Fedora 41’s critical PyPy 7.3.20 update patches 3 severe CVEs (CVE-2025-47273, CVE-2024-47081, CVE-2025-50181) in pip/setuptools. Learn how to secure your Python JIT compiler, fix path traversal risks, and apply the update via DNF. Includes changelog, bug references, and upgrade instructions.
Fedora 41 PyPy 7.3.20 Update: Critical Security Patches for Python JIT Compiler
Fedora 41 has rolled out an urgent update for PyPy, its high-performance Python implementation featuring a Just-In-Time (JIT) compiler.
This release (7.3.20) addresses three critical vulnerabilities—CVE-2025-47273, CVE-2024-47081, and CVE-2025-50181—impacting pip and setuptools. Here’s what developers and sysadmins need to know.
Key Security Fixes in PyPy 7.3.20
CVE-2025-47273: Path traversal vulnerability in
setuptools.PackageIndex(Bug #2367430).CVE-2024-47081:
.netrccredentials leak via malicious URLs inrequests(Bug #2372476).CVE-2025-50181:
urllib3redirects bypass retry-disabled settings (Bug #2373817).
Why this matters:
Exploits could lead to arbitrary code execution, data leaks, or supply-chain attacks
PyPy’s JIT optimizations (enabled in this build) make it a high-value target for attackers.
How to Apply the Update
Run the following command:
sudo dnf upgrade --advisory FEDORA-2025-9b8da6ad7e
For enterprise users: Test in staging first—PyPy’s JIT can impact performance-sensitive workloads.
Technical Deep Dive: PyPy’s Optimizations
PyPy isn’t just a Python interpreter—it’s a JIT-powered runtime that optimizes:
Standard types: Strings, dictionaries, and lists run 5–10x faster than CPython.
Memory usage: Ideal for long-running processes (e.g., web servers).
⚠️ Warning: JIT compilation is enabled by default in this build. Benchmark after upgrading.
Changelog & References
7.3.20-1: Fixes Bug #2376234 (upstream compatibility).
7.3.19-2: Initial CVE patches (Backported to Fedora 41).
Official Bug Reports:
FAQs
Q: Does this affect CPython?
A: No—these CVEs are specific to PyPy’s setuptools/urllib3 integrations.
Q: How urgent is the update?
A: Critical if you use PyPy with untrusted packages (e.g., public PyPI downloads).
Q: Can I disable JIT for security?
A: Yes, but performance drops significantly. Use PYTHONJIT=0 env var.

Nenhum comentário:
Postar um comentário