Setup "debugging" and misc cleanup

.github/actions/install-deps-action/action.yml

@@ -4,23 +4,35 @@ runs:
   using: 'composite'
   steps:
     ### OTHER REPOS ####
-
-    # Hard turn-off interactive mode
-    - run: echo 'debconf debconf/frontend select Noninteractive' | sudo debconf-set-selections
-      shell: bash
-
-    # Refresh packages list
-    - run: sudo apt update
+    # turn off interactive, refresh pkgs, use apt fast
+    - run: |
+        echo 'debconf debconf/frontend select Noninteractive' | sudo debconf-set-selections
+        sudo rm /var/lib/man-db/auto-update
+        echo "deb [signed-by=/etc/apt/keyrings/apt-fast.gpg] http://ppa.launchpad.net/apt-fast/stable/ubuntu $(source /etc/os-release && echo $UBUNTU_CODENAME) main" | sudo tee /etc/apt/sources.list.d/apt-fast.list
+        wget -q -O- "https://keyserver.ubuntu.com/pks/lookup?op=get&search=0xBC5934FD3DEBD4DAEA544F791E2824A7F22B44BD" | sudo gpg --dearmor -o /etc/apt/keyrings/apt-fast.gpg
+        sudo apt-get update -y
+        sudo apt-get install -y apt-fast aria2 tasksel
+        echo 'debconf apt-fast/maxdownloads string 100' |  sudo debconf-set-selections
+        echo 'debconf apt-fast/dlflag boolean true' | sudo debconf-set-selections  
+        echo 'debconf apt-fast/aptmanager string apt-get' | sudo debconf-set-selections
+        sudo tasksel remove ubuntu-desktop
       shell: bash
 
     ### DOWNLOAD AND INSTALL DEPENDENCIES ###
 
     # Download dependencies packaged by Ubuntu
-    - run: sudo apt -y install bison busybox-static cargo cmake coreutils cpio elfutils file flex gcc gcc-multilib git iproute2 jq kbd kmod libcap-dev libelf-dev libunwind-dev libvirt-clients libzstd-dev linux-headers-generic linux-tools-common linux-tools-generic make ninja-build pahole pkg-config python3-dev python3-pip python3-requests qemu-kvm rsync rustc stress-ng udev zstd libseccomp-dev libcap-ng-dev llvm clang python3-full pipx curl meson
+    - run: | 
+        sudo apt-fast install -f -y bison busybox-static cmake coreutils \
+        cpio elfutils file flex gcc gcc-multilib git iproute2 jq kbd kmod \
+        libcap-dev libelf-dev libunwind-dev libvirt-clients libzstd-dev \
+        linux-headers-generic linux-tools-common linux-tools-generic make \
+        ninja-build pahole pkg-config python3-dev python3-pip python3-requests \
+        qemu-kvm rsync stress-ng udev zstd libseccomp-dev libcap-ng-dev \
+        llvm clang python3-full curl meson bpftrace cargo rustc dwarves
       shell: bash
 
     # virtme-ng
-    - run: pip3 install virtme-ng --break-system-packages
+    - run: sudo pip3 install virtme-ng --break-system-packages
       shell: bash
 
     # Setup KVM support

.github/workflows/caching-build.yml

@@ -6,7 +6,7 @@ on:
       - cron: "0 * * * *"
     push:
     pull_request:
-      
+
 jobs:
   lint:
     runs-on: ubuntu-24.04
@@ -32,8 +32,8 @@ jobs:
       - run: sudo chown root /usr/bin/tar && sudo chmod u+s /usr/bin/tar
       # redundancy to exit fast
       - run: echo 'debconf debconf/frontend select Noninteractive' | sudo debconf-set-selections
-      - run: sudo apt update
-      - run: sudo apt install -y git --no-install-recommends
+      - run: sudo apt-get update
+      - run: sudo apt-get install -y git --no-install-recommends
       # get latest head commit of sched_ext for-next
       - run: echo "SCHED_EXT_KERNEL_COMMIT=$(git ls-remote https://git.kernel.org/pub/scm/linux/kernel/git/tj/sched_ext.git heads/for-next | awk '{print $1}')" >> $GITHUB_ENV
 
