On Tue, 2024-09-24 at 16:35 -0400, Rodrigo Vivi wrote: > We need to ensure this is called on every kind of runtime > resume and not only on d3cold is possible. Aligning with > the hpd sequence from i915. > > Cc: Vinod Govindapillai <vinod.govindapillai@xxxxxxxxx> > Signed-off-by: Rodrigo Vivi <rodrigo.vivi@xxxxxxxxx> > --- > drivers/gpu/drm/xe/display/xe_display.c | 4 ++-- > 1 file changed, 2 insertions(+), 2 deletions(-) > > diff --git a/drivers/gpu/drm/xe/display/xe_display.c b/drivers/gpu/drm/xe/display/xe_display.c > index 1ab4dd51094f..4d1dd5294b89 100644 > --- a/drivers/gpu/drm/xe/display/xe_display.c > +++ b/drivers/gpu/drm/xe/display/xe_display.c > @@ -330,8 +330,6 @@ static void xe_display_from_d3cold(struct xe_device *xe) > > intel_display_driver_init_hw(xe); > > - intel_hpd_init(xe); > - > intel_opregion_resume(display); > > intel_power_domains_enable(xe); > @@ -470,6 +468,8 @@ void xe_display_pm_runtime_resume(struct xe_device *xe) > > if (xe->d3cold.allowed) > xe_display_from_d3cold(xe); > + > + intel_hpd_init(xe); I would like to clarify, if the order of calling intel_hpd_init() and intel_hpd_poll_disable(xe) matter? Should the intel_hpd_init be called before the poll disable? Otherwise, this was missed in my earlier fix. Thanks for fixing this. Reviewed-by: Vinod Govindapillai <vinod.govindapillai@xxxxxxxxx> > } > > static void display_device_remove(struct drm_device *dev, void *arg)
