Withdraw, Then Withdraw Again — JavaScript Bug Hunt
Modelled on The DAO attack (June 2016): The DAO's contract sent ether to a caller before updating the caller's balance.
- Language: JavaScript
- Layer: Backend
- Difficulty: Hard
- Concepts: Security, Money, State
- Modelled on: The DAO · 2016
- Visible tests: an honest withdrawal pays the balance; re-entering withdraw pays nothing extra
- Reward: 50 XP for a complete fix
Incident
- Severity: SEV-1 (priority: Critical (funds drained))
- Ticket: BUG-REENTRY
- Reported by: audit
- Time to fix: 35 minutes, the target for a hard incident
Failing checks on the shipped build
- re-entering withdraw pays nothing extra — only mallory's own balance: expected 10 but got 40
How to work this incident
- Read the report: write down the expected behaviour and the actual one in a sentence each.
- Reproduce it: run the visible tests and find the failing check that matches the report.
- Read the logs and the failing checks for the first value that is wrong — the earliest wrong value is nearer the cause than the loudest error.
- Follow the request through the handler, the validation and the service it calls; find the first step whose output disagrees with the report.
- Fix the cause rather than the symptom, run the visible tests again, and submit against the hidden ones — then write the root cause down.
Briefing
Modelled on The DAO attack (June 2016): The DAO's contract sent ether to a caller before updating the caller's balance. The recipient was itself a contract whose receive hook called straight back into the withdrawal, and every nested call still saw the untouched balance. Roughly a third of The DAO's funds were drained, and Ethereum hard-forked to reverse it.
vault.js has the same order of operations. The re-entrant recipient is modelled deterministically in wallets.js: its receive calls withdraw again a fixed number of times.
Fix withdraw so a nested call can never pay out the same balance twice.
Bug report
BUG-REENTRY · Priority: Critical (funds drained) · Reported by: audit
vault.withdraw(who, wallet):
- pays the caller's whole balance by calling wallet.receive(amount, vault) once and returns amount
- a balance of 0 returns 0 without calling receive
- receive may call vault.withdraw again (re-entrancy); that nested call must see the balance as already paid out and return 0
- after any sequence of calls, vault.reserve equals the sum of the balances still held
- if receive throws, the withdrawal is undone — the balance and the reserve are restored — and the error propagates
Observed: an attacker with a balance of 10 re-entered three times, was paid 40, and the honest depositor's 30 could no longer be withdrawn.
Logs
[vault] withdraw mallory 10 (depth 4)
[vault] reserve 40 -> 0, balances { mallory: 0, alice: 30 }
[vault] withdraw alice: Error: insufficient reserveThe code as shipped
src/vault/vault.js (editable)
exports.createVault = function () {
var vault = { balances: {}, reserve: 0 };
vault.deposit = function (who, amount) {
vault.balances[who] = (vault.balances[who] || 0) + amount;
vault.reserve += amount;
};
vault.withdraw = function (who, wallet) {
var amount = vault.balances[who] || 0;
if (amount === 0) return 0;
if (amount > vault.reserve) throw new Error("insufficient reserve");
wallet.receive(amount, vault);
vault.reserve -= amount;
vault.balances[who] = 0;
return amount;
};
return vault;
};
Read-only context: src/vault/wallets.js.
Open the hunt to start the incident clock, edit the files, run the visible tests and submit against the hidden ones, then write the root cause down. More JavaScript bug hunts.