axi_dmac: Limit number of bursts on the source side

Currently the source side of the DMAC can issue requests for up to
2*FIFO_SIZE-1 bursts even though there is only room for FIFO_SIZE bursts in
the store and forward memory.

This can problematic for memory mapped buses. If the data is not read fast
enough from the DMAC back-pressure will propagate through the whole system
memory subsystem and can cause significant performance penalty or even a
deadlock halting the whole system.

To avoid this make sure that not more that than what fits into the
store-and-forward memory is requested by throttling the request ID based
on how much room is available in the memory.

Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>

library/axi_dmac/axi_dmac_burst_memory.v

@@ -48,6 +48,8 @@ module axi_dmac_burst_memory #(
   input [DATA_WIDTH_SRC-1:0] src_data,
   input src_data_last,
 
+  output [ID_WIDTH-1:0] src_data_request_id,
+
   input dest_clk,
   input dest_reset,
 
@@ -159,6 +161,8 @@ assign src_beat = src_mem_data_valid & src_mem_data_ready;
 assign src_last_beat = src_beat & src_mem_data_last;
 assign src_waddr = {src_id_reduced,src_beat_counter};
 
+assign src_data_request_id = src_dest_id;
+
 always @(*) begin
   if (src_last_beat == 1'b1) begin
     src_id_next <= inc_id(src_id);

library/axi_dmac/request_arb.v

@@ -251,6 +251,8 @@ wire [1:0] src_response_resp;
 */
 
 wire [ID_WIDTH-1:0] src_request_id;
+reg [ID_WIDTH-1:0] src_throttled_request_id;
+wire [ID_WIDTH-1:0] src_data_request_id;
 wire [ID_WIDTH-1:0] src_response_id;
 
 wire src_valid;
@@ -542,7 +544,7 @@ dmac_src_mm_axi #(
   .response_resp(src_response_resp),
 */
 
-  .request_id(src_request_id),
+  .request_id(src_throttled_request_id),
   .response_id(src_response_id),
   .address_id(src_address_id),
   .data_id(src_data_id),
@@ -615,7 +617,7 @@ dmac_src_axi_stream #(
   .req_sync_transfer_start(src_req_sync_transfer_start),
   .req_xlast(src_req_xlast),
 
-  .request_id(src_request_id),
+  .request_id(src_throttled_request_id),
   .response_id(src_response_id),
 
   .eot(src_eot),
@@ -671,7 +673,7 @@ dmac_src_fifo_inf #(
   .req_last_burst_length(src_req_last_burst_length),
   .req_sync_transfer_start(src_req_sync_transfer_start),
 
-  .request_id(src_request_id),
+  .request_id(src_throttled_request_id),
   .response_id(src_response_id),
 
   .eot(src_eot),
@@ -705,6 +707,39 @@ sync_bits #(
   .out(src_request_id)
 );
 
+`include "inc_id.h"
+
+function compare_id;
+  input [ID_WIDTH-1:0] a;
+  input [ID_WIDTH-1:0] b;
+  begin
+    compare_id = a[ID_WIDTH-1] == b[ID_WIDTH-1];
+    if (ID_WIDTH >= 2) begin
+      if (a[ID_WIDTH-2] == b[ID_WIDTH-2]) begin
+        compare_id = 1'b1;
+      end
+    end
+    if (ID_WIDTH >= 3) begin
+      if (a[ID_WIDTH-3:0] != b[ID_WIDTH-3:0]) begin
+        compare_id = 1'b1;
+      end
+    end
+  end
+endfunction
+
+/*
+ * Make sure that we do not request more data than what fits into the
+ * store-and-forward burst memory.
+ */
+always @(posedge src_clk) begin
+  if (src_resetn == 1'b0) begin
+    src_throttled_request_id <= 'h00;
+  end else if (src_throttled_request_id != src_request_id &&
+               compare_id(src_throttled_request_id, src_data_request_id)) begin
+    src_throttled_request_id <= inc_id(src_throttled_request_id);
+  end
+end
+
 sync_bits #(
   .NUM_OF_BITS(ID_WIDTH),
   .ASYNC_CLK(ASYNC_CLK_DEST_REQ)
@@ -744,6 +779,8 @@ axi_dmac_burst_memory #(
   .src_data(src_fifo_data),
   .src_data_last(src_fifo_last),
 
+  .src_data_request_id(src_data_request_id),
+
   .dest_clk(dest_clk),
   .dest_reset(~dest_resetn),
   .dest_data_valid(dest_fifo_valid),