VYPR
Unrated severityNVD Advisory· Published Aug 25, 2019· Updated Aug 5, 2024

CVE-2019-15540

CVE-2019-15540

Description

filters/filter-cso/filter-stream.c in the CSO filter in libMirage 3.2.2 in CDemu does not validate the part size, triggering a heap-based buffer overflow that can lead to root access by a local Linux user.

AI Insight

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

libMirage 3.2.2 CSO filter has a heap-based buffer overflow via a crafted CISO image, enabling local root privilege escalation.

Vulnerability

The CSO filter in libMirage version 3.2.2, located in filters/filter-cso/filter-stream.c, does not validate the part size field in a CISO (compressed ISO) image header. A crafted image can specify a part size larger than the declared block size, causing the inflate call to decompress data into an undersized heap buffer, triggering a heap-based buffer overflow [1][2]. The fix was committed as [0e9292] on 2019-08-25 [3].

Exploitation

An attacker with local user access can supply a malicious CISO image file to CDemu. The image header is parsed, and when the filter processes a compressed part, the oversized part size causes an out-of-bounds heap write during decompression. The provided proof-of-concept file causes a double free or corruption (!prev) abort, indicating exploitation of the heap metadata [1][2]. No special privileges beyond local user access are required.

Impact

Successful exploitation allows the attacker to corrupt heap memory, potentially leading to arbitrary code execution in the context of the CDemu process. Since CDemu often runs as root (to mount virtual drives), this can result in full root privilege escalation [1].

Mitigation

The vulnerability is fixed in libMirage 3.2.3 (or any version including commit [0e9292]) [3]. Users should update their libMirage package to the latest version available from the CDemu project. No workarounds are documented; the fix validates that the part size is either less than the block size (compressed) or equal to it (raw), rejecting malformed images.

AI Insight generated on May 26, 2026. Synthesized from this CVE's description and the cited reference URLs; citations are validated against the source bundle.

Affected products

6

Patches

0

No patches discovered yet.

Vulnerability mechanics

AI mechanics synthesis has not run for this CVE yet.

References

6

News mentions

0

No linked articles in our index yet.