hw: add offloadport for the FPU in snitch

Bender.lock

@@ -7,10 +7,10 @@ packages:
     dependencies:
     - common_cells
   axi:
-    revision: bec548fa2a9b18cbd7531105bb1fdf481ea8ad49
+    revision: 458c8c58cbe0acd0c5e6bb8592b52111bc9a5d84
     version: null
     source:
-      Git: https://github.com/colluca/axi.git
+      Git: https://github.com/Lura518/axi.git
     dependencies:
     - common_cells
     - common_verification

Bender.yml

@@ -19,7 +19,7 @@ package:
     - Matheus Cavalcante <[email protected]>
 
 dependencies:
-  axi:                { git: https://github.com/colluca/axi,                      rev:  multicast       }
+  axi:                { git: https://github.com/Lura518/axi,                      rev:  reduction       }
   axi_riscv_atomics:  { git: https://github.com/pulp-platform/axi_riscv_atomics,  version:  0.6.0       }
   common_cells:       { git: https://github.com/pulp-platform/common_cells,       rev:      snitch      }
   apb:                { git: https://github.com/pulp-platform/apb.git,            version:  0.2.2       }

docs/ug/system_integration.md

@@ -0,0 +1,51 @@
+# Integrating the Snitch cluster in an SoC
+
+While this repository provides many IPs that can be reused independently, we suggest to integrate the Snitch cluster as a whole, that is the `snitch_cluster` module, in derived systems.
+
+The `snitch_cluster` module is implemented in [snitch_cluster.sv](https://github.com/pulp-platform/{{ repo }}/blob/{{ branch }}/hw/snitch_cluster/src/snitch_cluster.sv).
+
+## Configurability
+
+A reference instantiation of the Snitch cluster can be found in the testbench used to test the cluster within this repository, see [testharness.sv](https://github.com/pulp-platform/{{ repo }}/blob/{{ branch }}/target/snitch_cluster/test/testharness.sv).
+
+As you may note, we do not instantiate the `snitch_cluster` directly but a so-called `snitch_cluster_wrapper`, with a much simplified interface. All parameters of the `snitch_cluster` module are set within the wrapper.
+
+The benefit of the wrapper is that it can be programmatically generated from a single source of truth, namely a JSON5 configuration file, from which the software hardware-abstraction layer (HAL), and all other sources dependent on the configuration within the repository, are also generated.
+
+This way, if you want to modify the cluster configuration, you don't have to go and manually change it in multiple places (the RTL, the HAL, etc.), but only in the single-source-of-truth cluster configuration file. More information on the configuration file can be found in the [tutorial](tutorial.md#configuring-the-hardware).
+
+We suggest that the same approach is used when integrating the Snitch cluster in an SoC. This allows you to easily test different configurations of the cluster inside your SoC.
+
+## Integrating the RTL
+
+We provide Make rules to generate the cluster wrapper and other RTL files. Include the following lines in a Makefile, to inherit Snitch's rules:
+```Makefile
+SN_ROOT = $(shell $(BENDER) path snitch_cluster)
+
+include $(SN_ROOT)/target/common/common.mk
+include $(SN_ROOT)/target/common/rtl.mk
+```
+
+!!! note
+    Snitch's Makefiles require `SN_ROOT` to be defined and to point to the root of the Snitch cluster repository. You can set this however you prefer, i.e. you don't have to use Bender if you manage your dependencies in a different way.
+
+You can then use the `sn-rtl` and `sn-clean-rtl` targets to respectively build and clean all of Snitch's generated RTL sources.
+<!-- TODO(colluca): In Picobello we explicitly use the $(SN_CLUSTER_WRAPPER) $(SN_CLUSTER_PKG) variables to build only the generated sources that depend on the cluster config. Find a common ground, probably define targets for only those files. -->
+
+## Integrating the software
+
+Similarly, Snitch comes with a collection of software tests and applications. These build on the functions provided by the Snitch runtime library, so they must be linked against an implementation of the latter. The runtime library abstracts away all the low-level characteristics of the system, allowing applications to be written in a mostly system-independent way, and to be portable to any multi-cluster Snitch-based system.
+To this end, every system must implement a hardware abstraction layer (HAL) for the Snitch runtime, which the mentioned infrastructure builds on.
+
+Given a path to the platform-specific HAL sources, you can reuse the Snitch cluster's Make rules to build the runtime, tests and applications for the target platform.
+Include the following lines in a Makefile, to inherit Snitch's rules:
+
+```Makefile
+SN_RUNTIME_HAL_DIR = sw/runtime/hal
+
+include $(SN_ROOT)/target/common/sw.mk
+```
+
+The included Makefile(s) can be customized to some extent by overriding some variables before the Makefile inclusion line.
+For example by setting `SNRT_BUILD_APPS = OFF` none of the default Snitch applications will be built.
+You can explicitly set the list of applications to be built via the `SNRT_APPS` variable, which can include additional system-dependent applications you may develop in the system repository. For further information on the available customization options you may want to take a look inside the recursively included Makefiles.

