CVE-2026-99769: drm/vc4: Drop the BO reservations outside of job_lock
In the Linux kernel, the following vulnerability has been resolved:
drm/vc4: Drop the BO reservations outside of joblock
vc4queuesubmit() calls drmexecfini() with joblock held and interrupts disabled. Besides unlocking the reservations, drmexecfini() frees the object array with kvfree(), which asks for preemptible task context or a non-NMI interrupt, and may reach mightsleep() through vfree().
Considering that an array that exceeds a page can take the vmalloc fallback, call drmexecfini() once joblock is dropped. The reservations are then held across the job kick, which is harmless as nothing in the completion path takes a BO reservation.
Note that vc4attachfences() has to stay inside the critical section, as it walks exec, which the job done worker is free to release as soon as joblock is dropped.
Affected Software
Remediation
Recommended actions to resolve this vulnerability, in priority order.
- Upgrade
Upgrade
linux/Kernelto a version that resolves this vulnerability.Fixed in 6.18.64 - Upgrade
Upgrade
linux/Kernelto a version that resolves this vulnerability.Fixed in 7.3.5 - Compensating control
In vc4_queue_submit(), call drm_exec_fini() after dropping job_lock, while keeping vc4_attach_fences() inside the critical section.
Frequently Asked Questions
Which systems are exposed to this issue?
Systems using the Linux VC4 DRM driver can encounter the issue when they execute the VC4 job-submission path through vc4_queue_submit().
What condition makes the unsafe allocation path possible?
The object array must exceed a page, allowing allocation to fall back to vmalloc. Cleanup then occurs while job_lock is held and interrupts are disabled, even though kvfree() may require a preemptible task context or non-NMI interrupt context.