CVE-2026-89988 PUBLISHED

kprobes: Protect kprobe_blacklist with RCU

Assigner: Linux
Reserved: 11.09.2026 Published: 16.09.2026 Updated: 16.09.2026

In the Linux kernel, the following vulnerability has been resolved:

kprobes: Protect kprobe_blacklist with RCU

__within_kprobe_blacklist() traverses kprobe_blacklist without holding kprobe_mutex. When a module is unloaded, kprobe_remove_area_blacklist() removes blacklist entries and immediately frees them with kfree(). A concurrent call to within_kprobe_blacklist() can therefore dereference freed memory.

Furthermore, within_kprobe_blacklist() can be called in atomic or non-preemptible contexts where the sleeping kprobe_mutex cannot be taken.

Protect kprobe_blacklist with RCU. Use guard(rcu)() and list_for_each_entry_rcu() for traversal, list_add_tail_rcu() for insertions, list_del_rcu() for deletions, and kfree_rcu() to reclaim entries safely after a grace period.

Metrics

CVSS Vector: CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H
CVSS Score: 7.8

AV:L - The UAF is reached only via local syscalls: x86 perf_event_open() hardware execute breakpoints and register_kprobe() (tracefs, perf kprobes, BPF). It is not invoked from network packet processing or a remote service. AC:L - An attacker can loop the blacklist walk while concurrently causing module COMING/GOING so kprobe_remove_area_blacklist() does list_del/kfree, creating both sides of the race. Attacker-driven use-after-free races are AC:L. PR:L - x86 hw_breakpoint_arch_parse() calls within_kprobe_blacklist() for kernel execute breakpoints during perf_event_open() before hw_breakpoint_parse()'s CAP_SYS_ADMIN check, so an unprivileged user can trigger the walk. Module autoload via request_module() can mutate the list without CAP_SYS_MODULE. UI:N - The attacker triggers the race with their own syscalls (perf_event_open and module autoload/unload). No separate victim action is required. S:U - The bug is kernel-heap use-after-free in kprobes. Effects stay inside the kernel's security authority and do not cross a VM, IOMMU, or other boundary. C:H - Concurrent walk versus kfree of struct kprobe_blacklist_entry is a heap use-after-free. Reclaiming the object yields an arbitrary kernel read primitive, so confidentiality impact is high. I:H - A heap use-after-free on a list node is exploitable via heap spraying for an arbitrary write and control-flow hijack, so integrity impact is high. A:H - Following the freed blacklist entry's next pointer or fields causes a kernel oops or panic, taking down the system.

Product Status

Vendor Linux
Product Linux
Versions Default: unaffected
  • affected from 376e242429bf8539ef39a080ac113c8799840b13 to 8f406205003f741225f620f0e357ca0de71d45e3 (excl.)
  • affected from 376e242429bf8539ef39a080ac113c8799840b13 to 630ed8734a02b992e8a97538179c2ece42c5210f (excl.)
  • affected from 376e242429bf8539ef39a080ac113c8799840b13 to 518d9b5568bb06f86e4434f7a30faa9612ea91ad (excl.)
  • affected from 376e242429bf8539ef39a080ac113c8799840b13 to 5da247e4d1e8661d4d6f997101d59967e8994e32 (excl.)
  • affected from 376e242429bf8539ef39a080ac113c8799840b13 to 6e62cf983a2bd547e6bbd3f935539233e5f97f8f (excl.)
  • affected from 376e242429bf8539ef39a080ac113c8799840b13 to 0c4256196b3a105307e2235fbfd85e768bbcdd0f (excl.)
Vendor Linux
Product Linux
Versions Default: affected
  • Version 3.16 is affected
  • unaffected from 0 to 3.16 (excl.)
  • unaffected from 6.1.188 to 6.1.* (incl.)
  • unaffected from 6.6.157 to 6.6.* (incl.)
  • unaffected from 6.12.110 to 6.12.* (incl.)
  • unaffected from 6.18.51 to 6.18.* (incl.)
  • unaffected from 7.2.5 to 7.2.* (incl.)
  • unaffected from 7.3-rc2 to * (incl.)

References