CVE-2025-38515

{"id": "CVE-2025-38515", "cveTags": [], "metrics": {}, "published": "2025-08-16T11:15:44.637", "references": [{"url": "https://git.kernel.org/stable/c/549a9c78c3ea6807d0dc4162a4f5ba59f217d5a0", "source": "416baaa9-dc9f-4396-8d5f-8c081fb06d67"}, {"url": "https://git.kernel.org/stable/c/8af39ec5cf2be522c8eb43a3d8005ed59e4daaee", "source": "416baaa9-dc9f-4396-8d5f-8c081fb06d67"}, {"url": "https://git.kernel.org/stable/c/c64f5310530baf75328292f9b9f3f2961d185183", "source": "416baaa9-dc9f-4396-8d5f-8c081fb06d67"}, {"url": "https://git.kernel.org/stable/c/e2d6547dc8b9b332f9bc00875197287a6a4db65a", "source": "416baaa9-dc9f-4396-8d5f-8c081fb06d67"}, {"url": "https://git.kernel.org/stable/c/e62f51d0ec8a9baf324caf9a564f8e318d36a551", "source": "416baaa9-dc9f-4396-8d5f-8c081fb06d67"}, {"url": "https://git.kernel.org/stable/c/ef58a95457466849fa7b31fd3953801a5af0f58b", "source": "416baaa9-dc9f-4396-8d5f-8c081fb06d67"}, {"url": "https://git.kernel.org/stable/c/ef841f8e4e1ff67817ca899bedc5ebb00847c0a7", "source": "416baaa9-dc9f-4396-8d5f-8c081fb06d67"}, {"url": "https://git.kernel.org/stable/c/f9a4f28a4fc4ee453a92a9abbe36e26224d17749", "source": "416baaa9-dc9f-4396-8d5f-8c081fb06d67"}], "vulnStatus": "Awaiting Analysis", "descriptions": [{"lang": "en", "value": "In the Linux kernel, the following vulnerability has been resolved:\n\ndrm/sched: Increment job count before swapping tail spsc queue\n\nA small race exists between spsc_queue_push and the run-job worker, in\nwhich spsc_queue_push may return not-first while the run-job worker has\nalready idled due to the job count being zero. If this race occurs, job\nscheduling stops, leading to hangs while waiting on the job\u2019s DMA\nfences.\n\nSeal this race by incrementing the job count before appending to the\nSPSC queue.\n\nThis race was observed on a drm-tip 6.16-rc1 build with the Xe driver in\nan SVM test case."}, {"lang": "es", "value": "En el kernel de Linux, se ha resuelto la siguiente vulnerabilidad: drm/sched: Incremento del recuento de trabajos antes de intercambiar la cola de cola de spsc. Existe una peque\u00f1a competencia entre spsc_queue_push y el trabajador de ejecuci\u00f3n de trabajos, en la que spsc_queue_push puede devolver un resultado no-first mientras el trabajador de ejecuci\u00f3n de trabajos ya est\u00e1 inactivo debido a que el recuento de trabajos es cero. Si se produce esta competencia, se detiene la programaci\u00f3n de trabajos, lo que provoca bloqueos mientras se espera en las barreras DMA del trabajo. Para solucionar esta competencia, incremente el recuento de trabajos antes de agregarlos a la cola de SPSC. Esta competencia se observ\u00f3 en una compilaci\u00f3n drm-tip 6.16-rc1 con el controlador Xe en un caso de prueba SVM."}], "lastModified": "2025-08-18T20:16:28.750", "sourceIdentifier": "416baaa9-dc9f-4396-8d5f-8c081fb06d67"}