Database/Kernel, userspace & hypervisor

sudo: TZ from the calling user shifts NOTBEFORE/NOTAFTER windows, reviving expired rules
Impact
Where sudoers grants access only inside a time window, an unprivileged local user can move that window by up to roughly 25 hours by setting TZ before invoking sudo, because timestamps written without a trailing 'Z' are evaluated against the inherited timezone of a setuid-root process. Rules that were supposed to have expired are treated as still valid, so a command that was only meant to be runnable during a maintenance shift becomes runnable at will. On GPU fleets this matters wherever time-boxed sudo is used to hand operators, on-call staff or job wrappers temporary root-adjacent actions on nodes - resetting GPUs, reloading the driver, flashing firmware, draining a node. Authentication is still enforced; only the time check is bypassed, so the exposure is limited to principals who already appear in sudoers with a time restriction.
Who can reach it
Local user who already has a sudoers entry carrying a NOTBEFORE or NOTAFTER restriction with a non-'Z' timestamp. Requires a shell on the node and whatever authentication the sudoers rule demands; no remote path.
What to do
Install the vendor sudo update (Red Hat ships fixes via RHSA-2026:71609 and RHSA-2026:75580 for RHEL 7 through 10; the upstream fix is commit 1820a349687522f51023d1ae5925125f59679a8c). No reboot or daemon restart is needed - sudo is invoked per command, so the new binary takes effect immediately. As an interim mitigation, rewrite every NOTBEFORE/NOTAFTER timestamp in sudoers to the explicit UTC form with the trailing 'Z', which removes the TZ dependency without patching.
References
Related entries
- Linux slab allocator: ABA race in the optimistic freelist return corrupts the partial listCVE-2026-97941 · Linux kernel mm/slab (__slab_try_return_freelist ABA race)High
- Linux AMD IOMMU: sign-discarding error check lets nested domains use an unallocated domain IDCVE-2026-98002 · Linux kernel iommu/amd (amd_iommu_alloc_domain_nested error check)High
- Linux kernel BPF: resilient lock release in rbtree callback causes use-after-freeCVE-2026-98060 · Linux kernel BPF (rbtree comparison callbacks, resilient lock kfuncs)High
- Linux kernel BPF: per-CPU map updates index by logical CPU id, reading past the update bufferCVE-2026-98149 · Linux kernel BPF per-CPU map update paths (array, hash, cgroup storage)High
- Linux kernel RDMA/rxe: ODP write paths accept read-only pages, letting RDMA traffic overwrite the page cacheCVE-2026-98361 · Linux kernel RDMA/rxe (soft-RoCE on-demand paging write path)High
- Linux kernel RDMA/rxe: MR re-registration swaps the PD before validating access flags, freeing a live PDCVE-2026-98366 · Linux kernel RDMA/rxe rxe_rereg_user_mr() (access-flag validation ordering)High
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.