diff --git a/RNS/Transport.py b/RNS/Transport.py index c54d61bd..cb5cb664 100755 --- a/RNS/Transport.py +++ b/RNS/Transport.py @@ -136,11 +136,10 @@ class Transport: USE_INBOUND_QUEUE = True USE_OUTBOUND_QUEUE = False - INBOUND_DA_QUEUE_LENGTH = 4096 - INBOUND_AN_QUEUE_LENGTH = 256 - INBOUND_PR_QUEUE_LENGTH = 256 - INBOUND_IL_QUEUE_LENGTH = 128 - OUTBOUND_QUEUE_LENGTH = 4096 + INBOUND_DA_QUEUE_LENGTH = 1024 + INBOUND_AN_QUEUE_LENGTH = 128 + INBOUND_PR_QUEUE_LENGTH = 128 + INBOUND_IL_QUEUE_LENGTH = 8 STATE_UNKNOWN = 0x00 STATE_UNRESPONSIVE = 0x01 diff --git a/docs/source/interfaces.rst b/docs/source/interfaces.rst index b35c1268..9a9374a8 100644 --- a/docs/source/interfaces.rst +++ b/docs/source/interfaces.rst @@ -1589,15 +1589,15 @@ Reticulum uses separate, prioritized queues for processing inbound packets. The You can tune the queue sizes from the defaults, for example if you're running a setup where you want to simply drop most traffic from ingress-limited interfaces. The queue sizes are configured in the ``[reticulum]`` section of the config * | The ``qlen_in_data`` option sets the maximum size of the data - queue. Defaults to ``4096``. + queue. Defaults to ``1024``. * | The ``qlen_in_announce`` option sets the maximum size of the announce - queue. Defaults to ``256``. - - * | The ``qlen_in_pr`` option sets the maximum size of the path request - queue. Defaults to ``256``. - - * | The ``qlen_in_il`` option sets the maximum size of the ingress limiter queue. Defaults to ``128``. + * | The ``qlen_in_pr`` option sets the maximum size of the path request + queue. Defaults to ``128``. + + * | The ``qlen_in_il`` option sets the maximum size of the ingress limiter + queue. Defaults to ``8``. + If an inbound packet arrives and its target queue is full, the packet is dropped. Reticulum will always completely drain higher-priority queues before starting to drain a lower-priority one. As an example, the ingress limiter queue will only start draining once all other queues are empty. \ No newline at end of file