Look up another CVE · open CVE-2026-93121 in the interactive tracker
CVE-2026-93121
usb: gadget: f_fs: Fix fence cleanup in ffs_dmabuf_transfer() error paths
Published 2026-09-17 (31 days ago) by the Linux kernel CNA. Subsystem drivers/usb. Debian urgency: unimportant.
What to do, per Debian release
Debian's own status for CVE-2026-93121 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 | fixedDSA-6528-1 · 2026-09-29 · text | 6.12.111-1 | Update the kernel and reboot. DSA-6528-1 shipped the fix on 2026-09-29, in 6.12.111-1. A running kernel keeps the old code until the machine restarts. |
| 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. |
| Debian 12bookworm · linux-6.12 | fixedDLA-4817-1 · 2026-10-04 · text | 6.12.111-1~deb12u1 | Update the kernel and reboot. DLA-4817-1 shipped the fix on 2026-10-04, in 6.12.111-1~deb12u1. A running kernel keeps the old code until the machine restarts. |
The vulnerability
What the bug is
The error paths for endpoint-disabled (ESHUTDOWN) and request-allocation failure (ENOMEM) in ffs_dmabuf_transfer() jump to err_fence_put which calls dma_fence_put() on the fence. However, at that point the fence has only been kmalloc'd — dma_fence_init() has not been called yet, so the refcount and the fence ops are uninitialized. Calling dma_fence_put() on such an object leads to undefined behavior. Use kfree() instead, since the fence is just a plain allocation at this stage, and rename the label to err_fence_free to reflect the actual cleanup action.
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/usb/gadget/function/f_fs.c
- What this is
- USB SUBSYSTEMMAINTAINERS section matched on drivers/usb/ · linux-usb@vger.kernel.org · Supported
- Kernel module
- usb_f_fs
Check it on the machine
grep -wE 'CONFIG_OF|CONFIG_PCI|CONFIG_USB|CONFIG_USB_F_FS|CONFIG_USB_GADGET|CONFIG_USB_PHY|CONFIG_USB_SUPPORT' /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 usb_f_fsPrints 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/amd64, amd64/cloud-amd64, amd64/rt/amd64, arm64/cloud-arm64, i386/686, i386/686-pae, i386/rt/686-pae, ppc64el/powerpc64le
- trixieMnot built on amd64/amd64, amd64/cloud-amd64, amd64/rt/amd64, arm64/cloud-arm64, armel/rpi, ppc64el/powerpc64le, ppc64el/powerpc64le-64k, s390x/s390x
- forky, sidMnot built on amd64/amd64, amd64/cloud-amd64, amd64/rt-amd64, arm64/cloud-arm64, loong64/loong64, ppc64el/powerpc64le, ppc64el/powerpc64le-64k, s390x/s390x
- bookworm · linux-6.12Mnot built on amd64/amd64, amd64/cloud-amd64, amd64/rt/amd64, arm64/cloud-arm64, ppc64el/powerpc64le, ppc64el/powerpc64le-64k
- M
- a module where it is built at all; some flavours do not build it
- Build switches
- CONFIG_OF, CONFIG_PCI, CONFIG_USB, CONFIG_USB_F_FS, CONFIG_USB_GADGET, CONFIG_USB_PHY, CONFIG_USB_SUPPORT
- 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
- 7.0 high
CVSS:3.1/AV:L/AC:H/PR:L/UI:N/S:U/C:H/I:H/A:Hsource recompute it - Attack vector
- local (AV:L)source
- EPSS
- 0.15%, higher than 4.2% of all CVEssource
- CISA KEV
- not listedthe KEV catalogue
- Debian urgency
- unimportantsource
- Subsystem
- drivers/usb
Upstream fixes
The versions the CNA records as introducing and fixing this, and the commit for each. source
| Introduced | Fixed in | Commit |
|---|---|---|
| 6.9 | 6.12.110 | be539138d9a1 |
| 6.9 | 6.18.52 | 5fd8baacc7dc |
| 6.9 | 7.2.6 | 58952c83dfe6 |
| 6.9 | 7.3 | 621707dc67c9 |
Debian advisories
| Advisory | Date | Shipped to | Debian |
|---|---|---|---|
| DLA-4817-1 | 2026-10-04 | bookworm 6.12.111-1~deb12u1 | tracker |
| DSA-6528-1 | 2026-09-29 | trixie 6.12.111-1 | tracker |
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 uninitialized fence is kmalloc'd and then dma_fence_put() in ffs_dmabuf_transfer(), reached only from ffs_epfile_ioctl()'s FUNCTIONFS_DMABUF_TRANSFER case after copy_from_user of usb_ffs_dmabuf_transfer_req; USB host traffic never supplies that ioctl or the fence object.
- AC:HAttack complexity, high
- err_fence_put runs only when epfile->ep != ep after ffs_epfile_wait_ep() (epfile->ep is NULLed solely by ffs_func_eps_disable() from gadget set_alt/disable/unbind) or when usb_ep_alloc_request(..., GFP_ATOMIC) fails; the ioctl caller cannot invoke those gadget callbacks and cannot reliably force that small GFP_ATOMIC allocation to fail.
- PR:LPrivileges required, low
- ffs_epfile_ioctl() has no capable() check; FUNCTIONFS_DMABUF_TRANSFER only needs the endpoint file open. FunctionFS inodes take uid/gid/mode from ffs_fs_parse_param() (uid=/gid=/fmode=), which deployed gadgets use to hand ep nodes to non-root daemons such as adbd.
- UI:NUser interaction, none
- The attacker issues FUNCTIONFS_DMABUF_TRANSFER themselves on an endpoint fd they already hold; ffs_epfile_wait_ep() needing a prior host SET_CONFIGURATION is ambient gadget enumeration, not a victim action at exploit time.
- S:UScope, unchanged
- dma_fence_put() on the kmalloc'd ffs_dma_fence runs dma_fence_release() in the same kernel that hosts FunctionFS, so impact stays inside that kernel's privilege boundary rather than crossing a VM or IOMMU authority.
- C:HConfidentiality impact, high
- kmalloc_obj(*fence) is not zeroed and dma_fence_init() has not run, so fence->refcount and fence->ops are leftover heap contents; dma_fence_put() calls dma_fence_release() if the sprayed refcount is 1, which rcu_dereference()s ops and reads other fence fields.
- I:HIntegrity impact, high
- dma_fence_release() calls ops->release() through the uninitialized fence->ops pointer (or walks garbage cb_list in the pending-signal path); a same-cache heap spray can point ops at attacker-controlled function pointers for control-flow hijack.
- A:HAvailability impact, high
- dma_fence_release() does list_empty(&fence->cb_list) and dma_fence_lock_irqsave() on uninitialized cb_list/lock and may invoke a wild ops->release(), any of which oopses or panics when the ESHUTDOWN or ENOMEM err_fence_put path runs.