CVE-2026-117678: idna: punycode labels are decoded before the 63-octet length check, so a long xn-- label causes super-linear CPU use
Inefficient Algorithmic Complexity vulnerability in benoitc idna allows an attacker who controls a hostname passed to the library to exhaust CPU via an over-long xn-- label. idna:ulabel/1 in src/idna.erl hands every ASCII label that starts with xn-- to punycode:decode/1 without applying checklabellength/1, the 63-octet DNS label limit. punycode:decode/1 in src/punycode.erl builds its output with list appends in insert/4, so its cost grows super-linearly with the label length. A label of a few kilobytes keeps a scheduler busy for seconds to minutes before the label is rejected as invalid punycode.
Every decode entry point reaches this path: decode/1, decode/2, tounicode/1 and fromascii/1. The encode path reaches it as well, because alabel/1 calls ulabel/1 on an ASCII xn-- label before its own length check, so encode/1, encode/2 and toascii/1 are affected for such input.
This issue affects idna: from 0.5.0 before 8.0.0.
Affected Software
Event History
Frequently Asked Questions
Which versions should be treated as affected?
idna versions from 0.5.0 up to, but not including, 8.0.0 are affected. Version 8.0.0 is outside the stated affected range.
What input and access does an attacker need?
An attacker needs to control a hostname supplied to the library and provide an ASCII label beginning with xn-- that exceeds the 63-octet DNS label limit. No privileges or user interaction are indicated by the supplied vector.
Which library calls can trigger the costly processing?
The decode-related entry points decode/1, decode/2, to_unicode/1, and from_ascii/1 reach the vulnerable path. encode/1, encode/2, and to_ascii/1 can also reach it when given an ASCII xn-- label.
What can be done before upgrading?
Reject or limit over-63-octet ASCII labels beginning with xn-- before passing hostnames to the library. This prevents the long label from reaching punycode decoding.