Summary
tornado: multipart split() creates huge temp list before max_parts check -> memory amplification DoS (httputil.py:34)
Advisory details
Description
Summary
parse_multipart_form_data (httputil.py:34) calls
data.split(b"--"+boundary+b"\r\n") before the max_parts check (:35).
A 600KB body with 100k parts creates a 100k-element transient list first,
then rejects transient memory amplification (each split element is a copy).
Pre-auth HTTP DoS.
Root cause
parts = data[:final_boundary_index].split(b"--" + boundary + b"\r\n") # :34 huge list first
if len(parts) > config.max_parts: # :35 check after
raise HTTPInputError("multipart/form-data has too many parts")
PoC
gist: https://gist.github.com/afldl/649861f25d39b53b7edbe0298e171617
poc.py + output.txt (100k parts from 600KB transient list).
Fix
Count separators without materializing the list (e.g. data.count(b"--"+boundary) first).
Credit
Reported by afldl, 2026-07.
References
- https://github.com/advisories/GHSA-8423-8fgw-73vq
- https://github.com/tornadoweb/tornado/security/advisories/GHSA-8423-8fgw-73vq
- https://github.com/tornadoweb/tornado/pull/3704
- https://github.com/tornadoweb/tornado/commit/de85b3f87446e323e881bbaa3d5a74f4b76e5f05
- https://gist.github.com/afldl/649861f25d39b53b7edbe0298e171617
- https://github.com/tornadoweb/tornado/releases/tag/v6.5.8
Related vulnerabilities
All Supply chain →- MEDIUMCVE-2026-71486
vLLM: Derender endpoints decode caller-supplied GenerateResponse token IDs without output bounds
- HIGHCVE-2026-79921
amqp091-go has a Potential Memory Exhaustion/Protocol Violation via Broker-Controlled Oversized Payload
- HIGHCVE-2026-67446
Mailpit: Thumbnail generation decodes unbounded image dimensions before scaling
- MEDIUMCVE-2026-82562
qs array-limit bypass via bracket-key comma parsing
- MEDIUMCVE-2026-73228
Django REST framework: Potential bypass of Django `DATA_UPLOAD_MAX_MEMORY_SIZE` when parsing oversized JSON and urlencoded request bodies via DRF `request.data`
- HIGHCVE-2026-73089
Browserslist: Unbounded memory growth (no cache eviction) via distinct query results, leading to eventual OOM