Merge pull request #1671 from firesim/sudo-once

Isolate `sudo` usage

.github/scripts/build-driver.py

@@ -0,0 +1,34 @@
+#!/usr/bin/env python3
+
+import argparse
+from enum import Enum
+from fabric.api import prefix, run, execute # type: ignore
+
+from ci_variables import ci_env
+
+class FpgaPlatform(Enum):
+    vitis = 'vitis'
+    f1 = 'f1'
+    xilinx_alveo_u250 = 'xilinx_alveo_u250'
+
+    def __str__(self):
+        return self.value
+
+parser = argparse.ArgumentParser(description='')
+parser.add_argument('--platform', type=FpgaPlatform, choices=list(FpgaPlatform), required=True)
+args = parser.parse_args()
+
+def build_driver():
+    """Runs compilation of FireSim driver for the make-default tuple"""
+
+    # assumptions:
+    #   - the firesim repo is already setup in a prior script
+    #   - machine-launch-script requirements are already installed
+
+    with prefix(f"cd {ci_env['REMOTE_WORK_DIR']}"):
+        with prefix('source sourceme-manager.sh --skip-ssh-setup'):
+            with prefix("cd ./sim"):
+                run(f"make PLATFORM={args.platform} {args.platform}")
+
+if __name__ == "__main__":
+    execute(build_driver, hosts=["localhost"])

.github/scripts/check-vitis-docs-generated-components.py → .github/scripts/check-docs-generated-components-local.py

@@ -1,10 +1,22 @@
 #!/usr/bin/env python3
 
-import sys
+import argparse
+from enum import Enum
 from fabric.api import prefix, run, settings, execute # type: ignore
 
 from ci_variables import ci_env
 
+class FpgaPlatform(Enum):
+    vitis = 'vitis'
+    xilinx_alveo_u250 = 'xilinx_alveo_u250'
+
+    def __str__(self):
+        return self.value
+
+parser = argparse.ArgumentParser(description='')
+parser.add_argument('--platform', type=FpgaPlatform, choices=list(FpgaPlatform), required=True)
+args = parser.parse_args()
+
 def run_docs_generated_components_check():
     """ Runs checks to make sure generated components of vitis docs have been
     updated. """
@@ -19,7 +31,13 @@ def run_docs_generated_components_check():
         with prefix('source sourceme-manager.sh --skip-ssh-setup'):
             with prefix("cd deploy"):
                 run("cat config_runtime.yaml")
-                path = "docs/Getting-Started-Guides/On-Premises-FPGA-Getting-Started/Running-Simulations/DOCS_EXAMPLE_config_runtime.yaml"
+                if args.platform == FpgaPlatform.vitis:
+                    subpath = 'AWS-EC2-F1-Getting-Started'
+                elif args.platform == FpgaPlatform.xilinx_alveo_u250:
+                    subpath = 'On-Premises-FPGA-Getting-Started'
+                else:
+                    assert False
+                path = f'docs/Getting-Started-Guides/{subpath}/Running-Simulations/DOCS_EXAMPLE_config_runtime.yaml'
                 run(f"cat ../{path}")
                 run(f"diff config_runtime.yaml ../{path}")
 

.github/scripts/check-docs-generated-components.py

@@ -1,17 +1,37 @@
 #!/usr/bin/env python3
 
+import argparse
+from enum import Enum
 from fabric.api import cd, prefix, run, execute # type: ignore
 
 from common import manager_fsim_dir, set_fabric_firesim_pem
 
+class FpgaPlatform(Enum):
+    vitis = 'vitis'
+    f1 = 'f1'
+    xilinx_alveo_u250 = 'xilinx_alveo_u250'
+
+    def __str__(self):
+        return self.value
+
+parser = argparse.ArgumentParser(description='')
+parser.add_argument('--platform', type=FpgaPlatform, choices=list(FpgaPlatform), required=True)
+args = parser.parse_args()
+
 def run_docs_generated_components_check():
     """ Runs checks to make sure generated components of docs have been
     updated. """
 
