amdgpu_gfx_parse_disable_cu(9)

πŸ“¦ version 7.3.1, in nixpkgs unstable

πŸ’… amdgpu_gfx_parse_disable_cu(9)

okay so bestie, welcome to the Kernel API Manual ✨ β€” this little page is all about amdgpu_gfx_parse_disable_cu, and yeah the name is a mouthful but she's literally just here to read a module parameter and stash the results. no cap. 🫢


🌸 NAME 🌸

amdgpu_gfx_parse_disable_cu β€” the function whose whole personality is parsing the disable_cu module parameter πŸŽ€


✨ SYNOPSIS ✨

void amdgpu_gfx_parse_disable_cu (struct amdgpu_device *adev , unsigned
int *mask , unsigned int max_se , unsigned int max_sh );

that's it, that's the whole signature, and yes she returns void β€” no receipts, no return value, she just does the thing and dips πŸ’…


πŸ¦‹ ARGUMENTS πŸ¦‹

  • adev β€” your amdgpu device pointer, the main character of this whole operation πŸ‘‘
  • mask β€” the array where all the per-shader array disable masks get stored, like little lockers for each shader array πŸ—„οΈ
  • max_se β€” the number of SEs, that's your count of shader engines, don't lose it 🌈
  • max_sh β€” the number of SHs, i.e. your shader array count, equally important, equally iconic ⭐

πŸ’– DESCRIPTION πŸ’–

so here's the tea β˜• β€” the bitmask of CUs that should be disabled, for the shader array identified by se and sh, gets stored right here:

mask[se * max_sh + sh]

it's giving "index math, but make it fashion" ✨ that's the entire lookup formula and she's not changing it for anyone. se, sh, max_sh β€” all up in there, all doing their part. no notes. 🧚


πŸ”₯ SEE ALSO πŸ”₯

like, that's the whole friend group right there β€” go say hi to all of them 🫢 literally every single one is listed, i checked πŸ’€


πŸŽ€ sign-off πŸŽ€

and that's the page!! sourced straight from ./drivers/gpu/drm/amd/amdgpu, dated October 2026, and filed under amdgpu_gfx_parse_disable_cu(9) β€” slay, go read the other pages, ily, bye πŸ’œπŸŒ·

πŸ“¦ where this page comes from