Look up another CVE · open CVE-2026-93154 in the interactive tracker
CVE-2026-93154
RDMA/irdma: Add refcounting to user ring MRs
Published 2026-09-17 (31 days 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-2026-93154 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 | vulnerable | 6.1.190-1 | No 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 13trixie | vulnerable | 6.12.113-1 | No fix published anywhere yet, neither in Debian nor upstream for the series trixie tracks. Watch the Debian tracker page for this CVE. Advisories for trixie have historically landed a median of 11 days after publication. |
| 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
Prevent userspace from deregistering the MRs that back QP/CQ/SRQ rings by bumping the MR's refcount upon association.
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/irdma/utils.c
drivers/infiniband/hw/irdma/verbs.c
- What this is
- INTEL ETHERNET PROTOCOL DRIVER FOR RDMAMAINTAINERS section matched on drivers/infiniband/hw/irdma/ · linux-rdma@vger.kernel.org · Supported
- Kernel module
- irdma
Check it on the machine
grep -wE 'CONFIG_INFINIBAND|CONFIG_INFINIBAND_IRDMA' /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 irdmaPrints 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 amd64/cloud-amd64, arm64/cloud-arm64
- trixieMnot built on amd64/cloud-amd64, arm64/cloud-arm64, armel/rpi
- forky, sidMnot built on amd64/cloud-amd64, arm64/cloud-arm64, armhf/armmp, armhf/armmp-lpae, loong64/loong64, ppc64el/powerpc64le, ppc64el/powerpc64le-64k, riscv64/riscv64, s390x/s390x
- M
- a module where it is built at all; some flavours do not build it
- Build switches
- CONFIG_INFINIBAND, CONFIG_INFINIBAND_IRDMA
- 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
- 2026-09-17 (31 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 6.3% 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 |
|---|---|---|
| 5.14 | 6.18.52 | 07974c267f60 |
| 5.14 | 7.2.6 | a6c0e693a7c8 |
| 5.14 | 7.3 | f67d8a08f60c |
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 trigger is a local IB_USER_VERBS_CMD_DEREG_MR write (ib_uverbs_dereg_mr / UVERBS_METHOD_MR_DESTROY) on /dev/infiniband/uverbs*. After irdma_get_pbl binds an IRDMA_MEMREG_TYPE_QP/CQ/SRQ ring MR to a live QP/CQ/SRQ, irdma_dereg_mr still frees it; no iWARP/RoCE message supplies that deregistration.
- AC:LAttack complexity, low
- The attacker drives every step with no race: irdma_reg_user_mr with reg_type QP/CQ/SRQ, then irdma_setup_umode_qp/irdma_create_cq/irdma_setup_umode_srq so irdma_get_pbl sets on_list=false, then irdma_dereg_mr. Pre-fix irdma_del_memlist is a no-op on an already-associated ring, so the free is deterministic.
- PR:LPrivileges required, low
- ib_uverbs_open() only checks rdma_dev_access_netns() (devices shared across netns by default) and ib_uverbs_dereg_mr() only requires ownership of the caller's MR uobject, with no capability gate. rdma-core udev rules expose uverbs as 0666; a small ring fits unprivileged RLIMIT_MEMLOCK without CAP_IPC_LOCK.
- UI:NUser interaction, none
- The attacker registers the ring MR via ib_uverbs_reg_mr, creates the QP/CQ/SRQ, and issues IB_USER_VERBS_CMD_DEREG_MR on that handle from their own process; no other user must mount a device, open a file, or otherwise act.
- S:UScope, unchanged
- irdma_dereg_mr kfree()s the irdma_mr and releases its umem/HMC PBLEs while the same host's QP/CQ/SRQ still holds iwpbl and the NIC still DMAs those rings. That is host-kernel memory corruption, not a KVM/Xen guest-to-host escape or an IOMMU isolation bypass.
- C:HConfidentiality impact, high
- irdma_dereg_mr calls ib_umem_release, irdma_free_pble, and kfree(iwmr) while hardware still walks those HMC PBLEs and DMA-reads SQ/RQ/CQ rings. Recycled umem pages and PBLE tables yield device DMA reads of kernel or other-process memory, and iwqp/iwcq->iwpbl dangles into the freed irdma_mr.
- I:HIntegrity impact, high
- The NIC keeps DMAing CQEs/WQEs into unpinned, reallocatable ring pages and walking freed PBLE page tables, producing device writes into recycled kernel memory. Combined with the irdma_mr heap UAF (struct irdma_pbl iwpbl is embedded in iwmr), this is an arbitrary-write primitive.
- A:HAvailability impact, high
- kfree of a still-referenced irdma_mr, irdma_free_pble of live HMC page tables, and DMA into unmapped SQ/RQ/CQ pages cause a kernel oops/panic or an irdma CQP/device reset when the QP/CQ/SRQ is next used or destroyed.