheme/netcfg/route/count:
cat /scheme/netcfg/route/count → 'routes: 4'
Complete connections view now covers TCP + UDP listening sockets,
plus route count summary for at-a-glance monitoring.
2026-07-08 18:19:42 +03:00
0725f81144
sockets: per-TCP-connection listing via netcfg
...
netcfg/sockets/list now enumerates each active TCP socket with:
- State (Established, Listen, SynSent, etc.)
- Local endpoint (e.g., 127.0.0.1:8080)
- Remote endpoint (e.g., 192.168.1.5:51234) or '-' if not connected
Per-protocol counts still shown:
sockets: N (tcp=N udp=N icmp=N raw=N)
tcp: Established 127.0.0.1:22 -> 10.0.0.5:54321
tcp: Listen 0.0.0.0:80 -> -
tcp: SynSent 192.168.1.1:1025 -> 8.8.8.8:53
...
Mirrors Linux 'ss -t' output format.
Uses smoltcp Socket enum pattern matching (was previously blocked
because Socket didn't expose type-check methods).
2026-07-08 18:16:05 +03:00
c015a375b1
mtu: per-interface MTU configuration via netcfg
...
LinkDevice trait gains set_mtu(mtu) with default no-op.
EthernetLink stores mtu in field, clamped to 576-9000 (RFC 791 min/max).
netcfg/ifaces/eth0/mtu becomes rw:
cat /scheme/netcfg/ifaces/eth0/mtu → '1500' (default)
echo 1400 > .../mtu → VPN MTU
echo 9000 > .../mtu → jumbo frames
Clamped to [576, 9000]:
- 576: minimum for IPv4 (RFC 791)
- 9000: standard jumbo frame MTU
Useful for VPNs (MTU 1400 to avoid fragmentation through tunnel overhead),
jumbo frames (MTU 9000 on datacenter networks), and constrained paths.
2026-07-08 17:57:16 +03:00
c1cfee3bea
qdisc: expose traffic shaping info via netcfg and LinkDevice trait
...
LinkDevice gains qdisc_info() → String defaulting to 'none'.
EthernetLink reports current qdisc configuration:
none / token_bucket rate=N burst=N tokens=N / priority_queue len=N max=N
TokenBucket and PriorityQueue gain public accessor methods:
rate(), burst(), tokens() / max_len()
netcfg exposes at /scheme/netcfg/ifaces/eth0/qdisc:
cat /scheme/netcfg/ifaces/eth0/qdisc → 'none' (default)
Enables monitoring tools to discover current traffic shaping.
Future: wo node for configuring qdisc type + parameters.
2026-07-08 17:50:31 +03:00
7abd45bf2e
netfilter: add /scheme/netfilter/stats node
...
Exposes per-chain filter statistics:
input: packets=0 bytes=0 policy=ACCEPT
output: packets=0 bytes=0 policy=ACCEPT
forward: packets=1423 bytes=456789 policy=DROP
prerouting: packets=0 bytes=0 policy=ACCEPT
postrouting: packets=0 bytes=0 policy=ACCEPT
rules: 5 active (snat=3 dnat=2)
conntrack: 127
Mirrors iptables -L -n -v output format.
Per-chain (packets, bytes) from FilterTable::chain_counters.
Rule breakdown shows active count + snat/dnat split.
Conntrack entry count from live table.
2026-07-08 17:20:24 +03:00
98c43529e7
cleanup: fix unused variable + unnecessary mut warnings (7 total)
...
- router/mod.rs: 4x |mut b| → |b| (&mut [u8] is already mutable)
- scheme/icmp.rs: handle_recvmsg stub params → _file, _how, _flags
- scheme/ip.rs: handle_recvmsg stub params → _file, _how, _flags
- scheme/socket.rs: get_sock_opt default + msg_send ctx → _ prefixed
- scheme/tun.rs: loop name → _name
2026-07-08 16:38:48 +03:00
e185e03d88
observer: add BPF-style filter + output path capture
...
Capture filter supports tcpdump-like expressions:
echo tcp > /scheme/netcfg/capture/filter
echo udp port 53 > /scheme/netcfg/capture/filter
echo icmp > /scheme/netcfg/capture/filter
echo tcp port 80 > /scheme/netcfg/capture/filter
cat /scheme/netcfg/capture/filter → reads current filter
Filter matches IP header protocol field + TCP/UDP port fields.
Parses IPv4 and IPv6 headers. Non-IP packets captured always.
Output path capture: dispatch() now captures packets going
to devices (outbound locally-generated traffic).
Full capture coverage: input (forward_packets) + output (dispatch).
2026-07-08 16:30:13 +03:00
c1eb0b3db0
observer: wire into forwarding/output paths for live capture
...
Router now captures packets flowing through the network stack:
- forward_packets(): capture all forwarded/local-delivered packets
- Observer injected via Router::new() from Smolnetd constructor
When /scheme/netcfg/capture/enable is written, all packets
traversing the router are captured into the ring buffer.
When disabled, zero overhead (AtomicBool check).