hw/reqrsp_interface/include/reqrsp_interface/typedef.svh

@@ -8,15 +8,15 @@
 `ifndef REQRSP_INTERFACE_TYPEDEF_SVH_
 `define REQRSP_INTERFACE_TYPEDEF_SVH_
 
-`define REQRSP_TYPEDEF_REQ_CHAN_T(__req_chan_t, __addr_t, __data_t, __strb_t) \
+`define REQRSP_TYPEDEF_REQ_CHAN_T(__req_chan_t, __addr_t, __data_t, __strb_t, __user_t) \
   typedef struct packed { \
-    __addr_t             addr;  \
-    __addr_t             mask;  \
-    logic                write; \
-    reqrsp_pkg::amo_op_e amo;   \
-    __data_t             data;  \
-    __strb_t             strb;  \
-    reqrsp_pkg::size_t   size;  \
+    __addr_t             addr;        \
+    logic                write;       \
+    reqrsp_pkg::amo_op_e amo;         \
+    __data_t             data;        \
+    __strb_t             strb;        \
+    __user_t             user;        \
+    reqrsp_pkg::size_t   size;        \
   } __req_chan_t;
 
 `define REQRSP_TYPEDEF_RSP_CHAN_T(__rsp_chan_t, __data_t) \
@@ -39,8 +39,8 @@
     logic      q_ready; \
   } __rsp_t;
 
-`define REQRSP_TYPEDEF_ALL(__name, __addr_t, __data_t, __strb_t) \
-  `REQRSP_TYPEDEF_REQ_CHAN_T(__name``_req_chan_t, __addr_t, __data_t, __strb_t) \
+`define REQRSP_TYPEDEF_ALL(__name, __addr_t, __data_t, __strb_t, __user_t) \
+  `REQRSP_TYPEDEF_REQ_CHAN_T(__name``_req_chan_t, __addr_t, __data_t, __strb_t, __user_t) \
   `REQRSP_TYPEDEF_RSP_CHAN_T(__name``_rsp_chan_t, __data_t) \
   `REQRSP_TYPEDEF_REQ_T(__name``_req_t, __name``_req_chan_t) \
   `REQRSP_TYPEDEF_RSP_T(__name``_rsp_t, __name``_rsp_chan_t)

hw/reqrsp_interface/src/axi_to_reqrsp.sv

@@ -448,7 +448,7 @@ module axi_to_reqrsp_intf #(
   typedef logic [IdWidth-1:0] id_t;
   typedef logic [UserWidth-1:0] user_t;
 
-  `REQRSP_TYPEDEF_ALL(reqrsp, addr_t, data_t, strb_t)
+  `REQRSP_TYPEDEF_ALL(reqrsp, addr_t, data_t, strb_t, user_t)
 
   `AXI_TYPEDEF_AW_CHAN_T(aw_chan_t, addr_t, id_t, user_t)
   `AXI_TYPEDEF_W_CHAN_T(w_chan_t, data_t, strb_t, user_t)

hw/reqrsp_interface/src/reqrsp_cut.sv

@@ -12,6 +12,8 @@ module reqrsp_cut #(
     parameter int unsigned AddrWidth = 0,
     /// Data width of the interface.
     parameter int unsigned DataWidth = 0,
+    /// User width of the interface.
+    parameter int unsigned UserWidth = 0,
     /// Request type.
     parameter type req_t             = logic,
     /// Response type.
@@ -32,8 +34,9 @@ module reqrsp_cut #(
   typedef logic [AddrWidth-1:0] addr_t;
   typedef logic [DataWidth-1:0] data_t;
   typedef logic [DataWidth/8-1:0] strb_t;
+  typedef logic [UserWidth-1:0] user_t;
 
-  `REQRSP_TYPEDEF_ALL(reqrsp, addr_t, data_t, strb_t)
+  `REQRSP_TYPEDEF_ALL(reqrsp, addr_t, data_t, strb_t, user_t)
 
   spill_register #(
     .T (reqrsp_req_chan_t),
@@ -74,6 +77,8 @@ module reqrsp_cut_intf #(
     parameter int unsigned AddrWidth = 0,
     /// Data width of the interface.
     parameter int unsigned DataWidth = 0,
+    /// User width of the interface.
+    parameter int unsigned UserWidth = 0,
     /// Bypass request channel.
     parameter bit  BypassReq         = 0,
     /// Bypass Response channel.
@@ -88,15 +93,17 @@ module reqrsp_cut_intf #(
   typedef logic [AddrWidth-1:0] addr_t;
   typedef logic [DataWidth-1:0] data_t;
   typedef logic [DataWidth/8-1:0] strb_t;
+  typedef logic [UserWidth-1:0] user_t;
 