@@ -45,8 +45,10 @@ jobs:
         uses: actions/cache@v4
         with:
           path: |
-            linux
-          key: kernel-build-${{ env.SCHED_EXT_KERNEL_COMMIT }}
+            linux/arch/x86/boot/bzImage
+            linux/usr/include
+            linux/**/*.h
+          key: kernel-build-${{ env.SCHED_EXT_KERNEL_COMMIT }}-4
 
       - if: ${{ steps.cache-kernel.outputs.cache-hit != 'true' }}
         uses: ./.github/actions/install-deps-action
@@ -62,14 +64,6 @@ jobs:
       - if: ${{ steps.cache-virtiofsd.outputs.cache-hit != 'true' && steps.cache-kernel.outputs.cache-hit != 'true' }}
         run: cargo install virtiofsd && sudo cp -a ~/.cargo/bin/virtiofsd /usr/lib/
 
-      # cache bzImage alone for rust tests (disk space limit workaround)
-      - name: Cache bzImage
-        id: cache-bzImage
-        uses: actions/cache@v4
-        with:
-          path: |
-            linux/arch/x86/boot/bzImage
-          key: kernel-bzImage-${{ env.SCHED_EXT_KERNEL_COMMIT }}
 
       - if: ${{ steps.cache-kernel.outputs.cache-hit != 'true' }}
         name: Clone Kernel
@@ -96,14 +90,18 @@ jobs:
   integration-test:
     runs-on: ubuntu-24.04
     needs: build-kernel
-    continue-on-error: true    
     strategy:
           matrix:
             scheduler: [ scx_bpfland, scx_lavd, scx_layered, scx_rlfifo, scx_rustland, scx_rusty ]
+          fail-fast: false
     steps:
       # prevent cache permission errors
       - run: sudo chown root /usr/bin/tar && sudo chmod u+s /usr/bin/tar      
       - uses: actions/checkout@v4
+      - uses: Swatinem/rust-cache@v2
+        with:
+          key: ${{ matrix.scheduler }}
+          prefix-key: "4"
       - uses: ./.github/actions/install-deps-action
       # cache virtiofsd (goes away w/ 24.04)
       - name: Cache virtiofsd
@@ -125,8 +123,10 @@ jobs:
         uses: actions/cache@v4
         with:
           path: |
-            linux
-          key: kernel-build-${{ env.SCHED_EXT_KERNEL_COMMIT }}
+            linux/arch/x86/boot/bzImage
+            linux/usr/include
+            linux/**/*.h
+          key: kernel-build-${{ env.SCHED_EXT_KERNEL_COMMIT }}-4
 
       # need to re-run job when kernel head changes between build and test running.
       - if: ${{ steps.cache-kernel.outputs.cache-hit != 'true' }}
@@ -139,7 +139,7 @@ jobs:
       - run: sudo chmod +x /usr/bin/veristat && sudo chmod 755 /usr/bin/veristat
 
       # The actual build:
-      - run: meson setup build -Dkernel=$(pwd)/linux -Dkernel_headers=./linux/usr/include -Denable_stress=true
+      - run: meson setup build -Dkernel=../linux/arch/x86/boot/bzImage -Dkernel_headers=../linux -Denable_stress=true -Dvng_rw_mount=true
       - run: meson compile -C build ${{ matrix.scheduler }}
 
