diff --git a/msm/sde/sde_connector.c b/msm/sde/sde_connector.c
index dc519fa9b2..90f551309d 100644
--- a/msm/sde/sde_connector.c
+++ b/msm/sde/sde_connector.c
@@ -2592,6 +2592,10 @@ static int sde_connector_populate_mode_info(struct drm_connector *conn,
 			continue;
 		}
 
+		sde_kms_info_add_keyint(info, "has_cwb_crop", sde_kms->catalog->has_cwb_crop);
+		sde_kms_info_add_keyint(info, "has_dedicated_cwb_support",
+			sde_kms->catalog->has_dedicated_cwb_support);
+
 		sde_kms_info_add_keyint(info, "mdp_transfer_time_us",
 			mode_info.mdp_transfer_time_us);
 
diff --git a/msm/sde/sde_encoder_phys_wb.c b/msm/sde/sde_encoder_phys_wb.c
index 6f4b15b105..e49ad684aa 100644
--- a/msm/sde/sde_encoder_phys_wb.c
+++ b/msm/sde/sde_encoder_phys_wb.c
@@ -314,6 +314,8 @@ static void sde_encoder_phys_wb_setup_fb(struct sde_encoder_phys *phys_enc,
 	struct sde_hw_wb_cfg *wb_cfg;
 	struct sde_hw_wb_cdp_cfg *cdp_cfg;
 	const struct msm_format *format;
+	struct sde_crtc_state *cstate;
+	struct sde_rect pu_roi = {0,};
 	int ret;
 	struct msm_gem_address_space *aspace;
 	u32 fb_mode;
@@ -324,6 +326,8 @@ static void sde_encoder_phys_wb_setup_fb(struct sde_encoder_phys *phys_enc,
 		return;
 	}
 
+	cstate = to_sde_crtc_state(wb_enc->crtc->state);
+
 	hw_wb = wb_enc->hw_wb;
 	wb_cfg = &wb_enc->wb_cfg;
 	cdp_cfg = &wb_enc->cdp_cfg;
@@ -373,22 +377,33 @@ static void sde_encoder_phys_wb_setup_fb(struct sde_encoder_phys *phys_enc,
 	}
 	wb_cfg->roi = *wb_roi;
 
-	if (hw_wb->caps->features & BIT(SDE_WB_XY_ROI_OFFSET)) {
-		ret = sde_format_populate_layout(aspace, fb, &wb_cfg->dest);
-		if (ret) {
-			SDE_DEBUG("failed to populate layout %d\n", ret);
-			return;
-		}
-		wb_cfg->dest.width = fb->width;
-		wb_cfg->dest.height = fb->height;
-		wb_cfg->dest.num_planes = wb_cfg->dest.format->num_planes;
-	} else {
-		ret = sde_format_populate_layout_with_roi(aspace, fb, wb_roi,
-			&wb_cfg->dest);
-		if (ret) {
-			/* this error should be detected during atomic_check */
-			SDE_DEBUG("failed to populate layout %d\n", ret);
-			return;
+	ret = sde_format_populate_layout(aspace, fb, &wb_cfg->dest);
+	if (ret) {
+		SDE_DEBUG("failed to populate layout %d\n", ret);
+		return;
+	}
+	wb_cfg->dest.width = fb->width;
+	wb_cfg->dest.height = fb->height;
+	wb_cfg->dest.num_planes = wb_cfg->dest.format->num_planes;
+
+	if (hw_wb->ops.setup_crop) {
+		wb_cfg->crop.x = wb_cfg->roi.x;
+		wb_cfg->crop.y = wb_cfg->roi.y;
+
+		if (cstate->user_roi_list.num_rects) {
+			sde_kms_rect_merge_rectangles(&cstate->user_roi_list, &pu_roi);
+
+			if ((wb_cfg->roi.w != pu_roi.w) || (wb_cfg->roi.h != pu_roi.h)) {
+				/* offset cropping region to PU region */
+				wb_cfg->crop.x = wb_cfg->crop.x - pu_roi.x;
+				wb_cfg->crop.y = wb_cfg->crop.y - pu_roi.y;
+				hw_wb->ops.setup_crop(hw_wb, wb_cfg, true);
+			}
+		} else if ((wb_cfg->roi.w != wb_cfg->dest.width) ||
+				(wb_cfg->roi.h != wb_cfg->dest.height)) {
+			hw_wb->ops.setup_crop(hw_wb, wb_cfg, true);
+		} else {
+			hw_wb->ops.setup_crop(hw_wb, wb_cfg, false);
 		}
 	}
 
