PT-2026-151679
CVE-2026-99468
·
Publicado
2026-11-19
·
Atualizado
2026-11-19
Descrição por
Pontuação
N/A
Não disponívelNão há classificações de severidade ou métricas disponíveis. Quando houver, atualizaremos as informações correspondentes na página.
In the Linux kernel, the following vulnerability has been resolved:
media: imx-jpeg: cancel timeout worker when streaming stops
Each per-fd context ctx owns a delayed work (ctx->task timer, callback
mxc jpeg device run timeout) armed via schedule delayed work() at the end
of mxc jpeg device run() to recover a stalled encode/decode job. The only
existing cancellation is cancel delayed work() in the frame-done IRQ
handler, which de-queues a pending work item but does not wait for a
callback that has already started, and it only runs when a frame completes.
When the fd is closed while a job is in flight (the frame-done IRQ has not
fired yet), nothing syncs the worker before mxc jpeg release() frees ctx
with kfree() after v4l2 m2m ctx release(). A queued or executing
mxc jpeg device run timeout() can then recover ctx through
container of(&ctx->task timer) and dereference it (ctx->mxc jpeg,
slot data, dev warn) after ctx has been freed.
Cancel the worker from mxc jpeg stop streaming(). The cancel cannot live
in mxc jpeg release(): mxc jpeg device run() arms the timer while holding
only hw lock, not the mxc jpeg->lock mutex that release holds, so a cancel
in release could still race a concurrent mxc jpeg device run() that
re-arms the timer afterwards. mxc jpeg stop streaming() instead runs inside
v4l2 m2m ctx release() -> vb2 queue release(), i.e. after
v4l2 m2m cancel job() has set TRANS ABORT and waited for any in-flight job
to finish (so v4l2 m2m try queue() will not queue and v4l2 m2m try run()
will not run any further job for this context, which prevents
mxc jpeg device run() from re-arming the timer) and before the m2m context
is freed. cancel delayed work sync() removes a pending work item and waits
for a running callback, so the worker can no longer race with the
subsequent kfree(). The cancel is placed before the buffer-release loop so
a concurrently running timeout callback cannot race with it over the same
buffers. If the frame-done IRQ canceled a still-pending timer, this cancel
is a no-op; if the timeout callback has already started, it waits for the
callback to finish. The same mxc jpeg stop streaming() call is also
reached from VIDIOC STREAMOFF, which drains the worker early, although
STREAMOFF itself does not free ctx -- the use-after-free arises only
when the fd is later closed.
This bug was found by static analysis.
Encontrou algum problema na descrição? Tem algo a acrescentar? Fique à vontade para nos escrever 👾