Merge tag 'drm-next-2019-07-16' of git://anongit.freedesktop.org/drm/drm

Pull drm updates from Dave Airlie:
 "The biggest thing in this is the AMD Navi GPU support, this again
  contains a bunch of header files that are large. These are the new AMD
  RX5700 GPUs that just recently became available.

  New drivers:
   - ST-Ericsson MCDE driver
   - Ingenic JZ47xx SoC

  UAPI change:
   - HDR source metadata property

  Core:
   - HDR inforframes and EDID parsing
   - drm hdmi infoframe unpacking
   - remove prime sg_table caching into dma-buf
   - New gem vram helpers to reduce driver code
   - Lots of drmP.h removal
   - reservation fencing fix
   - documentation updates
   - drm_fb_helper_connector removed
   - mode name command handler rewrite

  fbcon:
   - Remove the fbcon notifiers

  ttm:
   - forward progress fixes

  dma-buf:
   - make mmap call optional
   - debugfs refcount fixes
   - dma-fence free with pending signals fix
   - each dma-buf gets an inode

  Panels:
   - Lots of additional panel bindings

  amdgpu:
   - initial navi10 support
   - avoid hw reset
   - HDR metadata support
   - new thermal sensors for vega asics
   - RAS fixes
   - use HMM rather than MMU notifier
   - xgmi topology via kfd
   - SR-IOV fixes
   - driver reload fixes
   - DC use a core bpc attribute
   - Aux fixes for DC
   - Bandwidth calc updates for DC
   - Clock handling refactor
   - kfd VEGAM support

  vmwgfx:
   - Coherent memory support changes

  i915:
   - HDR Support
   - HDMI i2c link
   - Icelake multi-segmented gamma support
   - GuC firmware update
   - Mule Creek Canyon PCH support for EHL
   - EHL platform updtes
   - move i915.alpha_support to i915.force_probe
   - runtime PM refactoring
   - VBT parsing refactoring
   - DSI fixes
   - struct mutex dependency reduction
   - GEM code reorg

  mali-dp:
   - Komeda driver features

  msm:
   - dsi vs EPROBE_DEFER fixes
   - msm8998 snapdragon 835 support
   - a540 gpu support
   - mdp5 and dpu interconnect support

  exynos:
   - drmP.h removal

  tegra:
   - misc fixes

  tda998x:
   - audio support improvements
   - pixel repeated mode support
   - quantisation range handling corrections
   - HDMI vendor info fix

  armada:
   - interlace support fix
   - overlay/video plane register handling refactor
   - add gamma support

  rockchip:
   - RX3328 support

  panfrost:
   - expose perf counters via hidden ioctls

  vkms:
   - enumerate CRC sources list

  ast:
   - rework BO handling

  mgag200:
   - rework BO handling

  dw-hdmi:
   - suspend/resume support

  rcar-du:
   - R8A774A1 Soc Support
   - LVDS dual-link mode support
   - Additional formats
   - Misc fixes

  omapdrm:
   - DSI command mode display support

  stm
   - fb modifier support
   - runtime PM support

  sun4i:
   - use vmap ops

  vc4:
   - binner bo binding rework

  v3d:
   - compute shader support
   - resync/sync fixes
   - job management refactoring

  lima:
   - NULL pointer in irq handler fix
   - scheduler default timeout

  virtio:
   - fence seqno support
   - trace events

  bochs:
   - misc fixes

  tc458767:
   - IRQ/HDP handling

  sii902x:
   - HDMI audio support

  atmel-hlcdc:
   - misc fixes

  meson:
   - zpos support"

* tag 'drm-next-2019-07-16' of git://anongit.freedesktop.org/drm/drm: (1815 commits)
  Revert "Merge branch 'vmwgfx-next' of git://people.freedesktop.org/~thomash/linux into drm-next"
  Revert "mm: adjust apply_to_pfn_range interface for dropped token."
  mm: adjust apply_to_pfn_range interface for dropped token.
  drm/amdgpu/navi10: add uclk activity sensor
  drm/amdgpu: properly guard the generic discovery code
  drm/amdgpu: add missing documentation on new module parameters
  drm/amdgpu: don't invalidate caches in RELEASE_MEM, only do the writeback
  drm/amd/display: avoid 64-bit division
  drm/amdgpu/psp11: simplify the ucode register logic
  drm/amdgpu: properly guard DC support in navi code
  drm/amd/powerplay: vega20: fix uninitialized variable use
  drm/amd/display: dcn20: include linux/delay.h
  amdgpu: make pmu support optional
  drm/amd/powerplay: Zero initialize current_rpm in vega20_get_fan_speed_percent
  drm/amd/powerplay: Zero initialize freq in smu_v11_0_get_current_clk_freq
  drm/amd/powerplay: Use memset to initialize metrics structs
  drm/amdgpu/mes10.1: Fix header guard
  drm/amd/powerplay: add temperature sensor support for navi10
  drm/amdgpu: fix scheduler timeout calc
  drm/amdgpu: Prepare for hmm_range_register API change (v2)
  ...

11 files changed, 524 insertions, 213 deletions

diff --git a/drivers/gpu/drm/armada/armada_510.c b/drivers/gpu/drm/armada/armada_510.c
index dd5b5cc9a0d4..93d5c0a2d49a 100644
--- a/drivers/gpu/drm/armada/armada_510.c
+++ b/drivers/gpu/drm/armada/armada_510.c
@@ -11,18 +11,62 @@
 #include "armada_drm.h"
 #include "armada_hw.h"
 
+struct armada510_variant_data {
+	struct clk *clks[4];
+	struct clk *sel_clk;
+};
+
 static int armada510_crtc_init(struct armada_crtc *dcrtc, struct device *dev)
 {
+	struct armada510_variant_data *v;
 	struct clk *clk;
+	int idx;
+
+	v = devm_kzalloc(dev, sizeof(*v), GFP_KERNEL);
+	if (!v)
+		return -ENOMEM;
+
+	dcrtc->variant_data = v;
+
+	if (dev->of_node) {
+		struct property *prop;
+		const char *s;
+
+		of_property_for_each_string(dev->of_node, "clock-names", prop,
+					    s) {
+			if (!strcmp(s, "ext_ref_clk0"))
+				idx = 0;
+			else if (!strcmp(s, "ext_ref_clk1"))
+				idx = 1;
+			else if (!strcmp(s, "plldivider"))
+				idx = 2;
+			else if (!strcmp(s, "axibus"))
+				idx = 3;
+			else
+				continue;
+
+			clk = devm_clk_get(dev, s);
+			if (IS_ERR(clk))
+				return PTR_ERR(clk) == -ENOENT ? -EPROBE_DEFER :
+					PTR_ERR(clk);
+			v->clks[idx] = clk;
+		}
+	} else {
+		clk = devm_clk_get(dev, "ext_ref_clk1");
+		if (IS_ERR(clk))
+			return PTR_ERR(clk) == -ENOENT ? -EPROBE_DEFER :
+				PTR_ERR(clk);
+
+		v->clks[1] = clk;
+	}
 
-	clk = devm_clk_get(dev, "ext_ref_clk1");
-	if (IS_ERR(clk))
-		return PTR_ERR(clk) == -ENOENT ? -EPROBE_DEFER : PTR_ERR(clk);
-
-	dcrtc->extclk[0] = clk;
-
-	/* Lower the watermark so to eliminate jitter at higher bandwidths */
-	armada_updatel(0x20, (1 << 11) | 0xff, dcrtc->base + LCD_CFG_RDREG4F);
+	/*
+	 * Lower the watermark so to eliminate jitter at higher bandwidths.
+	 * Disable SRAM read wait state to avoid system hang with external
+	 * clock.
+	 */
+	armada_updatel(CFG_DMA_WM(0x20), CFG_SRAM_WAIT | CFG_DMA_WM_MASK,
+		       dcrtc->base + LCD_CFG_RDREG4F);
 