-  `REQRSP_TYPEDEF_ALL(reqrsp, addr_t, data_t, strb_t)
+  `REQRSP_TYPEDEF_ALL(reqrsp, addr_t, data_t, strb_t, user_t)
 
   reqrsp_req_t reqrsp_slv_req, reqrsp_mst_req;
   reqrsp_rsp_t reqrsp_slv_rsp, reqrsp_mst_rsp;
 
   reqrsp_cut #(
     .AddrWidth (AddrWidth),
     .DataWidth (DataWidth),
+    .UserWidth (UserWidth),
     .req_t (reqrsp_req_t),
     .rsp_t (reqrsp_rsp_t),
     .BypassReq (BypassReq),

hw/reqrsp_interface/src/reqrsp_demux.sv

@@ -104,6 +104,8 @@ module reqrsp_demux_intf #(
     parameter int unsigned AddrWidth    = 0,
     /// Data width of the interface.
     parameter int unsigned DataWidth    = 0,
+    /// User width of the interface.
+    parameter int unsigned UserWidth    = 0,
     /// Amount of outstanding responses. Determines the FIFO size.
     parameter int unsigned RespDepth    = 8,
     // Dependent parameters, DO NOT OVERRIDE!
@@ -120,8 +122,9 @@ module reqrsp_demux_intf #(
   typedef logic [AddrWidth-1:0] addr_t;
   typedef logic [DataWidth-1:0] data_t;
   typedef logic [DataWidth/8-1:0] strb_t;
+  typedef logic [UserWidth-1:0] user_t;
 
-  `REQRSP_TYPEDEF_ALL(reqrsp, addr_t, data_t, strb_t)
+  `REQRSP_TYPEDEF_ALL(reqrsp, addr_t, data_t, strb_t, user_t)
 
   reqrsp_req_t reqrsp_slv_req;
   reqrsp_rsp_t reqrsp_slv_rsp;

hw/reqrsp_interface/src/reqrsp_iso.sv

@@ -14,6 +14,8 @@ module reqrsp_iso #(
     parameter int unsigned AddrWidth = 0,
     /// Data width of the interface.
     parameter int unsigned DataWidth = 0,
+    /// User width of the interface.
+    parameter int unsigned UserWidth = 0,
     /// Request type.
     parameter type req_t             = logic,
     /// Response type.
@@ -43,8 +45,9 @@ module reqrsp_iso #(
   typedef logic [AddrWidth-1:0] addr_t;
   typedef logic [DataWidth-1:0] data_t;
   typedef logic [DataWidth/8-1:0] strb_t;
+  typedef logic [UserWidth-1:0] user_t;
 
-  `REQRSP_TYPEDEF_ALL(reqrsp, addr_t, data_t, strb_t)
+  `REQRSP_TYPEDEF_ALL(reqrsp, addr_t, data_t, strb_t, user_t)
 
   isochronous_spill_register #(
     .T (reqrsp_req_chan_t),
@@ -89,6 +92,8 @@ module reqrsp_iso_intf #(
     parameter int unsigned AddrWidth = 0,
     /// Data width of the interface.
     parameter int unsigned DataWidth = 0,
+    /// User width of the interface.
+    parameter int unsigned UserWidth = 0,
     /// Bypass.
     parameter bit  BypassReq         = 0,
     parameter bit  BypassRsp         = 0
@@ -110,15 +115,17 @@ module reqrsp_iso_intf #(
   typedef logic [AddrWidth-1:0] addr_t;
   typedef logic [DataWidth-1:0] data_t;
   typedef logic [DataWidth/8-1:0] strb_t;
+  typedef logic [UserWidth-1:0] user_t;
 
