Checkpoint conversion pull request

.gitignore

@@ -1,2 +1,10 @@
 checkpoints
 imgs/*.img
+*.img
+*.qcow2
+__pycache__/
+*.pyc
+sim_outs
+run_gen_snapshot.sh
+gem5
+bin

Notion.txt

@@ -0,0 +1,2 @@
+Find the Notion page for this project here with useful explanations for various aspects of the project:
+https://www.notion.so/gem5-Checkpoint-Conversion-2af46d7f056e804d9dc0e3a1c4758184?source=copy_link

README.md

@@ -1,80 +1,2 @@
-# QPoints
+Please follow the instructions in this [README](https://www.notion.so/README-2ea46d7f056e80d1aeb5f644d6bb0204?source=copy_link) to run gem5 simulation using QFlex snapshots.
 
-QPoints is a tool to generate a gem5 compatible checkpoints 
-using QEMU. This tool currently works for only ARM system. 
-This tool was created to reduce time spent in reaching
-a state state of long running applcations.
-
-### Requirements:
-1. QEMU
-2. Docker
-3. Python3
-   - Jinja2 python module
-4. gdb-multiarch
-5. A disk image
-6. (Optional) ARM Powered Apple Mac system. Needed this only to accelerate emulation.
-
-### Setup:
-1. Install docker, git, qemu-system packages on your host machine
-   ```
-   sudo apt-get install git docker qemu-system
-   ```
-2. Run setup script which sets up directoris and builds a docker image
-   ```
-   bash setup.sh
-   ```
-3. Download a disk image from Google Drive. Download script works in docker
-   environment. Follow below steps do get the image.
-   ```
-   bash run_docker.sh
-   bash imgs/download_image.sh
-   exit
-   ```
-   After running above set of commands `imgs` folder contains `ubuntu-arm.img` disk image.
-
-   Ubuntu image login user is `bgodala` and password is `1234`.
-
-### Steps to create a checkpoint:
-1. Launch QEMU emulation using the `run_ubuntu_linux.sh` file
-   ```
-   bash run_docker.sh
-   bash run_ubuntu.sh
-   ```
-2. After reaching a steady state invoke the following command to create a checkpoint.
-   `gen_snapshot.sh` command works in the docker environment. Start the docker and then
-   invoke checkpoint generation command.
-   ```
-   bash run_docker.sh
-   bash gen_snapshot.sh
-   exit
-   ```
-
-### Test Checkpoint:
-1. To test checkpoint gem5 needs to be built. Use the following command to build gem5.
-
-   Start docker environment.
-   ```
-   bash run_gem5_docker.sh
-   ```   
-   Build gem5.
-   ```
-   cd gem5
-   scons -j8 build/ARM/gem5.opt
-   exit
-   ```
-   If you encounter the following error while building gem5:
-   ```
-   No such file or directory: "/qpoints/gem5/fatal: detected dubious ownership in repository at '/qpoints/gem5'\nTo add an exception for this directory, call:\n\n\tgit config --global --add safe.directory /qpoints/gem5/hooks"
-   ```
-   You can fix this by following the information reported above:
-   ```
-   git config --global --add safe.directory /qpoints/gem5
-   ```
-
-
- 2. Run checkpoint in the docker environment.
-    ```
-    bash run_gem5_docker.sh
-    bash run_gem5.sh
-    exit
-    ```

docker/Dockerfile → archive/docker/Dockerfile

@@ -1,5 +1,5 @@
 #FROM arm64v8/ubuntu:latest
-FROM ubuntu:latest
+FROM ubuntu:24.04
 
 # Update package lists
 RUN apt-get update
@@ -8,10 +8,10 @@ RUN apt-get update
 RUN apt-get install -y gdb-multiarch
 RUN apt-get install -y python3
 RUN apt-get install -y python3-pip
-RUN apt-get install -y qemu
+RUN apt-get install -y qemu-utils
 RUN apt-get install -y qemu-system
-RUN pip3 install jinja2 
-RUN pip3 install gdown 
+RUN apt-get install -y python3-jinja2 
+RUN pip3 install --break-system-packages gdown 
 RUN apt-get install -y vim
 
