Crossbar read path – openenoc_axil_crossbar_rd

Routes AXI4-Lite read requests and returns their responses to the requesting initiators.

Source: openenoc_axil_crossbar_rd.sv.

Parameters

Parameters

Name	Default value	Allowed values	Description
S_COUNT	4	Integer >= 1	Number of upstream initiator interfaces.
M_COUNT	4	Integer >= 1	Number of downstream target/output interfaces.
ADDR_W	32	Integer >= 1	Byte-address width in bits; connected interfaces must agree.
S_ACCEPT	{S_COUNT{32'd16}}	One positive 32-bit count per source	Routing-context capacity per source; source 0 occupies the least-significant slice. Responses already buffered upstream are additional.
M_REGIONS	1	Integer >= 1	Number of decode regions per target.
M_BASE_ADDR	'0	Packed aligned addresses; zero selects automatic placement	One ADDR_W-bit base per target region; region index m*M_REGIONS+r starts in the least-significant slice.
M_ADDR_W	{M_COUNT{{ M_REGIONS{32'd24}}}}	Zero, or log2(STRB_W)..ADDR_W in crossbar integration	Region span is 2**width bytes; zero disables the region. Packing follows M_BASE_ADDR.
M_CONNECT	{M_COUNT{{ S_COUNT{1'b1}}}}	M_COUNT*S_COUNT-bit mask	Permitted paths; bit m*S_COUNT+s connects source s to target m.
M_ISSUE	{M_COUNT{32'd16}}	One positive 32-bit count per target	Routing-context capacity per target, including buffered addresses; target 0 occupies the least-significant slice.
M_SECURE	{M_COUNT{1'b0}}	M_COUNT-bit mask	A set bit disallows nonsecure accesses to that target.

Signals

Interface ports contain channels in both directions. The table identifies the selected modport or stream role; member directions follow that interface.

Public signals and interfaces

Name	Type	Direction / role	Description
clk	logic	input	Clock for this component or interface domain.
rst	logic	input	Active-high synchronous reset of routing, arbitration, and channel buffers.
s_axil_rd[S_COUNT]	taxi_axil_if.rd_slv	bidirectional (rd_slv)	Incoming AXI4-Lite read channels.
m_axil_rd[M_COUNT]	taxi_axil_if.rd_mst	bidirectional (rd_mst)	Outgoing AXI4-Lite read channels.

Architecture and operation

Each source decodes its AR address and access attributes. Each target has independent round-robin arbitration and a two-entry AXI4-Lite channel register slice (skid buffer) for outgoing AR. Admitting an address reserves one source and one target routing entry. The target may accept that buffered address later. A source can have reads pending at several targets.

The target routing queue identifies the source of each R response. The source routing queue preserves its own AR order across targets. A response transfers only when those two queue heads agree and the source R buffer has space. Once captured in that two-entry response buffer, routing entries are released even if the upstream initiator has not consumed the response. An R transfer into the buffer can admit a replacement AR on the same edge. The buffer presents the response upstream in the following interval.

An unmapped or disallowed AR consumes a source routing entry and creates an ordered local DECERR response with zero data. It consumes no target entry and produces no target AR transfer. It waits behind earlier reads from that source just as a target response would.

The map, connection matrix, security rules, and parameter packing follow the crossbar and decoder references. Source and target limits apply to routing entries; the R buffer adds response storage. Reset inhibits AR admission immediately and clears arbitration, routing queues, and channel VALID state at the sampling edge. Connected endpoints reset with the fabric.

Functional scenarios

These examples use the crossbar reference’s 2-source, 2-target configuration: 32-bit addresses/data, two routing entries per source and target, and one 64 KiB region per target. M0 begins at zero and M1 at 0x00010000. All accesses are secure. S0/S1 and M0/M1 identify the indexed public source and target interfaces. A tick is an interval ending at a rising sampling edge; x marks fields without VALID. Address/data labels are local to each example.

Continuous address and response pipelines

S0 offers four consecutive reads, A0 through A3, to M0. They return D0 through D3 in the same order. Source RREADY and target ARREADY remain high; the target supplies each response one interval after its AR handshake.

S0’s AR transfers occur in ticks 1 through 4, and M0 receives them in ticks 2 through 5. M0 returns data in ticks 3 through 6; S0 consumes it in ticks 4 through 7. Capturing each response frees routing capacity for the next address, allowing continuous progress with two routing entries.

An unstalled read pipeline sustains consecutive addresses and responses

A local decode error remains in read order

Reads from S0

Address	Destination	Response
A0	M0	Target returns D0 in tick 6
A1	Unmapped	Local DECERR with zero data
A2	M1	Target returns D2 in tick 8

The first two AR requests are admitted in ticks 1 and 2. S0 offers A2 from tick 3, but both source entries remain occupied until the earlier M0 response is captured in tick 6. That edge also admits A2.

The local error cannot overtake D0. S0 sees consecutive responses: D0/OKAY in tick 7, zero/DECERR in tick 8, and D2/OKAY in tick 9. Only A0 and A2 appear at target AR ports, in ticks 2 and 7.

A local decode error follows the earlier target response in source order