Merge pull request #63 from efabless/main

update verilator directory

.github/workflows/caravel_cocotb.yml

@@ -0,0 +1,49 @@
+name: Caravel Cocotb CI
+
+on:
+  push:
+    branches:
+      - '*'
+    paths:
+      - 'cocotb/caravel_cocotb/CI/**'
+      - '.github/workflows/caravel_cocotb.yml'
+      - 'cocotb/caravel_cocotb/scripts/**'
+      - '!cocotb/caravel_cocotb/**/'  # Exclude directories within the specified directory
+
+
+  pull_request:
+
+jobs:
+  build:
+    runs-on: ubuntu-latest
+
+    steps:
+    - name: Checkout code
+      uses: actions/checkout@v2
+
+    - name: Set up Python
+      uses: actions/setup-python@v2
+      with:
+        python-version: 3.8
+    - name: Set up QEMU
+      uses: docker/setup-qemu-action@v1
+
+    - name: Set up Docker Buildx
+      uses: docker/setup-buildx-action@v1
+
+    - name: Install dependencies
+      run: |
+        cd $GITHUB_WORKSPACE/cocotb
+        python3 -m pip install --upgrade pip
+        pip install -r requirements.txt
+        python3 -m pip install --upgrade --no-cache-dir volare
+    
+    - name: Install caravel_cocotb
+      run: |
+        cd $GITHUB_WORKSPACE/cocotb
+        pip install .
+
+    - name: Run main.py
+      run: |
+        cd $GITHUB_WORKSPACE/cocotb/caravel_cocotb/CI
+        python3 main.py

cocotb/README.md

@@ -21,9 +21,14 @@ This project aims to provide a user friendly environment for adding and running
 
 ## How to install caravel_cocotb
 
-This is temporary, it will soon be released to PyPi
+```bash
+ pip install caravel-cocotb
+```
+
+or to install from repo
 ```bash
    git clone [email protected]:efabless/caravel-sim-infrastructure.git
+   git checkout <release>
    cd caravel-sim-infrastructure/cocotb
    pip install .
    cd ../..
@@ -118,7 +123,7 @@ Commonly used APIs for firmware can be found in [`C_api`](docs/build/html/C_api.
 <!-- start create a test include3 -->
 # Test Examples
 
-Refer to this [directory](https://github.com/efabless/caravel_user_project/tree/cocotb_dev/verilog/dv/cocotb) for tests example generated for 16-bit counter
+Refer to this [directory](https://github.com/efabless/caravel_user_project/tree/main/verilog/dv/cocotb) for tests example generated for 16-bit counter
 <!-- end create a test include3 -->
 
 
@@ -145,7 +150,7 @@ usage: caravel_cocotb [-h] [-test TEST [TEST ...]] [-design_info DESIGN_INFO]
                       [-sdf_setup] [-clk CLK] [-lint]
                       [-macros MACROS [MACROS ...]] [-sim_path SIM_PATH]
                       [-verbosity VERBOSITY] [-openframe] [-check_commits]
-                      [-no_docker]
+                      [-no_docker] [-compile]
 
 Run cocotb tests
 
@@ -191,6 +196,7 @@ optional arguments:
   -openframe            use openframe for the simulation rather than caravel
   -check_commits        use to check if repos are up to date
   -no_docker            run iverilog without docker
+  -compile              force recompilation
 ```
 <!-- end run a test include -->
 
@@ -249,12 +255,13 @@ New directory named ``sim`` would be created under ``<repo root>/cocotb/`` or to
 
 | sim #  directory get generate when run a test
 │ ├── <tag> # tag of the run  
+│ │   ├── compilation # directory contain all logs and build files related to the RTL compilation
+│ │   │   └── compilation.log  # log file has all the commands used to run iverilog and any compilation error or warning
 │ │   ├── <sim type>-<test name> # test result directory contain all logs and wave related to the test
-│ │   │   └── <test name>.hex  # hex file used in running this test
+│ │   │   └── firmware.hex  # hex file used in running this test
 │ │   │   └── <test name>.log  # log file generated from cocotb 
-│ │   │   └── compilation.log  # log file has all the commands used to run iverilog and any compilation error or warning
 │ │   │   └── firmware.log     # log file has all the commands used to compile the C code and any compilation error or warning
-│ │   │   └── <test name>.vcd  # waves can be opened by gtkwave
+│ │   │   └── waves.vcd  # waves can be opened by gtkwave
 │ │   │   └── rerun.py         # script to rerun the test
 │ │   └── command.log    # command use for this run 
 │ │   └── repos_info.log # contain information about the repos used to run these tests 

