A vulnerability in the Linux kernel RDMA/umem subsystem affects handling of very large memory blocks when an IOMMU is in use. During linearization of an RDMA mapping, a single large block can be split across multiple scatter-gather entries. In __rdma_block_iter_next(), the logic that reassembles those split entries uses 32-bit stack values for data that can exceed 4 GB, causing integer truncation/overflow. As a result, the function can compute incorrect DMA addresses for blocks after truncation. The flaw can affect RDMA memory mapping correctness and may lead to invalid DMA operations in kernel context.
Mallory correlates every CVE against your assets, your vendors, and active adversary campaigns. Know which vulnerabilities matter for you, not just which ones are loud.
What it means. What to do now. Patch path, mitigations, and the assume-compromise checklist.
What an attacker gets, and what they’ve been doing with it.
If you can’t patch tonight, do this now.
Patch, then assume compromise.
No public exploits tracked yet. Mallory keeps watching.
No public exploit code observed for this vulnerability.
Products and vendors Mallory has correlated with this vulnerability. Open in Mallory to drill down to specific CPE configurations and version ranges.
Vendor-confirmed product mapping. Mallory continuously reconciles this list against your asset inventory.
34 sources tracked across advisories, community write-ups, and news. New activity surfaces here as Mallory finds it.
A Linux kernel RDMA/umem truncation vulnerability for memory block sizes of 4 GB or larger.
Linux kernel RDMA umem truncation vulnerability for block sizes of at least 4 GB.
Linux kernel RDMA umem block-size truncation vulnerability.
Linux kernel RDMA umem truncation vulnerability for memory block sizes of 4 GB or larger.
Query your assets running an affected version, and investigate the blast radius.
Every observed campaign linking this CVE to a named adversary.
Malware families riding this exploit, with evidence and IOCs.
YARA, Sigma, Snort, and vendor rules, auto-deployed to your SIEM.
Cross-references every affected SKU, including bundled OEM variants.
Community discussion across Reddit, Mastodon, and other social sources.