Skip to content

Latest commit

 

History

History
212 lines (164 loc) · 9.36 KB

fuzzing_config.md

File metadata and controls

212 lines (164 loc) · 9.36 KB

Fuzzing Configuration

The fuzzing configuration defines the parameters for the fuzzing campaign.

workers

  • Type: Integer
  • Description: The number of worker threads to parallelize fuzzing operations on.
  • Default: 10 workers

workerResetLimit

  • Type: Integer
  • Description: The number of call sequences a worker should process on its underlying chain before being fully reset, freeing memory. After resetting, the worker will be re-created and continue processing of call sequences.

    🚩 This setting, along with workers influence the speed and memory consumption of the fuzzer. Setting this value higher will result in greater memory consumption per worker. Setting it too high will result in the in-memory chain's database growing to a size that is slower to process. Setting it too low may result in frequent worker resets that are computationally expensive for complex contract deployments that need to be replayed during worker reconstruction.

  • Default: 50 sequences

timeout

  • Type: Integer
  • Description: The number of seconds before the fuzzing campaign should be terminated. If a zero value is provided, the timeout will not be enforced. The timeout begins after compilation succeeds and the fuzzing campaign has started.
  • Default: 0 seconds

testLimit

  • Type: Integer
  • Description: The number of function calls to make before the fuzzing campaign should be terminated. If a zero value is provided, no test limit will be enforced.
  • Default: 0 calls

callSequenceLength

  • Type: Integer
  • Description: The maximum number of function calls to generate in a single call sequence in the attempt to violate properties. After every callSequenceLength function calls, the blockchain is reset for the next sequence of transactions.
  • Default: 100 calls/sequence

coverageEnabled

  • Type: Boolean
  • Description: Whether coverage-increasing call sequences should be saved for the fuzzer to mutate/re-use. Enabling coverage allows for improved code exploration.
  • Default: true

corpusDirectory

  • Type: String
  • Description: The file path where the corpus should be saved. The corpus collects sequences during a fuzzing campaign that help drive fuzzer features (e.g. a call sequence that increases code coverage is stored in the corpus). These sequences can then be re-used/mutated by the fuzzer during the next fuzzing campaign.
  • Default: ""

targetContracts

  • Type: [String] (e.g. [FirstContract, SecondContract, ThirdContract])
  • Description: The list of contracts that will be deployed on the blockchain and then targeted for fuzzing by medusa. For single-contract compilations, this value can be left as []. This, however, is rare since most projects are multi-contract compilations.

    🚩 Note that the order specified in the array is the order in which the contracts are deployed to the blockchain. Thus, if you have a corpusDirectory set up, and you change the order of the contracts in the array, the corpus may no longer work since the contract addresses of the target contracts will change. This may render the entire corpus useless.

  • Default: []

predeployedContracts

  • Type: {"contractName": "contractAddress"} (e.g.{"TestContract": "0x1234"})
  • Description: This configuration parameter allows you to deterministically deploy contracts at predefined addresses.

    🚩 Predeployed contracts do not accept constructor arguments. This may be added in the future.

  • Default: {}

targetContractBalances

  • Type: [Base-16 Strings] (e.g. [0x123, 0x456, 0x789])
  • Description: The starting balance for each contract in targetContracts. If the constructor for a target contract is marked payable, this configuration option can be used to send ether during contract deployment. Note that this array has a one-to-one mapping to targetContracts. Thus, if targetContracts is [A, B, C] and targetContractsBalances is ["0", "0xff", "0"], then B will have a starting balance of 255 wei and A and C will have zero wei. Note that the wei-value has to be hex-encoded and cannot have leading zeros. For an improved user-experience, the balances may be encoded as base-10 format strings in the future.
  • Default: []

constructorArgs

  • Type: {"contractName": {"variableName": _value}}
  • Description: If a contract in the targetContracts has a constructor that takes in variables, these can be specified here. An example can be found here.
  • Default: {}

deployerAddress

  • Type: Address
  • Description: The address used to deploy contracts on startup, represented as a hex string.