 	/* Initialise SPU register */
 	writel_relaxed(ADV_HWC32ENABLE | ADV_HWC32ARGB | ADV_HWC32BLEND,
@@ -31,65 +75,77 @@ static int armada510_crtc_init(struct armada_crtc *dcrtc, struct device *dev)
 	return 0;
 }
 
+static const u32 armada510_clk_sels[] = {
+	SCLK_510_EXTCLK0,
+	SCLK_510_EXTCLK1,
+	SCLK_510_PLL,
+	SCLK_510_AXI,
+};
+
+static const struct armada_clocking_params armada510_clocking = {
+	/* HDMI requires -0.6%..+0.5% */
+	.permillage_min = 994,
+	.permillage_max = 1005,
+	.settable = BIT(0) | BIT(1),
+	.div_max = SCLK_510_INT_DIV_MASK,
+};
+
 /*
  * Armada510 specific SCLK register selection.
  * This gets called with sclk = NULL to test whether the mode is
  * supportable, and again with sclk != NULL to set the clocks up for
  * that.  The former can return an error, but the latter is expected
  * not to.
- *
- * We currently are pretty rudimentary here, always selecting
- * EXT_REF_CLK_1 for LCD0 and erroring LCD1.  This needs improvement!
  */
 static int armada510_crtc_compute_clock(struct armada_crtc *dcrtc,
 	const struct drm_display_mode *mode, uint32_t *sclk)
 {
-	struct clk *clk = dcrtc->extclk[0];
-	int ret;
+	struct armada510_variant_data *v = dcrtc->variant_data;
+	unsigned long desired_khz = mode->crtc_clock;
+	struct armada_clk_result res;
+	int ret, idx;
 
-	if (dcrtc->num == 1)
-		return -EINVAL;
+	idx = armada_crtc_select_clock(dcrtc, &res, &armada510_clocking,
+				       v->clks, ARRAY_SIZE(v->clks),
+				       desired_khz);
+	if (idx < 0)
+		return idx;
 
-	if (IS_ERR(clk))
-		return PTR_ERR(clk);
-
-	if (dcrtc->clk != clk) {
-		ret = clk_prepare_enable(clk);
-		if (ret)
-			return ret;
-		dcrtc->clk = clk;
-	}
+	ret = clk_prepare_enable(res.clk);
+	if (ret)
+		return ret;
 
 	if (sclk) {
-		uint32_t rate, ref, div;
+		clk_set_rate(res.clk, res.desired_clk_hz);
 
-		rate = mode->clock * 1000;
-		ref = clk_round_rate(clk, rate);
-		div = DIV_ROUND_UP(ref, rate);
-		if (div < 1)
-			div = 1;
+		*sclk = res.div | armada510_clk_sels[idx];
 
-		clk_set_rate(clk, ref);
-		*sclk = div | SCLK_510_EXTCLK1;
+		/* We are now using this clock */
+		v->sel_clk = res.clk;
+		swap(dcrtc->clk, res.clk);
 	}
 
+	clk_disable_unprepare(res.clk);
+
 	return 0;
 }
 
 static void armada510_crtc_disable(struct armada_crtc *dcrtc)
 {
-	if (!IS_ERR(dcrtc->clk)) {
+	if (dcrtc->clk) {
 		clk_disable_unprepare(dcrtc->clk);
-		dcrtc->clk = ERR_PTR(-EINVAL);
+		dcrtc->clk = NULL;
 	}
 }
 
 static void armada510_crtc_enable(struct armada_crtc *dcrtc,
 	const struct drm_display_mode *mode)
 {
-	if (IS_ERR(dcrtc->clk)) {
-		dcrtc->clk = dcrtc->extclk[0];
-		WARN_ON(clk_prepare_enable(dcrtc->clk));
+	struct armada510_variant_data *v = dcrtc->variant_data;
+
+	if (!dcrtc->clk && v->sel_clk) {
+		if (!WARN_ON(clk_prepare_enable(v->sel_clk)))
+			dcrtc->clk = v->sel_clk;
 	}
 }
 
diff --git a/drivers/gpu/drm/armada/armada_crtc.c b/drivers/gpu/drm/armada/armada_crtc.c
index 5d77e51e3181..d44fca4e1655 100644
--- a/drivers/gpu/drm/armada/armada_crtc.c
+++ b/drivers/gpu/drm/armada/armada_crtc.c
@@ -127,6 +127,70 @@ static void armada_drm_crtc_queue_state_event(struct drm_crtc *crtc)
 	}
 }
 
+static void armada_drm_update_gamma(struct drm_crtc *crtc)
+{
+	struct drm_property_blob *blob = crtc->state->gamma_lut;
+	void __iomem *base = drm_to_armada_crtc(crtc)->base;
+	int i;
+
+	if (blob) {
+		struct drm_color_lut *lut = blob->data;
+
+		armada_updatel(CFG_CSB_256x8, CFG_CSB_256x8 | CFG_PDWN256x8,
+			       base + LCD_SPU_SRAM_PARA1);
+
+		for (i = 0; i < 256; i++) {
+			writel_relaxed(drm_color_lut_extract(lut[i].red, 8),
+				       base + LCD_SPU_SRAM_WRDAT);
+			writel_relaxed(i | SRAM_WRITE | SRAM_GAMMA_YR,
+				       base + LCD_SPU_SRAM_CTRL);
+			readl_relaxed(base + LCD_SPU_HWC_OVSA_HPXL_VLN);
+			writel_relaxed(drm_color_lut_extract(lut[i].green, 8),
+				       base + LCD_SPU_SRAM_WRDAT);
+			writel_relaxed(i | SRAM_WRITE | SRAM_GAMMA_UG,
+				       base + LCD_SPU_SRAM_CTRL);
+			readl_relaxed(base + LCD_SPU_HWC_OVSA_HPXL_VLN);
+			writel_relaxed(drm_color_lut_extract(lut[i].blue, 8),
+				       base + LCD_SPU_SRAM_WRDAT);
+			writel_relaxed(i | SRAM_WRITE | SRAM_GAMMA_VB,
+				       base + LCD_SPU_SRAM_CTRL);
+			readl_relaxed(base + LCD_SPU_HWC_OVSA_HPXL_VLN);
+		}
+		armada_updatel(CFG_GAMMA_ENA, CFG_GAMMA_ENA,
+			       base + LCD_SPU_DMA_CTRL0);
+	} else {
+		armada_updatel(0, CFG_GAMMA_ENA, base + LCD_SPU_DMA_CTRL0);
+		armada_updatel(CFG_PDWN256x8, CFG_CSB_256x8 | CFG_PDWN256x8,
+			       base + LCD_SPU_SRAM_PARA1);
+	}
+}
+
+static enum drm_mode_status armada_drm_crtc_mode_valid(struct drm_crtc *crtc,
+	const struct drm_display_mode *mode)
+{
+	struct armada_crtc *dcrtc = drm_to_armada_crtc(crtc);
+
+	if (mode->vscan > 1)
+		return MODE_NO_VSCAN;
+
+	if (mode->flags & DRM_MODE_FLAG_DBLSCAN)
+		return MODE_NO_DBLESCAN;
+
+	if (mode->flags & DRM_MODE_FLAG_HSKEW)
+		return MODE_H_ILLEGAL;
+
+	/* We can't do interlaced modes if we don't have the SPU_ADV_REG */
+	if (!dcrtc->variant->has_spu_adv_reg &&
+	    mode->flags & DRM_MODE_FLAG_INTERLACE)
+		return MODE_NO_INTERLACE;
+
+	if (mode->flags & (DRM_MODE_FLAG_BCAST | DRM_MODE_FLAG_PIXMUX |
+			   DRM_MODE_FLAG_CLKDIV2))
+		return MODE_BAD;
+
+	return MODE_OK;
+}
+
 /* The mode_config.mutex will be held for this call */
 static bool armada_drm_crtc_mode_fixup(struct drm_crtc *crtc,
 	const struct drm_display_mode *mode, struct drm_display_mode *adj)
@@ -134,9 +198,18 @@ static bool armada_drm_crtc_mode_fixup(struct drm_crtc *crtc,
 	struct armada_crtc *dcrtc = drm_to_armada_crtc(crtc);
 	int ret;
 
