drm: mali-dp: Add plane upscaling support

Enable the scaling engine for upscaling a single plane using the polyphase
scaler. No image enhancement support or downscaling yet*, and composition
result scaling is not implemented.

* Downscaling a plane requires mclk > pxlclk.

Signed-off-by: Mihail Atanassov <mihail.atanassov@arm.com>
Signed-off-by: Liviu Dudau <Liviu.Dudau@arm.com>

7 files changed, 407 insertions, 17 deletions

diff --git a/drivers/gpu/drm/arm/malidp_crtc.c b/drivers/gpu/drm/arm/malidp_crtc.c
index 8f2e1ae51ba0..ceee2a77bfb4 100644
--- a/drivers/gpu/drm/arm/malidp_crtc.c
+++ b/drivers/gpu/drm/arm/malidp_crtc.c
@@ -247,6 +247,82 @@ static int malidp_crtc_atomic_check_ctm(struct drm_crtc *crtc,
 	return 0;
 }
 
+static int malidp_crtc_atomic_check_scaling(struct drm_crtc *crtc,
+					    struct drm_crtc_state *state)
+{
+	struct malidp_crtc_state *cs = to_malidp_crtc_state(state);
+	struct malidp_se_config *s = &cs->scaler_config;
+	struct drm_plane *plane;
+	const struct drm_plane_state *pstate;
+	u32 h_upscale_factor = 0; /* U16.16 */
+	u32 v_upscale_factor = 0; /* U16.16 */
+	u8 scaling = cs->scaled_planes_mask;
+
+	if (!scaling) {
+		s->scale_enable = false;
+		return 0;
+	}
+
+	/* The scaling engine can only handle one plane at a time. */
+	if (scaling & (scaling - 1))
+		return -EINVAL;
+
+	drm_atomic_crtc_state_for_each_plane_state(plane, pstate, state) {
+		struct malidp_plane *mp = to_malidp_plane(plane);
+		u64 crtc_w, crtc_h;
+		u32 phase;
+
+		if (!(mp->layer->id & scaling))
+			continue;
+
+		/*
+		 * Convert crtc_[w|h] to U32.32, then divide by U16.16 src_[w|h]
+		 * to get the U16.16 result.
+		 */
+		crtc_w = (u64)pstate->crtc_w << 32;
+		crtc_h = (u64)pstate->crtc_h << 32;
+		h_upscale_factor = (u32)(crtc_w / pstate->src_w);
+		v_upscale_factor = (u32)(crtc_h / pstate->src_h);
+
+		/* Downscaling won't work when mclk == pxlclk. */
+		if (!(h_upscale_factor >> 16) || !(v_upscale_factor >> 16))
+			return -EINVAL;
+
+		s->input_w = pstate->src_w >> 16;
+		s->input_h = pstate->src_h >> 16;
+		s->output_w = pstate->crtc_w;
+		s->output_h = pstate->crtc_h;
+
+#define SE_N_PHASE 4
+#define SE_SHIFT_N_PHASE 12
+		/* Calculate initial_phase and delta_phase for horizontal. */
+		phase = s->input_w;
+		s->h_init_phase =
+				((phase << SE_N_PHASE) / s->output_w + 1) / 2;
+
+		phase = s->input_w;
+		phase <<= (SE_SHIFT_N_PHASE + SE_N_PHASE);
+		s->h_delta_phase = phase / s->output_w;
+
+		/* Same for vertical. */
+		phase = s->input_h;
+		s->v_init_phase =
+				((phase << SE_N_PHASE) / s->output_h + 1) / 2;
+
+		phase = s->input_h;
+		phase <<= (SE_SHIFT_N_PHASE + SE_N_PHASE);
+		s->v_delta_phase = phase / s->output_h;
+#undef SE_N_PHASE
+#undef SE_SHIFT_N_PHASE
+		s->plane_src_id = mp->layer->id;
+	}
+
+	s->scale_enable = true;
+	s->hcoeff = malidp_se_select_coeffs(h_upscale_factor);
+	s->vcoeff = malidp_se_select_coeffs(v_upscale_factor);
+	return 0;
+}
+
 static int malidp_crtc_atomic_check(struct drm_crtc *crtc,
 				    struct drm_crtc_state *state)
 {
@@ -325,6 +401,7 @@ static int malidp_crtc_atomic_check(struct drm_crtc *crtc,
 
 	ret = malidp_crtc_atomic_check_gamma(crtc, state);
 	ret = ret ? ret : malidp_crtc_atomic_check_ctm(crtc, state);
+	ret = ret ? ret : malidp_crtc_atomic_check_scaling(crtc, state);
 