    🚩 Changing this address may render entries in the corpus invalid since the addresses of the target contracts will change.

  • Default: 0x30000

senderAddresses

  • Type: [Address]
  • Description: Defines the account addresses used to send function calls to deployed contracts in the fuzzing campaign.

    🚩 Changing these addresses may render entries in the corpus invalid since the sender(s) of corpus transactions may no longer be valid.

  • Default: [0x10000, 0x20000, 0x30000]

blockNumberDelayMax

  • Type: Integer
  • Description: Defines the maximum block number jump the fuzzer should make between test transactions. The fuzzer will use this value to make the next block's block.number between [1, blockNumberDelayMax] more than that of the previous block. Jumping block.number allows medusa to enter code paths that require a given number of blocks to pass.
  • Default: 60_480

blockTimestampDelayMax

  • Type: Integer
  • Description: The number of the maximum block timestamp jump the fuzzer should make between test transactions. The fuzzer will use this value to make the next block's block.timestamp between [1, blockTimestampDelayMax] more than that of the previous block. Jumping block.timestamptime allows medusa to enter code paths that require a given amount of time to pass.
  • Default: 604_800

blockGasLimit

  • Type: Integer
  • Description: The maximum amount of gas a block's transactions can use in total (thus defining max transactions per block).

    🚩 It is advised not to change this naively, as a minimum must be set for the chain to operate.

  • Default: 125_000_000

transactionGasLimit

  • Type: Integer
  • Description: Defines the amount of gas sent with each fuzzer-generated transaction.

    🚩 It is advised not to change this naively, as a minimum must be set for the chain to operate.

  • Default: 12_500_000

Using constructorArgs

There might be use cases where contracts in targetContracts have constructors that accept arguments. The constructorArgs configuration option allows you to specify those arguments. constructorArgs is a nested dictionary that maps contract name -> variable name -> variable value. Let's look at an example below:

// This contract is used to test deployment of contracts with constructor arguments.
contract TestContract {
    struct Abc {
        uint a;
        bytes b;
    }

    uint x;
    bytes2 y;
    Abc z;

    constructor(uint _x, bytes2 _y, Abc memory _z) {
        x = _x;
        y = _y;
        z = _z;
    }
}

contract DependentOnTestContract {
    address deployed;

    constructor(address _deployed) {
        deployed = _deployed;
    }
}

In the example above, we have two contracts TestContract and DependentOnTestContract. You will note that DependentOnTestContract requires the deployment of TestContract first so that it can accept the address of where TestContract was deployed. On the other hand, TestContract requires _x, _y, and _z. Here is what the constructorArgs value would look like for the above deployment:

Note: The example below has removed all the other project configuration options outside of targetContracts and constructorArgs

{
  "fuzzing": {
    "targetContracts": ["TestContract", "DependentOnTestContract"],
    "constructorArgs": {
      "TestContract": {
        "_x": "123456789",
        "_y": "0x5465",
        "_z": {
          "a": "0x4d2",
          "b": "0x54657374206465706c6f796d656e74207769746820617267756d656e7473"
        }
      },
      "DependentOnTestContract": {
        "_deployed": "DeployedContract:TestContract"
      }
    }
  }
}

First, let us look at targetContracts. As mentioned in the documentation for targetContracts, the order of the contracts in the array determine the order of deployment. This means that TestContract will be deployed first, which is what we want.

Now, let us look at constructorArgs. TestContract's dictionary specifies the exact name of the constructor argument (e.g. _x or _y) with their associated value. Since _z is of type TestContract.Abc, _z is also a dictionary that specifies each field in the TestContract.Abc struct.

For DependentOnTestContract, the _deployed key has a value of DeployedContract:TestContract. This tells medusa to look for a deployed contract that has the name TestContract and provide its address as the value for _deployed. Thus, whenever you need a deployed contract's address as an argument for another contract, you must follow the format DeployedContract:<ContractName>.