     with cd(manager_fsim_dir), prefix('source sourceme-manager.sh'):
         with prefix("cd deploy"):
             run("cat config_runtime.yaml")
-            path = 'docs/Getting-Started-Guides/AWS-EC2-F1-Getting-Started/Running-Simulations/DOCS_EXAMPLE_config_runtime.yaml'
+            if args.platform == FpgaPlatform.f1:
+                subpath = 'AWS-EC2-F1-Getting-Started'
+            elif args.platform == FpgaPlatform.xilinx_alveo_u250 or args.platform == FpgaPlatform.f1:
+                subpath = 'On-Premises-FPGA-Getting-Started'
+            else:
+                assert False
+            path = f'docs/Getting-Started-Guides/{subpath}/Running-Simulations/DOCS_EXAMPLE_config_runtime.yaml'
             run(f"cat ../{path}")
             run(f"diff config_runtime.yaml ../{path}")
             run("firesim --help")

.github/scripts/run-linux-poweroff-vitis.py → .github/scripts/run-linux-poweroff-local.py

@@ -1,35 +1,48 @@
 #!/usr/bin/env python3
 
 import sys
+import argparse
+from enum import Enum
 from fabric.api import prefix, run, settings, execute # type: ignore
 
 from ci_variables import ci_env
 from utils import search_match_in_last_workloads_output_file
 
-def run_linux_poweroff_vitis():
-    """ Runs Base Vitis Build """
+# vitis not supported since it should be removed eventually
+class FpgaPlatform(Enum):
+    xilinx_alveo_u250 = 'xilinx_alveo_u250'
+
+    def __str__(self):
+        return self.value
+
+parser = argparse.ArgumentParser(description='')
+parser.add_argument('--platform', type=FpgaPlatform, choices=list(FpgaPlatform), required=True)
+args = parser.parse_args()
+
+def run_linux_poweroff():
+    """ Runs Linux Poweroff Tests - All Single-node Tests (Single-core Rocket/BOOM, Multi-core Rocket)"""
 
     # assumptions:
     #   - machine-launch-script requirements are already installed
-    #   - XILINX_VITIS, XILINX_XRT, XILINX_VIVADO are setup (in env / LD_LIBRARY_PATH / path / etc)
+    #   - repo is already setup fully
 
     # repo should already be checked out
 
     with prefix(f"cd {ci_env['REMOTE_WORK_DIR']}"):
-        run("./build-setup.sh --skip-validate")
         with prefix('source sourceme-manager.sh --skip-ssh-setup'):
             with prefix('cd sw/firesim-software'):
                 # build outputs.yaml (use this workload since firemarshal can guestmount)
                 run("./marshal -v build test/outputs.yaml")
                 run("./marshal -v install test/outputs.yaml")
 
-            def run_w_timeout(workload_path, workload, timeout, num_passes):
+            def run_w_timeout(workload_path, config_runtime, workload, timeout, num_passes):
+                print(f"Starting workload run {workload}.")
                 log_tail_length = 300
                 rc = 0
                 with settings(warn_only=True):
                     # avoid logging excessive amounts to prevent GH-A masking secrets (which slows down log output)
                     # pty=False needed to avoid issues with screen -ls stalling in fabric
-                    rc = run(f"timeout {timeout} {workload_path}/run-workload.sh {workload_path}/config_runtime.yaml &> {workload}.log", pty=False).return_code
+                    rc = run(f"timeout {timeout} {workload_path}/run-workload.sh {workload_path}/{config_runtime} &> {workload}.log", pty=False).return_code
                     print(f" Printing last {log_tail_length} lines of log. See {workload}.log for full info.")
                     run(f"tail -n {log_tail_length} {workload}.log")
 
