ouroboros/src/ipcpd/unicast/dir/dht.c, branch 0.24.0

ipcpd: Rework congestion avoidance

2026-07-19T09:44:36+00:00

Congestion avoidance is a property of the layer, orthogonal to ARQ and to flow control: FRCP retransmits and lets the peer pace the sender, per flow, end-to-end; the IPCP paces path aggregates. Each signal means one thing: a loss triggers a retransmission, a mark means congestion, the peer window means a slow receiver. Every flow is paced by the same rate law whatever its QoS, so a greedy raw sender shares a bottleneck fairly with a reliable stream. The unit of control is the (destination address, QoS cube) aggregate: all flows toward that destination share one controller and one rate; a start-time fair-queuing pacer divides the rate across them by deadline instead of blocking the send path, and a new flow rides the aggregate's estimates at its current rate, with no probing of its own. Slow start runs once per aggregate. The congestion signal is a multi-bit magnitude: forwarders mark packets with their standing queue depth, MAX-combined across hops, so a packet carries the deepest queue on its path. The receiver feeds back a time-integral mean over a window that adapts to the flow's byte rate, measuring a slow flow with the same fidelity as a fast one. The sender runs AIMD scaled by elapsed wall-clock time, which makes the steady-state allocation RTT-independent. The PCI gains one byte: the path capacity as a quarter-log2 code. Forwarders estimate their egress rate from busy-period drain and MIN-stamp the byte, the receiver returns the window minimum with its feedback, and the sender scales its rate floor and additive slope to the bottleneck (C / 32). A deep cut implies a backlogged bottleneck and a backlogged bottleneck advertises its capacity, so the scaled floor is live exactly when recovery needs it: the probe heals a halving in seconds at any link rate, and the floor bounds the deepest hole to a factor 32 below the bottleneck. Signed-off-by: Dimitri Staessens Signed-off-by: Sander Vrijders

ipcpd: Fix NULL deref in DHT cleanup

2026-07-19T09:44:34+00:00

The dht_kv_find_node_rsp_msg function steals *contacts and NULLs it on success. When the subsequent msg->val allocation fails in dht_kv_find_value_rsp_msg(), the stolen contacts are freed along with the message, and do_dht_kv_find_value_req() is left with contacts == NULL while n_contacts is still positive: its fail_vals loop then dereferences NULL. Found by clang static analyzer. Signed-off-by: Dimitri Staessens Signed-off-by: Sander Vrijders

ipcpd: Store local names before DHT peering

2026-07-08T09:02:25+00:00

A name registered before the DHT had a peer stayed unpublished until it neared its expiry deadline. Make a never-stored local value (t_repl == 0) eligible for replication immediately, and flush replication once when the worker starts so such names are published as soon as the DHT is enrolled. Signed-off-by: Dimitri Staessens Signed-off-by: Sander Vrijders

ipcpd: Fix DHT bootstrap re-seeding

2026-07-08T09:02:24+00:00

The bootstrap seeder reported success as soon as the query was sent. Only skip seeding once the peer is in a bucket, and return -EAGAIN until it actually appears so the caller keeps retrying rather than assuming bootstrap completed. Signed-off-by: Dimitri Staessens Signed-off-by: Sander Vrijders

ipcpd: Fix unused values in unicast IPCP

2026-02-22T15:04:16+00:00

Use the key in the store operation in the DHT, to avoid static analyzer seeing it as being set without being used in release builds. The length of the packet buffer is only used when tracking DT stats, so should be set within the IPCP_FLOW_STATS guard. Signed-off-by: Dimitri Staessens Signed-off-by: Sander Vrijders

lib: Add struct llist for lists tracking len

2026-02-18T06:58:21+00:00

The DHT uses a struct {struct list_head, size_t len} pattern, which is also useful in the registry and other places. Having a struct llist (defined in list.h) with consistent macros for addition/deletion etc removes a lot of duplication and boilerplate and reduces the risk of inconsistent updates. The list management is now a macro-only implementation. Signed-off-by: Dimitri Staessens Signed-off-by: Sander Vrijders

build: Update copyright to 2026

2026-02-18T06:54:56+00:00

Signed-off-by: Dimitri Staessens Signed-off-by: Sander Vrijders

lib: Replace rdrbuff with a proper slab allocator

2026-01-26T06:50:33+00:00

This is a first step towards the Secure Shared Memory (SSM) infrastructure for Ouroboros, which will allow proper resource separation for non-privileged processes. This replaces the rdrbuff (random-deletion ring buffer) PoC allocator with a sharded slab allocator for the packet buffer pool to avoid the head-of-line blocking behaviour of the rdrb and reduce lock contention in multi-process scenarios. Each size class contains multiple independent shards, allowing parallel allocations without blocking. - Configurable shard count per size class (default: 4, set via SSM_POOL_SHARDS in CMake). The configured number of blocks are spread over the number of shards. As an example: SSM_POOL_512_BLOCKS = 768 blocks total These 768 blocks are shared among 4 shards (not 768 × 4 = 3072 blocks) - Lazy block distribution: all blocks initially reside in shard 0 and naturally migrate to process-local shards upon first allocation and subsequent free operations - Fallback with work stealing: processes attempt allocation from their local shard (pid % SSM_POOL_SHARDS) first, then steal from other shards if local is exhausted, eliminating fragmentation while maintaining low contention - Round-robin condvar signaling: blocking allocations cycle through all shard condition variables to ensure fairness - Blocks freed to allocator's shard: uses allocator_pid to determine target shard, enabling natural load balancing as process allocation patterns stabilize over time Maintains existing robust mutex semantics including EOWNERDEAD handling for dead process recovery. Internal structures exposed in ssm.h for testing purposes. Adds some tests (pool_test, pool_sharding_test.c. etc) verifying lazy distribution, migration, fallback stealing, and multiprocess behavior. Updates the ring buffer (rbuff) to use relaxed/acquire/release ordering on atomic indices. The ring buffer requires the (robust) mutex to ensure cross-structure synchronization between pool buffer writes and ring buffer index publication. Signed-off-by: Dimitri Staessens Signed-off-by: Sander Vrijders

build: Add option to toggle Ouroboros logging in tests

2026-01-19T07:14:55+00:00

The test output is a bit polluted with logs originating fomr the logging system, e.g.: 23: test_bind_prog started. 23: ==16636== reg/name(DB): Add prog reg_test to name testname. 23: ==16636== reg(EE): Removing from names. 23: test_bind_prog succeeded. This adds a flag DISABLE_TEST_LOGGING that suppresses log_* output in tests to keep them clean: 23: test_bind_prog started. 23: test_bind_prog succeeded. The status is printed in CMake output: -- Ouroboros logging in test output disabled -- Ouroboros logging in test output enabled By default the flag is ON (clean test output). Signed-off-by: Dimitri Staessens Signed-off-by: Sander Vrijders

ipcpd: Fix missing return in DHT

2025-08-24T08:23:47+00:00

The dht_kv_respond_req had a missing return statement. I'm not sure why the double mutex_unlock() never caused havoc in testing. Signed-off-by: Dimitri Staessens