@@ -467,8 +482,8 @@ static void _sde_encoder_phys_wb_setup_cwb(struct sde_encoder_phys *phys_enc,
 
 	hw_ctl = crtc->mixers[0].hw_ctl;
 	if (hw_ctl && hw_ctl->ops.setup_intf_cfg_v1 &&
-			test_bit(SDE_WB_CWB_CTRL | SDE_WB_DCWB_CTRL,
-					&hw_wb->caps->features)) {
+			(test_bit(SDE_WB_CWB_CTRL, &hw_wb->caps->features) ||
+			test_bit(SDE_WB_DCWB_CTRL, &hw_wb->caps->features))) {
 		struct sde_hw_intf_cfg_v1 intf_cfg = { 0, };
 
 		for (i = 0; i < crtc->num_mixers; i++)
@@ -621,6 +636,7 @@ static int _sde_enc_phys_wb_validate_cwb(struct sde_encoder_phys *phys_enc,
 	struct sde_rect pu_roi = {0,};
 	int out_width = 0, out_height = 0;
 	int ds_srcw = 0, ds_srch = 0, ds_outw = 0, ds_outh = 0;
+	const struct sde_format *fmt;
 	int data_pt;
 	int ds_in_use = false;
 	int i = 0;
@@ -632,6 +648,12 @@ static int _sde_enc_phys_wb_validate_cwb(struct sde_encoder_phys *phys_enc,
 		return 0;
 	}
 
+	fmt = sde_get_sde_format_ext(fb->format->format, fb->modifier);
+	if (!fmt) {
+		SDE_ERROR("unsupported output pixel format:%x\n", fb->format->format);
+		return -EINVAL;
+	}
+
 	ret = sde_wb_connector_state_get_output_roi(conn_state, &wb_roi);
 	if (ret) {
 		SDE_ERROR("failed to get roi %d\n", ret);
@@ -672,6 +694,10 @@ static int _sde_enc_phys_wb_validate_cwb(struct sde_encoder_phys *phys_enc,
 	 *	2.2) DSPP Tap point: same as No DS case
 	 *		a) wb-roi should be inside FB
 	 *		b) mode resolution & wb-roi should be same
+	 * 3) Partial Update case: additional stride check
+	 *      a) cwb roi should be inside PU region or FB
+	 *      b) cropping is only allowed for fully sampled data
+	 *      c) add check for stride and QOS setting by 256B
 	 */
 	if (ds_in_use && data_pt == CAPTURE_DSPP_OUT) {
 		out_width = ds_outw;
@@ -684,9 +710,23 @@ static int _sde_enc_phys_wb_validate_cwb(struct sde_encoder_phys *phys_enc,
 		out_height = mode->vdisplay;
 	}
 
-	if ((wb_roi.w != out_width) || (wb_roi.h != out_height)) {
+	if (SDE_FORMAT_IS_YUV(fmt) && ((wb_roi.w != out_width) || (wb_roi.h != out_height))) {
+		SDE_ERROR("invalid wb roi[%dx%d] with ds_use:%d out[%dx%d] fmt:%x\n",
+				wb_roi.w, wb_roi.h, ds_in_use, out_width, out_height,
+				fmt->base.pixel_format);
+		return -EINVAL;
+	}
+
+	if ((wb_roi.w > out_width) || (wb_roi.h > out_height)) {
 		SDE_ERROR("invalid wb roi[%dx%d] with ds_use:%d out[%dx%d]\n",
-				wb_roi.w, wb_roi.h, out_width, out_height);
+				wb_roi.w, wb_roi.h, ds_in_use, out_width, out_height);
+		return -EINVAL;
+	}
+
+	if (((wb_roi.w < out_width) || (wb_roi.h < out_height)) &&
+			(wb_roi.w * wb_roi.h * fmt->bpp) % 256) {
+		SDE_ERROR("invalid stride w = %d h = %d bpp =%d out_width = %d, out_height = %d\n",
+				wb_roi.w, wb_roi.h, fmt->bpp, out_width, out_height);
 		return -EINVAL;
 	}
 
