On Wed, Jul 27, 2016 at 04:10:45PM -0400, Alex Deucher wrote: > If we do it at enable time, it's too late for the feature > checks. > > bug: https://bugzilla.kernel.org/show_bug.cgi?id=115321 > Signed-off-by: Alex Deucher <alexander.deucher at amd.com> > Cc: Peter Wu <peter at lekensteyn.nl> > --- > drivers/gpu/drm/radeon/radeon_atpx_handler.c | 3 ++- > 1 file changed, 2 insertions(+), 1 deletion(-) > > diff --git a/drivers/gpu/drm/radeon/radeon_atpx_handler.c b/drivers/gpu/drm/radeon/radeon_atpx_handler.c > index 86dcdf3..b46b4c4 100644 > --- a/drivers/gpu/drm/radeon/radeon_atpx_handler.c > +++ b/drivers/gpu/drm/radeon/radeon_atpx_handler.c > @@ -536,7 +536,7 @@ static int radeon_atpx_get_client_id(struct pci_dev *pdev) > static const struct vga_switcheroo_handler radeon_atpx_handler = { > .switchto = radeon_atpx_switchto, > .power_state = radeon_atpx_power_state, > - .init = radeon_atpx_init, > + .init = NULL, Let's just remove this initialization, it allows for trivial removal of the init member later. radeon was the only user of this anyway. > .get_client_id = radeon_atpx_get_client_id, > }; > > @@ -572,6 +572,7 @@ static bool radeon_atpx_detect(void) > printk(KERN_INFO "vga_switcheroo: detected switching method %s handle\n", > acpi_method_name); > radeon_atpx_priv.atpx_detected = true; > + radeon_atpx_init(); > return true; > } > return false; > -- > 2.5.5 > Other than the command above, both patches are Reviewed-by: Peter Wu <peter at lekensteyn.nl> I'll be testing this (via your amd-staging-4.6 branch) tomorrow on an AMD/AMD hybrid (GCN 1.0 unfortunately, so restricted to radeon). -- Kind regards, Peter Wu https://lekensteyn.nl