 	return ret;
 }
@@ -353,6 +430,9 @@ static struct drm_crtc_state *malidp_crtc_duplicate_state(struct drm_crtc *crtc)
 	       sizeof(state->gamma_coeffs));
 	memcpy(state->coloradj_coeffs, old_state->coloradj_coeffs,
 	       sizeof(state->coloradj_coeffs));
+	memcpy(&state->scaler_config, &old_state->scaler_config,
+	       sizeof(state->scaler_config));
+	state->scaled_planes_mask = 0;
 
 	return &state->base;
 }
diff --git a/drivers/gpu/drm/arm/malidp_drv.c b/drivers/gpu/drm/arm/malidp_drv.c
index 24bc96f9e91c..baca2dce1b92 100644
--- a/drivers/gpu/drm/arm/malidp_drv.c
+++ b/drivers/gpu/drm/arm/malidp_drv.c
@@ -110,6 +110,52 @@ void malidp_atomic_commit_update_coloradj(struct drm_crtc *crtc,
 	}
 }
 
+static void malidp_atomic_commit_se_config(struct drm_crtc *crtc,
+					   struct drm_crtc_state *old_state)
+{
+	struct malidp_crtc_state *cs = to_malidp_crtc_state(crtc->state);
+	struct malidp_crtc_state *old_cs = to_malidp_crtc_state(old_state);
+	struct malidp_drm *malidp = crtc_to_malidp_device(crtc);
+	struct malidp_hw_device *hwdev = malidp->dev;
+	struct malidp_se_config *s = &cs->scaler_config;
+	struct malidp_se_config *old_s = &old_cs->scaler_config;
+	u32 se_control = hwdev->map.se_base +
+			 ((hwdev->map.features & MALIDP_REGMAP_HAS_CLEARIRQ) ?
+			 0x10 : 0xC);
+	u32 layer_control = se_control + MALIDP_SE_LAYER_CONTROL;
+	u32 scr = se_control + MALIDP_SE_SCALING_CONTROL;
+	u32 val;
+
+	/* Set SE_CONTROL */
+	if (!s->scale_enable) {
+		val = malidp_hw_read(hwdev, se_control);
+		val &= ~MALIDP_SE_SCALING_EN;
+		malidp_hw_write(hwdev, val, se_control);
+		return;
+	}
+
+	hwdev->se_set_scaling_coeffs(hwdev, s, old_s);
+	val = malidp_hw_read(hwdev, se_control);
+	val |= MALIDP_SE_SCALING_EN | MALIDP_SE_ALPHA_EN;
+
+	val |= MALIDP_SE_RGBO_IF_EN;
+	malidp_hw_write(hwdev, val, se_control);
+
+	/* Set IN_SIZE & OUT_SIZE. */
+	val = MALIDP_SE_SET_V_SIZE(s->input_h) |
+	      MALIDP_SE_SET_H_SIZE(s->input_w);
+	malidp_hw_write(hwdev, val, layer_control + MALIDP_SE_L0_IN_SIZE);
+	val = MALIDP_SE_SET_V_SIZE(s->output_h) |
+	      MALIDP_SE_SET_H_SIZE(s->output_w);
+	malidp_hw_write(hwdev, val, layer_control + MALIDP_SE_L0_OUT_SIZE);
+
+	/* Set phase regs. */
+	malidp_hw_write(hwdev, s->h_init_phase, scr + MALIDP_SE_H_INIT_PH);
+	malidp_hw_write(hwdev, s->h_delta_phase, scr + MALIDP_SE_H_DELTA_PH);
+	malidp_hw_write(hwdev, s->v_init_phase, scr + MALIDP_SE_V_INIT_PH);
+	malidp_hw_write(hwdev, s->v_delta_phase, scr + MALIDP_SE_V_DELTA_PH);
+}
+
 /*
  * set the "config valid" bit and wait until the hardware acts on it
  */
@@ -179,6 +225,7 @@ static void malidp_atomic_commit_tail(struct drm_atomic_state *state)
 	for_each_crtc_in_state(state, crtc, old_crtc_state, i) {
 		malidp_atomic_commit_update_gamma(crtc, old_crtc_state);
 		malidp_atomic_commit_update_coloradj(crtc, old_crtc_state);
+		malidp_atomic_commit_se_config(crtc, old_crtc_state);
 	}
 
 	drm_atomic_helper_commit_planes(drm, state, 0);
diff --git a/drivers/gpu/drm/arm/malidp_drv.h b/drivers/gpu/drm/arm/malidp_drv.h
index 75070efda2fc..040311ffcaec 100644
--- a/drivers/gpu/drm/arm/malidp_drv.h
+++ b/drivers/gpu/drm/arm/malidp_drv.h
@@ -53,6 +53,9 @@ struct malidp_crtc_state {
 	struct drm_crtc_state base;
 	u32 gamma_coeffs[MALIDP_COEFFTAB_NUM_COEFFS];
 	u32 coloradj_coeffs[MALIDP_COLORADJ_NUM_COEFFS];
+	struct malidp_se_config scaler_config;
+	/* Bitfield of all the planes that have requested a scaled output. */
+	u8 scaled_planes_mask;
 };
 