 # Set the entrypoint to start a shell

archive/run_docker.sh

@@ -0,0 +1,2 @@
+# We don't need a separate docker. You the docker contrainer for qflex!
+#docker run --rm --privileged  --net host -v $(pwd):/qpoints -it qpoints:1.0

archive/run_gem5_docker.sh

@@ -0,0 +1 @@
+docker run --volume $(pwd):/qpoints --rm -it cloudsuitetest/gem5-qpoints:1.0 

run_ubuntu_linux.sh → archive/run_ubuntu_linux.sh

@@ -2,9 +2,10 @@ qemu-system-aarch64 -nographic -machine virt,gic-version=3 -m 16384M -cpu cortex
 -snapshot \
 -serial mon:stdio \
 -qmp tcp:localhost:4444,server,nowait -monitor telnet::45454,server,nowait \
--drive file=imgs/ubuntu-arm.img,if=none,id=drive0,cache=writeback -device virtio-blk-device,indirect_desc=false,event_idx=false,drive=drive0,bootindex=0 \
+-drive file=/qpoints/imgs/ubuntu-arm.img,if=none,id=drive0,cache=writeback -device virtio-blk-device,indirect_desc=false,event_idx=false,drive=drive0,bootindex=0 \
 -nic none \
 -bios /usr/share/qemu-efi-aarch64/QEMU_EFI.fd
+
 #-kernel kernel_arm_built_img -append "root=/dev/vda2 ro  splash quiet" \
 #-initrd initrd.img \
 

