CVE-2025-71068 Overview
A vulnerability has been identified in the Linux kernel's svcrdma (Service RDMA) subsystem where the svc_rdma_copy_inline_range function fails to perform proper boundary validation when accessing the rqstp->rq_pages[rc_curpage] array. This missing bounds check allows the rc_curpage index to potentially exceed the allocated page array size, leading to out-of-bounds memory access.
Critical Impact
Memory corruption or system instability could occur when processing RDMA inline data paths due to missing array bounds validation in the svcrdma kernel subsystem.
Affected Products
- Linux kernel versions with vulnerable svcrdma implementation
- Systems utilizing NFS over RDMA (NFSoRDMA) services
- Enterprise storage and high-performance computing environments leveraging RDMA
Discovery Timeline
- 2026-01-13 - CVE CVE-2025-71068 published to NVD
- 2026-01-13 - Last updated in NVD database
Technical Details for CVE-2025-71068
Vulnerability Analysis
This vulnerability exists within the Linux kernel's svcrdma subsystem, specifically in the svc_rdma_copy_inline_range function. The function is responsible for copying inline data ranges during RDMA operations for NFS services. The core issue stems from the function indexing into rqstp->rq_pages[rc_curpage] without first verifying that rc_curpage remains within the bounds of the allocated page array.
When the kernel processes RDMA inline data, it iterates through pages to copy data. However, without proper validation, the rc_curpage counter can increment beyond the legitimate array boundaries. This out-of-bounds access can result in memory corruption, kernel crashes, or potential information disclosure from adjacent kernel memory regions.
Root Cause
The root cause is a missing boundary validation check in the svc_rdma_copy_inline_range function. The code accesses the rq_pages array using rc_curpage as an index without guards to ensure the index stays within the allocated array size. This oversight affects both the initial access and subsequent iterations when advancing to new pages during inline data processing.
Attack Vector
An attacker with the ability to send specially crafted RDMA requests to an affected NFS server could potentially trigger this vulnerability. The attack would involve manipulating RDMA inline data in a way that causes the rc_curpage index to exceed the page array boundaries. This is particularly relevant in environments where NFS over RDMA is exposed to untrusted networks or clients.
The vulnerability manifests in the svcrdma inline path processing. Rather than including synthetic code, the technical details of the vulnerable code path can be examined in the kernel git commits referenced below:
Detection Methods for CVE-2025-71068
Indicators of Compromise
- Kernel oops or panic messages referencing svc_rdma_copy_inline_range or related svcrdma functions
- Unexpected system crashes on servers running NFS over RDMA services
- Memory corruption errors in kernel logs associated with RDMA operations
- Abnormal RDMA request patterns in network traffic to NFS servers
Detection Strategies
- Monitor kernel logs for oops messages containing svcrdma or rq_pages references
- Deploy SentinelOne Singularity Platform for real-time kernel-level threat detection
- Implement audit rules to track RDMA-related system calls and NFS service anomalies
- Use kernel debugging tools to identify out-of-bounds memory access attempts
Monitoring Recommendations
- Enable kernel memory debugging options such as KASAN (Kernel Address Sanitizer) in development environments
- Configure centralized logging for all systems running NFS over RDMA services
- Implement network segmentation to isolate RDMA-enabled storage networks
- Deploy endpoint detection and response (EDR) solutions capable of monitoring kernel-level activity
How to Mitigate CVE-2025-71068
Immediate Actions Required
- Update the Linux kernel to a patched version containing the bounds check fix
- Review systems running NFS over RDMA services and prioritize patching
- Consider temporarily disabling NFS over RDMA if patching is not immediately possible
- Restrict network access to RDMA-enabled NFS services to trusted clients only
Patch Information
The Linux kernel maintainers have released patches that add proper boundary validation guards before the first use of rc_curpage and after advancing to a new page. The fix ensures that array access remains within allocated bounds throughout the inline data processing path.
Patches are available through the following kernel git commits:
Workarounds
- Disable NFS over RDMA (svcrdma) if not required for operations
- Implement network-level access controls to restrict RDMA traffic to trusted sources only
- Use standard TCP-based NFS connections as an alternative until patching is complete
# Disable svcrdma kernel module if not in use
sudo modprobe -r svcrdma
echo "blacklist svcrdma" | sudo tee /etc/modprobe.d/blacklist-svcrdma.conf
Disclaimer: This content was generated using AI. While we strive for accuracy, please verify critical information with official sources.