 #define to_malidp_crtc_state(x) container_of(x, struct malidp_crtc_state, base)
diff --git a/drivers/gpu/drm/arm/malidp_hw.c b/drivers/gpu/drm/arm/malidp_hw.c
index b8025b7c1aca..e444c23a8261 100644
--- a/drivers/gpu/drm/arm/malidp_hw.c
+++ b/drivers/gpu/drm/arm/malidp_hw.c
@@ -86,6 +86,80 @@ static const struct malidp_layer malidp550_layers[] = {
 	{ DE_SMART, MALIDP550_DE_LS_BASE, MALIDP550_DE_LS_PTR_BASE, MALIDP550_DE_LS_R1_STRIDE },
 };
 
+#define SE_N_SCALING_COEFFS	96
+static const u16 dp500_se_scaling_coeffs[][SE_N_SCALING_COEFFS] = {
+	[MALIDP_UPSCALING_COEFFS - 1] = {
+		0x0000, 0x0001, 0x0007, 0x0011, 0x001e, 0x002e, 0x003f, 0x0052,
+		0x0064, 0x0073, 0x007d, 0x0080, 0x007a, 0x006c, 0x0053, 0x002f,
+		0x0000, 0x3fc6, 0x3f83, 0x3f39, 0x3eea, 0x3e9b, 0x3e4f, 0x3e0a,
+		0x3dd4, 0x3db0, 0x3da2, 0x3db1, 0x3dde, 0x3e2f, 0x3ea5, 0x3f40,
+		0x0000, 0x00e5, 0x01ee, 0x0315, 0x0456, 0x05aa, 0x0709, 0x086c,
+		0x09c9, 0x0b15, 0x0c4a, 0x0d5d, 0x0e4a, 0x0f06, 0x0f91, 0x0fe5,
+		0x1000, 0x0fe5, 0x0f91, 0x0f06, 0x0e4a, 0x0d5d, 0x0c4a, 0x0b15,
+		0x09c9, 0x086c, 0x0709, 0x05aa, 0x0456, 0x0315, 0x01ee, 0x00e5,
+		0x0000, 0x3f40, 0x3ea5, 0x3e2f, 0x3dde, 0x3db1, 0x3da2, 0x3db0,
+		0x3dd4, 0x3e0a, 0x3e4f, 0x3e9b, 0x3eea, 0x3f39, 0x3f83, 0x3fc6,
+		0x0000, 0x002f, 0x0053, 0x006c, 0x007a, 0x0080, 0x007d, 0x0073,
+		0x0064, 0x0052, 0x003f, 0x002e, 0x001e, 0x0011, 0x0007, 0x0001
+	},
+	[MALIDP_DOWNSCALING_1_5_COEFFS - 1] = {
+		0x0059, 0x004f, 0x0041, 0x002e, 0x0016, 0x3ffb, 0x3fd9, 0x3fb4,
+		0x3f8c, 0x3f62, 0x3f36, 0x3f09, 0x3edd, 0x3eb3, 0x3e8d, 0x3e6c,
+		0x3e52, 0x3e3f, 0x3e35, 0x3e37, 0x3e46, 0x3e61, 0x3e8c, 0x3ec5,
+		0x3f0f, 0x3f68, 0x3fd1, 0x004a, 0x00d3, 0x0169, 0x020b, 0x02b8,
+		0x036e, 0x042d, 0x04f2, 0x05b9, 0x0681, 0x0745, 0x0803, 0x08ba,
+		0x0965, 0x0a03, 0x0a91, 0x0b0d, 0x0b75, 0x0bc6, 0x0c00, 0x0c20,
+		0x0c28, 0x0c20, 0x0c00, 0x0bc6, 0x0b75, 0x0b0d, 0x0a91, 0x0a03,
+		0x0965, 0x08ba, 0x0803, 0x0745, 0x0681, 0x05b9, 0x04f2, 0x042d,
+		0x036e, 0x02b8, 0x020b, 0x0169, 0x00d3, 0x004a, 0x3fd1, 0x3f68,
+		0x3f0f, 0x3ec5, 0x3e8c, 0x3e61, 0x3e46, 0x3e37, 0x3e35, 0x3e3f,
+		0x3e52, 0x3e6c, 0x3e8d, 0x3eb3, 0x3edd, 0x3f09, 0x3f36, 0x3f62,
+		0x3f8c, 0x3fb4, 0x3fd9, 0x3ffb, 0x0016, 0x002e, 0x0041, 0x004f
+	},
+	[MALIDP_DOWNSCALING_2_COEFFS - 1] = {
+		0x3f19, 0x3f03, 0x3ef0, 0x3edf, 0x3ed0, 0x3ec5, 0x3ebd, 0x3eb9,
+		0x3eb9, 0x3ebf, 0x3eca, 0x3ed9, 0x3eef, 0x3f0a, 0x3f2c, 0x3f52,
+		0x3f7f, 0x3fb0, 0x3fe8, 0x0026, 0x006a, 0x00b4, 0x0103, 0x0158,
+		0x01b1, 0x020d, 0x026c, 0x02cd, 0x032f, 0x0392, 0x03f4, 0x0455,
