Look up another CVE · open CVE-2026-74584 in the interactive tracker
CVE-2026-74584
RDMA/bnxt_re: zero shared page before exposing to userspace
Published 2026-08-22 (2 months ago) by the Linux kernel CNA. Subsystem drivers/infiniband. Debian urgency: not yet assigned.
What to do, per Debian release
Debian's own status for CVE-2026-74584 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 | 6.1.177-1 | Update the kernel and reboot. Fixed in 6.1.177-1. A running kernel keeps the old code until the machine restarts. |
| Debian 13trixie | fixed | 6.12.95-1 | Update the kernel and reboot. Fixed in 6.12.95-1. A running kernel keeps the old code until the machine restarts. |
| Debian 14forky | fixed | 7.0.14-1 | Update the kernel and reboot. Fixed in 7.0.14-1. A running kernel keeps the old code until the machine restarts. |
| Debian unstablesid | fixed | 7.0.14-1 | Update the kernel and reboot. Fixed in 7.0.14-1. A running kernel keeps the old code until the machine restarts. |
The vulnerability
What the bug is
bnxt_re_alloc_ucontext() allocates uctx->shpg via __get_free_page(GFP_KERNEL). The buddy allocator does not zero pages without __GFP_ZERO, so the page contains stale kernel data from whatever object most recently freed it. The page is then mapped into userspace via vm_insert_page() under BNXT_RE_MMAP_SH_PAGE in bnxt_re_mmap(). The driver only ever writes 4 bytes (a u32 AVID) at offset BNXT_RE_AVID_OFFT (0x10) inside bnxt_re_create_ah(); the remaining 4092 bytes of the page are exposed to userspace unsanitised, leaking kernel memory contents. Any user with access to /dev/infiniband/uverbsX on a host with a bnxt_re device (typically rdma group membership) can read this data via a single mmap() at pgoff 0 after IB_USER_VERBS_CMD_GET_CONTEXT. Other shared pages in the same file already use get_zeroed_page() correctly: drivers/infiniband/hw/bnxt_re/ib_verbs.c srq->uctx_srq_page = (void *)get_zeroed_page(GFP_KERNEL); cq->uctx_cq_page = (void *)get_zeroed_page(GFP_KERNEL); uctx->shpg is the only outlier. Bring it in line with the existing convention by switching to get_zeroed_page().
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/infiniband/hw/bnxt_re/ib_verbs.c
- What this is
- BROADCOM NETXTREME-E ROCE DRIVERMAINTAINERS section matched on drivers/infiniband/hw/bnxt_re/ · linux-rdma@vger.kernel.org · Supported
- Kernel module
- bnxt_re
Check it on the machine
grep -wE 'CONFIG_INFINIBAND|CONFIG_INFINIBAND_BNXT_RE' /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 bnxt_rePrints 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 does not build this code in 1 release. It does build it in 3 releases, so the answer depends on which release the machine runs.
- bookwormn
- trixieMnot built on amd64/cloud-amd64, arm64/cloud-arm64, armel/rpi, armhf/armmp, armhf/armmp-lpae, armhf/rt/armmp
- forky, sidMnot built on amd64/cloud-amd64, arm64/cloud-arm64, armhf/armmp, armhf/armmp-lpae
- n
- not enabled on any flavour - this release does not build the code
- M
- a module where it is built at all; some flavours do not build it
- Build switches
- CONFIG_INFINIBAND, CONFIG_INFINIBAND_BNXT_RE
- Reads as
- all on means the file is compiled; any one off means it is not.
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
- 2026-08-22 (2 months ago)source
- CVSS
- 7.1 high
CVSS:3.1/AV:L/AC:L/PR:N/UI:N/S:C/C:H/I:N/A:Nsource recompute it - Attack vector
- local (AV:L)source
- EPSS
- 0.18%, higher than 7.3% of all CVEssource
- CISA KEV
- not listedthe KEV catalogue
- Debian urgency
- not yet assignedsource
- Subsystem
- drivers/infiniband
Upstream fixes
The versions the CNA records as introducing and fixing this, and the commit for each. source
| Introduced | Fixed in | Commit |
|---|---|---|
| 4.11 | 5.10.260 | 53c97e9882f4 |
| 4.11 | 5.15.211 | 9896bdfd21d9 |
| 4.11 | 6.1.177 | 9128c2411b83 |
| 4.11 | 6.6.144 | c19b360fa10c |
| 4.11 | 6.12.95 | a3ed2daab02b |
| 4.11 | 6.18.37 | e2b143df2900 |
| 4.11 | 7.0.14 | c75f8ce4baa2 |
| 4.11 | 7.1 | f6b079629bec |
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 leak is triggered only through local access to /dev/infiniband/uverbs* via ioctl GET_CONTEXT and mmap(pgoff 0); bnxt_re_mmap maps the shared page with vm_insert_page, and remote RoCE/IB packet handling never reaches this uverbs allocation path.
- AC:LAttack complexity, low
- On a host with bnxt_re loaded, opening uverbs, issuing GET_CONTEXT, and mmap at offset 0 deterministically exposes the unstale page; the attacker controls all syscall parameters and needs no race, special memory layout, or victim-specific state.
- PR:NPrivileges required, none
- Kernel uverbs character devices are created world-accessible (mode 0666) via uverbs_devnode, and ib_uverbs_open performs no capability or root check beyond netns access, so any local process can open the device and allocate a bnxt_re ucontext.
- UI:NUser interaction, none
- Exploitation requires only attacker-driven open/ioctl/mmap/read syscalls on their own uverbs file descriptor; no cooperative victim action such as mounting filesystems, opening files, or clicking links is needed.
- S:CScope, changed
- On HPC/cloud hosts exposing bnxt_re to tenant containers or shared RDMA workloads, mmap of unstale host-kernel buddy pages leaks host kernel memory across the container/tenant security boundary into an unprivileged workload's address space.
- C:HConfidentiality impact, high
- __get_free_page without __GFP_ZERO leaves up to 4092 bytes of stale kernel heap/slack data readable per ucontext; repeated opens yield additional pages, disclosing kernel pointers, structures, and other sensitive material beyond a few-byte bounded leak.
- I:NIntegrity impact, none
- The flaw only exposes existing kernel memory contents to userspace via mmap; it does not corrupt, overwrite, or modify kernel objects, grant write primitives, or enable code execution through the shared page mapping.
- A:NAvailability impact, none
- Reading the unstale mapped page does not trigger kernel oops, panic, hang, or resource exhaustion; the vulnerable path completes normally and availability is unaffected aside from normal per-context page allocation.