Entropy Removed to Silence a Warning

Medium

· Backend

Incident

SEV-1

BUG-DSA1571

Priority: Critical · reported by security research

Time to fix: 20 min

Failing checks on the shipped build

  • different entropy gives different keys

    entropy must affect the key: expected a truthy value but got false

How to work this incident

  1. Read the report: write down the expected behaviour and the actual one in a sentence each.
  2. Reproduce it: run the visible tests and find the failing check that matches the report.
  3. 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.
  4. Follow the request through the handler, the validation and the service it calls; find the first step whose output disagrees with the report.
  5. Fix the cause rather than the symptom, run the visible tests again, and submit against the hidden ones — then write the root cause down.

Inspired by

Debian OpenSSL · CVE-2008-0166

Modelled on the Debian OpenSSL disaster (CVE-2008-0166): a maintainer removed a line that fed uninitialised memory into the random pool because it tripped a Valgrind warning. With it went almost all the entropy — the only varying input left was the process ID, so the whole distribution could only generate about 32,767 distinct keys.

keygen.js has the same shape: it accepts a seed of real entropy but never mixes it in.

Fix generateKey so the supplied entropy actually determines the key.

Code map

// Derives a key from the entropy pool and the process id.
exports.generateKey = function (entropy, pid) {
  var state = 5381;
  // NOTE: mixing the entropy pool tripped a memory checker, so it was removed.
  state = (state * 33 + pid) % 2147483647;
  state = (state * 33 + 17) % 2147483647;
  return state;
};

Sign in to edit the file, run the tests and submit a fix

A free account — no card, no trial clock. You come straight back here.

Entropy Removed to Silence a Warning — JavaScript Bug Hunt

Modelled on the Debian OpenSSL disaster (CVE-2008-0166): a maintainer removed a line that fed uninitialised memory into the random pool because it tripped a…

  • Language: JavaScript
  • Layer: Backend
  • Difficulty: Medium
  • Concepts: Security, Randomness
  • Modelled on: Debian OpenSSL · CVE-2008-0166
  • Visible tests: different entropy gives different keys; different pids still give different keys
  • Reward: 50 XP for a complete fix

Incident

  • Severity: SEV-1 (priority: Critical)
  • Ticket: BUG-DSA1571
  • Reported by: security research
  • Time to fix: 20 minutes, the target for a medium incident

Failing checks on the shipped build

  • different entropy gives different keys — entropy must affect the key: expected a truthy value but got false

How to work this incident

  1. Read the report: write down the expected behaviour and the actual one in a sentence each.
  2. Reproduce it: run the visible tests and find the failing check that matches the report.
  3. 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.
  4. Follow the request through the handler, the validation and the service it calls; find the first step whose output disagrees with the report.
  5. 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 Debian OpenSSL disaster (CVE-2008-0166): a maintainer removed a line that fed uninitialised memory into the random pool because it tripped a Valgrind warning. With it went almost all the entropy — the only varying input left was the process ID, so the whole distribution could only generate about 32,767 distinct keys.

keygen.js has the same shape: it accepts a seed of real entropy but never mixes it in.

Fix generateKey so the supplied entropy actually determines the key.

Bug report

BUG-DSA1571 · Priority: Critical · Reported by: security research

generateKey(entropy, pid):

  • must produce a key that depends on BOTH arguments
  • two different entropy values with the same pid must give different keys

Observed: every machine with pid 4242 generates the identical key regardless of the entropy handed to it. Scanning the keyspace takes seconds.

Logs

[keygen] collision: entropy=91117 and entropy=55051 produced the same key
[keygen] distinct keys observed across 10000 runs: 4

The code as shipped

src/crypto/keygen.js (editable)

// Derives a key from the entropy pool and the process id.
exports.generateKey = function (entropy, pid) {
  var state = 5381;
  // NOTE: mixing the entropy pool tripped a memory checker, so it was removed.
  state = (state * 33 + pid) % 2147483647;
  state = (state * 33 + 17) % 2147483647;
  return state;
};

Read-only context: src/crypto/POLICY.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.