@@ -41,7 +54,7 @@ def run_w_timeout(workload_path, workload, timeout, num_passes):
                     print(f"Printing last {log_tail_length} lines of all output files. See results-workload for more info.")
                     run(f"""cd deploy/results-workload/ && LAST_DIR=$(ls | tail -n1) && if [ -d "$LAST_DIR" ]; then tail -n{log_tail_length} $LAST_DIR/*/*; fi""")
 
-                    run(f"firesim terminaterunfarm -q -c {workload_path}/config_runtime.yaml")
+                    run(f"firesim terminaterunfarm -q -c {workload_path}/{config_runtime}")
 
                 if rc != 0:
                     print(f"Workload {workload} failed.")
@@ -64,7 +77,9 @@ def check(match_key, file_name = 'uartlog'):
 
                     print(f"Workload run {workload} successful.")
 
-            run_w_timeout(f"{ci_env['GITHUB_WORKSPACE']}/deploy/workloads/ci/vitis", "linux-poweroff-singlenode", "30m", 1)
+            run_w_timeout(f"{ci_env['GITHUB_WORKSPACE']}/deploy/workloads/ci/{args.platform}", "config_runtime_rocket_singlecore.yaml", "linux-poweroff-singlenode-rocketsinglecore", "30m", 1)
+            run_w_timeout(f"{ci_env['GITHUB_WORKSPACE']}/deploy/workloads/ci/{args.platform}", "config_runtime_rocket_quadcore.yaml", "linux-poweroff-singlenode-rocketquadcore", "30m", 1)
+            run_w_timeout(f"{ci_env['GITHUB_WORKSPACE']}/deploy/workloads/ci/{args.platform}", "config_runtime_boom_singlecore.yaml", "linux-poweroff-singlenode-boomsinglecore", "30m", 1)
 
 if __name__ == "__main__":
-    execute(run_linux_poweroff_vitis, hosts=["localhost"])
+    execute(run_linux_poweroff, hosts=["localhost"])

.github/scripts/run-local-buildbitstreams.py

@@ -7,7 +7,6 @@
 from github import Github
 import base64
 import time
-import argparse
 
 from ci_variables import ci_env
 
@@ -223,12 +222,16 @@ def replace_in_hwdb(hwdb_entry_name: str, link: str) -> None:
             # priority == roughly the more powerful and available
             # ipaddr, buildtool:version, use unique build dir, unique build dir path, priority (0 is highest)(unused by code but used to track which machine has most resources)
             hosts = [
+                (          "as4",  "vivado:2022.1", False, "", 0),
+                ("buildbot1@as4",  "vivado:2022.1",  True, "/scratch/buildbot1/FIRESIM_BUILD_DIR", 0),
+                ("buildbot2@as4",  "vivado:2022.1",  True, "/scratch/buildbot2/FIRESIM_BUILD_DIR", 0),
+                (    "localhost",   "vitis:2022.1", False, "", 0),
                 ("buildbot1@a17",   "vitis:2022.1",  True, "/scratch/buildbot1/FIRESIM_BUILD_DIR", 0),
-                (         "harp",   "vitis:2022.1", False, "", 2),
                 ("buildbot2@a17",   "vitis:2021.1",  True, "/scratch/buildbot2/FIRESIM_BUILD_DIR", 0),
                 ("buildbot3@a17",   "vitis:2021.1",  True, "/scratch/buildbot3/FIRESIM_BUILD_DIR", 0),
                 ("buildbot4@a17",   "vitis:2021.1",  True, "/scratch/buildbot4/FIRESIM_BUILD_DIR", 0),
                 (     "firesim1",   "vitis:2021.1", False, "", 1),
+                (         "harp",   "vitis:2022.1", False, "", 2),
                 (        "jktgz",  "vivado:2023.1", False, "", 3),
                 (       "jktqos",  "vivado:2023.1", False, "", 3),
             ]
@@ -278,16 +281,21 @@ def do_builds(batch_hwdbs):
 
