The fence-agents packages provide a collection of scripts for handling remote power management for cluster devices. They allow failed or unreachable nodes to be forcibly restarted and removed from the cluster.
Security Fix(es):
* urllib3: urllib3: Unbounded decompression chain leads to resource exhaustion (CVE-2025-66418)
* urllib3: urllib3 Streaming API improperly handles highly compressed data (CVE-2025-66471)
* urllib3: urllib3 vulnerable to decompression-bomb safeguard bypass when following HTTP redirects (streaming API) (CVE-2026-21441)
For more details about the security issue(s), including the impact, a CVSS score, acknowledgments, and other related information, refer to the CVE page(s) listed in the References section.
A deadlock vulnerability in the Linux kernel's drm/radeon driver was fixed by removing an attempt to progress the queue during fence signaling checks. The issue involved the fence lock doubling as the wait queue lock, which could cause self deadlock when signaled() tried to advance the queue. The fix avoids this by not advancing the queue, which is acceptable even if signaled() returns false incorrectly.
A high-severity double free vulnerability exists in the Linux kernel's drm/sched component in the drm_sched_job_add_resv_dependencies function. This flaw occurs when adding dependencies with drm_sched_job_add_dependency(), which improperly manages fence references on failure, leading to a potential double free. The issue has been present since the introduction of dependency tracking in the drm scheduler and was only recently fixed. Multiple Linux kernel versions prior to 6.8.0-1050.53~22.04.1 and related builds are affected.
CVE-2025-38703 is a high-severity vulnerability in the Linux kernel's drm/xe driver related to improper handling of dma-fences. The flaw allows use-after-free conditions when dma-fences are freed prematurely without respecting safe access rules, potentially leading to memory corruption. The issue arises because the driver frees data pointed to by dma-fences, such as timeline names, while these fences may still be referenced externally. The vulnerability has been fixed by ensuring compliance with dma-fence safe access rules, including introducing RCU grace periods before freeing associated data structures.
In the Linux kernel, the following vulnerability has been resolved: drm/msm: Fix a fence leak in submit error path In error paths, we could unref the submit without calling drm_sched_entity_push_job(), so msm_job_free() will never get called. Since drm_sched_job_cleanup() will NULL out the s_fence, we can use that to detect this case. Patchwork: https://patchwork.freedesktop.org/patch/653584/
A vulnerability in the Linux kernel's RISC-V uprobes implementation was fixed by adding a missing fence.i instruction after building the XOL buffer. Without this fence.i, stale or broken instructions could be executed incorrectly, causing instability or incorrect behavior. The issue was identified during BPF selftests on specific hardware. Multiple Linux kernel versions are affected, and patches have been released to address the flaw.
The fence-agents packages provide a collection of scripts for handling remote power management for cluster devices. They allow failed or unreachable nodes to be forcibly restarted and removed from the cluster.
Security Fix(es):
* jinja2: Jinja has a sandbox breakout through malicious filenames (CVE-2024-56201)
* jinja2: Jinja has a sandbox breakout through indirect reference to format method (CVE-2024-56326)
For more details about the security issue(s), including the impact, a CVSS score, acknowledgments, and other related information, refer to the CVE page(s) listed in the References section.
CVE-2024-36960 is a high-severity vulnerability in the Linux kernel's drm/vmwgfx subsystem where invalid reads occur due to incorrect length settings in fence signaled events. The drm_event length was set incorrectly, causing drm_read to copy more data than intended, leading to out-of-bounds reads. This flaw has been fixed in multiple Linux kernel versions starting from 5.4.0-1077.82~18.04.1 and others. The vulnerability has a CVSS score of 7.1 and does not have known exploits in the wild.