@@ -701,7 +741,7 @@ static int _sde_enc_phys_wb_validate_cwb(struct sde_encoder_phys *phys_enc,
 	/* validate wb roi against pu rect */
 	if (cstate->user_roi_list.num_rects) {
 		sde_kms_rect_merge_rectangles(&cstate->user_roi_list, &pu_roi);
-		if (wb_roi.w != pu_roi.w || wb_roi.h != pu_roi.h) {
+		if (wb_roi.w > pu_roi.w || wb_roi.h > pu_roi.h) {
 			SDE_ERROR("invalid wb roi with pu [%dx%d vs %dx%d]\n",
 					wb_roi.w, wb_roi.h, pu_roi.w, pu_roi.h);
 			return -EINVAL;
@@ -927,8 +967,8 @@ static void _sde_encoder_phys_wb_update_cwb_flush(
 		hw_ctl->ops.update_bitmask(hw_ctl, SDE_HW_FLUSH_CDM,
 				hw_cdm->idx, 1);
 
-	if (test_bit(SDE_WB_CWB_CTRL | SDE_WB_DCWB_CTRL,
-					&hw_wb->caps->features)) {
+	if (test_bit(SDE_WB_CWB_CTRL, &hw_wb->caps->features) ||
+			test_bit(SDE_WB_DCWB_CTRL, &hw_wb->caps->features)) {
 		for (i = 0; i < crtc->num_mixers; i++) {
 			src_pp_idx = (enum sde_cwb) (src_pp_idx + i);
 
diff --git a/msm/sde/sde_formats.c b/msm/sde/sde_formats.c
index 56b1e46f43..9f4e82325d 100644
--- a/msm/sde/sde_formats.c
+++ b/msm/sde/sde_formats.c
@@ -1146,69 +1146,6 @@ int sde_format_populate_layout(
 	return ret;
 }
 
-static void _sde_format_calc_offset_linear(struct sde_hw_fmt_layout *source,
-		u32 x, u32 y)
-{
-	if ((x == 0) && (y == 0))
-		return;
-
-	source->plane_addr[0] += y * source->plane_pitch[0];
-
-	if (source->num_planes == 1) {
-		source->plane_addr[0] += x * source->format->bpp;
-	} else {
-		uint32_t xoff, yoff;
-		uint32_t v_subsample = 1;
-		uint32_t h_subsample = 1;
-
-		_sde_get_v_h_subsample_rate(source->format->chroma_sample,
-				&v_subsample, &h_subsample);
-
-		xoff = x / h_subsample;
-		yoff = y / v_subsample;
-
-		source->plane_addr[0] += x;
-		source->plane_addr[1] += xoff +
-				(yoff * source->plane_pitch[1]);
-		if (source->num_planes == 2) /* pseudo planar */
-			source->plane_addr[1] += xoff;
-		else /* planar */
-			source->plane_addr[2] += xoff +
-				(yoff * source->plane_pitch[2]);
-	}
-}
-
-int sde_format_populate_layout_with_roi(
-		struct msm_gem_address_space *aspace,
-		struct drm_framebuffer *fb,
-		struct sde_rect *roi,
-		struct sde_hw_fmt_layout *layout)
-{
-	int ret;
-
-	ret = sde_format_populate_layout(aspace, fb, layout);
-	if (ret || !roi)
-		return ret;
-
-	if (!roi->w || !roi->h || (roi->x + roi->w > fb->width) ||
-			(roi->y + roi->h > fb->height)) {
-		DRM_ERROR("invalid roi=[%d,%d,%d,%d], fb=[%u,%u]\n",
-				roi->x, roi->y, roi->w, roi->h,
-				fb->width, fb->height);
-		ret = -EINVAL;
-	} else if (SDE_FORMAT_IS_LINEAR(layout->format)) {
-		_sde_format_calc_offset_linear(layout, roi->x, roi->y);
-		layout->width = roi->w;
-		layout->height = roi->h;
-	} else if (roi->x || roi->y || (roi->w != fb->width) ||
-			(roi->h != fb->height)) {
-		DRM_ERROR("non-linear layout with roi not supported\n");
-		ret = -EINVAL;
-	}
-
-	return ret;
-}
-
 int sde_format_check_modified_format(
 		const struct msm_kms *kms,
 		const struct msm_format *msm_fmt,
diff --git a/msm/sde/sde_formats.h b/msm/sde/sde_formats.h
index 3c47172b87..ddb907dbc9 100644
--- a/msm/sde/sde_formats.h
+++ b/msm/sde/sde_formats.h
@@ -1,6 +1,6 @@
 /* SPDX-License-Identifier: GPL-2.0-only */
 /*
- * Copyright (c) 2015-2019, The Linux Foundation. All rights reserved.
+ * Copyright (c) 2015-2019, 2021, The Linux Foundation. All rights reserved.
  */
 
