STRUCT DRM_SCHED_ENTITY_STATS(9)

💖 struct drm_...entity_stats(9) 💖

okay bestie, welcome to the Kernel API Manual — this page is literally all about struct drm_sched_entity_stats, aka the execution stats for an entity. no cap, it's giving "how much GPU time did this entity eat?" energy. 🍓

✨ NAME ✨

struct drm_sched_entity_stats — execution stats for an entity. 💅

🌸 SYNOPSIS 🌸

       struct drm_sched_entity_stats {
           struct kref     kref;
           spinlock_t lock;
           ktime_t runtime;
           ktime_t prev_runtime;
           ktime_t vruntime;
           struct ewma_drm_sched_avgtime   avg_job_us;
        };

👯 Members 👯

• kref — reference count for the object. 💖
• lock — lock guarding the runtime updates. 🔒
• runtime — time entity spent on the GPU. ⏳
• prev_runtime — previous runtime used to get the runtime delta. 🌀
• vruntime — virtual runtime as accumulated by the fair algorithm. 🧚
• avg_job_us — average job duration. 📊

💖 Description 💖

okay so here's the whole vibe: jobs and entities have decoupled lifetimes, ie. once the job has been de-queued we literally cannot access the entity anymore. 😭 but we do need know how much GPU time each entity has spent. so we track this in a separate object, and that object is reference counted by both entities and jobs. periodt. 🫶

🔥 SEE ALSO 🔥

• Kernel file: ./drivers/gpu/drm/scheduler/sched_internal.h
• drm_sched_entity_queue_pop(9)
• drm_sched_entity_queue_peek(9)
• drm_sched_entity_stats_get(9)
• drm_sched_entity_stats_put(9)

cute little sign-off bestie: this page lives under ./drivers/gpu/drm/scheduler, dated October 2026, and it's all about struct drm_...entity_stats(9). mwah 💋✨