[Intel-gfx] [PATCH 1/4] drm/i915/dsi/xelpd: Fix the bit mask for wakeup GB

[Ville Syrjälä]

On Tue, Oct 19, 2021 at 01:41:50PM +0300, Ville Syrjälä wrote:
> On Tue, Oct 19, 2021 at 01:28:10PM +0300, Jani Nikula wrote:
> > On Tue, 19 Oct 2021, Ville Syrjälä <ville.syrjala at linux.intel.com> wrote:
> > > On Tue, Oct 19, 2021 at 01:05:20PM +0300, Jani Nikula wrote:
> > >> On Mon, 18 Oct 2021, Vandita Kulkarni <vandita.kulkarni at intel.com> wrote:
> > >> 
> > >> Commit message goes here.
> > >> 
> > >> > Fixes: f87c46c43175 ("drm/i915/dsi/xelpd: Add WA to program LP to HS wakeup guardband")
> > >> > Signed-off-by: Vandita Kulkarni <vandita.kulkarni at intel.com>
> > >> > ---
> > >> >  drivers/gpu/drm/i915/display/icl_dsi.c | 2 +-
> > >> >  drivers/gpu/drm/i915/i915_reg.h        | 3 ++-
> > >> >  2 files changed, 3 insertions(+), 2 deletions(-)
> > >> >
> > >> > diff --git a/drivers/gpu/drm/i915/display/icl_dsi.c b/drivers/gpu/drm/i915/display/icl_dsi.c
> > >> > index 9ee62707ec72..8c166f92f8bd 100644
> > >> > --- a/drivers/gpu/drm/i915/display/icl_dsi.c
> > >> > +++ b/drivers/gpu/drm/i915/display/icl_dsi.c
> > >> > @@ -1271,7 +1271,7 @@ static void adlp_set_lp_hs_wakeup_gb(struct intel_encoder *encoder)
> > >> >  	if (DISPLAY_VER(i915) == 13) {
> > >> >  		for_each_dsi_port(port, intel_dsi->ports)
> > >> >  			intel_de_rmw(i915, TGL_DSI_CHKN_REG(port),
> > >> > -				     TGL_DSI_CHKN_LSHS_GB, 0x4);
> > >> > +				     TGL_DSI_CHKN_LSHS_GB_MASK, TGL_DSI_CHKN_LSHS_GB_MASK);
> > >> 
> > >> I think you mean the value should be TGL_DSI_CHKN_LSHS_GB.
> > >
> > > IMO the value should never be named that. It should be
> > > TGL_DSI_CHKN_LSHS_GB_<something>.
> > 
> > Alternatively,
> > 
> > #define TGL_DSI_CHKN_LSHS_GB(byte_clocks)		REG_FIELD_PREP(TGL_DSI_CHKN_LSHS_GB_MASK, (byte_clocks))
> > 
> > and
> > 
> > intel_de_rmw(i915, TGL_DSI_CHKN_REG(port),
> > 	     TGL_DSI_CHKN_LSHS_GB_MASK, TGL_DSI_CHKN_LSHS_GB(4));
> > 
> > ?
> > 
> > We're using the value in a specific place that references a w/a, so the
> > magic 4 isn't too bad.
> 
> Yeah, for parametrized defines I think the "_<something>" is
> not needed. Probably not even desired. The argument passed in
> is the "_<something>" essentially.

Oh and, yes, I think having the magic number in the code is fine
for cases like this. I'd say I probably even prefer it that way.
As long as the whole register value isn't a single magic hex constant
that I have to decode by hand to see what bitfields are getting what
values.

-- 
Ville Syrjälä
Intel