run_all.sh

@@ -0,0 +1,187 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+start_time="$(date +%s)"
+report_timing() {
+  local exit_code=$?
+  local end_time
+  end_time="$(date +%s)"
+  local elapsed=$((end_time - start_time))
+  echo "[${snapshot:-unknown}] run_all.sh completed in ${elapsed}s (exit code: ${exit_code})"
+}
+trap report_timing EXIT
+
+qemu_pid=""
+cleanup_children() {
+  if [[ -n "$qemu_pid" ]]; then
+    kill "$qemu_pid" >/dev/null 2>&1 || true
+  fi
+}
+trap 'cleanup_children; exit 130' INT TERM
+
+usage() {
+  cat <<'EOF'
+Usage: run_all.sh --qflex-ckp-dir DIR --gem5-ckp-dir DIR --cores N --mem MB \
+  --base IMAGE --snapshot NAME [--ssh-host HOST] [--ssh-user USER] \
+  [--monitor-base PORT] [--qmp-base PORT] [--ssh-base PORT]
+
+Example:
+  run_all.sh --qflex-ckp-dir qflex_checkpoints --gem5-ckp-dir gem5_checkpoints \
+    --cores 4 --mem 16384 --base web_search.qcow2 --snapshot snapshot_0
+  run_all.sh --qflex-ckp-dir qflex_checkpoints --gem5-ckp-dir gem5_checkpoints \
+    --cores 4 --mem 16384 --base web_search.qcow2 --snapshot snapshot_0 \
+    --ssh-host 127.0.0.1 --ssh-user ubuntu --monitor-base 45454 --qmp-base 4444 \
+    --ssh-base 2222
+EOF
+}
+
+qflex_ckp_dir=""
+gem5_ckp_dir=""
+cores=""
+mem=""
+base=""
+snapshot=""
+ssh_host="127.0.0.1"
+ssh_user="qflex"
+monitor_base="45454"
+qmp_base="4444"
+ssh_base="2222"
+
+while [[ $# -gt 0 ]]; do
+  case "$1" in
+    -h|--help)
+      usage
+      exit 0
+      ;;
+    --qflex-ckp-dir)
+      qflex_ckp_dir="${2:-}"
+      shift 2
+      ;;
+    --gem5-ckp-dir)
+      gem5_ckp_dir="${2:-}"
+      shift 2
+      ;;
+    --cores)
+      cores="${2:-}"
+      shift 2
+      ;;
+    --mem)
+      mem="${2:-}"
+      shift 2
+      ;;
+    --base)
+      base="${2:-}"
+      shift 2
+      ;;
+    --snapshot)
+      snapshot="${2:-}"
+      shift 2
+      ;;
+    --ssh-host)
+      ssh_host="${2:-}"
+      shift 2
+      ;;
+    --ssh-user)
+      ssh_user="${2:-}"
+      shift 2
+      ;;
+    --monitor-base)
+      monitor_base="${2:-}"
+      shift 2
+      ;;
+    --qmp-base)
+      qmp_base="${2:-}"
+      shift 2
+      ;;
+    --ssh-base)
+      ssh_base="${2:-}"
+      shift 2
+      ;;
+    *)
+      echo "Unknown argument: $1" >&2
+      usage
+      exit 1
+      ;;
+  esac
+done
+
+if [[ -z "$qflex_ckp_dir" || -z "$gem5_ckp_dir" || -z "$cores" || -z "$mem" || -z "$base" || -z "$snapshot" ]]; then
+  echo "Missing required arguments." >&2
+  usage
+  exit 1
+fi
+
+ROOT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+run_dir="${qflex_ckp_dir}/run"
+
+snapshot_idx=0
+if [[ "$snapshot" =~ _([0-9]+)$ ]]; then
+  snapshot_idx="${BASH_REMATCH[1]}"
+fi
+
+monitor_port=$((monitor_base + snapshot_idx))
+qmp_port=$((qmp_base + snapshot_idx))
+ssh_port=$((ssh_base + snapshot_idx))
+
+if [[ ! -d "$run_dir" ]]; then
+  echo "[${snapshot}] run directory not found: $run_dir" >&2
+  exit 1
+fi
+
+if [[ ! -f "$run_dir/run_qemu_emu.sh" ]]; then
+  cp -u "$ROOT_DIR/scripts/qflex/run_qemu_emu.sh" "$run_dir/"
+fi
+
+chmod +x "$run_dir/run_qemu_emu.sh"
+echo "[${snapshot}] start qemu in the background"
+(
+  cd "$run_dir"
+  exec ./run_qemu_emu.sh "$cores" "$mem" "$base" "$snapshot" \
+    "$monitor_port" "$qmp_port" "$ssh_port" > "qemu_emu_${snapshot}.log" 2>&1
+) &
+qemu_pid=$!
+
+# Wait for SSH to become available before proceeding.
+while true; do
+  if SSHPASS="qflex" sshpass -e ssh -o ConnectTimeout=2 -o StrictHostKeyChecking=no \
+    -o UserKnownHostsFile=/dev/null -p "$ssh_port" "${ssh_user}@${ssh_host}" "true" \
+    >/dev/null 2>&1; then
+    break
+  fi
+  echo "[${snapshot}] waiting for vm ssh (${ssh_user}@${ssh_host}:${ssh_port})..."
+  sleep 0.5
+done
+
+img_dest_dir="${gem5_ckp_dir}/${snapshot}"
+tmp_log="${run_dir}/gen_snapshot_${snapshot}.log"
+
+echo "[${snapshot}] start generating gem5 checkpoint"
+"$ROOT_DIR/gen_snapshot.sh" "$gem5_ckp_dir" "$snapshot" "" 0 "$cores" "$monitor_port" \
+  2> "$tmp_log"
+
+if [[ -d "$img_dest_dir" ]]; then
+  mv "$tmp_log" "$img_dest_dir/gen_snapshot_${snapshot}.log"
+else
+  echo "[${snapshot}] Destination directory does not exist: $img_dest_dir" >&2
+  exit 1
+fi
+
+if [[ ! -f "$run_dir/convert.sh" ]]; then
+  cp -u "$ROOT_DIR/scripts/qflex/convert.sh" "$run_dir/"
+fi
+
+(
+  cd "$run_dir"
+  chmod +x convert.sh
+  echo "[${snapshot}] start converting the disk image"
+  ./convert.sh "$base" "$snapshot"
+)
+
+img_src="${run_dir}/${snapshot}.img"
+if [[ -f "$img_src" ]]; then
+  echo "[${snapshot}] moving the raw disk image to the destination folder"
+  mv "$img_src" "$img_dest_dir/."
+else
+  echo "[${snapshot}] Converted image not found: $img_src" >&2
+  exit 1
+fi