 #ifndef _SDE_FORMATS_H
@@ -108,22 +108,6 @@ int sde_format_populate_layout(
 		struct drm_framebuffer *fb,
 		struct sde_hw_fmt_layout *fmtl);
 
-/**
- * sde_format_populate_layout_with_roi - populate the given format layout
- *                     based on mmu, fb, roi, and format found in the fb
- * @aspace:            address space pointer
- * @fb:                framebuffer pointer
- * @roi:               region of interest (optional)
- * @fmtl:              format layout structure to populate
- *
- * Return: error code on failure, 0 on success
- */
-int sde_format_populate_layout_with_roi(
-		struct msm_gem_address_space *aspace,
-		struct drm_framebuffer *fb,
-		struct sde_rect *roi,
-		struct sde_hw_fmt_layout *fmtl);
-
 /**
  * sde_format_get_framebuffer_size - get framebuffer memory size
  * @format:            DRM pixel format
diff --git a/msm/sde/sde_hw_catalog.c b/msm/sde/sde_hw_catalog.c
index f4da11486e..fc4d64b2e4 100644
--- a/msm/sde/sde_hw_catalog.c
+++ b/msm/sde/sde_hw_catalog.c
@@ -2530,6 +2530,9 @@ static int sde_wb_parse_dt(struct device_node *np, struct sde_mdss_cfg *sde_cfg)
 		if (sde_cfg->has_wb_ubwc)
 			set_bit(SDE_WB_UBWC, &wb->features);
 
+		if (sde_cfg->has_cwb_crop)
+			set_bit(SDE_WB_CROP, &wb->features);
+
 		set_bit(SDE_WB_XY_ROI_OFFSET, &wb->features);
 