+		0x04b4, 0x051e, 0x0585, 0x05eb, 0x064c, 0x06a8, 0x06fe, 0x074e,
+		0x0796, 0x07d5, 0x080c, 0x0839, 0x085c, 0x0875, 0x0882, 0x0887,
+		0x0881, 0x0887, 0x0882, 0x0875, 0x085c, 0x0839, 0x080c, 0x07d5,
+		0x0796, 0x074e, 0x06fe, 0x06a8, 0x064c, 0x05eb, 0x0585, 0x051e,
+		0x04b4, 0x0455, 0x03f4, 0x0392, 0x032f, 0x02cd, 0x026c, 0x020d,
+		0x01b1, 0x0158, 0x0103, 0x00b4, 0x006a, 0x0026, 0x3fe8, 0x3fb0,
+		0x3f7f, 0x3f52, 0x3f2c, 0x3f0a, 0x3eef, 0x3ed9, 0x3eca, 0x3ebf,
+		0x3eb9, 0x3eb9, 0x3ebd, 0x3ec5, 0x3ed0, 0x3edf, 0x3ef0, 0x3f03
+	},
+	[MALIDP_DOWNSCALING_2_75_COEFFS - 1] = {
+		0x3f51, 0x3f60, 0x3f71, 0x3f84, 0x3f98, 0x3faf, 0x3fc8, 0x3fe3,
+		0x0000, 0x001f, 0x0040, 0x0064, 0x008a, 0x00b1, 0x00da, 0x0106,
+		0x0133, 0x0160, 0x018e, 0x01bd, 0x01ec, 0x021d, 0x024e, 0x0280,
+		0x02b2, 0x02e4, 0x0317, 0x0349, 0x037c, 0x03ad, 0x03df, 0x0410,
+		0x0440, 0x0468, 0x048f, 0x04b3, 0x04d6, 0x04f8, 0x0516, 0x0533,
+		0x054e, 0x0566, 0x057c, 0x0590, 0x05a0, 0x05ae, 0x05ba, 0x05c3,
+		0x05c9, 0x05c3, 0x05ba, 0x05ae, 0x05a0, 0x0590, 0x057c, 0x0566,
+		0x054e, 0x0533, 0x0516, 0x04f8, 0x04d6, 0x04b3, 0x048f, 0x0468,
+		0x0440, 0x0410, 0x03df, 0x03ad, 0x037c, 0x0349, 0x0317, 0x02e4,
+		0x02b2, 0x0280, 0x024e, 0x021d, 0x01ec, 0x01bd, 0x018e, 0x0160,
+		0x0133, 0x0106, 0x00da, 0x00b1, 0x008a, 0x0064, 0x0040, 0x001f,
+		0x0000, 0x3fe3, 0x3fc8, 0x3faf, 0x3f98, 0x3f84, 0x3f71, 0x3f60
+	},
+	[MALIDP_DOWNSCALING_4_COEFFS - 1] = {
+		0x0094, 0x00a9, 0x00be, 0x00d4, 0x00ea, 0x0101, 0x0118, 0x012f,
+		0x0148, 0x0160, 0x017a, 0x0193, 0x01ae, 0x01c8, 0x01e4, 0x01ff,
+		0x021c, 0x0233, 0x024a, 0x0261, 0x0278, 0x028f, 0x02a6, 0x02bd,
+		0x02d4, 0x02eb, 0x0302, 0x0319, 0x032f, 0x0346, 0x035d, 0x0374,
+		0x038a, 0x0397, 0x03a3, 0x03af, 0x03bb, 0x03c6, 0x03d1, 0x03db,
+		0x03e4, 0x03ed, 0x03f6, 0x03fe, 0x0406, 0x040d, 0x0414, 0x041a,
+		0x0420, 0x041a, 0x0414, 0x040d, 0x0406, 0x03fe, 0x03f6, 0x03ed,
+		0x03e4, 0x03db, 0x03d1, 0x03c6, 0x03bb, 0x03af, 0x03a3, 0x0397,
+		0x038a, 0x0374, 0x035d, 0x0346, 0x032f, 0x0319, 0x0302, 0x02eb,
+		0x02d4, 0x02bd, 0x02a6, 0x028f, 0x0278, 0x0261, 0x024a, 0x0233,
+		0x021c, 0x01ff, 0x01e4, 0x01c8, 0x01ae, 0x0193, 0x017a, 0x0160,
+		0x0148, 0x012f, 0x0118, 0x0101, 0x00ea, 0x00d4, 0x00be, 0x00a9
+	},
+};
+
 #define MALIDP_DE_DEFAULT_PREFETCH_START	5
 
 static int malidp500_query_hw(struct malidp_hw_device *hwdev)
@@ -211,6 +285,55 @@ static int malidp500_rotmem_required(struct malidp_hw_device *hwdev, u16 w, u16
 	return w * drm_format_plane_cpp(fmt, 0) * 8;
 }
 
