Skip to content

Latest commit

 

History

History
117 lines (104 loc) · 4.4 KB

docker.md

File metadata and controls

117 lines (104 loc) · 4.4 KB
Raw

Using DR.CHECKER from docker

Thanks to Sebastiano, we created a dockerized environment for DR.CHECKER. We have three images main, server and client:

  • main: This is the main analysis image, that has all the LLVM setup and can the actual DR.CHECKER analysis.
  • server: This is the visualization server, that processes the JSONs and serves them.
  • client: This is the visualization client, that serves the webpage for viewing DR.CHECKER results.

By default,

  • The folder: <repo_path>/docker/dockershare will be mounted at /dockershare in main and server image.
  • This git repo will be present in the folder /dr_checker/gitrepo of all the images.
  • server image has its port 5000 mapped to port 5000 of the host machine.
  • client image has its port 5000 mapped to port 8080 of the host machine.

Please follow the below instructions on how to use it.

Installing Docker

If you already have docker and docker-compose on your machine. Please ignore this step.

cd <repo_path>/docker/
./install_docker.sh

Setting up folders

First, copy the kernel sources to <repo_path>/docker/dockershare folder.

Getting images

  • Pull the images 
    cd <repo_path>/docker/
docker-compose pull

All the building of kernel and the LLVM bitcode files have to be done on the main docker image.

Running the analysis

First, run and log into the main container:

cd <repo_path>/docker/
docker-compose run main /bin/bash

From inside the main container follow all the instructions from: Building to Running Analysis

NOTE: ALL THE FILES (like toolchains) NEEDED TO BUILD THE KERNEL SHOULD BE IN THE FOLDER: <repo_path>/docker/dockershare

Viewing results

For this we need to run both the server and client images:

Running server

Run and log into the server container:

cd <repo_path>/docker/
docker-compose run server /bin/bash

Follow the Steps 1 and 2 of Running Visualizer

Running client

Run and log into the client container:

cd <repo_path>/docker/
docker-compose run client /bin/bash
## inside the new shell
cd /dr_checker/gitrepo/visualizer
serve -s ./client/build

Accessing the web-interface

You can view all the results from: http:://localhost:8080

Example

We have uploaded a mediatek kernel 33.2.A.3.123.tar.bz2. First download and extract the above file.

Extract the above file into the folder: <repo_path>/docker/dockershare/mediatek_kernel

Building and Running the analysis

cd <repo_path>/docker/
docker-compose run main /bin/bash
# inside the new shell
cd /dockershare/mediatek_kernel
source ./env.sh
cd kernel-3.18
# the following step may not be needed depending on the kernel
mkdir out
make O=out ARCH=arm64 tubads_defconfig

# this following command copies all the compilation commands to makeout.txt (This might take time)
make V=1 -j8 O=out ARCH=arm64 > makeout.txt 2>&1
cd /dr_checker/gitrepo/helper_scripts/runner_scripts

python run_all.py -l /dockershare/mediatek_kernel/llvm_bitcode_out -a 1 -m /dockershare/mediatek_kernel/kernel-3.18/makeout.txt -g aarch64-linux-android-gcc -n 2 -o /dockershare/mediatek_kernel/kernel-3.18/out -k /dockershare/mediatek_kernel/kernel-3.18 -f /dockershare/mediatek_kernel/dr_checker_out

The above command takes quite some time (30 min - 1hr). All the results JSONs will be present at /dockershare/mediatek_kernel/dr_checker_out (value of the option -l)

Visualizing the results

Server

Run the server container on your machine:

cd <repo_path>/docker/
docker-compose run server /bin/bash

With in the running container, change the results dir in server config file: /dr_checker/gitrepo/visualizer/server/config.py

RESULTS_DIR=/dockershare/mediatek_kernel/dr_checker_out

Run the server (in the container shell)

cd /dr_checker/gitrepo/visualizer
python ./server/app.py

Client

Open a new terminal on your machine and run the client container:

cd <repo_path>/docker/
docker-compose run client /bin/bash
# with in the container shell
cd /dr_checker/gitrepo/visualizer
serve -s ./client/build

On your machine, open a browser and go to : http://localhost:8080. Enjoy 😄