[Intel-gfx] [RFC][PATCH 5/5] drm/i915/display: Add Nearest-neighbor based integer scaling support
Laxminarayan Bharadiya, Pankaj
pankaj.laxminarayan.bharadiya at intel.com
Thu Mar 12 09:13:24 UTC 2020
> -----Original Message-----
> From: Ville Syrjälä <ville.syrjala at linux.intel.com>
> Sent: 10 March 2020 21:47
> To: Laxminarayan Bharadiya, Pankaj
> <pankaj.laxminarayan.bharadiya at intel.com>
> Cc: jani.nikula at linux.intel.com; daniel at ffwll.ch; intel-
> gfx at lists.freedesktop.org; dri-devel at lists.freedesktop.org; airlied at linux.ie;
> maarten.lankhorst at linux.intel.com; tzimmermann at suse.de;
> mripard at kernel.org; mihail.atanassov at arm.com; Joonas Lahtinen
> <joonas.lahtinen at linux.intel.com>; Vivi, Rodrigo <rodrigo.vivi at intel.com>;
> Chris Wilson <chris at chris-wilson.co.uk>; Souza, Jose
> <jose.souza at intel.com>; De Marchi, Lucas <lucas.demarchi at intel.com>;
> Roper, Matthew D <matthew.d.roper at intel.com>; Deak, Imre
> <imre.deak at intel.com>; Shankar, Uma <uma.shankar at intel.com>; linux-
> kernel at vger.kernel.org; Nautiyal, Ankit K <ankit.k.nautiyal at intel.com>
> Subject: Re: [RFC][PATCH 5/5] drm/i915/display: Add Nearest-neighbor
> based integer scaling support
>
> On Tue, Feb 25, 2020 at 12:35:45PM +0530, Pankaj Bharadiya wrote:
> > Integer scaling (IS) is a nearest-neighbor upscaling technique that
> > simply scales up the existing pixels by an integer (i.e., whole
> > number) multiplier.Nearest-neighbor (NN) interpolation works by
> > filling in the missing color values in the upscaled image with that of
> > the coordinate-mapped nearest source pixel value.
> >
> > Both IS and NN preserve the clarity of the original image. Integer
> > scaling is particularly useful for pixel art games that rely on sharp,
> > blocky images to deliver their distinctive look.
> >
> > Program the scaler filter coefficients to enable the NN filter if
> > scaling filter property is set to DRM_SCALING_FILTER_NEAREST_NEIGHBOR
> > and enable integer scaling.
> >
> > Bspec: 49247
> >
> > Signed-off-by: Pankaj Bharadiya
> > <pankaj.laxminarayan.bharadiya at intel.com>
> > Signed-off-by: Ankit Nautiyal <ankit.k.nautiyal at intel.com>
> > ---
> > drivers/gpu/drm/i915/display/intel_display.c | 83
> > +++++++++++++++++++- drivers/gpu/drm/i915/display/intel_display.h |
> > 2 + drivers/gpu/drm/i915/display/intel_sprite.c | 20 +++--
> > 3 files changed, 97 insertions(+), 8 deletions(-)
> >
> > diff --git a/drivers/gpu/drm/i915/display/intel_display.c
> > b/drivers/gpu/drm/i915/display/intel_display.c
> > index b5903ef3c5a0..6d5f59203258 100644
> > --- a/drivers/gpu/drm/i915/display/intel_display.c
> > +++ b/drivers/gpu/drm/i915/display/intel_display.c
> > @@ -6237,6 +6237,73 @@ void skl_scaler_disable(const struct
> intel_crtc_state *old_crtc_state)
> > skl_detach_scaler(crtc, i);
> > }
> >
> > +/**
> > + * Theory behind setting nearest-neighbor integer scaling:
> > + *
> > + * 17 phase of 7 taps requires 119 coefficients in 60 dwords per set.
> > + * The letter represents the filter tap (D is the center tap) and
> > +the number
> > + * represents the coefficient set for a phase (0-16).