2026-07-08 16:25:50 +03:00
ff66b96266
observer: packet capture facility (tcpdump-like) via netcfg
...
Adds packet observer with ring buffer (256 packets default):
- Observer::capture(packet) hooks into forwarding/output paths
- AtomicBool enable/disable toggle (no capture overhead when off)
- capture/count: live stats (total captured, buffered, enabled)
- capture/read: drain hex dump of buffered packets
- capture/enable|disable: toggle capture on/off
Usage:
echo > /scheme/netcfg/capture/enable # start capture
cat /scheme/netcfg/capture/read # dump captured packets (hex)
cat /scheme/netcfg/capture/count # stats
echo > /scheme/netcfg/capture/disable # stop
Mirrors Linux AF_PACKET + tcpdump facility.
2026-07-08 16:12:15 +03:00
998593f243
stats: add statistics() to bridge, bond, and tun devices
...
Bridge: statistics() aggregates rx/tx bytes+packets from all member ports.
arp_stats() delegates to each port, showing per-port breakdown.
Bond: statistics() aggregates from all slaves (active+standby).
arp_stats() delegates to each slave.
TUN: statistics() now returns live counters tracked during send()/recv().
rx_bytes, rx_packets, tx_bytes, tx_packets increment per-packet.
Previously these devices returned Stats::default() (all zeros).
Now netcfg /ifaces/*/stats shows real data for bridge, bond, tun.
2026-07-08 16:00:26 +03:00
5cfbe23b5c
cleanup: fix unreachable pattern warnings in TCP/UDP socket options
...
TCP and UDP get_sock_opt/set_sock_opt had duplicate match arms
due to constant collisions (Linux-level values share namespace):
TCP: TCP_MAXSEG=2, IP_TTL=2, SO_REUSEADDR=2 → kept TCP_MAXSEG
UDP: IP_TTL=2, SO_REUSEADDR=2 → kept IP_TTL (more useful)
Removed unreachable arms with explanatory comments. The
collision is inherent — Linux uses different option levels
(SOL_SOCKET vs IPPROTO_TCP vs IPPROTO_IP) but Redox scheme
has a flat namespace. Applications that use multi-level
getsockopt() would need richer level dispatching.
2026-07-08 15:43:42 +03:00
27f8e351e6
arp: add statistics counters and netcfg exposure
...
EthernetLink gains per-interface ARP counters:
- arp_requests: ARP requests sent (incremented in send_arp)
- arp_replies: ARP replies received (incremented on cache insert)
- arp_cache_hits: successful neighbor cache lookups
- arp_cache_misses: cache misses (vacant or expired entries)
LinkDevice trait gains arp_stats() → String method.
EthernetLink implementation: 'requests=N replies=N hits=N misses=N entries=N'
netcfg exposes at /scheme/netcfg/ifaces/eth0/arp/stats:
echo /scheme/netcfg/ifaces/eth0/arp/stats
→ requests=5 replies=3 hits=142 misses=8 entries=4
Existing arp/list (MAC→IP) and arp/flush (clear) are unchanged.
Useful for monitoring ARP churn, detecting ARP storms, and debugging
neighbor discovery issues.
2026-07-08 15:38:49 +03:00
ba5a7cd3fc
sysctl: add /scheme/netcfg/sysctl tree with ip_forward toggle
...
Router gains Rc<Cell<bool>> ip_forward flag (default: true).
When false, forward_packets() returns immediately — no packets forwarded
between interfaces. Security best practice for non-router hosts.
netcfg scheme gains sysctl subtree:
/scheme/netcfg/sysctl/net/ipv4/ip_forward (rw: 0 or 1)
Read: echo /scheme/netcfg/sysctl/net/ipv4/ip_forward → 1
Write: echo 0 > /scheme/netcfg/sysctl/net/ipv4/ip_forward (disable)
Restore: echo 1 > /scheme/netcfg/sysctl/net/ipv4/ip_forward
Mirrors Linux /proc/sys/net/ipv4/ip_forward.
Shared via Rc<Cell<bool>> between Router and NetCfgScheme.
2026-07-08 15:20:21 +03:00
25d6decc85
route: add route types — blackhole, unreachable, prohibit
...
RouteTable Rule gains route_type: RouteType field:
- Unicast (default): normal forwarding
- Blackhole: silently drop, no ICMP error
- Unreachable: drop + send ICMP Destination Unreachable
- Prohibit: drop + send ICMP Administratively Prohibited
Route type checked in all forwarding/outbound paths:
- Forward path (IPv4): blackhole/unreachable/prohibit drop + error
- Output path (IPv4): blackhole/unreachable/prohibit drop + error
- Output path (IPv6): blackhole/unreachable/prohibit drop
Usage: Rule::with_type(cidr, via, dev, src, RouteType::Blackhole)
Existing Rule::new() creates Unicast routes (backward compatible).
Display: blackhole/unreachable/prohibit prefixed in route list output.
Mirrors Linux iproute2 'ip route add blackhole 10.0.0.0/8'.