-	/* We can't do interlaced modes if we don't have the SPU_ADV_REG */
-	if (!dcrtc->variant->has_spu_adv_reg &&
-	    adj->flags & DRM_MODE_FLAG_INTERLACE)
+	/*
+	 * Set CRTC modesetting parameters for the adjusted mode.  This is
+	 * applied after the connectors, bridges, and encoders have fixed up
+	 * this mode, as described above drm_atomic_helper_check_modeset().
+	 */
+	drm_mode_set_crtcinfo(adj, CRTC_INTERLACE_HALVE_V);
+
+	/*
+	 * Validate the adjusted mode in case an encoder/bridge has set
+	 * something we don't support.
+	 */
+	if (armada_drm_crtc_mode_valid(crtc, adj) != MODE_OK)
 		return false;
 
 	/* Check whether the display mode is possible */
@@ -275,16 +348,9 @@ static void armada_drm_crtc_mode_set_nofb(struct drm_crtc *crtc)
 
 	armada_reg_queue_set(regs, i, sclk, LCD_CFG_SCLK_DIV);
 
-	if (interlaced ^ dcrtc->interlaced) {
-		if (adj->flags & DRM_MODE_FLAG_INTERLACE)
-			drm_crtc_vblank_get(&dcrtc->crtc);
-		else
-			drm_crtc_vblank_put(&dcrtc->crtc);
-		dcrtc->interlaced = interlaced;
-	}
-
 	spin_lock_irqsave(&dcrtc->irq_lock, flags);
 
+	dcrtc->interlaced = interlaced;
 	/* Even interlaced/progressive frame */
 	dcrtc->v[1].spu_v_h_total = adj->crtc_vtotal << 16 |
 				    adj->crtc_htotal;
@@ -342,6 +408,20 @@ static void armada_drm_crtc_mode_set_nofb(struct drm_crtc *crtc)
 	spin_unlock_irqrestore(&dcrtc->irq_lock, flags);
 }
 
+static int armada_drm_crtc_atomic_check(struct drm_crtc *crtc,
+					struct drm_crtc_state *state)
+{
+	DRM_DEBUG_KMS("[CRTC:%d:%s]\n", crtc->base.id, crtc->name);
+
+	if (state->gamma_lut && drm_color_lut_size(state->gamma_lut) != 256)
+		return -EINVAL;
+
+	if (state->color_mgmt_changed)
+		state->planes_changed = true;
+
+	return 0;
+}
+
 static void armada_drm_crtc_atomic_begin(struct drm_crtc *crtc,
 					 struct drm_crtc_state *old_crtc_state)
 {
@@ -349,6 +429,9 @@ static void armada_drm_crtc_atomic_begin(struct drm_crtc *crtc,
 
 	DRM_DEBUG_KMS("[CRTC:%d:%s]\n", crtc->base.id, crtc->name);
 
+	if (crtc->state->color_mgmt_changed)
+		armada_drm_update_gamma(crtc);
+
 	dcrtc->regs_idx = 0;
 	dcrtc->regs = dcrtc->atomic_regs;
 }
@@ -387,6 +470,9 @@ static void armada_drm_crtc_atomic_disable(struct drm_crtc *crtc,
 
 	DRM_DEBUG_KMS("[CRTC:%d:%s]\n", crtc->base.id, crtc->name);
 
+	if (old_state->adjusted_mode.flags & DRM_MODE_FLAG_INTERLACE)
+		drm_crtc_vblank_put(crtc);
+
 	drm_crtc_vblank_off(crtc);
 	armada_drm_crtc_update(dcrtc, false);
 
@@ -431,12 +517,17 @@ static void armada_drm_crtc_atomic_enable(struct drm_crtc *crtc,
 	armada_drm_crtc_update(dcrtc, true);
 	drm_crtc_vblank_on(crtc);
 
+	if (crtc->state->adjusted_mode.flags & DRM_MODE_FLAG_INTERLACE)
+		WARN_ON(drm_crtc_vblank_get(crtc));
+
 	armada_drm_crtc_queue_state_event(crtc);
 }
 
 static const struct drm_crtc_helper_funcs armada_crtc_helper_funcs = {
+	.mode_valid	= armada_drm_crtc_mode_valid,
 	.mode_fixup	= armada_drm_crtc_mode_fixup,
 	.mode_set_nofb	= armada_drm_crtc_mode_set_nofb,
+	.atomic_check	= armada_drm_crtc_atomic_check,
 	.atomic_begin	= armada_drm_crtc_atomic_begin,
 	.atomic_flush	= armada_drm_crtc_atomic_flush,
 	.atomic_disable	= armada_drm_crtc_atomic_disable,
@@ -457,6 +548,13 @@ static void armada_load_cursor_argb(void __iomem *base, uint32_t *pix,
 		for (x = 0; x < width; x++, p++) {
 			uint32_t val = *p;
 
+			/*
+			 * In "ARGB888" (HWC32) mode, writing to the SRAM
+			 * requires these bits to contain:
+			 * 31:24 = alpha 23:16 = blue 15:8 = green 7:0 = red
+			 * So, it's actually ABGR8888.  This is independent
+			 * of the SWAPRB bits in DMA control register 0.
+			 */
 			val = (val & 0xff00ff00) |
 			      (val & 0x000000ff) << 16 |
 			      (val & 0x00ff0000) >> 16;
@@ -673,6 +771,14 @@ static void armada_drm_crtc_destroy(struct drm_crtc *crtc)
 	kfree(dcrtc);
 }
 
+static int armada_drm_crtc_late_register(struct drm_crtc *crtc)
+{
+	if (IS_ENABLED(CONFIG_DEBUG_FS))
+		armada_drm_crtc_debugfs_init(drm_to_armada_crtc(crtc));
+
+	return 0;
+}
+
 /* These are called under the vbl_lock. */
 static int armada_drm_crtc_enable_vblank(struct drm_crtc *crtc)
 {
@@ -700,14 +806,93 @@ static const struct drm_crtc_funcs armada_crtc_funcs = {
 	.cursor_set	= armada_drm_crtc_cursor_set,
 	.cursor_move	= armada_drm_crtc_cursor_move,
 	.destroy	= armada_drm_crtc_destroy,
+	.gamma_set	= drm_atomic_helper_legacy_gamma_set,
 	.set_config	= drm_atomic_helper_set_config,
 	.page_flip	= drm_atomic_helper_page_flip,
 	.atomic_duplicate_state = drm_atomic_helper_crtc_duplicate_state,
 	.atomic_destroy_state = drm_atomic_helper_crtc_destroy_state,
+	.late_register	= armada_drm_crtc_late_register,
 	.enable_vblank	= armada_drm_crtc_enable_vblank,
 	.disable_vblank	= armada_drm_crtc_disable_vblank,
 };
 