> > + *
> > + * +------------+------------------------+------------------------+
> > + * |Index value | Data value coeffient 1 | Data value coeffient 2 |
> > + * +------------+------------------------+------------------------+
> > + * | 00h | B0 | A0 |
> > + * +------------+------------------------+------------------------+
> > + * | 01h | D0 | C0 |
> > + * +------------+------------------------+------------------------+
> > + * | 02h | F0 | E0 |
> > + * +------------+------------------------+------------------------+
> > + * | 03h | A1 | G0 |
> > + * +------------+------------------------+------------------------+
> > + * | 04h | C1 | B1 |
> > + * +------------+------------------------+------------------------+
> > + * | ... | ... | ... |
> > + * +------------+------------------------+------------------------+
> > + * | 38h | B16 | A16 |
> > + * +------------+------------------------+------------------------+
> > + * | 39h | D16 | C16 |
> > + * +------------+------------------------+------------------------+
> > + * | 3Ah | F16 | C16 |
> > + * +------------+------------------------+------------------------+
> > + * | 3Bh | Reserved | G16 |
> > + * +------------+------------------------+------------------------+
> > + *
> > + * To enable nearest-neighbor scaling: program scaler coefficents
> > +with
> > + * the center tap (Dxx) values set to 1 and all other values set to
> > +0 as per
> > + * SCALER_COEFFICIENT_FORMAT
> > + *
> > + */
> > +void skl_setup_nearest_neighbor_filter(struct drm_i915_private
> *dev_priv,
> > + enum pipe pipe, int scaler_id)
>
> skl_scaler_...
>
> > +{
> > +
> > + int coeff = 0;
> > + int phase = 0;
> > + int tap;
> > + int val = 0;
>
> Needlessly wide scope for most of these.
>
> > +
> > + /*enable the index auto increment.*/
> > + intel_de_write_fw(dev_priv, SKL_PS_COEF_INDEX_SET0(pipe,
> scaler_id),
> > + _PS_COEE_INDEX_AUTO_INC);
> > +
> > + for (phase = 0; phase < 17; phase++) {
> > + for (tap = 0; tap < 7; tap++) {
> > + coeff++;
>
> Can be part of the % check.
OK.
>
> > + if (tap == 3)
> > + val = (phase % 2) ? (0x800) : (0x800 << 16);
>
> Parens overload.
OK. Will remove.
>
> > +
> > + if (coeff % 2 == 0) {
> > + intel_de_write_fw(dev_priv,
> SKL_PS_COEF_DATA_SET0(pipe, scaler_id), val);
> > + val = 0;
>
> Can drop this val=0 if you move the variable into tight scope and initialize
> there.
Moving val=0 initialization to the tight scope will not work here as we need
to retain "val" and write only when 2 coefficients are ready (since 2
coefficients are packed in 1 dword).
e.g. for (12th , 11th) coefficients, coefficient reg value should be ( (0 << 16) | 0x800).
If we initialize val = 0 in tight loop, 0 will be written to coefficient register.
>
> I was trying to think of a bit more generic way to do this, but couldn't really
> think of anything apart from pre-filling the entire coefficient set and the
> programming blindly. And that seems a bit wasteful if we only care about
> nearest neighbour.
>
> > + }
> > +
> > + }
> > +
> > + }
> > +
> > + intel_de_write_fw(dev_priv, SKL_PS_COEF_DATA_SET0(pipe,
> scaler_id),
> > +0); }
> > +
> > static void skl_pfit_enable(const struct intel_crtc_state
> > *crtc_state) {
> > struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
> > @@ -6260,9 +6327,23 @@ static void skl_pfit_enable(const struct
> intel_crtc_state *crtc_state)
> > pfit_w = (crtc_state->pch_pfit.size >> 16) & 0xFFFF;
> > pfit_h = crtc_state->pch_pfit.size & 0xFFFF;
> >
> > + id = scaler_state->scaler_id;
> > +
> > if (state->scaling_filter ==
> > DRM_SCALING_FILTER_NEAREST_NEIGHBOR) {
> > scaling_filter = PS_FILTER_PROGRAMMED;
> > + skl_setup_nearest_neighbor_filter(dev_priv, pipe,
> id);
>
> This should be sitting alongside the other register writes.
I missed this, thanks for pointing out.
>
> > +
> > + /* Make the scaling window size to integer multiple
> of
> > + * source.
> > + *
> > + * TODO: Should userspace take desision to round
> > + * scaling window to integer multiple?
>
> To give userspace actual control of the pfit window size we need the border
> props (or something along those lines). Step 1 is
> https://patchwork.freedesktop.org/series/68409/. There are further steps in
> my branch after that, but it's still missing the border props for eDP/LVDS/DSI
> since I was too lazy to think how they should interact with the existing scaling
> mode prop.
>
> > + */
> > + pfit_w = rounddown(pfit_w,
> > + (crtc_state->pipe_src_w << 16));
> > + pfit_h = rounddown(pfit_h,
> > + (crtc_state->pipe_src_h << 16));
> > }
>
> This part should be dropped as Daniel mentioned.