 		if (IS_SDE_CTL_REV_100(sde_cfg->ctl_rev))
@@ -5038,6 +5041,7 @@ static int _sde_hardware_pre_caps(struct sde_mdss_cfg *sde_cfg, uint32_t hw_rev)
 	} else if (IS_WAIPIO_TARGET(hw_rev)) {
 		sde_cfg->has_dedicated_cwb_support = true;
 		sde_cfg->has_wb_ubwc = true;
+		sde_cfg->has_cwb_crop = true;
 		sde_cfg->has_qsync = true;
 		sde_cfg->perf.min_prefill_lines = 40;
 		sde_cfg->vbif_qos_nlvl = 8;
diff --git a/msm/sde/sde_hw_catalog.h b/msm/sde/sde_hw_catalog.h
index d0dd3212fa..0c67b9a37e 100644
--- a/msm/sde/sde_hw_catalog.h
+++ b/msm/sde/sde_hw_catalog.h
@@ -528,6 +528,7 @@ enum {
  *                          data arrives.
  * @SDE_WB_HAS_CWB          Writeback block supports concurrent writeback
  * @SDE_WB_HAS_DCWB         Writeback block supports dedicated CWB
+ * @SDE_WB_CROP             CWB supports cropping
  * @SDE_WB_CWB_CTRL         Separate CWB control is available for configuring
  * @SDE_WB_DCWB_CTRL        Separate DCWB control is available for configuring
  * @SDE_WB_MAX              maximum value
@@ -551,6 +552,7 @@ enum {
 	SDE_WB_INPUT_CTRL,
 	SDE_WB_HAS_CWB,
 	SDE_WB_HAS_DCWB,
+	SDE_WB_CROP,
 	SDE_WB_CWB_CTRL,
 	SDE_WB_DCWB_CTRL,
 	SDE_WB_MAX
@@ -1447,6 +1449,7 @@ struct sde_perf_cfg {
  * @has_src_split      source split feature status
  * @has_cdp            Client driven prefetch feature status
  * @has_wb_ubwc        UBWC feature supported on WB
+ * @has_cwb_crop       CWB cropping is supported
  * @has_cwb_support    indicates if device supports primary capture through CWB
  * @has_dedicated_cwb_support    indicates if device supports dedicated path for CWB capture
  * @cwb_blk_off        CWB offset address
@@ -1530,6 +1533,7 @@ struct sde_mdss_cfg {
 	bool has_cdp;
 	bool has_dim_layer;
 	bool has_wb_ubwc;
+	bool has_cwb_crop;
 	bool has_cwb_support;
 	bool has_dedicated_cwb_support;
 	u32 cwb_blk_off;
diff --git a/msm/sde/sde_hw_wb.c b/msm/sde/sde_hw_wb.c
index 8a8b65a5e2..2715797698 100644
--- a/msm/sde/sde_hw_wb.c
+++ b/msm/sde/sde_hw_wb.c
@@ -42,6 +42,8 @@
 #define WB_CREQ_LUT_1			0x09C
 #define WB_UBWC_STATIC_CTRL		0x144
 #define WB_MUX				0x150
+#define WB_CROP_CTRL			0x154
+#define WB_CROP_OFFSET			0x158
 #define WB_CSC_BASE			0x260
 #define WB_DST_ADDR_SW_STATUS		0x2B0
 #define WB_CDP_CNTL			0x2B4
@@ -233,6 +235,21 @@ static void sde_hw_wb_roi(struct sde_hw_wb *ctx, struct sde_hw_wb_cfg *wb)
 	SDE_REG_WRITE(c, WB_OUT_SIZE, out_size);
 }
 
+static void sde_hw_wb_crop(struct sde_hw_wb *ctx, struct sde_hw_wb_cfg *wb, bool crop)
+{
+	struct sde_hw_blk_reg_map *c = &ctx->hw;
+	u32 crop_xy;
+
+	crop_xy = (wb->crop.y << 16) | wb->crop.x;
+
+	if (crop) {
+		SDE_REG_WRITE(c, WB_CROP_CTRL, 0x1);
+		SDE_REG_WRITE(c, WB_CROP_OFFSET, crop_xy);
+	} else {
+		SDE_REG_WRITE(c, WB_CROP_CTRL, 0x0);
+	}
+}
+
 static void sde_hw_wb_setup_qos_lut(struct sde_hw_wb *ctx,
 		struct sde_hw_wb_qos_cfg *cfg)
 {
@@ -366,6 +383,9 @@ static void _setup_wb_ops(struct sde_hw_wb_ops *ops,
 	if (test_bit(SDE_WB_XY_ROI_OFFSET, &features))
 		ops->setup_roi = sde_hw_wb_roi;
 
+	if (test_bit(SDE_WB_CROP, &features))
+		ops->setup_crop = sde_hw_wb_crop;
+
 	if (test_bit(SDE_WB_QOS, &features))
 		ops->setup_qos_lut = sde_hw_wb_setup_qos_lut;
 
diff --git a/msm/sde/sde_hw_wb.h b/msm/sde/sde_hw_wb.h
index 34485fb3f3..61ced9c939 100644
--- a/msm/sde/sde_hw_wb.h
+++ b/msm/sde/sde_hw_wb.h
@@ -17,6 +17,7 @@ struct sde_hw_wb_cfg {
 	struct sde_hw_fmt_layout dest;
 	enum sde_intf_mode intf_mode;
 	struct sde_rect roi;
+	struct sde_rect crop;
 	bool is_secure;
 };
 
@@ -88,6 +89,9 @@ struct sde_hw_wb_ops {
 	void (*setup_roi)(struct sde_hw_wb *ctx,
 		struct sde_hw_wb_cfg *wb);
 
+	void (*setup_crop)(struct sde_hw_wb *ctx,
+		struct sde_hw_wb_cfg *wb, bool crop);
+
 	/**
 	 * setup_qos_lut - setup danger, safe, creq, etc. LUTs
 	 * @ctx: Pointer to pipe context