+static void malidp500_se_write_pp_coefftab(struct malidp_hw_device *hwdev,
+					   u32 direction,
+					   u16 addr,
+					   u8 coeffs_id)
+{
+	int i;
+	u16 scaling_control = MALIDP500_SE_CONTROL + MALIDP_SE_SCALING_CONTROL;
+
+	malidp_hw_write(hwdev,
+			direction | (addr & MALIDP_SE_COEFFTAB_ADDR_MASK),
+			scaling_control + MALIDP_SE_COEFFTAB_ADDR);
+	for (i = 0; i < ARRAY_SIZE(dp500_se_scaling_coeffs); ++i)
+		malidp_hw_write(hwdev, MALIDP_SE_SET_COEFFTAB_DATA(
+				dp500_se_scaling_coeffs[coeffs_id][i]),
+				scaling_control + MALIDP_SE_COEFFTAB_DATA);
+}
+
+static int malidp500_se_set_scaling_coeffs(struct malidp_hw_device *hwdev,
+					   struct malidp_se_config *se_config,
+					   struct malidp_se_config *old_config)
+{
+	/* Get array indices into dp500_se_scaling_coeffs. */
+	u8 h = (u8)se_config->hcoeff - 1;
+	u8 v = (u8)se_config->vcoeff - 1;
+
+	if (WARN_ON(h >= ARRAY_SIZE(dp500_se_scaling_coeffs) ||
+		    v >= ARRAY_SIZE(dp500_se_scaling_coeffs)))
+		return -EINVAL;
+
+	if ((h == v) && (se_config->hcoeff != old_config->hcoeff ||
+			 se_config->vcoeff != old_config->vcoeff)) {
+		malidp500_se_write_pp_coefftab(hwdev,
+					       (MALIDP_SE_V_COEFFTAB |
+						MALIDP_SE_H_COEFFTAB),
+					       0, v);
+	} else {
+		if (se_config->vcoeff != old_config->vcoeff)
+			malidp500_se_write_pp_coefftab(hwdev,
+						       MALIDP_SE_V_COEFFTAB,
+						       0, v);
+		if (se_config->hcoeff != old_config->hcoeff)
+			malidp500_se_write_pp_coefftab(hwdev,
+						       MALIDP_SE_H_COEFFTAB,
+						       0, h);
+	}
+
+	return 0;
+}
+
 static int malidp550_query_hw(struct malidp_hw_device *hwdev)
 {
 	u32 conf = malidp_hw_read(hwdev, MALIDP550_CONFIG_ID);
@@ -384,6 +507,20 @@ static int malidp550_rotmem_required(struct malidp_hw_device *hwdev, u16 w, u16
 	return w * bytes_per_col;
 }
 
+static int malidp550_se_set_scaling_coeffs(struct malidp_hw_device *hwdev,
+					   struct malidp_se_config *se_config,
+					   struct malidp_se_config *old_config)
+{
+	u32 mask = MALIDP550_SE_CTL_VCSEL(MALIDP550_SE_CTL_SEL_MASK) |
+		   MALIDP550_SE_CTL_HCSEL(MALIDP550_SE_CTL_SEL_MASK);
+	u32 new_value = MALIDP550_SE_CTL_VCSEL(se_config->vcoeff) |
+			MALIDP550_SE_CTL_HCSEL(se_config->hcoeff);
+
+	malidp_hw_clearbits(hwdev, mask, MALIDP550_SE_CONTROL);
+	malidp_hw_setbits(hwdev, new_value, MALIDP550_SE_CONTROL);
+	return 0;
+}
+
 static int malidp650_query_hw(struct malidp_hw_device *hwdev)
 {
 	u32 conf = malidp_hw_read(hwdev, MALIDP550_CONFIG_ID);
@@ -448,6 +585,7 @@ const struct malidp_hw_device malidp_device[MALIDP_MAX_DEVICES] = {
 		.set_config_valid = malidp500_set_config_valid,
 		.modeset = malidp500_modeset,
 		.rotmem_required = malidp500_rotmem_required,
+		.se_set_scaling_coeffs = malidp500_se_set_scaling_coeffs,
 		.features = MALIDP_DEVICE_LV_HAS_3_STRIDES,
 	},
 	[MALIDP_550] = {
@@ -483,6 +621,7 @@ const struct malidp_hw_device malidp_device[MALIDP_MAX_DEVICES] = {
 		.set_config_valid = malidp550_set_config_valid,
 		.modeset = malidp550_modeset,
 		.rotmem_required = malidp550_rotmem_required,
+		.se_set_scaling_coeffs = malidp550_se_set_scaling_coeffs,
 		.features = 0,
 	},
 	[MALIDP_650] = {
@@ -519,6 +658,7 @@ const struct malidp_hw_device malidp_device[MALIDP_MAX_DEVICES] = {
 		.set_config_valid = malidp550_set_config_valid,
 		.modeset = malidp550_modeset,
 		.rotmem_required = malidp550_rotmem_required,
+		.se_set_scaling_coeffs = malidp550_se_set_scaling_coeffs,
 		.features = 0,
 	},
 };
diff --git a/drivers/gpu/drm/arm/malidp_hw.h b/drivers/gpu/drm/arm/malidp_hw.h
index 470fe71724a3..52a65a8a056f 100644
--- a/drivers/gpu/drm/arm/malidp_hw.h
+++ b/drivers/gpu/drm/arm/malidp_hw.h
@@ -61,6 +61,25 @@ struct malidp_layer {
 	u16 stride_offset;	/* Offset to the first stride register. */
 };
 
