Look up another CVE · open CVE-2026-89764 in the interactive tracker
CVE-2026-89764
rust: devres: fix race between concurrent revokers
Published 2026-09-11 (37 days ago) by the Linux kernel CNA. Subsystem rust/kernel. Debian urgency: unimportant.
What to do, per Debian release
Debian's own status for CVE-2026-89764 in every release this site tracks, and what each one means in practice. Every status below is the one the Debian Security Team publishes for this CVE. source
| Release | Status | Version | What this means |
|---|---|---|---|
| Debian 12bookworm | not affected | 6.1.190-1 | Nothing to do. bookworm never shipped the vulnerable code. |
| Debian 13trixie | not affected | 6.12.113-1 | Nothing to do. trixie never shipped the vulnerable code. |
| Debian 14forky | fixed | 7.2.6-1 | Update the kernel and reboot. Fixed in 7.2.6-1. A running kernel keeps the old code until the machine restarts. |
| Debian unstablesid | fixed | 7.2.6-1 | Update the kernel and reboot. Fixed in 7.2.6-1. A running kernel keeps the old code until the machine restarts. |
The vulnerability
What the bug is
There is a potential race condition when two paths try to revoke a Devres concurrently. The driver core's devres_release_all() calls Revocable::revoke() via the release callback, while Devres::drop() calls revoke_nosync() on another CPU. The revoker that does not claim the is_available swap returns immediately, but the revoker that did may still be executing drop_in_place() on the inner data. This can cause a use-after-free when the other revoker's caller proceeds to drop adjacent resources that drop_in_place() still references (e.g., Devres<DmaMappedSgt> racing with SGTable freeing the backing sg_table and pages). Fix this by adding a Completion. The release callback signals the Completion after revoke() finishes, and Devres::drop() waits for it when it loses the is_available swap. This ensures the wrapped object is fully torn down before Devres::drop() returns.
Where it lives
The fix touches these files. If the subsystem is one you do not use, the practical exposure is lower, though the package is still the vulnerable one.
rust/kernel/devres.rs
- What this is
- DRIVER CORE, KOBJECTS, DEBUGFS AND SYSFSMAINTAINERS section matched on rust/kernel/devres.rs · driver-core@lists.linux.dev · Supported
Show whether Debian builds this, per release
Not established. This site could not find the build rule for the affected file in the kernel trees Debian ships, so it has no answer about whether Debian compiles this code. Read that as unknown, never as absent.
- bookworm, trixie, forky, sid?
- ?
- not established - no Makefile entry for the affected file
Where a release above reads "not established", no build rule for the affected file was found in that release's kernel tree. That is a gap in this lookup, not a finding that the code is absent.
From the configuration Debian publishes for every kernel flavour it builds. It covers only the architectures Debian ships a linux-config package for, and nothing about a kernel you built yourself. Debian's own status above stays authoritative about the source package.MAINTAINERS
Signals
- Published
- 2026-09-11 (37 days ago)source
- CVSS
- 7.8 high
CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:Hsource recompute it - Attack vector
- local (AV:L)source
- EPSS
- 0.17%, higher than 5.8% of all CVEssource
- CISA KEV
- not listedthe KEV catalogue
- Debian urgency
- unimportantsource
- Subsystem
- rust/kernel
Upstream fixes
The versions the CNA records as introducing and fixing this, and the commit for each. source
| Introduced | Fixed in | Commit |
|---|---|---|
| 6.18 | 6.18.56 | e7930d857cc7 |
| 6.18 | 7.2.4 | c7e3d57c705a |
| 6.18 | 7.3 | acc516dfa197 |
Why the CNA scored it this way
Quoted from the kernel CNA's own scoring file, word for word. A highlighted metric carries the most severe value CVSS v3.1 defines for it, which is what pushes the score up. source
- AV:LAttack vector, local
- The race is in the Rust Devres teardown path (Devres::drop vs devres_release_all release callback), reached by local driver unbind/rmmod/sysfs unbind or hot-unplug of Rust-using devices (nova-core NVIDIA GPU, TH1520 PWM, PCI/platform samples). It is not reachable from network packets or a remote protocol.
- AC:LAttack complexity, low
- The attacker can drive both sides: one CPU in device_unbind_cleanup()/devres_release_all() calling Revocable::revoke() (which sleeps in synchronize_rcu(), widening the window), and another dropping the Devres-containing object (SGTable, Arc<irq::Registration>, PWM Chip data). That is an attacker-controlled race, not an uncontrolled layout condition.
- PR:LPrivileges required, low
- Unbind/rmmod is privileged, but per driver-removal UAF scoring an unprivileged local user with the device (DRM render node, PWM sysfs on TH1520 boards, IRQ-generating I/O) can hold/drop Devres-backed objects or keep the other CPU in teardown while the device is unbound. This does not require init-namespace root or a user-namespace capability.
- UI:NUser interaction, none
- Exploitation does not need a separate victim action such as opening a file or clicking UI. The attacker performs device I/O or holds extra refs themselves while routine unbind, module unload, or hot-unplug runs.
- S:UScope, unchanged
- Impact is kernel memory corruption and privilege escalation on the same host. It does not cross a VM, IOMMU, or other distinct security-authority boundary.
- C:HConfidentiality impact, high
- The bug is a use-after-free: the losing revoker returns while drop_in_place() still uses inner data (e.g. dma_unmap_sgtable on SGTable pages, free_irq cookie, IoMem). UAF of those objects yields an arbitrary kernel read primitive.
- I:HIntegrity impact, high
- The same UAF of sg_table/pages, IRQ Registration handler state, or MMIO mappings is heap corruption exploitable for writes and control-flow hijack, not a bounded or integrity-neutral fault.
- A:HAvailability impact, high
- Use-after-free of DMA mappings, IRQ registrations, or iomem during unbind can oops/panic the kernel even if not fully exploited, so availability impact is complete.