Will remove.
Thanks,
Pankaj
>
> >
> > hscale = (crtc_state->pipe_src_w << 16) / pfit_w; @@ -
> 6271,8
> > +6352,6 @@ static void skl_pfit_enable(const struct intel_crtc_state
> *crtc_state)
> > uv_rgb_hphase = skl_scaler_calc_phase(1, hscale, false);
> > uv_rgb_vphase = skl_scaler_calc_phase(1, vscale, false);
> >
> > - id = scaler_state->scaler_id;
> > -
> > spin_lock_irqsave(&dev_priv->uncore.lock, irqflags);
> >
> > intel_de_write_fw(dev_priv, SKL_PS_CTRL(pipe, id),
>
> I think we should also explicitly indicate here which cofficient set(s) we're
> going to use, even if using set0 does mean those bits will be 0.
>
> > diff --git a/drivers/gpu/drm/i915/display/intel_display.h
> > b/drivers/gpu/drm/i915/display/intel_display.h
> > index f92efbbec838..49f58d3c98fe 100644
> > --- a/drivers/gpu/drm/i915/display/intel_display.h
> > +++ b/drivers/gpu/drm/i915/display/intel_display.h
> > @@ -586,6 +586,8 @@ void intel_crtc_arm_fifo_underrun(struct
> > intel_crtc *crtc,
> > u16 skl_scaler_calc_phase(int sub, int scale, bool chroma_center);
> > int skl_update_scaler_crtc(struct intel_crtc_state *crtc_state); void
> > skl_scaler_disable(const struct intel_crtc_state *old_crtc_state);
> > +void skl_setup_nearest_neighbor_filter(struct drm_i915_private
> *dev_priv,
> > + enum pipe pipe, int scaler_id);
> > void ilk_pfit_disable(const struct intel_crtc_state *old_crtc_state);
> > u32 glk_plane_color_ctl(const struct intel_crtc_state *crtc_state,
> > const struct intel_plane_state *plane_state); diff --
> git
> > a/drivers/gpu/drm/i915/display/intel_sprite.c
> > b/drivers/gpu/drm/i915/display/intel_sprite.c
> > index fd7b31a21723..5bef5c031374 100644
> > --- a/drivers/gpu/drm/i915/display/intel_sprite.c
> > +++ b/drivers/gpu/drm/i915/display/intel_sprite.c
> > @@ -415,18 +415,26 @@ skl_program_scaler(struct intel_plane *plane,
> > u16 y_vphase, uv_rgb_vphase;
> > int hscale, vscale;
> > const struct drm_plane_state *state = &plane_state->uapi;
> > + u32 src_w = drm_rect_width(&plane_state->uapi.src) >> 16;
> > + u32 src_h = drm_rect_height(&plane_state->uapi.src) >> 16;
> > u32 scaling_filter = PS_FILTER_MEDIUM;
> > + struct drm_rect dst;
> >
> > if (state->scaling_filter ==
> DRM_SCALING_FILTER_NEAREST_NEIGHBOR) {
> > scaling_filter = PS_FILTER_PROGRAMMED;
> > + skl_setup_nearest_neighbor_filter(dev_priv, pipe,
> scaler_id);
> > +
> > + /* Make the scaling window size to integer multiple of source
> > + * TODO: Should userspace take desision to round scaling
> window
> > + * to integer multiple?
> > + */
> > + crtc_w = rounddown(crtc_w, src_w);
> > + crtc_h = rounddown(crtc_h, src_h);
> > }
> >
> > - hscale = drm_rect_calc_hscale(&plane_state->uapi.src,
> > - &plane_state->uapi.dst,
> > - 0, INT_MAX);
> > - vscale = drm_rect_calc_vscale(&plane_state->uapi.src,
> > - &plane_state->uapi.dst,
> > - 0, INT_MAX);
> > + drm_rect_init(&dst, crtc_x, crtc_y, crtc_w, crtc_h);
>
> Drop as well.
>
> > + hscale = drm_rect_calc_hscale(&plane_state->uapi.src, &dst, 0,
> INT_MAX);
> > + vscale = drm_rect_calc_vscale(&plane_state->uapi.src, &dst, 0,
> > +INT_MAX);
> >
> > /* TODO: handle sub-pixel coordinates */
> > if (intel_format_info_is_yuv_semiplanar(fb->format, fb->modifier)
> &&
> > --
> > 2.23.0
>
> --
> Ville Syrjälä
> Intel
More information about the Intel-gfx
mailing list