+enum malidp_scaling_coeff_set {
+	MALIDP_UPSCALING_COEFFS = 1,
+	MALIDP_DOWNSCALING_1_5_COEFFS = 2,
+	MALIDP_DOWNSCALING_2_COEFFS = 3,
+	MALIDP_DOWNSCALING_2_75_COEFFS = 4,
+	MALIDP_DOWNSCALING_4_COEFFS = 5,
+};
+
+struct malidp_se_config {
+	u8 scale_enable : 1;
+	u8 hcoeff : 3;
+	u8 vcoeff : 3;
+	u8 plane_src_id;
+	u16 input_w, input_h;
+	u16 output_w, output_h;
+	u32 h_init_phase, h_delta_phase;
+	u32 v_init_phase, v_delta_phase;
+};
+
 /* regmap features */
 #define MALIDP_REGMAP_HAS_CLEARIRQ	(1 << 0)
 
@@ -153,6 +172,10 @@ struct malidp_hw_device {
 	 */
 	int (*rotmem_required)(struct malidp_hw_device *hwdev, u16 w, u16 h, u32 fmt);
 
+	int (*se_set_scaling_coeffs)(struct malidp_hw_device *hwdev,
+				     struct malidp_se_config *se_config,
+				     struct malidp_se_config *old_config);
+
 	u8 features;
 
 	u8 min_line_size;
@@ -250,6 +273,28 @@ static inline bool malidp_hw_pitch_valid(struct malidp_hw_device *hwdev,
 	return !(pitch & (hwdev->map.bus_align_bytes - 1));
 }
 
+/* U16.16 */
+#define FP_1_00000	0x00010000	/* 1.0 */
+#define FP_0_66667	0x0000AAAA	/* 0.6667 = 1/1.5 */
+#define FP_0_50000	0x00008000	/* 0.5 = 1/2 */
+#define FP_0_36363	0x00005D17	/* 0.36363 = 1/2.75 */
+#define FP_0_25000	0x00004000	/* 0.25 = 1/4 */
+
+static inline enum malidp_scaling_coeff_set
+malidp_se_select_coeffs(u32 upscale_factor)
+{
+	return (upscale_factor >= FP_1_00000) ? MALIDP_UPSCALING_COEFFS :
+	       (upscale_factor >= FP_0_66667) ? MALIDP_DOWNSCALING_1_5_COEFFS :
+	       (upscale_factor >= FP_0_50000) ? MALIDP_DOWNSCALING_2_COEFFS :
+	       (upscale_factor >= FP_0_36363) ? MALIDP_DOWNSCALING_2_75_COEFFS :
+	       MALIDP_DOWNSCALING_4_COEFFS;
+}
+
+#undef FP_0_25000
+#undef FP_0_36363
+#undef FP_0_50000
+#undef FP_0_66667
+#undef FP_1_00000
 /*
  * background color components are defined as 12bits values,
  * they will be shifted right when stored on hardware that
diff --git a/drivers/gpu/drm/arm/malidp_planes.c b/drivers/gpu/drm/arm/malidp_planes.c
index 8106e22f9349..814fda23cead 100644
--- a/drivers/gpu/drm/arm/malidp_planes.c
+++ b/drivers/gpu/drm/arm/malidp_planes.c
@@ -25,6 +25,9 @@
 #define MALIDP_LAYER_FORMAT		0x000
 #define MALIDP_LAYER_CONTROL		0x004
 #define   LAYER_ENABLE			(1 << 0)
+#define   LAYER_FLOWCFG_MASK		7
+#define   LAYER_FLOWCFG(x)		(((x) & LAYER_FLOWCFG_MASK) << 1)
+#define     LAYER_FLOWCFG_SCALE_SE	3
 #define   LAYER_ROT_OFFSET		8
 #define   LAYER_H_FLIP			(1 << 10)
 #define   LAYER_V_FLIP			(1 << 11)
@@ -133,16 +136,50 @@ static const struct drm_plane_funcs malidp_de_plane_funcs = {
 	.atomic_print_state = malidp_plane_atomic_print_state,
 };
 
