VYPR
← All advisories
Moderate severityNVD Advisory· Published Dec 17, 2021· Updated Aug 4, 2024

CVE-2021-41496

CVE-2021-41496

Description

Buffer overflow in NumPy <1.19 in array_from_pyobj via crafted array with negative dimensions can cause denial of service.

AI Insight

LLM-synthesized narrative grounded in this CVE's description and references.

Buffer overflow in NumPy <1.19 in `array_from_pyobj` via crafted array with negative dimensions can cause denial of service.

Vulnerability

A buffer overflow vulnerability exists in the array_from_pyobj function of fortranobject.c in NumPy versions prior to 1.19 [1]. The issue occurs when constructing an error message string: the code uses strcpy, sprintf, and strcat without bounds checking on the stack-allocated buffer mess (200 bytes). An attacker-provided array with negative dimensions can cause the formatted string to exceed the buffer size, leading to a buffer overflow [3]. The affected code path is triggered when the array has intent(cache|hide) or optional attributes and contains negative dimensions [3].

Exploitation

An attacker needs the ability to craft a NumPy array with negative dimension values that is then passed to an internal routine calling array_from_pyobj. According to the vendor, this requires the attacker to already be a privileged user or the action must be performed internally [1]. No specific user interaction beyond array construction is required; the overflow occurs during the error-handling path when negative dimensions are detected [3].

Impact

Successful exploitation can cause a denial of service due to a buffer overflow, potentially crashing the Python interpreter or corrupting memory [1]. The impact is limited to Denial of Service; the vendor does not regard the issue as a typical vulnerability because the necessary preconditions imply the attacker already has significant access [1].

Mitigation

The issue was resolved in NumPy version 1.19 by implementing proper bounds checking and replacing unsafe string functions (e.g., strcpy replaced with strncpy) [3]. Users should upgrade to NumPy 1.19 or later to mitigate the vulnerability. No workaround is provided for earlier versions.

References
  1. NVD - CVE-2021-41496
  2. Potential buffer-overflow from string operations in function array_from_pyobj of fortranobject.c

AI Insight generated on May 21, 2026. Synthesized from this CVE's description and the cited reference URLs; citations are validated against the source bundle.

Affected packages

Versions sourced from the GitHub Security Advisory.

PackageAffected versionsPatched versions
numpyPyPI
< 1.191.19

Affected products

70

Patches

0

No patches discovered yet.

Vulnerability mechanics

AI mechanics synthesis has not run for this CVE yet.

References

4

News mentions

0

No linked articles in our index yet.