2026-07-08 15:02:40 +03:00
107e3b6851
stp: enforce blocking on unicast send + enhanced stats display
...
BridgeDevice STP hardening:
- send(): check STP blocking before unicast forwarding (host-originated)
- recv(): check STP blocking before unicast forwarding (switched frames)
- Previously only flood() checked STP; unicast forwarding bypassed it.
A blocked port must never forward any traffic — STP semantics now correct.
netcfg stats enhancement:
- Per-interface stats now include mtu= and link= fields alongside counters
- Applies to both eth0 and loopback
- Enables bandwidth monitoring tools to self-discover MTU/link state
2026-07-08 14:53:08 +03:00
0baceb0ef6
tcp: extend TcpInfo with send/recv queue, congestion window, MSS
...
TCP_INFO socket option now returns real-time connection metrics:
- tcpi_snd_queuelen: bytes queued for transmission
- tcpi_rcv_queuelen: bytes waiting in receive buffer
- tcpi_snd_cwnd: send capacity (congestion window proxy)
- tcpi_rcv_wnd: receive capacity (advertised window proxy)
- tcpi_snd_mss: maximum segment size (1460 for Ethernet)
- tcpi_state, tcpi_rto preserved from prior version
Struct layout: #[repr(C)] 28 bytes, compatible with Linux TCP_INFO consumers.
2026-07-08 14:30:05 +03:00
e05315fc38
udp: socket option completeness — SO_REUSEADDR, SO_BROADCAST, IP_TTL
...
- UDP port allocation now falls back to claim_port_reuse() (SO_REUSEADDR)
- SO_REUSEADDR get/set added to both UDP and TCP schemes
- SO_BROADCAST getter added to UDP (always returns 1)
- IP_TTL getter/setter added to UDP (get/set hop_limit)
- TCP: SO_REUSEADDR get/set added for API completeness
- All new options return known values for application compatibility
2026-07-08 13:57:00 +03:00
2b278390ee
tun: wire TUN scheme into event loop + SLAAC RS/RA protocol
...
TUN integration:
- Smolnetd gains tun_scheme: TunScheme field and tun_file parameter
- on_tun_scheme_event() handler added (scheme event → poll)
- main.rs: EventSource::TunScheme, subscription, dispatch
- TUN devices can now receive and transmit packets through the netstack
SLAAC (RFC 4862):
- build_router_solicitation(): ICMPv6 Type 133 with source LL address option
- parse_router_advertisement(): ICMPv6 Type 134 with Prefix Information
option extraction (on-link, autonomous, lifetimes)
- Slacd state machine: Idle → Solicited → Configured
tick() drives RS retransmit (3 retries, 5s timeout)
process_ra() extracts autonomous /64 prefixes
- ParsedRa, RaPrefix public structs for integration with IPv6 stack
Reference: Linux 7.1 ndisc_send_rs() / ndisc_router_discovery() /
addrconf_prefix_rcv()
2026-07-08 13:47:04 +03:00
30db94c970
stp: integrate 802.1D Spanning Tree Protocol into BridgeDevice
...
Adds loop prevention to Ethernet bridges:
- BridgeDevice gains stp: Option<StpState> field
- enable_stp(priority, mac) method initializes STP per-bridge
- BPDU frames (dst 01:80:c2:00:00:00) intercepted in recv(),
processed locally, never forwarded
- STP hello timer sends periodic BPDUs on all ports (root bridge)
- flood() skips STP-blocked ports
- build_bpdu() made public for bridge integration
- stp module declared in link/mod.rs
The recv() flow now: age MACs → check STP hello timer →
poll ports → detect BPDU (absorb) → normal frame (learn + forward).
Reference: Linux 7.1 net/bridge/br_stp.c, br_stp_bpdu.c, br_stp_timer.c
2026-07-08 13:38:38 +03:00
bb3e36e4e0
restore: networking stack files from reflog (Phases 1-6)
...
Recovered from reflog commits 1c80937e and d0ecc067
2026-07-08 13:27:49 +03:00
4506bfe02a
stp: add IEEE 802.1D Spanning Tree Protocol for bridge loop prevention
2026-07-08 13:26:39 +03:00
cb1c326645
netcfg: add sockets/list node for active connection count
2026-07-08 09:42:03 +03:00
e653ef10d6
base: revert oneshot_async service types, fix local fork deps, migrate remote to RedBear-OS
...
- Revert initfs/rootfs service types from oneshot_async to blocking
Scheme/Notify/Oneshot to fix init ordering races.
- Add /scheme/pci retry loop in pcid-spawner.
- Bump redox_event to 0.4.8; use local paths for redox-ioctl and redox-rt.
- Regenerate Cargo.lock / bootstrap/Cargo.lock with only local forks.
- Update submodule origin from redbear-os-base.git to RedBear-OS.git
branch submodule/base per single-repo policy.
2026-07-05 22:25:00 +03:00
dd08b76a39
Red Bear OS base baseline from 0.1.0 pre-patched archive
2026-06-27 09:21:43 +03:00
Red Bear OS