Disclaimer: This content was generated using AI. While we strive for accuracy, please verify critical information with official sources.
Sources: National Vulnerability Database, CISA KEV, Exploit-DB
CVE-2026-111273 Overview
CVE-2026-111273 is an off-by-one heap-based buffer overflow in the unarr decompression library through version 1.1.1. The flaw resides in tar_get_name() within tar/parse-tar.c and is reachable through the public ar_entry_get_name() API. A crafted TAR archive that supplies a fully populated ustar prefix and name field triggers a one-byte NUL write past the heap allocation. The result is heap corruption and process termination for any application linking unarr to parse untrusted archives. The weakness is classified under CWE-193: Off-by-one Error.
Critical Impact
Processing an attacker-supplied TAR archive corrupts adjacent heap memory and crashes the host process, enabling local denial-of-service against archive viewers, document readers, and e-book tools that embed unarr.
Affected Products
- unarr library through version 1.1.1
- Applications embedding unarr for TAR archive parsing (for example, document and e-book readers)
- Downstream distributions packaging vulnerable unarr builds
Discovery Timeline
- 2026-10-15 - CVE-2026-111273 published to NVD
- 2026-10-16 - Last updated in NVD database
Technical Details for CVE-2026-111273
Vulnerability Analysis
The unarr library parses TAR archives using the historical ustar format, which splits long paths across a 155-byte prefix field and a 100-byte name field. tar_get_name() concatenates these fields with a path separator and a terminating NUL byte. When both fields are fully populated with 155 and 100 non-NUL bytes respectively, the function calculates an allocation size that fails to account for either the separator or the trailing NUL. The subsequent write places a NUL byte one position past the end of the heap buffer, which is the defining characteristic of CWE-193.
The overflow corrupts heap metadata or adjacent chunk contents depending on the allocator in use. On modern libc implementations with hardened allocators, the single-byte overwrite typically triggers an assertion or abort, producing a reliable denial-of-service primitive. Attack prerequisites are limited to local delivery of a crafted archive and user interaction to open it. For deeper analysis, see the HackMD write-up for CVE-2026-111273 and CVE-2026-111274 and the VulnCheck advisory.
Root Cause
The allocation size computation in tar_get_name() underestimates the required buffer by one byte when the ustarprefix and name fields are both fully occupied. The function then writes a terminating NUL at buffer[prefix_len + 1 + name_len], which equals the allocation length rather than length - 1. The vulnerable source lines are referenced in the upstream snippet on GitHub.
Attack Vector
An attacker crafts a TAR archive whose header contains a 155-byte prefix with no NUL terminator and a 100-byte name field also fully populated. The victim opens the archive in an application that calls ar_entry_get_name(). During name resolution, tar_get_name() performs the off-by-one write, corrupting the heap and terminating the process. Exploitation is local and requires user interaction, consistent with the published CVSS 4.0 attack-vector metadata.
// No verified exploit code is published. See the HackMD analysis
// linked above for a technical walk-through of the crafted ustar
// header layout that triggers the off-by-one write.
Detection Methods for CVE-2026-111273
Indicators of Compromise
- Repeated unexpected crashes or SIGABRT terminations in processes that link unarr immediately after opening a .tar file
- malloc_consolidate(): invalid chunk size or similar glibc heap assertion messages in system logs
- TAR archives whose header blocks contain a 155-byte non-NUL prefix field paired with a 100-byte non-NUL name field
Detection Strategies
- Inventory binaries that statically or dynamically link unarr 1.1.1 or earlier and flag them for patch tracking
- Scan archive repositories and mail gateways for TAR files whose ustar headers lack NUL terminators in both prefix and name fields
- Fuzz embedded unarr builds with AFL++ or libFuzzer harnesses targeting ar_entry_get_name() to confirm patch application
Monitoring Recommendations
- Forward application crash telemetry and core dumps to a central log pipeline for correlation with recent archive-open events
- Alert on repeated process restarts for document, comic, and e-book reader applications on end-user endpoints
- Monitor package-management events for installations of unarr versions at or below 1.1.1 across managed Linux fleets
How to Mitigate CVE-2026-111273
Immediate Actions Required
- Upgrade unarr to a release that includes the fix for CVE-2026-111273 once published upstream at the selmf/unarr repository
- Rebuild and redistribute any application that statically links unarr, including bundled copies in third-party readers
- Instruct users to avoid opening TAR archives from untrusted sources until patched builds are deployed
Patch Information
At publication, no fixed version is listed in the NVD entry. Track the VulnCheck advisory and the upstream repository for the corrected release. The fix must adjust the allocation length in tar_get_name() to include both the path separator and the terminating NUL, or bound the write index to length - 1.
Workarounds
- Disable TAR handling in applications that embed unarr where the format is not required
- Pre-filter archives at the gateway by rejecting TAR headers in which both the prefix and name fields are fully occupied without NUL terminators
- Run archive-processing applications under sandboxing frameworks such as firejail, bubblewrap, or seccomp to contain crashes and limit blast radius
# Example: wrap an archive viewer with firejail to contain crashes
firejail --seccomp --private-tmp --net=none /usr/bin/your-archive-viewer
Disclaimer: This content was generated using AI. While we strive for accuracy, please verify critical information with official sources.
Sources: National Vulnerability Database, CISA KEV, Exploit-DB