+int armada_crtc_select_clock(struct armada_crtc *dcrtc,
+			     struct armada_clk_result *res,
+			     const struct armada_clocking_params *params,
+			     struct clk *clks[], size_t num_clks,
+			     unsigned long desired_khz)
+{
+	unsigned long desired_hz = desired_khz * 1000;
+	unsigned long desired_clk_hz;	// requested clk input
+	unsigned long real_clk_hz;	// actual clk input
+	unsigned long real_hz;		// actual pixel clk
+	unsigned long permillage;
+	struct clk *clk;
+	u32 div;
+	int i;
+
+	DRM_DEBUG_KMS("[CRTC:%u:%s] desired clock=%luHz\n",
+		      dcrtc->crtc.base.id, dcrtc->crtc.name, desired_hz);
+
+	for (i = 0; i < num_clks; i++) {
+		clk = clks[i];
+		if (!clk)
+			continue;
+
+		if (params->settable & BIT(i)) {
+			real_clk_hz = clk_round_rate(clk, desired_hz);
+			desired_clk_hz = desired_hz;
+		} else {
+			real_clk_hz = clk_get_rate(clk);
+			desired_clk_hz = real_clk_hz;
+		}
+
+		/* If the clock can do exactly the desired rate, we're done */
+		if (real_clk_hz == desired_hz) {
+			real_hz = real_clk_hz;
+			div = 1;
+			goto found;
+		}
+
+		/* Calculate the divider - if invalid, we can't do this rate */
+		div = DIV_ROUND_CLOSEST(real_clk_hz, desired_hz);
+		if (div == 0 || div > params->div_max)
+			continue;
+
+		/* Calculate the actual rate - HDMI requires -0.6%..+0.5% */
+		real_hz = DIV_ROUND_CLOSEST(real_clk_hz, div);
+
+		DRM_DEBUG_KMS("[CRTC:%u:%s] clk=%u %luHz div=%u real=%luHz\n",
+			dcrtc->crtc.base.id, dcrtc->crtc.name,
+			i, real_clk_hz, div, real_hz);
+
+		/* Avoid repeated division */
+		if (real_hz < desired_hz) {
+			permillage = real_hz / desired_khz;
+			if (permillage < params->permillage_min)
+				continue;
+		} else {
+			permillage = DIV_ROUND_UP(real_hz, desired_khz);
+			if (permillage > params->permillage_max)
+				continue;
+		}
+		goto found;
+	}
+
+	return -ERANGE;
+
+found:
+	DRM_DEBUG_KMS("[CRTC:%u:%s] selected clk=%u %luHz div=%u real=%luHz\n",
+		dcrtc->crtc.base.id, dcrtc->crtc.name,
+		i, real_clk_hz, div, real_hz);
+
+	res->desired_clk_hz = desired_clk_hz;
+	res->clk = clk;
+	res->div = div;
+
+	return i;
+}
+
 static int armada_drm_crtc_create(struct drm_device *drm, struct device *dev,
 	struct resource *res, int irq, const struct armada_variant *variant,
 	struct device_node *port)
@@ -734,7 +919,6 @@ static int armada_drm_crtc_create(struct drm_device *drm, struct device *dev,
 	dcrtc->variant = variant;
 	dcrtc->base = base;
 	dcrtc->num = drm->mode_config.num_crtc;
-	dcrtc->clk = ERR_PTR(-EINVAL);
 	dcrtc->cfg_dumb_ctrl = DUMB24_RGB888_0;
 	dcrtc->spu_iopad_ctrl = CFG_VSCALE_LN_EN | CFG_IOPAD_DUMB24;
 	spin_lock_init(&dcrtc->irq_lock);
@@ -791,6 +975,12 @@ static int armada_drm_crtc_create(struct drm_device *drm, struct device *dev,
 
 	drm_crtc_helper_add(&dcrtc->crtc, &armada_crtc_helper_funcs);
 
+	ret = drm_mode_crtc_set_gamma_size(&dcrtc->crtc, 256);
+	if (ret)
+		return ret;
+
+	drm_crtc_enable_color_mgmt(&dcrtc->crtc, 0, false, 256);
+
 	return armada_overlay_plane_create(drm, 1 << dcrtc->num);
 
 err_crtc_init:
diff --git a/drivers/gpu/drm/armada/armada_crtc.h b/drivers/gpu/drm/armada/armada_crtc.h
index 86e1ad30ee40..b21267d1745f 100644
--- a/drivers/gpu/drm/armada/armada_crtc.h
+++ b/drivers/gpu/drm/armada/armada_crtc.h
@@ -36,10 +36,10 @@ struct armada_variant;
 struct armada_crtc {
 	struct drm_crtc		crtc;
 	const struct armada_variant *variant;
+	void			*variant_data;
 	unsigned		num;
 	void __iomem		*base;
 	struct clk		*clk;
-	struct clk		*extclk[2];
 	struct {
 		uint32_t	spu_v_h_total;
 		uint32_t	spu_v_porch;
@@ -72,6 +72,25 @@ struct armada_crtc {
 
 void armada_drm_crtc_update_regs(struct armada_crtc *, struct armada_regs *);
 
+struct armada_clocking_params {
+	unsigned long permillage_min;
+	unsigned long permillage_max;
+	u32 settable;
+	u32 div_max;
+};
+
+struct armada_clk_result {
+	unsigned long desired_clk_hz;
+	struct clk *clk;
+	u32 div;
+};
+
+int armada_crtc_select_clock(struct armada_crtc *dcrtc,
+			     struct armada_clk_result *res,
+			     const struct armada_clocking_params *params,
+			     struct clk *clks[], size_t num_clks,
+			     unsigned long desired_khz);
+
 extern struct platform_driver armada_lcd_platform_driver;
 
 #endif
diff --git a/drivers/gpu/drm/armada/armada_debugfs.c b/drivers/gpu/drm/armada/armada_debugfs.c
index 4784e5d99d53..dc3716dbb2c0 100644
--- a/drivers/gpu/drm/armada/armada_debugfs.c
+++ b/drivers/gpu/drm/armada/armada_debugfs.c
@@ -25,50 +25,33 @@ static int armada_debugfs_gem_linear_show(struct seq_file *m, void *data)
 	return 0;
 }
 
-static int armada_debugfs_reg_show(struct seq_file *m, void *data)
+static int armada_debugfs_crtc_reg_show(struct seq_file *m, void *data)
 {
-	struct drm_device *dev = m->private;
-	struct armada_private *priv = dev->dev_private;
-	int n, i;
-
-	if (priv) {
-		for (n = 0; n < ARRAY_SIZE(priv->dcrtc); n++) {
-			struct armada_crtc *dcrtc = priv->dcrtc[n];
-			if (!dcrtc)
-				continue;
-
-			for (i = 0x84; i <= 0x1c4; i += 4) {
-				uint32_t v = readl_relaxed(dcrtc->base + i);
-				seq_printf(m, "%u: 0x%04x: 0x%08x\n", n, i, v);
-			}
-		}
+	struct armada_crtc *dcrtc = m->private;
+	int i;
+
+	for (i = 0x84; i <= 0x1c4; i += 4) {
+		u32 v = readl_relaxed(dcrtc->base + i);
+		seq_printf(m, "0x%04x: 0x%08x\n", i, v);
 	}
 
 	return 0;
 }
 
-static int armada_debugfs_reg_r_open(struct inode *inode, struct file *file)
+static int armada_debugfs_crtc_reg_open(struct inode *inode, struct file *file)
 {
-	return single_open(file, armada_debugfs_reg_show, inode->i_private);
+	return single_open(file, armada_debugfs_crtc_reg_show,
+			   inode->i_private);
 }
 