             # hwdb_entry_name, platform_name, buildtool:version
             batch_hwdbs_in = [
+                # hwdb's to verify FPGA builds
                 ("vitis_firesim_rocket_singlecore_no_nic", "vitis", "vitis:2022.1"),
-
                 ("nitefury_firesim_rocket_singlecore_no_nic", "rhsresearch_nitefury_ii", "vitis:2022.1"),
-
-                ("alveo_u250_firesim_rocket_singlecore_no_nic", "xilinx_alveo_u250", "vitis:2021.1"),
-                ("alveo_u250_firesim_gemmini_rocket_singlecore_no_nic", "xilinx_alveo_u250", "vitis:2021.1"),
                 ("alveo_u200_firesim_rocket_singlecore_no_nic", "xilinx_alveo_u200", "vitis:2021.1"),
+                ("alveo_u250_firesim_rocket_singlecore_no_nic", "xilinx_alveo_u250", "vitis:2021.1"),
                 ("alveo_u280_firesim_rocket_singlecore_no_nic", "xilinx_alveo_u280", "vitis:2021.1"),
-
                 ("xilinx_vcu118_firesim_rocket_singlecore_4GB_no_nic", "xilinx_vcu118", "vivado:2023.1"),
+
+                # extra hwdb's to run CI with
+                ("alveo_u250_firesim_rocket_quadcore_no_nic", "xilinx_alveo_u250", "vivado:2022.1"),
+                ("alveo_u250_firesim_boom_singlecore_no_nic", "xilinx_alveo_u250", "vivado:2022.1"),
+                ("alveo_u250_firesim_rocket_singlecore_nic", "xilinx_alveo_u250", "vivado:2022.1"),
+
+                # extra hwdb's
+                ("alveo_u250_firesim_gemmini_rocket_singlecore_no_nic", "xilinx_alveo_u250", "vitis:2021.1"),
             ]
 
             do_builds(batch_hwdbs_in)

.github/workflows/firesim-aws-run-tests.yml

@@ -191,7 +191,7 @@ jobs:
     steps:
       - uses: actions/checkout@v4
       - name: Check docs components that require manual re-generation (e.g. config_runtime.yaml example, help output)
-        run: ./.github/scripts/check-docs-generated-components.py
+        run: ./.github/scripts/check-docs-generated-components.py --platform f1
 
   build-f1-driver:
     name: build-f1-driver
@@ -311,7 +311,7 @@ jobs:
       - uses: actions/checkout@v4
       - name: Run buildbitstream command and update sample AGFIs
         run: .github/scripts/run-agfi-buildbitstream.py
-      - uses: peter-evans/create-pull-request@v5
+      - uses: peter-evans/create-pull-request@v6
         with:
           base: ${{ github.head_ref }}
           add-paths: "deploy/sample-backup-configs/sample_config_hwdb.yaml"

.github/workflows/firesim-non-aws-run-tests.yml

@@ -33,7 +33,7 @@ env:
   LC_ALL: "en_US.UTF-8"
   CI_LABEL_DEBUG: ${{ contains(github.event.pull_request.labels.*.name, 'ci:debug') }}
   CI_LABEL_PERSIST: ${{ contains(github.event.pull_request.labels.*.name, 'ci:persist-prior-workflows') }}
-  REMOTE_WORK_DIR: /scratch/buildbot/fs-shared/fs-${{ github.sha }}
+  REMOTE_WORK_DIR: /scratch/buildbot/fs-shared/fs-${{ github.sha }}-non-aws
   TERM: xterm-256-color
 
 jobs:
@@ -191,20 +191,20 @@ jobs:
           cd ${{ env.REMOTE_WORK_DIR }}
           ./build-setup.sh --skip-validate
           source sourceme-manager.sh --skip-ssh-setup
-          firesim managerinit --platform vitis
+          firesim managerinit --platform xilinx_alveo_u250
           cd sw/firesim-software && ./init-submodules.sh
 
