CVE-2026-68418
NoneCVSS
—
EPSS
0.2%
Published
Aug 10, 2026
CISA due date
—
Cloudanix analysis
A flaw in the Linux kernel's remote direct memory access code lets a local user crash the entire system by sending a specially crafted request.
- Exploitation
- Local Privilege
- Attacker needs
- Low Privilege
- Complexity
- Low
- Impact
- Denial Of Service
- Component
- OS kernel
- Class
- Memory Corruption
- Remediation
- Patch
- Effort
- Medium
- Patch risk
- Low
What it actually does
Null pointer dereference in the RDMA/irdma QP creation path when a user supplies a zero value for user_wqe_bufs, because iwqp->iwpbl is only populated conditionally yet is later unconditionally dereferenced inside irdma_setup_virt_qp.
Why it matters to the business
Local users on systems using the irdma RDMA driver can trigger a kernel crash, causing denial of service and potential loss of availability for RDMA-dependent workloads; no data disclosure or code execution is possible.
How an attack unfolds
An attacker with local access crafts a user-space QP creation request that sets req.user_wqe_bufs to zero; the kernel skips the PBL population step, later dereferences the still-null iwqp->iwpbl pointer, and crashes the system.
How to fix it
Apply the upstream Linux kernel fix that unconditionally invokes irdma_get_pbl during QP creation and aborts on failure, matching the behavior already present for CQ and SRQ paths; rebuild and boot the patched kernel.
How to confirm the fix worked
Confirm the patch is present by inspecting the irdma QP creation code path for an unconditional irdma_get_pbl call that returns early on failure; test by attempting a zero user_wqe_bufs QP create and verifying the operation now fails gracefully instead of dereferencing NULL.
Typical attack path
- Obtain local user privileges on a system with the irdma RDMA driver loaded.
- Issue an RDMA QP creation request via the RDMA user-space API with user_wqe_bufs set to zero.
- Trigger the unconditional dereference inside irdma_setup_virt_qp, causing a kernel NULL pointer dereference and panic.
How to detect it
Search kernel logs for oops or panic messages referencing irdma_setup_virt_qp or null pointer dereference during QP creation; correlate with user-space processes performing RDMA QP create syscalls immediately prior to the crash.
What your security team should do
- Prioritize patching kernel packages that include the irdma driver on any system exposing RDMA interfaces.
- Restrict untrusted local accounts from accessing RDMA devices via device permissions or cgroups until patches are deployed.
- Monitor EPSS and upstream stable kernel releases for any change in exploit maturity.
The record
In the Linux kernel, the following vulnerability has been resolved: RDMA/irdma: Prevent user-triggered null deref on QP create Previously, the user QP creation path would only attempt to populate iwqp->iwpbl if the user-provided req.user_wqe_bufs field was non-zero. The problem is that iwqp->iwpbl is unconditionally dereferenced later on in irdma_setup_virt_qp. While there was a check for iwqp->iwpbl != NULL, this check would only occur if req.user_wqe_bufs was non-zero. The end result is that a user could send a zero user_wqe_bufs value and trigger a null ptr deref. Fix this by unconditionally calling irdma_get_pbl and bailing if it fails, similar to the CQ and SRQ paths.
Affected packages
| Ecosystem | Package | Introduced | Fixed |
|---|