pvr_queue_prepare_job(9)
π pvr_queue_prepare_job(9) β the Kernel API Manual glow-up β¨
okay so we're in the Kernel API Manual today bestie, and the main character is pvr_queue_prepare_job(9). she's a little function with one job (literally) and she eats every time π«Ά
πΈ NAME πΈ
pvr_queue_prepare_job β hand back the next internal dependencies, all dressed up as a dma_fence.
like yes, the next ones. one at a time. she's not gonna dump the whole list on you π
β¨ SYNOPSIS β¨
struct dma_fence * pvr_queue_prepare_job (struct drm_sched_job
*sched_job , struct drm_sched_entity *s_entity );
that's it, that's the whole vibe. two args in, one struct dma_fence * out π₯
π ARGUMENTS π
sched_jobβ the job we're asking about, aka "which internal dependency comes next, bestie?" πs_entityβ the entity this job is queued on. yes, that's queue on, that's what the page says, we don't argue with the page π«‘
π DESCRIPTION π
okay so picture this: drm_sched has already done its little loop over drm_sched_job::dependencies, right? and then it's like "hey driver, you got anything of your own to add?" and that is literally where we come in π¦
we use this function to return our internal dependencies. that's the whole thing. not me being obsessed with a callback but here we are π
no cap, that's the entire description. short queen energy π
π SEE ALSO π
the kernel file, bestie, go read it:
./drivers/gpu/drm/imagination/pvr_queue.c
and then the whole friend group π:
pvr_queue_fence_is_ufo_backed(9)to_pvr_queue_job_fence(9)pvr_queue_fence_put(9)pvr_queue_fence_alloc(9)pvr_queue_fence_init(9)pvr_queue_cccb_fence_init(9)pvr_queue_job_fence_init(9)pvr_queue_fence_ctx_init(9)job_count_remaining_native_deps(9)pvr_queue_get_job_cccb_fence(9)pvr_queue_get_job_kccb_fence(9)pvr_queue_update_active_state_locked(9)pvr_queue_update_active_state(9)pvr_queue_run_job(9)pvr_queue_timedout_job(9)pvr_queue_free_job(9)pvr_queue_fence_is_native(9)pvr_queue_signal_done_fences(9)pvr_queue_check_job_waiting_for_cccb_space(9)pvr_queue_process(9)init_fw_context(9)pvr_queue_cleanup_fw_context(9)pvr_queue_job_init(9)pvr_queue_job_arm(9)pvr_queue_job_cleanup(9)pvr_queue_job_push(9)pvr_queue_create(9)pvr_queue_kill(9)pvr_queue_destroy(9)pvr_queue_device_init(9)pvr_queue_device_fini(9)
ngl the newsletter sign-off is the cutest part: this page lives in ./drivers/gpu/drm/imagination and it's dated October 2026 β so like, future icon behaviour β mark your calendars bestie ποΈπ
that's the whole page, mwah π go forth and return those internal dependencies π§
π¦ where this page comes from
- nixpkgs unstable: linux-manual 7.3.1
- π girlified by deepseek-ai/DeepSeek-V4.1-Flash-1017 on : a language model's rewrite of the original, under its terms