-static const struct file_operations fops_reg_r = {
-	.owner = THIS_MODULE,
-	.open = armada_debugfs_reg_r_open,
-	.read = seq_read,
-	.llseek = seq_lseek,
-	.release = single_release,
-};
-
-static int armada_debugfs_write(struct file *file, const char __user *ptr,
-	size_t len, loff_t *off)
+static int armada_debugfs_crtc_reg_write(struct file *file,
+	const char __user *ptr, size_t len, loff_t *off)
 {
-	struct drm_device *dev = file->private_data;
-	struct armada_private *priv = dev->dev_private;
-	struct armada_crtc *dcrtc = priv->dcrtc[0];
-	char buf[32], *p;
-	uint32_t reg, val;
+	struct armada_crtc *dcrtc;
+	unsigned long reg, mask, val;
+	char buf[32];
 	int ret;
+	u32 v;
 
 	if (*off != 0)
 		return 0;
@@ -81,24 +64,35 @@ static int armada_debugfs_write(struct file *file, const char __user *ptr,
 		return ret;
 	buf[len] = '\0';
 
-	reg = simple_strtoul(buf, &p, 16);
-	if (!isspace(*p))
+	if (sscanf(buf, "%lx %lx %lx", &reg, &mask, &val) != 3)
 		return -EINVAL;
-	val = simple_strtoul(p + 1, NULL, 16);
+	if (reg < 0x84 || reg > 0x1c4 || reg & 3)
+		return -ERANGE;
 
-	if (reg >= 0x84 && reg <= 0x1c4)
-		writel(val, dcrtc->base + reg);
+	dcrtc = ((struct seq_file *)file->private_data)->private;
+	v = readl(dcrtc->base + reg);
+	v &= ~mask;
+	v |= val & mask;
+	writel(v, dcrtc->base + reg);
 
 	return len;
 }
 
-static const struct file_operations fops_reg_w = {
+static const struct file_operations armada_debugfs_crtc_reg_fops = {
 	.owner = THIS_MODULE,
-	.open = simple_open,
-	.write = armada_debugfs_write,
-	.llseek = noop_llseek,
+	.open = armada_debugfs_crtc_reg_open,
+	.read = seq_read,
+	.write = armada_debugfs_crtc_reg_write,
+	.llseek = seq_lseek,
+	.release = single_release,
 };
 
+void armada_drm_crtc_debugfs_init(struct armada_crtc *dcrtc)
+{
+	debugfs_create_file("armada-regs", 0600, dcrtc->crtc.debugfs_entry,
+			    dcrtc, &armada_debugfs_crtc_reg_fops);
+}
+
 static struct drm_info_list armada_debugfs_list[] = {
 	{ "gem_linear", armada_debugfs_gem_linear_show, 0 },
 };
@@ -106,24 +100,8 @@ static struct drm_info_list armada_debugfs_list[] = {
 
 int armada_drm_debugfs_init(struct drm_minor *minor)
 {
-	struct dentry *de;
-	int ret;
-
-	ret = drm_debugfs_create_files(armada_debugfs_list,
-				       ARMADA_DEBUGFS_ENTRIES,
-				       minor->debugfs_root, minor);
-	if (ret)
-		return ret;
-
-	de = debugfs_create_file("reg", S_IFREG | S_IRUSR,
-				 minor->debugfs_root, minor->dev, &fops_reg_r);
-	if (!de)
-		return -ENOMEM;
-
-	de = debugfs_create_file("reg_wr", S_IFREG | S_IWUSR,
-				 minor->debugfs_root, minor->dev, &fops_reg_w);
-	if (!de)
-		return -ENOMEM;
+	drm_debugfs_create_files(armada_debugfs_list, ARMADA_DEBUGFS_ENTRIES,
+				 minor->debugfs_root, minor);
 
 	return 0;
 }
diff --git a/drivers/gpu/drm/armada/armada_drm.h b/drivers/gpu/drm/armada/armada_drm.h
index 3e60cec4e4b0..c7794c8bdd90 100644
--- a/drivers/gpu/drm/armada/armada_drm.h
+++ b/drivers/gpu/drm/armada/armada_drm.h
@@ -75,6 +75,7 @@ void armada_fbdev_fini(struct drm_device *);
 
 int armada_overlay_plane_create(struct drm_device *, unsigned long);
 
+void armada_drm_crtc_debugfs_init(struct armada_crtc *dcrtc);
 int armada_drm_debugfs_init(struct drm_minor *);
 
 #endif
diff --git a/drivers/gpu/drm/armada/armada_drv.c b/drivers/gpu/drm/armada/armada_drv.c
index 4a6e8ed05925..521464f08ccd 100644
--- a/drivers/gpu/drm/armada/armada_drv.c
+++ b/drivers/gpu/drm/armada/armada_drv.c
@@ -97,6 +97,17 @@ static int armada_drm_bind(struct device *dev)
 		return ret;
 	}
 
+	/* Remove early framebuffers */
+	ret = drm_fb_helper_remove_conflicting_framebuffers(NULL,
+							    "armada-drm-fb",
+							    false);
+	if (ret) {
+		dev_err(dev, "[" DRM_NAME ":%s] can't kick out simple-fb: %d\n",
+			__func__, ret);
+		kfree(priv);
+		return ret;
+	}
+
 	priv->drm.dev_private = priv;
 
 	dev_set_drvdata(dev, &priv->drm);
@@ -168,6 +179,8 @@ static void armada_drm_unbind(struct device *dev)
 
 	drm_dev_unregister(&priv->drm);
 
+	drm_atomic_helper_shutdown(&priv->drm);
+
 	component_unbind_all(dev, &priv->drm);
 
 	drm_mode_config_cleanup(&priv->drm);
@@ -188,23 +201,15 @@ static int compare_dev_name(struct device *dev, void *data)
 }
 
 static void armada_add_endpoints(struct device *dev,
-	struct component_match **match, struct device_node *port)
+	struct component_match **match, struct device_node *dev_node)
 {
 	struct device_node *ep, *remote;
 
-	for_each_child_of_node(port, ep) {
+	for_each_endpoint_of_node(dev_node, ep) {
 		remote = of_graph_get_remote_port_parent(ep);
-		if (!remote || !of_device_is_available(remote)) {
-			of_node_put(remote);
-			continue;
-		} else if (!of_device_is_available(remote->parent)) {
-			dev_warn(dev, "parent device of %pOF is not available\n",
-				 remote);
-			of_node_put(remote);
-			continue;
-		}
-
-		drm_of_component_match_add(dev, match, compare_of, remote);
+		if (remote && of_device_is_available(remote))
+			drm_of_component_match_add(dev, match, compare_of,
+						   remote);
 		of_node_put(remote);
 	}
 }
@@ -226,7 +231,6 @@ static int armada_drm_probe(struct platform_device *pdev)
 
 	if (dev->platform_data) {
 		char **devices = dev->platform_data;
-		struct device_node *port;
 		struct device *d;
 		int i;
 
