diff --git a/artiq/gateware/rtio/sed/lane_distributor.py b/artiq/gateware/rtio/sed/lane_distributor.py
new file mode 100644
index 000000000..2cdf025b8
--- /dev/null
+++ b/artiq/gateware/rtio/sed/lane_distributor.py
@@ -0,0 +1,118 @@
+from migen import *
+
+from artiq.gateware.rtio import cri
+
+
+def layout_lane_io(seqn_width, layout_payload):
+    return [
+        ("we", 1, DIR_M_TO_S),
+        ("writable", 1, DIR_S_TO_M),
+        ("seqn", seqn_width, DIR_M_TO_S),
+        ("payload", layout_payload, DIR_M_TO_S)
+    ]
+
+
+# CRI write happens in 3 cycles:
+# 1. set timestamp
+# 2. set other payload elements and issue write command
+# 3. check status
+
+class LaneDistributor(Module):
+    def __init__(self, lane_count, fifo_size, layout_payload, fine_ts_width):
+        if lane_count & (lane_count - 1):
+            raise NotImplementedError("lane count must be a power of 2")
+
+        seqn_width = 4*bits_for(fifo_size-1)*lane_count
+
+        self.cri = cri.Interface()
+        self.minimum_coarse_timestamp = Signal(64-fine_ts_width)
+        self.lane_io = [Record(layout_lane_io(seqn_width, layout_payload))
+                        for _ in range(lane_count)]
+
+        # # #
+
+        o_status_wait = Signal()
+        o_status_underflow = Signal()
+        o_status_sequence_error = Signal()
+        self.comb += self.cri.o_status.eq(Cat(o_status_wait, o_status_underflow,
+                                              o_status_sequence_error))
+
+        # internal state
+        current_lane = Signal(max=lane_count)
+        last_coarse_timestamp = Signal(64-fine_ts_width)
+        last_lane_coarse_timestamps = Array(Signal(64-fine_ts_width)
+                                            for _ in range(lane_count))
+        seqn = Signal(seqn_width)
+
+        # distribute data to lanes
+        for lio in lane_io:
+            self.comb += [
+                lio.seqn.eq(seqn),
+                lio.payload.channel.eq(self.cri.chan_sel[:16]),
+                lio.payload.timestamp.eq(self.cri.timestamp),
+            ]
+            if hasattr(lio.payload, "address"):
+                self.comb += lio.payload.address.eq(self.cri.address)
+            if hasattr(lio.payload, "data"):
+                self.comb += lio.payload.data.eq(self.cri.data)
+
+        # when timestamp arrives in cycle #1, prepare computations
+        coarse_timestamp = self.cri.timestamp[fine_ts_width:]
+        timestamp_above_min = Signal()
+        timestamp_above_laneA_min = Signal()
+        timestamp_above_laneB_min = Signal()
+        use_laneB = Signal()
+        use_lanen = Signal(max=lane_count)
+        self.sync += [
+            timestamp_above_min.eq(coarse_timestamp > self.minimum_coarse_timestamp),
+            timestamp_above_laneA_min.eq(coarse_timestamp > last_lane_coarse_timestamps[current_lane]),
+            timestamp_above_laneB_min.eq(coarse_timestamp > last_lane_coarse_timestamps[current_lane + 1]),
+            If(coarse_timestamp <= last_coarse_timestamp,
+                use_lanen.eq(current_lane + 1),
+                use_laneB.eq(1)
+            ).Else(
+                use_lanen.eq(current_lane),
+                use_laneB.eq(0)
+            )
+        ]
+
+        # cycle #2, write
+        timestamp_above_lane_min = Signal()
+        do_write = Signal()
+        do_underflow = Signal()
+        do_sequence_error = Signal()
+        self.comb += [
+            timestamp_above_lane_min.eq(Mux(use_laneB, timestamp_above_laneB_min, timestamp_above_laneA_min)),
+            do_write.eq((self.cri.cmd == cri.commands["write"]) & timestamp_above_min & timestamp_above_lane_min),
+            do_underflow.eq((self.cri.cmd == cri.commands["write"]) & ~timestamp_above_min),
+            do_sequence_error((self.cri.cmd == cri.commands["write"]) & timestamp_above_min & ~timestamp_above_lane_min),
+            Array(lio.we for lio in lane_io)[use_lanen].eq(do_write)
+        ]
+        self.sync += [
+            If(do_write,
+                current_lane.eq(current_lane + 1),
+                last_coarse_timestamp.eq(coarse_timestamp),
+                last_lane_coarse_timestamps[use_lanen].eq(coarse_timestamp),
+                seqn.eq(seqn + 1),
+            )
+        ]
+
+        # cycle #3, read status
+        current_lane_writable = Signal()
+        self.comb += [
+            current_lane_writable.eq((lio.writable for lio in lane_io)[current_lane]),
+            o_status_wait.eq(~current_lane_writable)
+        ]
+        self.sync += [
+            o_status_underflow.eq(do_underflow),
+            o_status_sequence_error(do_sequence_error)
+        ]
+
+        # current lane has been full, spread events by switching to the next.
+        current_lane_writable_r = Signal()
+        self.sync += [
+            current_lane_writable_r.eq(current_lane_writable),
+            If(~current_lane_writable_r & current_lane_writable,
+                current_lane.eq(current_lane + 1)
+            )
+        ]