cik_pcie_gart_enable(9)

📦 version 7.3.1, in nixpkgs unstable

💅 cik_pcie_gart_enable — like, the gart enable one, from the Kernel API Manual (section 9)

okay bestie, so this one's called cik_pcie_gart_enable(9) and honestly it's giving "quietly essential" — no cap, it's the little function that makes the GART actually, you know, gart. Let's get into it 🎀

✨ SYNOPSIS ✨

int cik_pcie_gart_enable (struct radeon_device *rdev );

that's it, that's the whole signature. one argument, one int back. iconic minimalism 💅

🌸 ARGUMENTS 🌸

  • rdev — a radeon_device pointer. that's the whole description, bestie, the manpage really said "you know what it is" 💀 so yeah: pointer to the radeon device struct. that's your main character right there.

🦋 DESCRIPTION 🦋

okay so here's the tea ☕ — this function does a bunch of setup all at once:

  • it sets up the TLBs
  • it programs the page tables for VMID0
  • it sets up the hw for VMIDs 1-15, which are allocated on demand (not upfront, they're ✨situational✨)
  • it sets up the global locations for the LDS, the GDS, and the GPUVM for FSA64 clients (CIK)

and the return value? it returns 0 for success, and errors for failure. so like, check your return value, bestie, don't just ignore it 😭

no cap, failure here means errors come back — this is not a "fire and forget" situation, it's giving "please actually read the return code" 🫶

👑 SEE ALSO 👑

bestie, the cross-reference list here is long and it ATE. every single one of these is a real, separate manpage (section 9) — go read them, they're all in the kernel file ./drivers/gpu/drm/radeon/cik.c:

and the kernel file itself lives at ./drivers/gpu/drm/radeon/cik.c — that's the source of truth, go say hi 🌈

and that's the whole page, bestie!! it ships out of ./drivers/gpu/drm/radeon, dated October 2026, and yeah — the footer really does say cik_pcie_gart_enable(9) again because the manpage tooling is nothing if not consistent 💀💖 slay, go enable some garts 🧚✨

📦 where this page comes from