@@ -242,10 +246,8 @@ static int armada_drm_probe(struct platform_device *pdev)
 		for (i = 0; devices[i]; i++) {
 			d = bus_find_device_by_name(&platform_bus_type, NULL,
 						    devices[i]);
-			if (d && d->of_node) {
-				for_each_child_of_node(d->of_node, port)
-					armada_add_endpoints(dev, &match, port);
-			}
+			if (d && d->of_node)
+				armada_add_endpoints(dev, &match, d->of_node);
 			put_device(d);
 		}
 	}
diff --git a/drivers/gpu/drm/armada/armada_fb.c b/drivers/gpu/drm/armada/armada_fb.c
index 2029d5f3c0a2..de030cb0aa90 100644
--- a/drivers/gpu/drm/armada/armada_fb.c
+++ b/drivers/gpu/drm/armada/armada_fb.c
@@ -84,6 +84,7 @@ struct armada_framebuffer *armada_framebuffer_create(struct drm_device *dev,
 struct drm_framebuffer *armada_fb_create(struct drm_device *dev,
 	struct drm_file *dfile, const struct drm_mode_fb_cmd2 *mode)
 {
+	const struct drm_format_info *info = drm_get_format_info(dev, mode);
 	struct armada_gem_object *obj;
 	struct armada_framebuffer *dfb;
 	int ret;
@@ -94,7 +95,7 @@ struct drm_framebuffer *armada_fb_create(struct drm_device *dev,
 		mode->pitches[2]);
 
 	/* We can only handle a single plane at the moment */
-	if (drm_format_num_planes(mode->pixel_format) > 1 &&
+	if (info->num_planes > 1 &&
 	    (mode->handles[0] != mode->handles[1] ||
 	     mode->handles[0] != mode->handles[2])) {
 		ret = -EINVAL;
diff --git a/drivers/gpu/drm/armada/armada_hw.h b/drivers/gpu/drm/armada/armada_hw.h
index 85b998e8ac7a..9c88b38a4003 100644
--- a/drivers/gpu/drm/armada/armada_hw.h
+++ b/drivers/gpu/drm/armada/armada_hw.h
@@ -85,6 +85,16 @@ enum {
 	ADV_VSYNC_H_OFF	= 0xfff << 0,
 };
 
+/* LCD_CFG_RDREG4F - Armada 510 only */
+enum {
+	CFG_SRAM_WAIT	= BIT(11),
+	CFG_SMPN_FASTTX	= BIT(10),
+	CFG_DMA_ARB	= BIT(9),
+	CFG_DMA_WM_EN	= BIT(8),
+	CFG_DMA_WM_MASK	= 0xff,
+#define CFG_DMA_WM(x)	((x) & CFG_DMA_WM_MASK)
+};
+
 enum {
 	CFG_565		= 0,
 	CFG_1555	= 1,
@@ -166,6 +176,10 @@ enum {
 	SRAM_READ	= 0 << 14,
 	SRAM_WRITE	= 2 << 14,
 	SRAM_INIT	= 3 << 14,
+	SRAM_GAMMA_YR	= 0x0 << 8,
+	SRAM_GAMMA_UG	= 0x1 << 8,
+	SRAM_GAMMA_VB	= 0x2 << 8,
+	SRAM_PALETTE	= 0x3 << 8,
 	SRAM_HWC32_RAM1	= 0xc << 8,
 	SRAM_HWC32_RAM2	= 0xd << 8,
 	SRAM_HWC32_RAMR	= SRAM_HWC32_RAM1,
@@ -313,19 +327,4 @@ enum {
 	PWRDN_IRQ_LEVEL		= 1 << 0,
 };
 
-static inline u32 armada_rect_hw_fp(struct drm_rect *r)
-{
-	return (drm_rect_height(r) & 0xffff0000) | drm_rect_width(r) >> 16;
-}
-
-static inline u32 armada_rect_hw(struct drm_rect *r)
-{
-	return drm_rect_height(r) << 16 | (drm_rect_width(r) & 0x0000ffff);
-}
-
-static inline u32 armada_rect_yx(struct drm_rect *r)
-{
-	return (r)->y1 << 16 | ((r)->x1 & 0x0000ffff);
-}
-
 #endif
diff --git a/drivers/gpu/drm/armada/armada_overlay.c b/drivers/gpu/drm/armada/armada_overlay.c
index 7857d86a6736..e8060216b389 100644
--- a/drivers/gpu/drm/armada/armada_overlay.c
+++ b/drivers/gpu/drm/armada/armada_overlay.c
@@ -24,7 +24,7 @@
 #define DEFAULT_ENCODING	DRM_COLOR_YCBCR_BT601
 
 struct armada_overlay_state {
-	struct drm_plane_state base;
+	struct armada_plane_state base;
 	u32 colorkey_yr;
 	u32 colorkey_ug;
 	u32 colorkey_vb;
@@ -35,7 +35,7 @@ struct armada_overlay_state {
 	u16 saturation;
 };
 #define drm_to_overlay_state(s) \
-	container_of(s, struct armada_overlay_state, base)
+	container_of(s, struct armada_overlay_state, base.base)
 
 static inline u32 armada_spu_contrast(struct drm_plane_state *state)
 {
@@ -91,41 +91,39 @@ static void armada_drm_overlay_plane_atomic_update(struct drm_plane *plane,
 		armada_reg_queue_mod(regs, idx,
 				     0, CFG_PDWN16x66 | CFG_PDWN32x66,
 				     LCD_SPU_SRAM_PARA1);
-	val = armada_rect_hw_fp(&state->src);
-	if (armada_rect_hw_fp(&old_state->src) != val)
+	val = armada_src_hw(state);
+	if (armada_src_hw(old_state) != val)
 		armada_reg_queue_set(regs, idx, val, LCD_SPU_DMA_HPXL_VLN);
-	val = armada_rect_yx(&state->dst);
-	if (armada_rect_yx(&old_state->dst) != val)
+	val = armada_dst_yx(state);
+	if (armada_dst_yx(old_state) != val)
 		armada_reg_queue_set(regs, idx, val, LCD_SPU_DMA_OVSA_HPXL_VLN);
-	val = armada_rect_hw(&state->dst);
-	if (armada_rect_hw(&old_state->dst) != val)
+	val = armada_dst_hw(state);
+	if (armada_dst_hw(old_state) != val)
 		armada_reg_queue_set(regs, idx, val, LCD_SPU_DZM_HPXL_VLN);
 	/* FIXME: overlay on an interlaced display */
 	if (old_state->src.x1 != state->src.x1 ||
 	    old_state->src.y1 != state->src.y1 ||
-	    old_state->fb != state->fb) {
+	    old_state->fb != state->fb ||
+	    state->crtc->state->mode_changed) {
 		const struct drm_format_info *format;
-		u16 src_x, pitches[3];
-		u32 addrs[2][3];
+		u16 src_x;
 
