drm_gem_fb_end_cpu_access(9)

📦 version 7.3.1, in nixpkgs unstable

💅 drm_gem_fb_end_cpu_access

okay so this one lives in the Kernel API Manual, and it's lowkey the "we're so done here" bestie of the DRM framebuffer world. Let's get into it 🫶

✨ NAME ✨

drm_gem_fb_end_cpu_access — signals end of CPU access to GEM buffer objects

like, it's the closing ceremony for when your CPU was touching GEM buffer objects and now it's NOT. periodt. 🎀

🌸 SYNOPSIS 🌸

void drm_gem_fb_end_cpu_access(struct drm_framebuffer *fb, enum dma_data_direction dir);

no return value, it just does the thing and dips 💅

🦋 ARGUMENTS 🦋

  • fb — the framebuffer. the main character here, honestly 👑
  • dir — access mode. which way the data was flowing, iykyk

💖 DESCRIPTION 💖

So basically, this function signals the end of CPU access to the given framebuffer's GEM buffer objects. That's it, that's the vibe ✨

Here's the part you CANNOT skip, bestie: this function must be paired with a corresponding call to drm_gem_fb_begin_cpu_access. it's giving matched set, it's giving two halves of one soul, it's giving "don't you dare call the end without the begin" 😭 fr fr, they go together.

And for imported BOs? the function calls dma_buf_end_cpu_access. so it's out here delegating, as it should 🫶

See also drm_gem_fb_begin_cpu_access — the other half of the duo, go say hi 🌈

⭐ SEE ALSO ⭐

the whole friend group, all tagged and accounted for 💜

quick vibe check: begin + end are a package deal. call the end without the begin and you're just out here causing chaos, and not the cute kind 💀

and that's the page!! shoutout to ./drivers/gpu/drm, dated October 2026, living under drm_gem_fb_end_cpu_access(9) — ily, go pair your calls bestie 🍓🧚💅

📦 where this page comes from