+static int malidp_se_check_scaling(struct malidp_plane *mp,
+				   struct drm_plane_state *state)
+{
+	struct drm_crtc_state *crtc_state =
+		drm_atomic_get_existing_crtc_state(state->state, state->crtc);
+	struct malidp_crtc_state *mc;
+	struct drm_rect clip = { 0 };
+	u32 src_w, src_h;
+	int ret;
+
+	if (!crtc_state)
+		return -EINVAL;
+
+	clip.x2 = crtc_state->adjusted_mode.hdisplay;
+	clip.y2 = crtc_state->adjusted_mode.vdisplay;
+	ret = drm_plane_helper_check_state(state, &clip, 0, INT_MAX, true, true);
+	if (ret)
+		return ret;
+
+	src_w = state->src_w >> 16;
+	src_h = state->src_h >> 16;
+	if ((state->crtc_w == src_w) && (state->crtc_h == src_h)) {
+		/* Scaling not necessary for this plane. */
+		mc->scaled_planes_mask &= ~(mp->layer->id);
+		return 0;
+	}
+
+	if (mp->layer->id & (DE_SMART | DE_GRAPHICS2))
+		return -EINVAL;
+
+	mc = to_malidp_crtc_state(crtc_state);
+
+	mc->scaled_planes_mask |= mp->layer->id;
+	/* Defer scaling requirements calculation to the crtc check. */
+	return 0;
+}
+
 static int malidp_de_plane_check(struct drm_plane *plane,
 				 struct drm_plane_state *state)
 {
 	struct malidp_plane *mp = to_malidp_plane(plane);
 	struct malidp_plane_state *ms = to_malidp_plane_state(state);
-	struct drm_crtc_state *crtc_state;
 	struct drm_framebuffer *fb;
-	struct drm_rect clip = { 0 };
 	int i, ret;
-	u32 src_w, src_h;
 
 	if (!state->crtc || !state->fb)
 		return 0;
@@ -163,9 +200,6 @@ static int malidp_de_plane_check(struct drm_plane *plane,
 		}
 	}
 
-	src_w = state->src_w >> 16;
-	src_h = state->src_h >> 16;
-
 	if ((state->crtc_w > mp->hwdev->max_line_size) ||
 	    (state->crtc_h > mp->hwdev->max_line_size) ||
 	    (state->crtc_w < mp->hwdev->min_line_size) ||
@@ -182,22 +216,16 @@ static int malidp_de_plane_check(struct drm_plane *plane,
 	    (state->fb->pitches[1] != state->fb->pitches[2]))
 		return -EINVAL;
 
+	ret = malidp_se_check_scaling(mp, state);
+	if (ret)
+		return ret;
+
 	/* packed RGB888 / BGR888 can't be rotated or flipped */
 	if (state->rotation != DRM_ROTATE_0 &&
 	    (fb->format->format == DRM_FORMAT_RGB888 ||
 	     fb->format->format == DRM_FORMAT_BGR888))
 		return -EINVAL;
 
-	crtc_state = drm_atomic_get_existing_crtc_state(state->state, state->crtc);
-	clip.x2 = crtc_state->adjusted_mode.hdisplay;
-	clip.y2 = crtc_state->adjusted_mode.vdisplay;
-	ret = drm_plane_helper_check_state(state, &clip,
-					   DRM_PLANE_HELPER_NO_SCALING,
-					   DRM_PLANE_HELPER_NO_SCALING,
-					   true, true);
-	if (ret)
-		return ret;
-
 	ms->rotmem_size = 0;
 	if (state->rotation & MALIDP_ROTATED_MASK) {
 		int val;
@@ -302,6 +330,16 @@ static void malidp_de_plane_update(struct drm_plane *plane,
 	val &= ~LAYER_COMP_MASK;
 	val |= LAYER_COMP_PIXEL;
 
+	val &= ~LAYER_FLOWCFG(LAYER_FLOWCFG_MASK);
+	if (plane->state->crtc) {
+		struct malidp_crtc_state *m =
+			to_malidp_crtc_state(plane->state->crtc->state);
+
+		if (m->scaler_config.scale_enable &&
+		    m->scaler_config.plane_src_id == mp->layer->id)
+			val |= LAYER_FLOWCFG(LAYER_FLOWCFG_SCALE_SE);
+	}
+
 	/* set the 'enable layer' bit */
 	val |= LAYER_ENABLE;
 
