Post on the oss-sec list about password authentication bypass in OpenSSH with musl libc

What happened and what to do

The post, published on the oss-sec list on 5 November by Jeffrey Walton, addresses the "sshole" discussion, a password authentication bypass in OpenSSH combined with musl libc. The author goes beyond the specific case: he argues that using a single "*" character, or a short string of random characters, so that hashes do not match likely creates conditions for a side channel or timing attack. According to him, cryptographic routines should operate in as constant a time as possible, and the early deviation caused by differing string lengths provides the side channel.

In practice, a company running SSH servers on musl-based images can review how its authentication code compares password hashes, checking for early exits that depend on content or input length. It can also measure login response time for invalid, locked or placeholder-hash accounts in a controlled lab with repeated measurements. As a fix, the team can adopt constant-time comparisons, standardize handling of accounts without a valid password, and include these tests in its build and staging pipeline. Monitoring authentication latency in production also helps detect anomalies.

How the consultancy can help

Diagnosis of exposure in servers and container images that run OpenSSH with musl libc, including a review of the password authentication flow and hash handling. Then a scoped implementation of timing tests in a lab, recommended fixes, and continuous monitoring of login latency.

Next step

If your company runs OpenSSH on musl libc environments, send a short description of your infrastructure and password access flows. Based on it, Wendelmaques will prepare a scoped proposal for diagnosis and implementation.