-  `REQRSP_TYPEDEF_ALL(reqrsp, addr_t, data_t, strb_t)
+  `REQRSP_TYPEDEF_ALL(reqrsp, addr_t, data_t, strb_t, user_t)
 
   reqrsp_req_t reqrsp_src_req, reqrsp_dst_req;
   reqrsp_rsp_t reqrsp_src_rsp, reqrsp_dst_rsp;
 
   reqrsp_iso #(
     .AddrWidth (AddrWidth),
     .DataWidth (DataWidth),
+    .UserWidth (UserWidth),
     .req_t     (reqrsp_req_t),
     .rsp_t     (reqrsp_rsp_t),
     .BypassReq (BypassReq),

hw/reqrsp_interface/src/reqrsp_mux.sv

@@ -14,6 +14,8 @@ module reqrsp_mux #(
     parameter int unsigned               AddrWidth    = 0,
     /// Data width of the interface.
     parameter int unsigned               DataWidth    = 0,
+    /// User width of the interface.
+    parameter int unsigned               UserWidth    = 0,
     /// Request type.
     parameter type                       req_t        = logic,
     /// Response type.
@@ -39,8 +41,9 @@ module reqrsp_mux #(
   typedef logic [AddrWidth-1:0] addr_t;
   typedef logic [DataWidth-1:0] data_t;
   typedef logic [DataWidth/8-1:0] strb_t;
+  typedef logic [UserWidth-1:0] user_t;
 
-  `REQRSP_TYPEDEF_REQ_CHAN_T(req_chan_t, addr_t, data_t, strb_t)
+  `REQRSP_TYPEDEF_REQ_CHAN_T(req_chan_t, addr_t, data_t, strb_t, user_t)
 
   localparam int unsigned LogNrPorts = cf_math_pkg::idx_width(NrPorts);
 
@@ -159,6 +162,8 @@ module reqrsp_mux_intf #(
     parameter int unsigned      AddrWidth    = 0,
     /// Data width of the interface.
     parameter int unsigned      DataWidth    = 0,
+    /// User width of the interface.
+    parameter int unsigned      UserWidth    = 0,
     /// Amount of outstanding responses. Determines the FIFO size.
     parameter int unsigned      RespDepth    = 8,
     /// Cut timing paths on the request path. Incurs a cycle additional latency.
@@ -175,8 +180,9 @@ module reqrsp_mux_intf #(
   typedef logic [AddrWidth-1:0] addr_t;
   typedef logic [DataWidth-1:0] data_t;
   typedef logic [DataWidth/8-1:0] strb_t;
+  typedef logic [UserWidth-1:0] user_t;
 
-  `REQRSP_TYPEDEF_ALL(reqrsp, addr_t, data_t, strb_t)
+  `REQRSP_TYPEDEF_ALL(reqrsp, addr_t, data_t, strb_t, user_t)
 
   reqrsp_req_t [NrPorts-1:0] reqrsp_slv_req;
   reqrsp_rsp_t [NrPorts-1:0] reqrsp_slv_rsp;
@@ -188,6 +194,7 @@ module reqrsp_mux_intf #(
     .NrPorts (NrPorts),
     .AddrWidth (AddrWidth),
     .DataWidth (DataWidth),
+    .UserWidth (UserWidth),
     .req_t (reqrsp_req_t),
     .rsp_t (reqrsp_rsp_t),
     .RespDepth (RespDepth),

hw/reqrsp_interface/src/reqrsp_to_axi.sv

@@ -48,15 +48,13 @@ module reqrsp_to_axi import reqrsp_pkg::*; #(
   parameter int unsigned ID = 0,
   /// Data width of bus, must be 32 or 64.
   parameter int unsigned DataWidth = 32'b0,
-  parameter int unsigned UserWidth = 32'b0,
   parameter type reqrsp_req_t = logic,
   parameter type reqrsp_rsp_t = logic,
   parameter type axi_req_t = logic,
   parameter type axi_rsp_t = logic
 ) (
   input  logic clk_i,
   input  logic rst_ni,
-  input  logic [UserWidth-1:0] user_i,
   input  reqrsp_req_t reqrsp_req_i,
   output reqrsp_rsp_t reqrsp_rsp_o,
   output axi_req_t axi_req_o,
@@ -175,7 +173,7 @@ module reqrsp_to_axi import reqrsp_pkg::*; #(
   assign axi_req_o.ar.lock   = (reqrsp_req_i.q.amo == AMOLR);
   assign axi_req_o.ar.cache  = axi_pkg::CACHE_MODIFIABLE;
   assign axi_req_o.ar.id     = $unsigned(ID);
-  assign axi_req_o.ar.user   = user_i;
+  assign axi_req_o.ar.user   = reqrsp_req_i.q.user;
   assign axi_req_o.ar_valid  = q_valid_read;
   assign q_ready_read        = axi_rsp_i.ar_ready;
 
