ttm_bo_alloc_at_place(9)

📦 version 7.3.1, in nixpkgs unstable

💖 NAME 💖

okay bestie, this is ttm_bo_alloc_at_place(9) from the Kernel API Manual, and it is literally the page about attempting to allocate a BO’s backing store in a place. no cap, that’s the whole mission. 💅

✨ SYNOPSIS ✨

       int ttm_bo_alloc_at_place (struct ttm_buffer_object *bo , const struct
       ttm_place *place , bool force_space , struct ttm_resource **res , struct
       ttm_bo_alloc_state *alloc_state );

🎀 ARGUMENTS 🎀

okay, the cast of characters — every one of these matters, so don’t skip the vibe check:

  • bo — the buffer whose backing store we’re trying to allocate. main character energy fr fr 🫶
  • place — the place we’re attempting the allocation in. location, location, location ✨
  • force_space — a bool saying if we should evict buffers to force space. yes, it’s giving “make room” 💅
  • res — on allocation success, this is where the resulting struct ttm_resource lands. success delivery moment 🎁
  • alloc_state — the object holding allocation state, such as charged cgroups. state-keeping icon 👑

💅 RETURN 💅

the return values are giving error-code realness, so here’s the exact rundown:

  • -EBUSY — no space available, but the allocation should be retried with ttm_bo_evict_alloc. don’t give up yet 👀
  • -ENOSPC — no space available, and the allocation should not be retried. this one’s a hard no 💀
  • -ERESTARTSYS — an interruptible sleep was interrupted by a signal. rude interruption behavior 😭

🦋 SEE ALSO 🦋

Kernel file: ./drivers/gpu/drm/ttm/ttm_bo.c — that’s where this lives, iykyk. the source is the moment 🌈

the SEE ALSO squad, all icon status:

okay, that’s the whole page, bestie. source/version: ./drivers/gpu/drm/ttm; date on this baddie: October 2026. mwah, periodt 💜🫶

📦 where this page comes from