-		armada_drm_plane_calc(state, addrs, pitches, false);
-
-		armada_reg_queue_set(regs, idx, addrs[0][0],
+		armada_reg_queue_set(regs, idx, armada_addr(state, 0, 0),
 				     LCD_SPU_DMA_START_ADDR_Y0);
-		armada_reg_queue_set(regs, idx, addrs[0][1],
+		armada_reg_queue_set(regs, idx, armada_addr(state, 0, 1),
 				     LCD_SPU_DMA_START_ADDR_U0);
-		armada_reg_queue_set(regs, idx, addrs[0][2],
+		armada_reg_queue_set(regs, idx, armada_addr(state, 0, 2),
 				     LCD_SPU_DMA_START_ADDR_V0);
-		armada_reg_queue_set(regs, idx, addrs[1][0],
+		armada_reg_queue_set(regs, idx, armada_addr(state, 1, 0),
 				     LCD_SPU_DMA_START_ADDR_Y1);
-		armada_reg_queue_set(regs, idx, addrs[1][1],
+		armada_reg_queue_set(regs, idx, armada_addr(state, 1, 1),
 				     LCD_SPU_DMA_START_ADDR_U1);
-		armada_reg_queue_set(regs, idx, addrs[1][2],
+		armada_reg_queue_set(regs, idx, armada_addr(state, 1, 2),
 				     LCD_SPU_DMA_START_ADDR_V1);
 
-		val = pitches[0] << 16 | pitches[0];
+		val = armada_pitch(state, 0) << 16 | armada_pitch(state, 0);
 		armada_reg_queue_set(regs, idx, val, LCD_SPU_DMA_PITCH_YC);
-		val = pitches[1] << 16 | pitches[2];
+		val = armada_pitch(state, 1) << 16 | armada_pitch(state, 2);
 		armada_reg_queue_set(regs, idx, val, LCD_SPU_DMA_PITCH_UV);
 
 		cfg = CFG_DMA_FMT(drm_fb_to_armada_fb(state->fb)->fmt) |
@@ -143,6 +141,8 @@ static void armada_drm_overlay_plane_atomic_update(struct drm_plane *plane,
 		src_x = state->src.x1 >> 16;
 		if (format->num_planes == 1 && src_x & (format->hsub - 1))
 			cfg ^= CFG_DMA_MOD(CFG_SWAPUV);
+		if (to_armada_plane_state(state)->interlace)
+			cfg |= CFG_DMA_FTOGGLE;
 		cfg_mask = CFG_CBSH_ENA | CFG_DMAFORMAT |
 			   CFG_DMA_MOD(CFG_SWAPRB | CFG_SWAPUV |
 				       CFG_SWAPYU | CFG_YUV2RGB) |
@@ -304,13 +304,10 @@ static void armada_overlay_reset(struct drm_plane *plane)
 	if (plane->state)
 		__drm_atomic_helper_plane_destroy_state(plane->state);
 	kfree(plane->state);
+	plane->state = NULL;
 
 	state = kzalloc(sizeof(*state), GFP_KERNEL);
 	if (state) {
-		state->base.plane = plane;
-		state->base.color_encoding = DEFAULT_ENCODING;
-		state->base.color_range = DRM_COLOR_YCBCR_LIMITED_RANGE;
-		state->base.rotation = DRM_MODE_ROTATE_0;
 		state->colorkey_yr = 0xfefefe00;
 		state->colorkey_ug = 0x01010100;
 		state->colorkey_vb = 0x01010100;
@@ -320,8 +317,10 @@ static void armada_overlay_reset(struct drm_plane *plane)
 		state->brightness = DEFAULT_BRIGHTNESS;
 		state->contrast = DEFAULT_CONTRAST;
 		state->saturation = DEFAULT_SATURATION;
+		__drm_atomic_helper_plane_reset(plane, &state->base.base);
+		state->base.base.color_encoding = DEFAULT_ENCODING;
+		state->base.base.color_range = DRM_COLOR_YCBCR_LIMITED_RANGE;
 	}
-	plane->state = &state->base;
 }
 
 struct drm_plane_state *
@@ -334,8 +333,9 @@ armada_overlay_duplicate_state(struct drm_plane *plane)
 
 	state = kmemdup(plane->state, sizeof(*state), GFP_KERNEL);
 	if (state)
-		__drm_atomic_helper_plane_duplicate_state(plane, &state->base);
-	return &state->base;
+		__drm_atomic_helper_plane_duplicate_state(plane,
+							  &state->base.base);
+	return &state->base.base;
 }
 
 static int armada_overlay_set_property(struct drm_plane *plane,
diff --git a/drivers/gpu/drm/armada/armada_plane.c b/drivers/gpu/drm/armada/armada_plane.c
index 09d9400edd7b..f08b4f37816d 100644
--- a/drivers/gpu/drm/armada/armada_plane.c
+++ b/drivers/gpu/drm/armada/armada_plane.c
@@ -76,23 +76,6 @@ void armada_drm_plane_calc(struct drm_plane_state *state, u32 addrs[2][3],
 	}
 }
 
-static unsigned armada_drm_crtc_calc_fb(struct drm_plane_state *state,
-	struct armada_regs *regs, bool interlaced)
-{
-	u16 pitches[3];
-	u32 addrs[2][3];
-	unsigned i = 0;
-
-	armada_drm_plane_calc(state, addrs, pitches, interlaced);
-
-	/* write offset, base, and pitch */
-	armada_reg_queue_set(regs, i, addrs[0][0], LCD_CFG_GRA_START_ADDR0);
-	armada_reg_queue_set(regs, i, addrs[1][0], LCD_CFG_GRA_START_ADDR1);
-	armada_reg_queue_mod(regs, i, pitches[0], 0xffff, LCD_CFG_GRA_PITCH);
-
-	return i;
-}
-
 int armada_drm_plane_prepare_fb(struct drm_plane *plane,
 	struct drm_plane_state *state)
 {
@@ -123,20 +106,50 @@ void armada_drm_plane_cleanup_fb(struct drm_plane *plane,
 int armada_drm_plane_atomic_check(struct drm_plane *plane,
 	struct drm_plane_state *state)
 {
-	if (state->fb && !WARN_ON(!state->crtc)) {
-		struct drm_crtc *crtc = state->crtc;
-		struct drm_crtc_state *crtc_state;
-
-		if (state->state)
-			crtc_state = drm_atomic_get_existing_crtc_state(state->state, crtc);
-		else
-			crtc_state = crtc->state;
-		return drm_atomic_helper_check_plane_state(state, crtc_state,
-							   0, INT_MAX,
-							   true, false);
-	} else {
+	struct armada_plane_state *st = to_armada_plane_state(state);
+	struct drm_crtc *crtc = state->crtc;
+	struct drm_crtc_state *crtc_state;
+	bool interlace;
+	int ret;
+
+	if (!state->fb || WARN_ON(!state->crtc)) {
 		state->visible = false;
+		return 0;
+	}
+
+	if (state->state)
+		crtc_state = drm_atomic_get_existing_crtc_state(state->state, crtc);
+	else
+		crtc_state = crtc->state;
+
+	ret = drm_atomic_helper_check_plane_state(state, crtc_state, 0,
+						  INT_MAX, true, false);
+	if (ret)
+		return ret;
+
+	interlace = crtc_state->adjusted_mode.flags & DRM_MODE_FLAG_INTERLACE;
+	if (interlace) {
+		if ((state->dst.y1 | state->dst.y2) & 1)
+			return -EINVAL;
+		st->src_hw = drm_rect_height(&state->src) >> 17;
+		st->dst_yx = state->dst.y1 >> 1;
+		st->dst_hw = drm_rect_height(&state->dst) >> 1;
+	} else {
+		st->src_hw = drm_rect_height(&state->src) >> 16;
+		st->dst_yx = state->dst.y1;
+		st->dst_hw = drm_rect_height(&state->dst);
 	}
+
+	st->src_hw <<= 16;
+	st->src_hw |= drm_rect_width(&state->src) >> 16;
+	st->dst_yx <<= 16;
+	st->dst_yx |= state->dst.x1 & 0x0000ffff;
+	st->dst_hw <<= 16;
+	st->dst_hw |= drm_rect_width(&state->dst) & 0x0000ffff;
+
+	armada_drm_plane_calc(state, st->addrs, st->pitches, interlace);
+	st->interlace = interlace;
+
 	return 0;
 }
 
