CVE-2026-98366
7.8 HIGHPublished 2026-10-06 · Updated 2026-10-07
AI risk analysis
- Summary
- The Linux kernel's RDMA/rxe driver has a race condition where memory region access flags are validated after the protection domain (PD) is already reassigned, potentially leading to use-after-free vulnerabilities when a PD is freed while still referenced by active memory windows.
- Exploitability
- Exploitation requires careful crafting of memory registration operations with specific flag combinations, likely needing kernel-level access or a local user context with RDMA capabilities, but the vulnerability is relatively straightforward to trigger once the right conditions are met.
- Blast radius
- If exploited, this vulnerability could lead to kernel crashes, memory corruption, and potentially privilege escalation in systems running RDMA-enabled kernels with active memory window operations.
- Detection
- No reliable host or network indicator is derivable from the published description.
- Prioritized remediation
- Apply the kernel patch that fixes the order of PD reassignment and access flag validation in rxe_rereg_user_mr() function, specifically addressing the race condition between PD reference counting and memory window cleanup.
Analysis generated locally by garnet-coder (no data left the box). AI-assisted — verify against primary sources before acting.
NVD description
In the Linux kernel, the following vulnerability has been resolved: RDMA/rxe: validate access flags before swapping the MR's PD rxe_rereg_user_mr() reassigns mr->ibmr.pd first and only then validates the IB_MR_REREG_ACCESS argument: if (flags & IB_MR_REREG_PD) { rxe_put(old_pd); rxe_get(pd); mr->ibmr.pd = ibpd; } if (flags & IB_MR_REREG_ACCESS) { if (access & ~RXE_ACCESS_SUPPORTED_MR) return ERR_PTR(-EOPNOTSUPP); mr->access = access; } Both flags pass the entry check because RXE_MR_REREG_SUPPORTED is IB_MR_REREG_PD | IB_MR_REREG_ACCESS, so a caller can reach the access check with mr->ibmr.pd already reassigned. mr->ibmr.pd is owned by the core, which adjusts pd->usecnt only on the success path: ib_uverbs_rereg_mr() jumps to put_new_uobj on a driver error without undoing the reassignment, so mr->pd == new_pd while the usecnts still charge the MR to orig_pd. ib_dereg_mr_user() then decrements new_pd, whose count can reach zero while a memory window still references it; uverbs_free_pd() frees the PD on that count alone and rxe_mw_cleanup() writes to freed memory: BUG: KASAN: slab-use-after-free in __rxe_put+0x31/0xa0 Write of size 4 at addr ffff8881301dd690 by task rxe_poc/591 __rxe_put+0x31/0xa0 rxe_mw_cleanup+0x42/0x200 __rxe_cleanup+0x115/0x370 rxe_dealloc_mw+0x4c/0x80 Allocated by task 591: ib_uverbs_alloc_pd+0x258/0x540 Freed by task 591: ib_dealloc_pd_user+0x174/0x210 uverbs_free_pd+0x8d/0xc0 ib_uverbs_dealloc_pd+0x18e/0x1d0 Validate the access flags before mutating any state so the callback either applies every requested change or none.
CVSS vector
CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H
All references
- https://git.kernel.org/stable/c/08f12745cb72981aa2cabe214af0c7a42a856f0a
- https://git.kernel.org/stable/c/4dd7a53f1c5b44693c26dac4b9b5bd5bb9d604c8
- https://git.kernel.org/stable/c/7230cc456d4bb2221c20c5f1d22b38b8576aa16f
- https://git.kernel.org/stable/c/ae36a5b609ae79f4de966328b78d2584be9719a4
- https://git.kernel.org/stable/c/fe602c91a52e161ace4afd5bdb8f33270c554c6f
Source data: NVD (nvd.nist.gov), public domain. Exploit-DB.ai adds local AI analysis for defensive use only.
Related CVEs
- HIGHCVE-2026-97421
- CRITICALCVE-2026-98323
- HIGHCVE-2026-98367
- CRITICALCVE-2026-100075
- HIGHCVE-2026-13087
- HIGHCVE-2026-64561PoC
- CRITICALCVE-2026-64566
- HIGHCVE-2026-64574
Related by shared AI tags and CWE weakness class. Browse the full archive.