       # Print CPU model before running the tests (this can be useful for
@@ -148,13 +148,35 @@ jobs:
 
       # Test schedulers
       - run: meson compile -C build test_sched_${{ matrix.scheduler }}
+      # this is where errors we want logs on start occurring, so always generate debug info and save logs
+        if: always()
       # Stress schedulers
       - uses: cytopia/[email protected]
         name: stress test
+        if: always()
         with:
           retries: 3
           command: meson compile -C build stress_tests_${{ matrix.scheduler }}
       - run: meson compile -C build veristat_${{ matrix.scheduler }}
+        if: always()
+      - run: sudo cat /var/log/dmesg > host-dmesg.ci.log
+        if: always()
+      - run: echo "NICE_REF=${{ github.event.pull_request && github.head_ref || github.ref_name }}" >> $GITHUB_ENV
+        if: always()
+      - run: mkdir -p ./log_save/
+        if: always()
+      # no symlink following here (to avoid cycles)
+      - run: sudo find '/home/runner/' -iname '*.ci.log' -exec mv {} ./log_save/ \;
+        if: always()
+      - name: upload debug logs, bpftrace, veristat, dmesg, etc.
+        if: always()
+        uses: actions/upload-artifact@v4
+        with:
+          name: ${{ matrix.scheduler }}_logs_${{ env.NICE_REF }}_${{ github.run_id }}_${{ github.run_attempt }}
+          path: ./log_save/*.ci.log
+          # it's all txt files w/ 90 day retention, lets be nice.
+          compression-level: 9
+
 
   rust-test-core:
     runs-on: ubuntu-24.04
@@ -166,6 +188,10 @@ jobs:
       # prevent cache permission errors
       - run: sudo chown root /usr/bin/tar && sudo chmod u+s /usr/bin/tar      
       - uses: actions/checkout@v4
+      - uses: Swatinem/rust-cache@v2
+        with:
+          key: ${{ matrix.component }}
+          prefix-key: "4"
       - uses: ./.github/actions/install-deps-action
       # cache virtiofsd (goes away w/ 24.04)
       - name: Cache virtiofsd
@@ -180,28 +206,24 @@ jobs:
 
       # get latest head commit of sched_ext for-next
       - run: echo "SCHED_EXT_KERNEL_COMMIT=$(git ls-remote https://git.kernel.org/pub/scm/linux/kernel/git/tj/sched_ext.git heads/for-next | awk '{print $1}')" >> $GITHUB_ENV
-      # cache bzImage alone for rust tests
-      - name: Cache bzImage
-        id: cache-bzImage
+
+      - name: Cache Kernel
+        id: cache-kernel
         uses: actions/cache@v4
         with:
           path: |
             linux/arch/x86/boot/bzImage
-          key: kernel-bzImage-${{ env.SCHED_EXT_KERNEL_COMMIT }}
+            linux/usr/include
+            linux/**/*.h
+          key: kernel-build-${{ env.SCHED_EXT_KERNEL_COMMIT }}-4
 
       # need to re-run job when kernel head changes between build and test running.
-      - if: ${{ steps.cache-bzImage.outputs.cache-hit != 'true' }}
+      - if: ${{ steps.cache-kernel.outputs.cache-hit != 'true' }}
         name: exit if cache stale
         run: exit -1
 
-      - uses: Swatinem/rust-cache@v2
-        with:
-          workspaces: rust
-          key: ${{ matrix.component }}
-          prefix-key: "1"
       - run: cargo build  --manifest-path rust/${{ matrix.component }}/Cargo.toml
-      - run: cargo test --manifest-path rust/${{ matrix.component }}/Cargo.toml --no-run
-      - run: vng -v --memory 10G --cpu 8 -r linux/arch/x86/boot/bzImage  --net user -- cargo test --manifest-path rust/${{ matrix.component }}/Cargo.toml
+      - run: vng -v --rw --memory 10G --cpu 8 -r linux/arch/x86/boot/bzImage  --net user -- cargo test --manifest-path rust/${{ matrix.component }}/Cargo.toml
 
   rust-test-schedulers:
     runs-on: ubuntu-24.04
@@ -213,6 +235,10 @@ jobs:
       # prevent cache permission errors
       - run: sudo chown root /usr/bin/tar && sudo chmod u+s /usr/bin/tar      
       - uses: actions/checkout@v4
+      - uses: Swatinem/rust-cache@v2
+        with:
+          key: ${{ matrix.scheduler }}
+          prefix-key: "4"
       - uses: ./.github/actions/install-deps-action
       # cache virtiofsd (goes away w/ 24.04)
       - name: Cache virtiofsd
@@ -227,28 +253,24 @@ jobs:
 
