CVE-2026-54890

BEAM VM crash via integer underflow in binary_to_term BIT_BINARY_EXT decoding

HIGH
7.5
CVSS Score
Published: July 27, 2026 at 03:39 PM UTC
Updated: October 24, 2026 at 01:10 PM UTC
2 products affected across 1 vendorErlang /otpErts

Using OTP by Erlang?

Get instant alerts when new vulnerabilities affect OTP - before attackers find them.

Real-time alertsTrack your software stackFree plan available
Severity
HIGH7.5
Exploit prediction
EPSS 0.4%
SSVC
No known exploit

Description

Integer Underflow (Wrap or Wraparound) vulnerability in erlang otp erlang/otp (erts modules), erlang otp erts (erts modules) allows Forced Integer Overflow, Excessive Allocation. This vulnerability is associated with program files erts/emulator/beam/external.c, emulator/beam/external.c. The BIT_BINARY_EXT tag (77) handler in the External Term Format (ETF) decoder accepts an encoding with both length and trailing-bits fields set to zero. The subsequent computation of the bitstring size underflows an unsigned integer, producing a value of roughly 2^64 that is then passed as a memory allocation size. The allocator aborts the entire node with a message such as "Cannot allocate 2305843009213693951 bytes of memory (of type binary)". The crash is a VM-level abort, not an Erlang-level exception. It cannot be intercepted by supervision trees, by try/catch, or by passing the [safe] option to binary_to_term/2 (which only restricts atom creation and does not perform structural validation of binary encodings). Any application that decodes ETF from untrusted sources via binary_to_term/1,2 or enif_binary_to_term() is exposed. The Erlang distribution protocol also decodes incoming terms through the same code path, but distribution is expected to run on trusted networks per the OTP Secure Coding Guidelines (DSG-011). This issue affects OTP from OTP 27.0 before OTP 29.0.4, OTP 28.5.0.4 and OTP 27.3.4.15, corresponding to erts from 15.0 before 17.0.4, 16.4.0.4 and 15.2.7.11.

Affected

Our own answer: ranges merged, affected separated from platform-only, with confidence and review flags.

27.0 ≤ v < 27.3.4.1528.0 ≤ v < 28.5.0.429.0 ≤ v < 29.0.415.024ef4cbaeda9b9c26682cba75f2f15b0c58722aa27.0
Vendor states this is fixed in dc1bf9344c0ce62717cf60866590cea0242780fd (fixed)
Affected range stops below 27.3.4.15, so it looks safe — our inference, not a vendor statement
Affected range stops below 28.5.0.4, so it looks safe — our inference, not a vendor statement
Affected range stops below 29.0.4, so it looks safe — our inference, not a vendor statement
15.0 ≤ v < 15.2.7.1116.0 ≤ v < 16.4.0.417.0 ≤ v < 17.0.4
Affected range stops below 15.2.7.11, so it looks safe — our inference, not a vendor statement
Affected range stops below 16.4.0.4, so it looks safe — our inference, not a vendor statement
Affected range stops below 17.0.4, so it looks safe — our inference, not a vendor statement

References (5)

RelatedPatchThird Party Advisory
Source: cverecord
https://cna.erlef.org/cves/CVE-2026-54890.html
RelatedPatchThird Party Advisory
Source: cverecord
https://osv.dev/vulnerability/EEF-CVE-2026-54890
x_version-schemeRelease Notes
Source: cverecord
https://www.erlang.org/doc/system/versions.html#order-of-versions

SSVC — Action Prioritization

SSVCNo known exploitNot automatablePartial impact

SSVC signals action urgency (what to patch first), separate from CVSS which measures severity.

CVSS

Attack Vector
NETWORK
Complexity
LOW
Privileges
NONE
User Interaction
NONE
Scope
UNCHANGED
Confidentiality
NONE
Integrity
NONE
Availability
HIGH
Vector?

Quick Info

StatusANALYZED
CWE
CWE-191CWE-789CWE-1284

Get alerted

Stop reading CVE feeds looking for your own stack

Track the products you run and VulnTracker tells you when a new CVE affects them — with what is known about fixing it, and nothing about the rest.

Free plan, no card