> Date: Thu, 15 Aug 2024 15:02:08 +0200 > From: Mark Kettenis <mark.kette...@xs4all.nl> > > > From: Greg Steuck <gne...@openbsd.org> > > Date: Wed, 14 Aug 2024 21:16:08 -0700 > > > > Jonathan Gray <j...@jsg.id.au> writes: > > > > > for in_s0ix to be set the activate function needs to > > > call the pmops wrappers > > > > > > that may help > > > > > > amdgpu_pm_ops contains other function pointers that may > > > also be of interest > > > > Applying this (and undoing the return true stampede) gets me into a > > state where I can reliably zzz and wakeup with a caveat: the monitor > > remains ON :) > > Well, you'll need to keep the return true stampede. Because: > > > The trace has this new diagnostics: > > > > > drm:pid29637:amdgpu_acpi_is_s0ix_active *ERROR* Power consumption will be > > > higher as BIOS has not been configured for suspend-to-idle. > > > To use suspend-to-idle change the sleep mode in BIOS setup. > > This will make amdgpu_acpi_is_s0ix_active return false, which means > the driver will skip any attempts to suspend your GPU. > > The fact that it prints that message is a bit worrying though. It > suggests that they took away S3 but didn't actually gave you S0i? > > Or maybe the code that checks S0i support is buggy?
Can you send me the files from /var/db/acpi for this machine?