       # get latest head commit of sched_ext for-next
       - run: echo "SCHED_EXT_KERNEL_COMMIT=$(git ls-remote https://git.kernel.org/pub/scm/linux/kernel/git/tj/sched_ext.git heads/for-next | awk '{print $1}')" >> $GITHUB_ENV
-      # cache bzImage alone for rust tests
-      - name: Cache bzImage
-        id: cache-bzImage
+      # Cache Kernel alone for rust tests
+      - name: Cache Kernel
+        id: cache-kernel
         uses: actions/cache@v4
         with:
           path: |
             linux/arch/x86/boot/bzImage
-          key: kernel-bzImage-${{ env.SCHED_EXT_KERNEL_COMMIT }}
+            linux/usr/include
+            linux/**/*.h
+          key: kernel-build-${{ env.SCHED_EXT_KERNEL_COMMIT }}-4
 
       # need to re-run job when kernel head changes between build and test running.
-      - if: ${{ steps.cache-bzImage.outputs.cache-hit != 'true' }}
+      - if: ${{ steps.cache-kernel.outputs.cache-hit != 'true' }}
         name: exit if cache stale
         run: exit -1
 
-      - uses: Swatinem/rust-cache@v2
-        with:
-          workspaces: scheds/rust
-          key: ${{ matrix.scheduler }}
-          prefix-key: "1"
       - run: cargo build  --manifest-path scheds/rust/${{ matrix.scheduler }}/Cargo.toml
-      - run: cargo test --manifest-path scheds/rust/${{ matrix.scheduler }}/Cargo.toml --no-run
-      - run: vng -v --memory 10G --cpu 8 -r linux/arch/x86/boot/bzImage  --net user -- cargo test --manifest-path scheds/rust/${{ matrix.scheduler }}/Cargo.toml
+      - run: vng -v --rw --memory 10G --cpu 8 -r linux/arch/x86/boot/bzImage  --net user -- cargo test --manifest-path scheds/rust/${{ matrix.scheduler }}/Cargo.toml
 
   pages:
     runs-on: ubuntu-24.04
@@ -270,8 +292,7 @@ jobs:
           rustup install nightly
           export PATH="~/.cargo/bin:$PATH"
           RUSTDOCFLAGS="--enable-index-page -Zunstable-options" ~/.cargo/bin/cargo +nightly doc --workspace --no-deps --bins --lib --examples --document-private-items --all-features
-          sudo apt update
-          sudo apt install build-essential graphviz sphinx-doc python3-sphinx-rtd-theme texlive-latex-recommended python3-yaml -y
+          sudo apt-fast install build-essential graphviz sphinx-doc python3-sphinx-rtd-theme texlive-latex-recommended python3-yaml -y
           cargo install htmlq
           git clone --single-branch -b for-next --depth 1 https://git.kernel.org/pub/scm/linux/kernel/git/tj/sched_ext.git linux
           cd linux

.github/workflows/sched-ext.config

@@ -32,3 +32,23 @@ CONFIG_PREEMPT_RCU=y
 #
 CONFIG_DEBUG_LOCKDEP=y
 CONFIG_DEBUG_ATOMIC_SLEEP=y
+
+# Bpftrace headers (for additional debug info)
+CONFIG_BPF=y
+CONFIG_BPF_SYSCALL=y
+CONFIG_BPF_JIT=y
+CONFIG_HAVE_EBPF_JIT=y
+CONFIG_BPF_EVENTS=y
+CONFIG_FTRACE_SYSCALLS=y
+CONFIG_FUNCTION_TRACER=y
+CONFIG_HAVE_DYNAMIC_FTRACE=y
+CONFIG_DYNAMIC_FTRACE=y
+CONFIG_HAVE_KPROBES=y
+CONFIG_KPROBES=y
+CONFIG_KPROBE_EVENTS=y
+CONFIG_ARCH_SUPPORTS_UPROBES=y
+CONFIG_UPROBES=y
+CONFIG_UPROBE_EVENTS=y
+CONFIG_DEBUG_FS=y
+# more bpftrace to make that work
+CONFIG_IKHEADERS=y