cocotb/caravel_cocotb/CI/base_class.py

@@ -0,0 +1,14 @@
+import logging
+
+
+class BaseClass:
+    def __init__(self):
+        self.configure_logger()
+
+    def configure_logger(self):
+        self.logger = logging.getLogger(f"{self.__class__.__name__}")
+        self.logger.setLevel(logging.INFO)
+        formatter = logging.Formatter("%(levelname)s@%(asctime)s: [%(name)s] %(message)s")
+        console_handler = logging.StreamHandler()
+        console_handler.setFormatter(formatter)
+        self.logger.addHandler(console_handler)

cocotb/caravel_cocotb/CI/checker.py

@@ -0,0 +1,169 @@
+from base_class import BaseClass
+import yaml
+import os
+import xml.etree.ElementTree as ET
+
+
+class Checker(BaseClass):
+    def __init__(self):
+        super().__init__()
+        pass
+
+    def check_command(self, command):
+        self.command = command
+        self.logger.info(f"[Checker] Check test {self.command}")
+        all_tests_paths = self.get_expected_tests_dir()
+        is_pass = self.check_exist_pass(all_tests_paths)
+        if not is_pass:
+            # don't do other checks in sdf sim because it didn't run
+            return
+        self.check_configs()
+        self.check_seed(all_tests_paths)
+        self.check_dump_wave(all_tests_paths)
+        self.check_compile(all_tests_paths)
+        self.check_macros(all_tests_paths)
+        return
+
+    def get_expected_tests_dir(self):
+        self.logger.debug(f"[get_expected_tests_dir] Checking tests exist and pass test = {self.command.test} testlist = {self.command.test_list} sim = {self.command.sim}")
+        if self.command.sim is None:
+            self.logger.debug("Sim is None")
+            sims = ["RTL"]
+        else:
+            sims = self.command.sim.split()
+        expected_tests = list()
+        # to get tests run by -t 
+        if self.command.test is None:
+            self.logger.debug("Test is None")
+            tests = None
+        else:
+            tests = self.command.test.split()
+            for sim_type in sims:
+                for test_name in tests:
+                    if sim_type != "GL_SDF": 
+                        expected_tests.append(f"{sim_type}-{test_name}")
+                    else:
+                        expected_tests.append(f"{sim_type}-{test_name}-{'nom-t' if self.command.corner is None else self.command.corner}")
+        # to get tests run by -tl
+        if self.command.test_list is not None:
+            with open(self.command.test_list, 'r') as yaml_file:
+                yaml_data = yaml_file.read()
+            tests = yaml.safe_load(yaml_data).get("Tests", [])
+            for test in tests:
+                test_name = test.get("name")
+                sim_type = test.get("sim")
+                if sim_type != "GL_SDF": 
+                    expected_tests.append(f"{sim_type}-{test_name}")
+                else:
+                    expected_tests.append(f"{sim_type}-{test_name}-{'nom-t' if self.command.corner is None else self.command.corner}")
+        self.tag_path = f"{self.command.run_location if self.command.sim_path is None else self.command.sim_path}/sim/{self.command.tag}/"
+        full_tests_paths = [self.tag_path+expected_test for expected_test in expected_tests]
+        self.logger.info(f"[get_expected_tests_dir] Expected tests: {full_tests_paths}")
+        return full_tests_paths
+
+    def check_exist_pass(self, all_tests_paths):
+        for test_path in all_tests_paths:
+            if os.path.exists(test_path):
+                self.logger.info(f"[check_exist_pass] Test {test_path} exists")
+            else:
+                raise ValueError(f"[check_exist_pass] Test {test_path} doesn't exist")
+                return False
+
+            if os.path.exists(f"{test_path}/passed"):
+                self.logger.info(f"[check_exist_pass] Test {test_path} passed")
+            else:
+                if self.command.sim != "GL_SDF":
+                    raise ValueError(f"[check_exist_pass] Test {test_path} failed")
+                return False
+        return True
+
+    def check_configs(self):
+        # read design_info.yaml
+        design_info_path = self.command.design_info if self.command.design_info is not None else f"{self.command.run_location}/design_info.yaml"
+        with open(design_info_path, 'r') as yaml_file:
+            yaml_data = yaml_file.read()
+        design_info = yaml.safe_load(yaml_data)
+        caravel_root_exp = design_info.get("CARAVEL_ROOT")
+        mgmt_core_exp = design_info.get("MCW_ROOT")
+        pdk_root_exp = design_info.get("PDK_ROOT")+"/"+design_info.get("PDK")
+        pdk_exp = design_info.get("PDK")[:-1]
+        clk_exp = self.command.clk if self.command.clk is not None else design_info.get("clk")
+        max_err_exp = int(self.command.max_error) if self.command.max_error is not None else 3
