Reviewed and applied. Thanks! Alex
On Mon, Nov 9, 2026 at 9:21 PM Joshua Horne via B4 Relay <[email protected]> wrote: > > From: Joshua Horne <[email protected]> > > smu8_force_clock_level() passes the pp_dpm_sclk level mask straight to > PPSMC_MSG_SetSclkSoftMin and PPSMC_MSG_SetSclkSoftMax, which take a > level index, as the other smu8 callers do through smu8_get_sclk_level(). > In manual mode, "echo 0 > pp_dpm_sclk" thus selects level 1, "echo 2" > selects level 4, and masks above the highest level index are out of > range for the SMU, so e.g. "echo 3", "echo 7" or "echo 0 1 2 3 4 5 6 7" > do not select the requested levels. The original single level version > sent 1 << level and had the same problem. > > Send the lowest and highest level in the mask, as smu10 does, and reject > levels above those the SMU has enabled, as smu10 and vega20 do. > > Fixes: d7341ef66870 ("drm/amd/powerplay: add some hwmgr functions for sysfs > interface on Carrizo") > Signed-off-by: Joshua Horne <[email protected]> > --- > I found this while working on overdrive support for Carrizo/Stoney on my > laptop (ASUS X555DA, A10-8700P); that series will follow separately. > > Tested on that laptop, 7.3.1 with this patch applied to amdgpu. In manual > mode: > - "echo N > pp_dpm_sclk" for each N from 0 to 7 moves the current level > marker to N > - under GPU load, "echo 2 3 4 5" keeps the clock between levels 2 and 5, > and "echo 0 1 2 3 4 5 6 7" lets it reach level 7 again > - "echo 8" and "echo 0 8" fail with -EINVAL and leave the level unchanged > Without the patch, "echo 0" selected level 1, "echo 3" and "echo 7" did not > select the requested level, and "echo 8" was accepted. > On amd-staging-drm-next the patch applies unchanged and builds without new > warnings with W=1, sparse and smatch. > > I used an AI coding assistant (Claude Code) while working on this. It > found the bug, wrote the change and a draft of the changelog, and wrote > and ran the build checks and test scripts. I reviewed the change and the > changelog, ran the tests on my hardware, and I am responsible for the > patch. > --- > drivers/gpu/drm/amd/pm/powerplay/hwmgr/smu8_hwmgr.c | 14 ++++++++++++-- > 1 file changed, 12 insertions(+), 2 deletions(-) > > diff --git a/drivers/gpu/drm/amd/pm/powerplay/hwmgr/smu8_hwmgr.c > b/drivers/gpu/drm/amd/pm/powerplay/hwmgr/smu8_hwmgr.c > index 8f82d7c07e8c..5a3f697a37a2 100644 > --- a/drivers/gpu/drm/amd/pm/powerplay/hwmgr/smu8_hwmgr.c > +++ b/drivers/gpu/drm/amd/pm/powerplay/hwmgr/smu8_hwmgr.c > @@ -1483,15 +1483,25 @@ static int smu8_store_cc6_data(struct pp_hwmgr > *hwmgr, uint32_t separation_time, > static int smu8_force_clock_level(struct pp_hwmgr *hwmgr, > enum pp_clock_type type, uint32_t mask) > { > + struct phm_clock_voltage_dependency_table *table = > + hwmgr->dyn_state.vddc_dependency_on_sclk; > + uint32_t low, high; > + > switch (type) { > case PP_SCLK: > + /* the SMU takes level indices, not a mask of levels */ > + low = mask ? (ffs(mask) - 1) : 0; > + high = mask ? (fls(mask) - 1) : 0; > + if (high >= min(table->count, smu8_get_max_sclk_level(hwmgr))) > + return -EINVAL; > + > smum_send_msg_to_smc_with_parameter(hwmgr, > PPSMC_MSG_SetSclkSoftMin, > - mask, > + low, > NULL); > smum_send_msg_to_smc_with_parameter(hwmgr, > PPSMC_MSG_SetSclkSoftMax, > - mask, > + high, > NULL); > break; > default: > > --- > base-commit: 47df73aa49d1988c8f684b7caf8be6570e966844 > change-id: 20261109-smu8-sclk-fix-2eacba2cbfc0 > >