@@ -170,21 +183,25 @@ static void armada_drm_primary_plane_atomic_update(struct drm_plane *plane,
 			val |= CFG_PDWN256x24;
 		armada_reg_queue_mod(regs, idx, 0, val, LCD_SPU_SRAM_PARA1);
 	}
-	val = armada_rect_hw_fp(&state->src);
-	if (armada_rect_hw_fp(&old_state->src) != val)
+	val = armada_src_hw(state);
+	if (armada_src_hw(old_state) != val)
 		armada_reg_queue_set(regs, idx, val, LCD_SPU_GRA_HPXL_VLN);
-	val = armada_rect_yx(&state->dst);
-	if (armada_rect_yx(&old_state->dst) != val)
+	val = armada_dst_yx(state);
+	if (armada_dst_yx(old_state) != val)
 		armada_reg_queue_set(regs, idx, val, LCD_SPU_GRA_OVSA_HPXL_VLN);
-	val = armada_rect_hw(&state->dst);
-	if (armada_rect_hw(&old_state->dst) != val)
+	val = armada_dst_hw(state);
+	if (armada_dst_hw(old_state) != val)
 		armada_reg_queue_set(regs, idx, val, LCD_SPU_GZM_HPXL_VLN);
 	if (old_state->src.x1 != state->src.x1 ||
 	    old_state->src.y1 != state->src.y1 ||
 	    old_state->fb != state->fb ||
 	    state->crtc->state->mode_changed) {
-		idx += armada_drm_crtc_calc_fb(state, regs + idx,
-					       dcrtc->interlaced);
+		armada_reg_queue_set(regs, idx, armada_addr(state, 0, 0),
+				     LCD_CFG_GRA_START_ADDR0);
+		armada_reg_queue_set(regs, idx, armada_addr(state, 1, 0),
+				     LCD_CFG_GRA_START_ADDR1);
+		armada_reg_queue_mod(regs, idx, armada_pitch(state, 0), 0xffff,
+				     LCD_CFG_GRA_PITCH);
 	}
 	if (old_state->fb != state->fb ||
 	    state->crtc->state->mode_changed) {
@@ -194,7 +211,7 @@ static void armada_drm_primary_plane_atomic_update(struct drm_plane *plane,
 			cfg |= CFG_PALETTE_ENA;
 		if (state->visible)
 			cfg |= CFG_GRA_ENA;
-		if (dcrtc->interlaced)
+		if (to_armada_plane_state(state)->interlace)
 			cfg |= CFG_GRA_FTOGGLE;
 		cfg_mask = CFG_GRAFORMAT |
 			   CFG_GRA_MOD(CFG_SWAPRB | CFG_SWAPUV |
@@ -245,7 +262,7 @@ static void armada_drm_primary_plane_atomic_disable(struct drm_plane *plane,
 	/* Disable plane and power down most RAMs and FIFOs */
 	armada_reg_queue_mod(regs, idx, 0, CFG_GRA_ENA, LCD_SPU_DMA_CTRL0);
 	armada_reg_queue_mod(regs, idx, CFG_PDWN256x32 | CFG_PDWN256x24 |
-			     CFG_PDWN256x8 | CFG_PDWN32x32 | CFG_PDWN64x66,
+			     CFG_PDWN32x32 | CFG_PDWN64x66,
 			     0, LCD_SPU_SRAM_PARA1);
 
 	dcrtc->regs_idx += idx;
@@ -259,12 +276,37 @@ static const struct drm_plane_helper_funcs armada_primary_plane_helper_funcs = {
 	.atomic_disable	= armada_drm_primary_plane_atomic_disable,
 };
 
+void armada_plane_reset(struct drm_plane *plane)
+{
+	struct armada_plane_state *st;
+	if (plane->state)
+		__drm_atomic_helper_plane_destroy_state(plane->state);
+	kfree(plane->state);
+	st = kzalloc(sizeof(*st), GFP_KERNEL);
+	if (st)
+		__drm_atomic_helper_plane_reset(plane, &st->base);
+}
+
+struct drm_plane_state *armada_plane_duplicate_state(struct drm_plane *plane)
+{
+	struct armada_plane_state *st;
+
+	if (WARN_ON(!plane->state))
+		return NULL;
+
+	st = kmemdup(plane->state, sizeof(*st), GFP_KERNEL);
+	if (st)
+		__drm_atomic_helper_plane_duplicate_state(plane, &st->base);
+
+	return &st->base;
+}
+
 static const struct drm_plane_funcs armada_primary_plane_funcs = {
 	.update_plane	= drm_atomic_helper_update_plane,
 	.disable_plane	= drm_atomic_helper_disable_plane,
 	.destroy	= drm_primary_helper_destroy,
-	.reset		= drm_atomic_helper_plane_reset,
-	.atomic_duplicate_state = drm_atomic_helper_plane_duplicate_state,
+	.reset		= armada_plane_reset,
+	.atomic_duplicate_state = armada_plane_duplicate_state,
 	.atomic_destroy_state = drm_atomic_helper_plane_destroy_state,
 };
 
diff --git a/drivers/gpu/drm/armada/armada_plane.h b/drivers/gpu/drm/armada/armada_plane.h
index ff4281ba7fad..2707ec781941 100644
--- a/drivers/gpu/drm/armada/armada_plane.h
+++ b/drivers/gpu/drm/armada/armada_plane.h
@@ -1,6 +1,24 @@
 #ifndef ARMADA_PLANE_H
 #define ARMADA_PLANE_H
 
+struct armada_plane_state {
+	struct drm_plane_state base;
+	u32 src_hw;
+	u32 dst_yx;
+	u32 dst_hw;
+	u32 addrs[2][3];
+	u16 pitches[3];
+	bool interlace;
+};
+
+#define to_armada_plane_state(st) \
+	container_of(st, struct armada_plane_state, base)
+#define armada_src_hw(state) to_armada_plane_state(state)->src_hw
+#define armada_dst_yx(state) to_armada_plane_state(state)->dst_yx
+#define armada_dst_hw(state) to_armada_plane_state(state)->dst_hw
+#define armada_addr(state, f, p) to_armada_plane_state(state)->addrs[f][p]
+#define armada_pitch(state, n) to_armada_plane_state(state)->pitches[n]
+
 void armada_drm_plane_calc(struct drm_plane_state *state, u32 addrs[2][3],
 	u16 pitches[3], bool interlaced);
 int armada_drm_plane_prepare_fb(struct drm_plane *plane,
@@ -9,6 +27,11 @@ void armada_drm_plane_cleanup_fb(struct drm_plane *plane,
 	struct drm_plane_state *old_state);
 int armada_drm_plane_atomic_check(struct drm_plane *plane,
 	struct drm_plane_state *state);
+void armada_plane_reset(struct drm_plane *plane);
+struct drm_plane_state *armada_plane_duplicate_state(struct drm_plane *plane);
+void armada_plane_destroy_state(struct drm_plane *plane,
+				struct drm_plane_state *state);
+
 int armada_drm_primary_plane_init(struct drm_device *drm,
 	struct drm_plane *primary);