vega10_start_dpm(9)
π NAME π
vega10_start_dpm β bestie, this is the one that pokes the SMC and is like "hey girl, turn on the DPMs you actually support" β¨
(it's from the kernel API manual, vega10_start_dpm(9), if you were wondering where this whole vibe lives π)
πΈ SYNOPSIS πΈ
int vega10_start_dpm (struct pp_hwmgr *hwmgr , uint32_t bitmap );
okay so it takes two things, returns an int, and that's literally it. no cap. π«Ά
β¨ ARGUMENTS β¨
hwmgrβ this is the address of the powerplay hardware manager. like, the actual address. hand it over and we're cooking π₯bitmapβ the bitmap for the features to enabled. yes, "to enabled." iconic grammar moment, we love her, we keep her π basically it tells us which features we're flipping on.
π RETURN π
It gives you back 0 on at least one DPM is successfully enabled. so a 0 here is a W, not an error β one DPM making it through is enough to count π
lowkey important: the 0 means at least one suceeded, so don't go reading it like "everything worked perfectly," because that's not the claim bestie π
π¦ SEE ALSO π¦
The kernel file lives here, go be nosy:
./drivers/gpu/drm/amd/pm/powerplay/hwmgr/vega10_hwmgr.c
and these are the sibling functions, the whole friend group fr fr π«Ά
vega10_get_evv_voltages(9)vega10_patch_with_vdd_leakage(9)vega10_patch_lookup_table_with_leakage(9)vega10_trim_voltage_table(9)vega10_construct_voltage_tables(9)vega10_populate_single_gfx_level(9)vega10_populate_single_soc_level(9)vega10_populate_all_graphic_levels(9)vega10_populate_all_memory_levels(9)vega10_init_smc_table(9)
and that's the whole page, no notes π
cute lil sign-off: this one ships from ./drivers/gpu/drm/amd/pm/power... (yes the source line gets truncated, she's mysterious like that π§) dated October 2026, and the page is literally called vega10_start_dpm(9) β same as the top, we're consistent, we're slaying 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