meson-scripts/run_stress_tests

@@ -40,21 +40,36 @@ def run_stress_test(
     vng_path: str,
     kernel: str,
     verbose: bool,
+    rw: bool,
+    headers: str,
 ) -> int:
     scheduler_args = config.get('scheduler_args')
     stress_cmd = config.get('stress_cmd')
     s_path = sched_path(build_dir, config.get('sched'))
     sched_cmd = s_path + " " + config.get('sched_args')
     timeout_sec = int(config.get("timeout_sec"))
     if vng_path:
+        cmd = [vng_path, "--user", "root", "-v", "-r", kernel]
+        if config.get("qemu_opts"):
+            cmd += ['--qemu-opts']
+            cmd += [f"'{config.get("qemu_opts")}'"]
         vm_input = f"{stress_cmd} & timeout --foreground --preserve-status {timeout_sec} {sched_cmd}"
         if bpftrace_scripts := config.get('bpftrace_scripts'):
             vm_input = f"\"{build_dir}/bpftrace_stress_wrapper.sh\" '{stress_cmd}' '{sched_cmd}' '{timeout_sec}' '{bpftrace_scripts}'"
-        cmd = [vng_path, "--user", "root", "-v", "--", vm_input]
+            if headers:
+                vm_input += f" '{headers}'"
+        if rw and os.getenv('CI'):
+            print('mounting VNG as RW because CI')
+            cmd += ["--rw"]
+        elif rw:
+            print('not mounting VNG as RW because not CI')
+        cmd += ["--"]
+        cmd += [vm_input]
         err = sys.stderr if output == "-" else open(output, "w")
         out = sys.stdout if output == "-" else err
+        print(f"vng cmd is {cmd}")
         proc = subprocess.Popen(
-            cmd, env=os.environ, cwd=kernel, shell=False, stdout=out,
+            cmd, env=os.environ, shell=False, stdout=out,
             stderr=err, stdin=subprocess.PIPE, text=True)
         proc.wait()
         return proc.returncode
@@ -85,6 +100,8 @@ def stress_tests(args: Namespace) -> None:
             vng_path,
             args.kernel,
             args.verbose,
+            args.rw,
+            args.headers
         )
     for test_name, ret in return_codes.items():
         if ret not in (143, 0):
@@ -114,6 +131,12 @@ if __name__ == "__main__":
     parser.add_argument(
         '--sched', default='', help='Scheduler to test (default: all)'
     )
+    parser.add_argument(
+        '--rw', default=False, help='Mount VNG Directories as RW (dangerous)'
+    )
+    parser.add_argument(
+        '--headers', default='', help='Kernel Headers Path'
+    )
 
     args = parser.parse_args()
     if args.verbose:

meson-scripts/test_sched

@@ -17,6 +17,16 @@ GUEST_TIMEOUT=60
 #
 declare -A SCHEDS
 
+VNG_RW=''
+
+# Enable vng rw for when on ci.
+if [ $# -ge 3 ] ; then
+  if [ "$3" == "VNG_RW=true" ]; then
+    echo 'setting vng to mount rw'
+    VNG_RW=' --rw '
+  fi
+fi
+
 # enable running tests on individual schedulers
 if [ $# -ge 2 ] ; then
     SCHEDS[$2]=""
@@ -63,7 +73,7 @@ for sched in ${!SCHEDS[@]}; do
 
     rm -f /tmp/output
     timeout --preserve-status ${GUEST_TIMEOUT} \
-        vng -m 2G -v -r ${kernel} -- \
+        vng --user root -m 10G --cpu 8 $VNG_RW -v -r ${kernel} -- \
             "timeout --foreground --preserve-status ${TEST_TIMEOUT} ${sched_path} ${args}" \
                 2> >(tee /tmp/output) </dev/null
         grep -v " Speculative Return Stack Overflow" /tmp/output | \
@@ -79,4 +89,5 @@ for sched in ${!SCHEDS[@]}; do
     else
         echo "OK: ${sched}"
     fi
+    cp /tmp/output test_log.ci.log
 done