CVE-2026-99614: drm/nouveau/nvkm: clear use.enabled when subdev init fails
In the Linux kernel, the following vulnerability has been resolved:
drm/nouveau/nvkm: clear use.enabled when subdev init fails
On Kepler boards without the video decode falcon firmware installed (msvld/mspdec/msppp), every other VA-API decode attempt binds to an engine with no ucode loaded. The submitted work never completes, which leads to a context switch timeout, a channel recovery that cannot reset the engine, and eviction waits on fences that never signal. The process ends up stuck in D state, or killed by SIGBUS when it touches a buffer evicted after the wait times out:
msvld: unable to load firmware data msvld: init failed, -19 fifo: SCHEDERROR 0a [CTXSWTIMEOUT] vo[5444]: channel 14 killed! WARNING: drivers/gpu/drm/nouveau/nvkm/engine/fifo/runl.c:117 at nvkmrunlrc+0x2a4/0x3b0 [nouveau] [TTM] Buffer eviction failed WARNING: drivers/gpu/drm/nouveau/nouveaubo.c:1092 at nouveaubomoventfy.part.0+0x9b/0x100 [nouveau]
Tracing nvkmsubdevref() and nvkmsubdevinit() shows the pattern: a reference taken with enabled=0 runs nvkmfalconinit(), which returns -ENODEV; the next reference sees enabled=1 and succeeds without running it.
Clear use.enabled when init() fails, so that every later reference retries init and fails cleanly. The other readers of use.enabled are fini() handlers that use it to skip teardown, which they now also do after a failed init().
On a GK107 (Quadro K1100M) without video firmware, playing H.264 with mpv --hwdec=auto: without this patch the GPU channel is killed within a minute and mpv hangs; with it, VA-API decoder creation fails, mpv falls back to software decoding, and no faults are logged.
Affected Software
Remediation
Recommended actions to resolve this vulnerability, in priority order.
- Upgrade
Upgrade
linux/Kernelto a version that resolves this vulnerability.Fixed in 6.12.118 - 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 - Upgrade
Upgrade
debian/linuxto a version that resolves this vulnerability.Fixed in 6.1.176-1Fixed in 6.1.192-2Fixed in 7.3.5-1 - Compensating control
In Nouveau, clear use.enabled when subdevice init() fails so later references retry initialization and fail cleanly.
Event History
Frequently Asked Questions
Which systems are affected in practice?
The issue is described on Kepler boards using Nouveau where video decode falcon firmware for msvld, mspdec, or msppp is not installed. It is triggered during VA-API decode attempts.
What symptoms indicate this issue is occurring?
Logs can show firmware loading and initialization failures such as "msvld: unable to load firmware data" and "msvld: init failed, -19," followed by a FIFO CTXSW_TIMEOUT and channel kill. Processes may become stuck in D state, and buffer access after eviction timeouts may result in SIGBUS.
What happens after the initial firmware initialization failure?
A later reference can incorrectly treat the subdevice as enabled without rerunning firmware initialization, binding decode work to an engine with no microcode loaded. The submitted work then never completes, which can cause failed channel recovery and fence waits that never signal.