-  run-vitis-check-docs-generated-components:
-    name: run-vitis-check-docs-generated-components
+  run-xilinx_alveo_u250-check-docs-generated-components:
+    name: run-xilinx_alveo_u250-check-docs-generated-components
     needs: [setup-local-fpga-repo]
     runs-on: local-fpga
     steps:
       - uses: actions/checkout@v4
-      - name: Check Vitis docs components that require manual re-generation (e.g. config_runtime.yaml example)
-        run: ./.github/scripts/check-vitis-docs-generated-components.py
+      - name: Check Xilinx Alveo U250 docs components that require manual re-generation (e.g. config_runtime.yaml example)
+        run: ./.github/scripts/check-docs-generated-components-local.py --platform xilinx_alveo_u250
 
-  run-parallel-vcs-metasims-and-vitis-driver:
-    name: run-parallel-vcs-metasims-and-vitis-driver
+  run-parallel-vcs-metasims-and-xilinx_alveo_u250-driver:
+    name: run-parallel-vcs-metasims-and-xilinx_alveo_u250-driver
     needs: [setup-local-fpga-repo]
     runs-on: local-fpga
     steps:
@@ -217,13 +217,13 @@ jobs:
       - uses: actions/checkout@v4
       - name: Run parallel VCS metasims
         run: .github/scripts/run-parallel-vcs-metasims.py
-      - name: Build Vitis driver
-        run: .github/scripts/build-vitis-driver.py
+      - name: Build Xilinx Alveo U250 driver
+        run: .github/scripts/build-driver.py --platform xilinx_alveo_u250
 
-  run-basic-linux-poweroff-vitis:
+  run-basic-linux-poweroffs-xilinx_alveo_u250:
     if: contains(github.event.pull_request.labels.*.name, 'ci:fpga-deploy')
     needs: [setup-local-fpga-repo]
-    name: run-basic-linux-poweroff-vitis
+    name: run-basic-linux-poweroffs-xilinx_alveo_u250
     runs-on: local-fpga
     steps:
       # This forces a fresh clone of the repo during the `checkout` step
@@ -233,8 +233,8 @@ jobs:
           rm -rf ${{ github.workspace }}/* || true
           rm -rf ${{ github.workspace }}/.* || true
       - uses: actions/checkout@v4
-      - name: Run simple linux poweroff test w/ vitis
-        run: .github/scripts/run-linux-poweroff-vitis.py
+      - name: Run non-networked single-node Linux poweroff tests (single-core Rocket/BOOM, multi-core Rocket) w/ Xilinx Alveo U250
+        run: .github/scripts/run-linux-poweroff-local.py --platform xilinx_alveo_u250
 
   run-local-fpga-buildbitstream:
     if: contains(github.event.pull_request.labels.*.name, 'ci:local-fpga-buildbitstream-deploy')
@@ -252,7 +252,7 @@ jobs:
       - uses: actions/checkout@v4
       - name: Run buildbitstream command and update sample local bitstreams
         run: .github/scripts/run-local-buildbitstreams.py
-      - uses: peter-evans/create-pull-request@v5
+      - uses: peter-evans/create-pull-request@v6
         with:
           base: ${{ github.head_ref }}
           add-paths: "deploy/sample-backup-configs/sample_config_hwdb.yaml"
@@ -263,7 +263,7 @@ jobs:
 
   cleanup-local-fpga-repo:
     name: cleanup-local-fpga-repo
-    needs: [run-parallel-vcs-metasims-and-vitis-driver, run-basic-linux-poweroff-vitis, run-vitis-check-docs-generated-components, run-local-fpga-buildbitstream]
+    needs: [run-xilinx_alveo_u250-check-docs-generated-components, run-parallel-vcs-metasims-and-xilinx_alveo_u250-driver, run-basic-linux-poweroffs-xilinx_alveo_u250, run-local-fpga-buildbitstream]
     # uses a separate runner to cleanup (irrespective, of other jobs cancelled, running, etc)
     runs-on: local-fpga-cleanup
     if: ${{ always() }}