On Fri, Sep 20, 2019 at 03:29:06PM +0300, Ville Syrjälä wrote:
> On Thu, Sep 19, 2019 at 03:16:40PM -0700, James Ausmus wrote:
> > The memory type values have changed in TGL, so we need to translate them
> > differently than ICL.
> >
> > BSpec: 53998
> >
> > Cc: Ville Syrjälä <[email protected]>
> > Cc: Stanislav Lisovskiy <[email protected]>
> > Signed-off-by: James Ausmus <[email protected]>
> > ---
> > drivers/gpu/drm/i915/display/intel_bw.c | 59 ++++++++++++++++++-------
> > 1 file changed, 43 insertions(+), 16 deletions(-)
> >
> > diff --git a/drivers/gpu/drm/i915/display/intel_bw.c
> > b/drivers/gpu/drm/i915/display/intel_bw.c
> > index 688858ebe4d0..11224d9a6752 100644
> > --- a/drivers/gpu/drm/i915/display/intel_bw.c
> > +++ b/drivers/gpu/drm/i915/display/intel_bw.c
> > @@ -35,22 +35,49 @@ static int icl_pcode_read_mem_global_info(struct
> > drm_i915_private *dev_priv,
> > if (ret)
> > return ret;
> >
> > - switch (val & 0xf) {
> > - case 0:
> > - qi->dram_type = INTEL_DRAM_DDR4;
> > - break;
> > - case 1:
> > - qi->dram_type = INTEL_DRAM_DDR3;
> > - break;
> > - case 2:
> > - qi->dram_type = INTEL_DRAM_LPDDR3;
> > - break;
> > - case 3:
> > - qi->dram_type = INTEL_DRAM_LPDDR3;
>
> This should be LPDDR4 actually. Doesn't really matter but would be nice
> to fix as well.
Either my git send-email config or the ML seems to be eating my original
patch mail, and it's not hitting the list, patchwork, or CI, so will
have to send a v2 anyway and I will fix this up in that.
>
> > - break;
> > - default:
> > - MISSING_CASE(val & 0xf);
> > - break;
> > + if (IS_GEN(dev_priv, 12)) {
> > + switch (val & 0xf) {
> > + case 0:
> > + qi->dram_type = INTEL_DRAM_DDR4;
> > + break;
> > + case 3:
> > + qi->dram_type = INTEL_DRAM_LPDDR4;
> > + break;
> > + case 4:
> > + qi->dram_type = INTEL_DRAM_DDR3;
> > + break;
> > + case 5:
> > + qi->dram_type = INTEL_DRAM_LPDDR3;
> > + break;
> > + case 1:
> > + case 2:
> > + /* Unimplemented */
>
> Seems pointless to list these.
Will drop in v2.
>
> The numbers match bspec. Unfortunatley I can't get tgl
> configdb to cooperate so can't double check against the
> MC register definition.
>
> Reviewed-by: Ville Syrjälä <[email protected]>
Thanks!
-James
>
> > + /* fall through */
> > + default:
> > + MISSING_CASE(val & 0xf);
> > + break;
> > + }
> > + } else if (IS_GEN(dev_priv, 11)) {
> > + switch (val & 0xf) {
> > + case 0:
> > + qi->dram_type = INTEL_DRAM_DDR4;
> > + break;
> > + case 1:
> > + qi->dram_type = INTEL_DRAM_DDR3;
> > + break;
> > + case 2:
> > + qi->dram_type = INTEL_DRAM_LPDDR3;
> > + break;
> > + case 3:
> > + qi->dram_type = INTEL_DRAM_LPDDR3;
> > + break;
> > + default:
> > + MISSING_CASE(val & 0xf);
> > + break;
> > + }
> > + } else {
> > + MISSING_CASE(INTEL_GEN(dev_priv));
> > + qi->dram_type = INTEL_DRAM_LPDDR3; /* Conservative default */
> > }
> >
> > qi->num_channels = (val & 0xf0) >> 4;
> > --
> > 2.22.1
>
> --
> Ville Syrjälä
> Intel
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