@@ -190,11 +188,11 @@ module reqrsp_to_axi import reqrsp_pkg::*; #(
   assign axi_req_o.aw.lock   = (reqrsp_req_i.q.amo == AMOSC);
   assign axi_req_o.aw.cache  = axi_pkg::CACHE_MODIFIABLE;
   assign axi_req_o.aw.id     = $unsigned(ID);
-  assign axi_req_o.aw.user   = user_i;
+  assign axi_req_o.aw.user   = reqrsp_req_i.q.user;
   assign axi_req_o.w.data    = write_data;
   assign axi_req_o.w.strb    = reqrsp_req_i.q.strb;
   assign axi_req_o.w.last    = 1'b1;
-  assign axi_req_o.w.user    = user_i;
+  assign axi_req_o.w.user    = reqrsp_req_i.q.user;
 
   // Both channels need to handshake (independently).
   stream_fork #(
@@ -305,12 +303,11 @@ module reqrsp_to_axi_intf #(
   parameter int unsigned AddrWidth = 32'd0,
   /// AXI and REQRSP data width.
   parameter int unsigned DataWidth = 32'd0,
-  /// AXI user width.
-  parameter int unsigned AxiUserWidth = 32'd0
+  /// AXI and REQRSP user width.
+  parameter int unsigned UserWidth = 32'd0
 ) (
   input logic clk_i,
   input logic rst_ni,
-  input logic [AxiUserWidth-1:0] user_i,
   REQRSP_BUS  reqrsp,
   AXI_BUS     axi
 );
@@ -319,9 +316,9 @@ module reqrsp_to_axi_intf #(
   typedef logic [DataWidth-1:0] data_t;
   typedef logic [DataWidth/8-1:0] strb_t;
   typedef logic [AxiIdWidth-1:0] id_t;
-  typedef logic [AxiUserWidth-1:0] user_t;
+  typedef logic [UserWidth-1:0] user_t;
 
-  `REQRSP_TYPEDEF_ALL(reqrsp, addr_t, data_t, strb_t)
+  `REQRSP_TYPEDEF_ALL(reqrsp, addr_t, data_t, strb_t, user_t)
 
   `AXI_TYPEDEF_AW_CHAN_T(aw_chan_t, addr_t, id_t, user_t)
   `AXI_TYPEDEF_W_CHAN_T(w_chan_t, data_t, strb_t, user_t)
@@ -347,7 +344,6 @@ module reqrsp_to_axi_intf #(
   ) i_reqrsp_to_axi (
     .clk_i,
     .rst_ni,
-    .user_i,
     .reqrsp_req_i (reqrsp_req),
     .reqrsp_rsp_o (reqrsp_rsp),
     .axi_req_o (axi_req),

hw/reqrsp_interface/test/reqrsp_to_axi_tb.sv

@@ -58,11 +58,10 @@ module reqrsp_to_axi_tb import reqrsp_pkg::*; #(
     .AxiIdWidth (IW),
     .AddrWidth (AW),
     .DataWidth (DW),
-    .AxiUserWidth (UW)
+    .UserWidth (UW)
   ) i_reqrsp_to_axi (
     .clk_i (clk),
     .rst_ni (rst_n),
-    .user_i ('0),
     .reqrsp (master),
     .axi (slave)
   );

hw/snitch/src/riscv_instr.sv

@@ -327,7 +327,7 @@ package riscv_instr;
   localparam logic [31:0] DMSTAT             = 32'b0000101?????00000000?????0101011;
   localparam logic [31:0] DMSTR              = 32'b0000110??????????000000000101011;
   localparam logic [31:0] DMREP              = 32'b000011100000?????000000000101011;
-  localparam logic [31:0] DMMCAST            = 32'b000100000000?????000000000101011;
+  localparam logic [31:0] DMUSER             = 32'b0001000??????????000000000101011;
   localparam logic [31:0] FREP_O             = 32'b????????????????????????10001011;
   localparam logic [31:0] IREP               = 32'b?????????????????????????0111111;
   localparam logic [31:0] SCFGRI             = 32'b????????????00000001?????0101011;
@@ -1140,7 +1140,8 @@ package riscv_instr;
   localparam logic [11:0] CSR_FPMODE = 12'h7c1;
   localparam logic [11:0] CSR_BARRIER = 12'h7c2;
   localparam logic [11:0] CSR_SC = 12'h7c3;
-  localparam logic [11:0] CSR_MCAST = 12'h7c4;
+  localparam logic [11:0] CSR_USER_LOW = 12'h7c4;
+  localparam logic [11:0] CSR_USER_HIGH = 12'h7c5;
   localparam logic [11:0] CSR_HTIMEDELTAH = 12'h615;
   localparam logic [11:0] CSR_CYCLEH = 12'hc80;
   localparam logic [11:0] CSR_TIMEH = 12'hc81;