CVE-2025-12105
Description
A flaw was found in the asynchronous message queue handling of the libsoup library, widely used by GNOME and WebKit-based applications to manage HTTP/2 communications. When network operations are aborted at specific timing intervals, an internal message queue item may be freed twice due to missing state synchronization. This leads to a use-after-free memory access, potentially crashing the affected application. Attackers could exploit this behavior remotely by triggering specific HTTP/2 read and cancel sequences, resulting in a denial-of-service condition.
AI Insight
LLM-synthesized narrative grounded in this CVE's description and references.
A use-after-free vulnerability in libsoup's HTTP/2 message queue handling allows remote denial of service via crafted abort sequences.
Vulnerability
Description
A use-after-free vulnerability exists in libsoup's handling of asynchronous message queues during HTTP/2 communications. When network operations are aborted at specific timing intervals, an internal message queue item may be freed twice due to missing state synchronization, leading to a double-free condition [2][4]. This flaw originates from an incorrect assumption in the queue management logic, where a 'finished' callback can be invoked twice on the same item.
Exploitation
An attacker can exploit this vulnerability remotely by sending specially crafted HTTP/2 requests and triggering specific read and cancel sequences. The attack requires no authentication and can be performed over the network, making it a low-complexity exploit that only needs the ability to initiate HTTP/2 connections to the affected service [2]. The double-free results in a use-after-free memory access, which can corrupt the application's memory state.
Impact
Successful exploitation leads to an application crash, resulting in a denial of service condition. This can render services using libsoup (such as GNOME and WebKit-based applications) unavailable [1][2]. The vulnerability has a CVSS v3 base score of 7.5 (High), indicating significant potential for disruption.
Mitigation
Red Hat has released errata RHSA-2025:23437 and RHSA-2025:23139 to address this issue for Red Hat Enterprise Linux [1][3]. The upstream libsoup project has also provided a fix in merge request !481, which is included in subsequent releases [4]. Users are strongly advised to update their libsoup packages to the patched versions.
AI Insight generated on May 19, 2026. Synthesized from this CVE's description and the cited reference URLs; citations are validated against the source bundle.
Affected products
1Patches
0No patches discovered yet.
Vulnerability mechanics
AI mechanics synthesis has not run for this CVE yet.
References
5News mentions
0No linked articles in our index yet.