@@ -314,7 +352,8 @@ static void malidp_de_plane_disable(struct drm_plane *plane,
 {
 	struct malidp_plane *mp = to_malidp_plane(plane);
 
-	malidp_hw_clearbits(mp->hwdev, LAYER_ENABLE,
+	malidp_hw_clearbits(mp->hwdev,
+			    LAYER_ENABLE | LAYER_FLOWCFG(LAYER_FLOWCFG_MASK),
 			    mp->layer->base + MALIDP_LAYER_CONTROL);
 }
 
diff --git a/drivers/gpu/drm/arm/malidp_regs.h b/drivers/gpu/drm/arm/malidp_regs.h
index 4bf338d4172e..91d2ba5729bf 100644
--- a/drivers/gpu/drm/arm/malidp_regs.h
+++ b/drivers/gpu/drm/arm/malidp_regs.h
@@ -106,6 +106,40 @@
 #define MALIDP_COEF_TABLE_ADDR		0x00030
 #define MALIDP_COEF_TABLE_DATA		0x00034
 
+/* Scaling engine registers and masks. */
+#define   MALIDP_SE_SCALING_EN			(1 << 0)
+#define   MALIDP_SE_ALPHA_EN			(1 << 1)
+#define   MALIDP_SE_RGBO_IF_EN			(1 << 4)
+#define   MALIDP550_SE_CTL_SEL_MASK		7
+#define   MALIDP550_SE_CTL_VCSEL(x) \
+		(((x) & MALIDP550_SE_CTL_SEL_MASK) << 20)
+#define   MALIDP550_SE_CTL_HCSEL(x) \
+		(((x) & MALIDP550_SE_CTL_SEL_MASK) << 16)
+
+/* Blocks with offsets from SE_CONTROL register. */
+#define MALIDP_SE_LAYER_CONTROL			0x14
+#define   MALIDP_SE_L0_IN_SIZE			0x00
+#define   MALIDP_SE_L0_OUT_SIZE			0x04
+#define   MALIDP_SE_SET_V_SIZE(x)		(((x) & 0x1fff) << 16)
+#define   MALIDP_SE_SET_H_SIZE(x)		(((x) & 0x1fff) << 0)
+#define MALIDP_SE_SCALING_CONTROL		0x24
+#define   MALIDP_SE_H_INIT_PH			0x00
+#define   MALIDP_SE_H_DELTA_PH			0x04
+#define   MALIDP_SE_V_INIT_PH			0x08
+#define   MALIDP_SE_V_DELTA_PH			0x0c
+#define   MALIDP_SE_COEFFTAB_ADDR		0x10
+#define     MALIDP_SE_COEFFTAB_ADDR_MASK	0x7f
+#define     MALIDP_SE_V_COEFFTAB		(1 << 8)
+#define     MALIDP_SE_H_COEFFTAB		(1 << 9)
+#define     MALIDP_SE_SET_V_COEFFTAB_ADDR(x) \
+		(MALIDP_SE_V_COEFFTAB | ((x) & MALIDP_SE_COEFFTAB_ADDR_MASK))
+#define     MALIDP_SE_SET_H_COEFFTAB_ADDR(x) \
+		(MALIDP_SE_H_COEFFTAB | ((x) & MALIDP_SE_COEFFTAB_ADDR_MASK))
+#define   MALIDP_SE_COEFFTAB_DATA		0x14
+#define     MALIDP_SE_COEFFTAB_DATA_MASK	0x3fff
+#define     MALIDP_SE_SET_COEFFTAB_DATA(x) \
+		((x) & MALIDP_SE_COEFFTAB_DATA_MASK)
+
 /* register offsets and bits specific to DP500 */
 #define MALIDP500_ADDR_SPACE_SIZE	0x01000
 #define MALIDP500_DC_BASE		0x00000
@@ -146,6 +180,7 @@
 #define MALIDP500_DE_LG2_BASE		0x00300
 #define MALIDP500_DE_LG2_PTR_BASE	0x0031c
 #define MALIDP500_SE_BASE		0x00c00
+#define MALIDP500_SE_CONTROL		0x00c0c
 #define MALIDP500_SE_PTR_BASE		0x00e0c
 #define MALIDP500_DC_IRQ_BASE		0x00f00
 #define MALIDP500_CONFIG_VALID		0x00f00
@@ -175,6 +210,7 @@
 #define MALIDP550_DE_LS_PTR_BASE	0x0042c
 #define MALIDP550_DE_PERF_BASE		0x00500
 #define MALIDP550_SE_BASE		0x08000
+#define MALIDP550_SE_CONTROL		0x08010
 #define MALIDP550_DC_BASE		0x0c000
 #define MALIDP550_DC_CONTROL		0x0c010
 #define   MALIDP550_DC_CONFIG_REQ	(1 << 16)