data built 2026-10-19T13:16:53Z

Look up another CVE · open CVE-2025-21927 in the interactive tracker

CVE-2025-21927

nvme-tcp: fix potential memory corruption in nvme_tcp_recv_pdu()

9.8 EPSS 0.5% AV:N

Published 2025-04-01 (19 months ago) by the Linux kernel CNA. Subsystem drivers/nvme. Debian urgency: not yet assigned.

What to do, per Debian release

Debian's own status for CVE-2025-21927 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

ReleaseStatus VersionWhat this means
Debian 12bookwormvulnerable6.1.190-1No fix published anywhere yet, neither in Debian nor upstream for the series bookworm tracks. Watch the Debian tracker page for this CVE. Advisories for bookworm have historically landed a median of 23 days after publication.
Debian 13trixiefixed6.12.19-1Update the kernel and reboot. Fixed in 6.12.19-1. A running kernel keeps the old code until the machine restarts.
Debian 14forkyfixed6.12.19-1Update the kernel and reboot. Fixed in 6.12.19-1. A running kernel keeps the old code until the machine restarts.
Debian unstablesidfixed6.12.19-1Update the kernel and reboot. Fixed in 6.12.19-1. A running kernel keeps the old code until the machine restarts.

The vulnerability

What the bug is

nvme_tcp_recv_pdu() doesn't check the validity of the header length. When header digests are enabled, a target might send a packet with an invalid header length (e.g. 255), causing nvme_tcp_verify_hdgst() to access memory outside the allocated area and cause memory corruptions by overwriting it with the calculated digest. Fix this by rejecting packets with an unexpected header length.

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.

drivers/nvme/host/tcp.c

What this is
NVM EXPRESS DRIVERMAINTAINERS section matched on drivers/nvme/host/ · linux-nvme@lists.infradead.org · Supported
Kernel module
nvme_tcp

Check it on the machine

grep -w CONFIG_NVME_TCP /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 nvme_tcp

Prints 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 as a loadable module. Every release below compiles it as a module on every flavour, so whether the code is in the running kernel depends on whether that module is loaded.

  • bookworm, trixie, forky, sidm
m
a loadable module on every flavour - check lsmod, can be blacklisted
Build switch
CONFIG_NVME_TCP

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
2025-04-01 (19 months ago)source
CVSS
9.8 critical
CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:Hsource recompute it
Attack vector
network (AV:N)source
EPSS
0.50%, higher than 41.1% of all CVEssource
CISA KEV
not listedthe KEV catalogue
Debian urgency
not yet assignedsource
Subsystem
drivers/nvme

Upstream fixes

The versions the CNA records as introducing and fixing this, and the commit for each. source

IntroducedFixed inCommit
5.06.12.199fbc953d6b38
5.06.13.722b06c89aa6b
5.06.14ad95bab0cd28

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:NAttack vector, network
The malformed PDU arrives over a routable TCP connection (NVMe/TCP, typically port 4420/8009) and is parsed by the nvme-tcp host driver's socket receive path; a malicious/compromised NVMe-oF target or an on-path attacker injecting into the stream triggers it from anywhere on the network.
AC:LAttack complexity, low
The attacker simply sets hdr->hlen to an arbitrary value (e.g. 255) with the HDGST flag in any rsp/r2t/c2h_data PDU — the out-of-bounds read and 4-byte out-of-bounds write occur deterministically, with no race, no memory-layout dependence, and unlimited retries thanks to the driver's automatic error-recovery reconnect.
PR:NPrivileges required, none
The vulnerable parsing happens on any PDU received right after the ICReq/ICResp handshake, before the NVMe Connect command or any DH-HMAC-CHAP authentication, so the remote peer needs no credentials or privileges on the victim host whatsoever.
UI:NUser interaction, none
No victim action is needed at attack time — NVMe-oF connections are persistent system configuration established at boot and re-established automatically by nvme_tcp_error_recovery(), so a malicious target or on-path attacker can send the crafted PDU at will.
S:UScope, unchanged
The out-of-bounds access corrupts kernel slab memory within the same host kernel security authority; no VM, IOMMU, or sandbox boundary is crossed.
C:HConfidentiality impact, high
nvme_tcp_verify_hdgst() reads 4 raw heap bytes at an attacker-chosen offset up to 255 past a 28-byte kmalloc object and dev_err() prints them verbatim ("recv %#x"), giving a repeatable adjacent-heap disclosure primitive; the CRC input additionally hashes up to 227 bytes of out-of-bounds memory.
I:HIntegrity impact, high
The CRC32C result is written out of bounds at pdu + hlen, an attacker-chosen offset up to 227 bytes past a kmalloc-32 allocation, corrupting neighbouring slab objects; since 24 of the hashed bytes are attacker-controlled and CRC32C is invertible, heap grooming makes both the offset and the written 32-bit value effectively attacker-chosen — a controlled heap write suitable for control-flow hijacking.
A:HAvailability impact, high
The out-of-bounds slab write corrupts adjacent kernel objects and slab metadata, causing oops/panic (immediate BUG under KASAN/hardened SLUB), and the attacker can repeat it across auto-reconnects to reliably crash the host.