CVE-2026-54872

Issue summary: The generic elliptic-curve scalar multiplication used for
ECDSA and SM2 signature operations with curves that do not have a dedicated
implementation leaks information about the secret nonce through timing.

Impact summary: An attacker able to measure signing times may learn
information about the per-signature secret nonce, which over many signatures
can, via a lattice / Hidden Number Problem attack, lead to recovery of the
private key.

CWE: CWE-208: Observable Timing Discrepancy

Description: The generic elliptic-curve scalar multiplication used for
curves that do not have a dedicated constant-time implementation pads the
secret scalar with non-constant-time BIGNUM operations, so the time taken
depends on the value of the secret scalar derived from the ECDSA and SM2 nonce.

The leak is very small; observing it requires a large number of
measurements. The effect is largest for curves whose group order lies
on a machine-word boundary, such as brainpoolP384r1.

Applications using ECDSA signing over the Brainpool and other generic prime
curves, and SM2 signing on platforms that use the generic implementation,
are vulnerable to this issue.

The NIST curves P-256, P-384 and P-521 use dedicated constant-time
implementations and are not affected.

FIPS Impact: no
The FIPS modules are not affected: the approved NIST curves used in the FIPS
provider have dedicated constant-time implementations and do not use the
affected code path.