Look up another CVE · open CVE-2026-74365 in the interactive tracker
CVE-2026-74365
nvdimm/btt: Handle preemption in BTT lane acquisition
Published 2026-08-15 (2 months ago) by the Linux kernel CNA. Subsystem Documentation/driver-api. Debian urgency: not yet assigned.
What to do, per Debian release
Debian's own status for CVE-2026-74365 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 | fixed | 6.1.180-1 | Update the kernel and reboot. Fixed in 6.1.180-1. A running kernel keeps the old code until the machine restarts. |
| Debian 13trixie | fixed | 6.12.100-1 | Update the kernel and reboot. Fixed in 6.12.100-1. A running kernel keeps the old code until the machine restarts. |
| Debian 14forky | fixed | 7.1.5-1 | Update the kernel and reboot. Fixed in 7.1.5-1. A running kernel keeps the old code until the machine restarts. |
| Debian unstablesid | fixed | 7.1.5-1 | Update the kernel and reboot. Fixed in 7.1.5-1. A running kernel keeps the old code until the machine restarts. |
The vulnerability
What the bug is
BTT lanes serialize access to per-lane metadata and workspace state during BTT I/O. The btt-check unit test reports data mismatches during BTT writes due to a race in lane acquisition that can lead to silent data corruption. The existing lane model uses a spinlock together with a per-CPU recursion count. That recursion model stopped being valid after BTT lanes became preemptible: another task can run on the same CPU, observe a non-zero recursion count, bypass locking, and use the same lane concurrently. BTT lanes are also held across arena_write_bytes() calls. That path reaches nsio_rw_bytes(), which flushes writes with nvdimm_flush(). Some provider flush callbacks can sleep, making a spinlock the wrong primitive for the lane lifetime. Replace the spinlock-based recursion model with a dynamically allocated per-lane mutex array and take the lane lock unconditionally. Add might_sleep() to catch any future atomic-context caller. Found with the ndctl unit test btt-check.sh.
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.
Documentation/driver-api/nvdimm/btt.rst
drivers/nvdimm/nd.h
drivers/nvdimm/region_devs.c
- What this is
- DOCUMENTATIONMAINTAINERS section matched on Documentation/ · linux-doc@vger.kernel.org · Maintained
- Kernel module
- libnvdimm
Check it on the machine
grep -w CONFIG_LIBNVDIMM /boot/config-$(uname -r)Reads the booted kernel's own configuration, on the machine and nowhere else. =y is built in, =m is a module, and an is not set line means the code is not there.
lsmod | grep -w libnvdimmPrints a line if that module is loaded right now, and nothing if it is not. One that is not loaded can still be loaded later by anything that needs it.
Show whether Debian builds this, per release
Debian builds this, but not on every flavour. At least one release below builds the code on some of its kernel flavours and not on others, so the answer depends on the architecture and flavour installed.
- bookwormMnot built on armhf/armmp, armhf/armmp-lpae, armhf/rt/armmp, i386/686, ppc64el/powerpc64le
- trixieMnot built on armel/rpi, armhf/armmp, armhf/armmp-lpae, armhf/rt/armmp, ppc64el/powerpc64le, ppc64el/powerpc64le-64k, s390x/s390x
- forky, sidMnot built on armhf/armmp, armhf/armmp-lpae, loong64/loong64, ppc64el/powerpc64le, ppc64el/powerpc64le-64k, s390x/s390x
- M
- a module where it is built at all; some flavours do not build it
- Build switch
- CONFIG_LIBNVDIMM
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 kbuild Makefiles Debian linux-config packages
Signals
- Published
- 2026-08-15 (2 months ago)source
- CVSS
- 7.3 high
CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:Lsource recompute it - Attack vector
- local (AV:L)source
- EPSS
- 0.13%, higher than 2.5% of all CVEssource
- CISA KEV
- not listedthe KEV catalogue
- Debian urgency
- not yet assignedsource
- Subsystem
- Documentation/driver-api
Upstream fixes
The versions the CNA records as introducing and fixing this, and the commit for each. source
| Introduced | Fixed in | Commit |
|---|---|---|
| 6.1.63 | 6.1.178 | fd7a97b2514c |
| 6.6.2 | 6.6.145 | 73e35c1bdfa1 |
| 6.7 | 6.12.97 | 417918783bcf |
| 6.7 | 6.18.40 | 5c53406098b5 |
| 6.7 | 7.1.5 | 4eafa810b042 |
| 6.7 | 7.2 | 8d4b989d9c9a |
| 4.19.299 | 0 | 0 |
| 5.4.261 | 0 | 0 |
| 5.10.201 | 0 | 0 |
| 5.15.139 | 0 | 0 |
| 6.5.12 | 0 | 0 |
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 flaw is reached only through local block I/O on a BTT-backed /dev/pmem* device via vfs_write/pwrite/O_DIRECT paths into btt_submit_bio->btt_write_pg/btt_read_pg->nd_region_acquire_lane; it is not in a network protocol handler.
- AC:LAttack complexity, low
- After BTT lanes became preemptible, a second task on the same CPU can bypass the per-CPU recursion spinlock while the first still holds the lane across sleeping arena_write_bytes/nvdimm_flush; an attacker reliably wins by issuing concurrent writes (e.g., many threads) to the BTT block device.
- PR:LPrivileges required, low
- No capability checks exist on the bio submission path; any local principal with write access to the BTT block device or a filesystem mounted on it (common for database/service accounts on PMEM servers) can trigger the race without real init-namespace root.
- UI:NUser interaction, none
- Exploitation requires only automated concurrent block writes from the attacker-controlled process and does not depend on any victim user opening files, clicking prompts, or performing other interactive actions.
- S:UScope, unchanged
- Impact is confined to corrupting BTT metadata and data within the attacker's reachable persistent-memory namespace on the same host; it does not cross VM, container, or IOMMU security boundaries to affect other security authorities.
- C:HConfidentiality impact, high
- Unserialized lane use corrupts BTT map/log/freelist and in-memory arena->rtt/freelist state, so reads can return stale or mis-mapped sector contents (cross-LBA disclosure); the ndctl btt-check test observed data mismatches confirming readable integrity loss.
- I:HIntegrity impact, high
- Concurrent writers can silently corrupt on-media BTT translation tables and user data blocks, producing arbitrary persistent storage modification rather than a bounded or detectable single-field change, matching the fix commit's silent data-corruption description.
- A:LAvailability impact, low
- The primary failure mode is silent corruption and subsequent I/O errors on the affected namespace rather than a kernel panic or hang, but corrupted BTT metadata can persistently degrade or deny reliable access to the volume until recovery or rebuild.