GPU VulnDB

Database/Kernel, userspace & hypervisor

Linux kernel idxd: stale saved-config pointer after a failed second FLR allows invalid memory access

CVSS 7.8CVE-2026-31442Kernel, userspace & hypervisorcurated

Impact

In the idxd driver for Intel Data Streaming Accelerator devices, the scratch area holding the saved configuration was freed without being set to NULL, so if a first Function Level Reset completes and a second FLR then fails to allocate the scratch area, the driver can access the stale pointer. The exposure is narrow - it needs two FLRs with an allocation failure in between, and FLR is an administrative operation on the host - but the consequence is invalid memory access in the host kernel, which on a server where DSA is used for memory-move offload or exposed to guests via VFIO means a crash during device reset. The record gives no evidence of a tenant-reachable path.

Who can reach it

Local and privileged: triggering FLR on the idxd device requires root on the host (or the hypervisor-side reset path for an assigned device). Not reachable by an unprivileged tenant; realistically hit during device reset or driver recovery under memory pressure.

What to do

Update the host kernel to a build carrying the stable fix commits below and reboot the affected nodes; the fix simply NULLs the scratch pointer after FLR. Only hosts whose Intel accelerator devices are driven by idxd are affected, and the window is a reset operation, so this can ride along with the next scheduled kernel reboot rather than an emergency one.

References

Related entries

All Kernel, userspace & hypervisor entries

This entry is curated: imported from vendor advisories with machine assistance, not yet individually verified. Confirm against your vendor's advisory before acting, and report anything wrong.