[rtl] Add 16 external performance counters at mhpmcounter[28:13]

rtl/ibex_core.sv

@@ -71,6 +71,9 @@ module ibex_core #(
     output logic        irq_x_ack_o,
     output logic [4:0]  irq_x_ack_id_o,
 
+    // External performance counters
+    input  logic [15:0] external_perf_i, // Bind to zero if unused
+
     // Debug Interface
     input  logic        debug_req_i,
 
@@ -1073,7 +1076,8 @@ module ibex_core #(
       .mem_store_i             ( perf_store                   ),
       .dside_wait_i            ( perf_dside_wait              ),
       .mul_wait_i              ( perf_mul_wait                ),
-      .div_wait_i              ( perf_div_wait                )
+      .div_wait_i              ( perf_div_wait                ),
+      .external_perf_i         ( external_perf_i              )
   );
 
   // These assertions are in top-level as instr_valid_id required as the enable term

rtl/ibex_core_tracing.sv

@@ -65,6 +65,8 @@ module ibex_core_tracing #(
     output logic        irq_x_ack_o,
     output logic [4:0]  irq_x_ack_id_o,
 
+    input  logic [15:0] external_perf_i,
+
     // Debug Interface
     input  logic        debug_req_i,
 
@@ -162,6 +164,8 @@ module ibex_core_tracing #(
     .irq_x_ack_o,
     .irq_x_ack_id_o,
 
+    .external_perf_i,
+
     .debug_req_i,
 
     .rvfi_valid,

rtl/ibex_cs_registers.sv

@@ -118,7 +118,8 @@ module ibex_cs_registers #(
     input  logic                 mem_store_i,            // store to memory in this cycle
     input  logic                 dside_wait_i,           // core waiting for the dside
     input  logic                 mul_wait_i,             // core waiting for multiply
-    input  logic                 div_wait_i              // core waiting for divide
+    input  logic                 div_wait_i,             // core waiting for divide
+    input  logic [15:0]          external_perf_i         // external performance counters
 );
 
   import ibex_pkg::*;
@@ -1166,6 +1167,10 @@ module ibex_cs_registers #(
     mhpmcounter_incr[10] = instr_ret_compressed_i; // num of compressed instr
     mhpmcounter_incr[11] = mul_wait_i;             // cycles waiting for multiply
     mhpmcounter_incr[12] = div_wait_i;             // cycles waiting for divide
+
+    for (int unsigned i=0; i<16; i++) begin : gen_mhpmcounter_incr_external
+      mhpmcounter_incr[13+i] = external_perf_i[i]; // Start at ID=13
+    end
   end
 
   // event selector (hardwired, 0 means no event)