Look up another CVE · open CVE-2026-74328 in the interactive tracker
CVE-2026-74328
iommufd: Destroy the pages content after detaching from dmabuf
Published 2026-08-15 (2 months ago) by the Linux kernel CNA. Subsystem drivers/iommu. Debian urgency: unimportant.
What to do, per Debian release
Debian's own status for CVE-2026-74328 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 | not affected | 6.12.113-1 | Nothing to do. trixie never shipped the vulnerable code. |
| 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
Sashiko points out this has gotten out of order, the mutex could still be in use through the dmabuf invalidation callbacks. Don't destroy any of the pages content until the dmabuf is fully detached.
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/iommu/iommufd/pages.c
- What this is
- IOMMUFDMAINTAINERS section matched on drivers/iommu/iommufd/ · iommu@lists.linux.dev · Maintained
Check it on the machine
grep -w CONFIG_IOMMUFD /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.
Show whether Debian builds this, per release
Debian does not build this code in 3 releases. The build rule could not be established for 1 release, so nothing is claimed about those.
- bookworm?
- trixie, forky, sidn
- ?
- not established - no Makefile entry for the affected file
- n
- not enabled on any flavour - this release does not build the code
Where a release above reads "not established", no build rule for the affected file was found in that release's kernel tree. That is a gap in this lookup, not a finding that the code is absent.
- Build switch
- CONFIG_IOMMUFD
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
- 8.8 high
CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:C/C:H/I:H/A:Hsource recompute it - Attack vector
- local (AV:L)source
- EPSS
- 0.17%, higher than 5.4% of all CVEssource
- CISA KEV
- not listedthe KEV catalogue
- Debian urgency
- unimportantsource
- Subsystem
- drivers/iommu
Upstream fixes
The versions the CNA records as introducing and fixing this, and the commit for each. source
| Introduced | Fixed in | Commit |
|---|---|---|
| 6.19 | 7.1.5 | 0507fcedbdcc |
| 6.19 | 7.2 | f2d70dbd3dce |
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
- Reached via local ioctl on /dev/iommu or /dev/vfio/vfio: open iommufd, map a dma-buf with IOMMU_IOAS_MAP_FILE to create iopt_pages, then unmap/destroy to hit iopt_release_pages(). No network packet path exists.
- AC:LAttack complexity, low
- Attacker controls both sides of the race: one thread issues IOMMU_IOAS_UNMAP/IOMMU_DESTROY while another triggers dma_buf_invalidate_mappings() (e.g. VFIO PCI dmabuf revoke or IOMMU_TEST_OP_DMABUF_REVOKE), reliably hitting mutex_destroy before detach.
- PR:LPrivileges required, low
- Exploitation needs only local access to iommufd and a compatible dma-buf fd (VFIO PCI exporter or iommufd selftest), not init-namespace root; cloud GPU/VFIO tenants with device group membership qualify as PR:L.
- UI:NUser interaction, none
- No victim interaction is required; the attacker programmatically creates the IOAS dmabuf mapping and orchestrates concurrent teardown and invalidation through their own file descriptors and ioctl calls.
- S:CScope, changed
- The flaw corrupts dmabuf revocation teardown while IOMMU domains may still hold mappings; iopt_revoke_notify unmaps tracked domains and a failed/partial revoke can leave stale IOMMU PTEs, breaking DMA isolation across the IOMMU security boundary.
- C:HConfidentiality impact, high
- iopt_release_pages() calls mutex_destroy() and later kfree(pages) while dma_buf attachments remain and invalidate_mappings (iopt_revoke_notify) can still lock pages->mutex, causing kernel heap UAF with arbitrary read primitives.
- I:HIntegrity impact, high
- Use-after-free of the embedded mutex and entire iopt_pages structure during concurrent invalidate_mappings enables heap corruption and, via incomplete IOMMU unmapping on revoke, can redirect device DMA to attacker-influenced physical pages.
- A:HAvailability impact, high
- Concurrent access to a destroyed mutex or freed iopt_pages during dmabuf invalidation or detach provokes KASAN-reported UAF, kernel oops/panic, and deterministic DoS on affected GPU/VFIO+iommufd workloads.