+
+        # read configs.yaml generated
+        config_path = f"{self.tag_path}/configs.yaml"
+        with open(config_path, 'r') as yaml_file:
+            yaml_data = yaml_file.read()
+        configs = yaml.safe_load(yaml_data)
+        if clk_exp != configs.get("clock"):
+            raise ValueError(f"[check_configs] Clock mismatch: {clk_exp} != {configs.get('clock')}")
+        if max_err_exp != int(configs.get("max_err")):
+            raise ValueError(f"[check_configs] Max error mismatch: {max_err_exp} != {configs.get('max_err')}")
+        if caravel_root_exp != configs.get("CARAVEL_ROOT"):
+            raise ValueError(f"[check_configs] Caravel root mismatch: {caravel_root_exp} != {configs.get('CARAVEL_ROOT')}")
+        if mgmt_core_exp != configs.get("MCW_ROOT"):
+            raise ValueError(f"[check_configs] Management core mismatch: {mgmt_core_exp} != {configs.get('MCW_ROOT')}")
+        if pdk_root_exp != configs.get("PDK_ROOT"):
+            raise ValueError(f"[check_configs] PDK root mismatch: {pdk_root_exp} != {configs.get('PDK_ROOT')}")
+        if pdk_exp != configs.get("PDK"):
+            raise ValueError(f"[check_configs] PDK mismatch: {pdk_exp} != {configs.get('PDK')}")
+
+    def check_seed(self, all_tests_paths):
+        if self.command.seed is not None:
+            for test_path in all_tests_paths:
+                seed_file_path = f"{test_path}/seed.xml"
+                with open(seed_file_path, 'r') as xml_file:
+                    xml_content = xml_file.read()
+
+                # Parse the XML content
+                root = ET.fromstring(xml_content)
+
+                # Find the random_seed property and extract its value
+                seed_value = None
+                for testsuite in root.findall(".//testsuite"):
+                    seed_elem = testsuite.find(".//property[@name='random_seed']")
+                    if seed_elem is not None:
+                        seed_value = int(seed_elem.get("value"))
+                        break
+                if seed_value == self.command.seed:
+                    self.logger.info(f"[check_seed] Test run with correct seed {seed_value}")
+                else:
+                    raise ValueError(f"[check_seed] Test run with incorrect seed {seed_value} instead of {self.command.seed}")
+
+    def check_dump_wave(self, all_tests_paths):
+        dump_wave_exp = False if self.command.no_wave is not None else True
+        for test_path in all_tests_paths:
+            if os.path.exists(f"{test_path}/waves.vcd") and not dump_wave_exp:
+                raise ValueError(f"[check_dump_wave] Test {test_path} dump waves while -no_wave switch is used")
+            elif not os.path.exists(f"{test_path}/waves.vcd") and dump_wave_exp:
+                raise ValueError(f"[check_dump_wave] Test {test_path} doesn't dump waves while -no_wave switch isn't used")
+            else:
+                self.logger.info(f"[check_dump_wave] Test {test_path} has wave {'dumped' if dump_wave_exp else 'not dumped'} waves as expected")
+
+    def check_compile(self, all_tests_paths):
+        is_compile_shared = True if self.command.compile is None else False
+        for test_path in all_tests_paths:
+            simvpp_exist = os.path.exists(f"{test_path}/sim.vvp")
+            if simvpp_exist and is_compile_shared:
+                raise ValueError(f"[check_compile] Test {test_path} compile is not shared while -compile switch is used")
+            elif not simvpp_exist and not is_compile_shared:
+                raise ValueError(f"[check_compile] Test {test_path} shared compile while -compile switch isn't used simvpp exist = {simvpp_exist} is shared = {is_compile_shared}")
+            else:
+                self.logger.info(f"[check_compile] Test {test_path} has compile {'shared' if is_compile_shared else 'not shared'} as expected")
+
+    def check_macros(self, all_tests_paths):
+        macro_used = self.command.macro
+        if macro_used is not None:
+            pattern_to_search = f"Found {macro_used} effect"
+            for test_path in all_tests_paths:
+                # search pattern in all .log files
+                for filename in os.listdir(test_path):
+                    if filename.endswith('.log'):
+                        file_log = os.path.join(test_path, filename)
+                        with open(file_log, 'r') as log_file:
+                            content = log_file.read()
+                            if pattern_to_search in content:
+                                self.logger.info(f"[check_macros] Test {test_path} uses macro {macro_used} correctly")
+                                return True
+                raise ValueError(f"[check_macros] Test {test_path} doesn't use macro {macro_used}")
+                return False