media: vicodec: add support for CROP and COMPOSE selection

index a6fd0477633b..927d7b417a2d 100644
@@ -11,6 +11,7 @@
 
 #include <linux/string.h>
 #include "codec-fwht.h"
+#include <linux/kernel.h>
 
 /*
  * Note: bit 0 of the header must always be 0. Otherwise it cannot

@@ -226,7 +227,7 @@ static void dequantize_inter(s16 *coeff)
 			*coeff <<= *quant;
 }
 
-static void fwht(const u8 *block, s16 *output_block, unsigned int stride,
+static void fwht(const u8 *block, s16 *output_block, unsigned int coded_width,
 		 unsigned int input_step, bool intra)
 {
 	/* we'll need more than 8 bits for the transformed coefficients */

@@ -237,9 +238,9 @@ static void fwht(const u8 *block, s16 *output_block, unsigned int stride,
 	unsigned int i;
 
 	/* stage 1 */
-	stride *= input_step;
+	coded_width *= input_step;
 
-	for (i = 0; i < 8; i++, tmp += stride, out += 8) {
+	for (i = 0; i < 8; i++, tmp += coded_width, out += 8) {
 		switch (input_step) {
 		case 1:
 			workspace1[0]  = tmp[0] + tmp[1] - add;

@@ -361,7 +362,7 @@ static void fwht(const u8 *block, s16 *output_block, unsigned int stride,
  * Furthermore values can be negative... This is just a version that
  * works with 16 signed data
  */
-static void fwht16(const s16 *block, s16 *output_block, int stride, int intra)
+static void fwht16(const s16 *block, s16 *output_block, int coded_width, int intra)
 {
 	/* we'll need more than 8 bits for the transformed coefficients */
 	s32 workspace1[8], workspace2[8];

@@ -369,7 +370,7 @@ static void fwht16(const s16 *block, s16 *output_block, int stride, int intra)
 	s16 *out = output_block;
 	int i;
 
-	for (i = 0; i < 8; i++, tmp += stride, out += 8) {
+	for (i = 0; i < 8; i++, tmp += coded_width, out += 8) {
 		/* stage 1 */
 		workspace1[0]  = tmp[0] + tmp[1];
 		workspace1[1]  = tmp[0] - tmp[1];

@@ -555,14 +556,14 @@ static void ifwht(const s16 *block, s16 *output_block, int intra)
 }
 
 static void fill_encoder_block(const u8 *input, s16 *dst,
-			       unsigned int stride, unsigned int input_step)
+			       unsigned int coded_width, unsigned int input_step)
 {
 	int i, j;
 
 	for (i = 0; i < 8; i++) {
 		for (j = 0; j < 8; j++, input += input_step)
 			*dst++ = *input;
-		input += (stride - 8) * input_step;
+		input += (coded_width - 8) * input_step;
 	}
 }
 

@@ -593,7 +594,7 @@ static int var_inter(const s16 *old, const s16 *new)
 }
 
 static int decide_blocktype(const u8 *cur, const u8 *reference,
-			    s16 *deltablock, unsigned int stride,
+			    s16 *deltablock, unsigned int coded_width,
 			    unsigned int input_step)
 {
 	s16 tmp[64];

@@ -603,7 +604,7 @@ static int decide_blocktype(const u8 *cur, const u8 *reference,
 	int vari;
 	int vard;
 
-	fill_encoder_block(cur, tmp, stride, input_step);
+	fill_encoder_block(cur, tmp, coded_width, input_step);
 	fill_encoder_block(reference, old, 8, 1);
 	vari = var_intra(tmp);
 

@@ -620,7 +621,7 @@ static int decide_blocktype(const u8 *cur, const u8 *reference,
 	return vari <= vard ? IBLOCK : PBLOCK;
 }
 
-static void fill_decoder_block(u8 *dst, const s16 *input, int stride)
+static void fill_decoder_block(u8 *dst, const s16 *input, int coded_width)
 {
 	int i, j;
 

@@ -633,11 +634,11 @@ static void fill_decoder_block(u8 *dst, const s16 *input, int stride)
 			else
 				*dst = *input;
 		}
-		dst += stride - 8;
+		dst += coded_width - 8;
 	}
 }
 
-static void add_deltas(s16 *deltas, const u8 *ref, int stride)
+static void add_deltas(s16 *deltas, const u8 *ref, int coded_width)
 {
 	int k, l;
 

@@ -654,12 +655,12 @@ static void add_deltas(s16 *deltas, const u8 *ref, int stride)
 				*deltas = 255;
 			deltas++;
 		}
-		ref += stride - 8;
+		ref += coded_width - 8;
 	}
 }
 
 static u32 encode_plane(u8 *input, u8 *refp, __be16 **rlco, __be16 *rlco_max,
-			struct fwht_cframe *cf, u32 height, u32 width,
+			struct fwht_cframe *cf, u32 height, u32 width, u32 coded_width,
 			unsigned int input_step,
 			bool is_intra, bool next_is_intra)
 {

@@ -671,7 +672,11 @@ static u32 encode_plane(u8 *input, u8 *refp, __be16 **rlco, __be16 *rlco_max,
 	unsigned int last_size = 0;
 	unsigned int i, j;
 
+	width = round_up(width, 8);
+	height = round_up(height, 8);
+
 	for (j = 0; j < height / 8; j++) {
+		input = input_start + j * 8 * coded_width * input_step;
 		for (i = 0; i < width / 8; i++) {
 			/* intra code, first frame is always intra coded. */
 			int blocktype = IBLOCK;

@@ -679,9 +684,9 @@ static u32 encode_plane(u8 *input, u8 *refp, __be16 **rlco, __be16 *rlco_max,
 
 			if (!is_intra)
 				blocktype = decide_blocktype(input, refp,
-					deltablock, width, input_step);
+					deltablock, coded_width, input_step);
 			if (blocktype == IBLOCK) {
-				fwht(input, cf->coeffs, width, input_step, 1);
+				fwht(input, cf->coeffs, coded_width, input_step, 1);
 				quantize_intra(cf->coeffs, cf->de_coeffs,
 					       cf->i_frame_qp);
 			} else {

@@ -722,7 +727,6 @@ static u32 encode_plane(u8 *input, u8 *refp, __be16 **rlco, __be16 *rlco_max,
 			}
 			last_size = size;
 		}
-		input += width * 7 * input_step;
 	}
 
 exit_loop:

@@ -747,29 +751,31 @@ static u32 encode_plane(u8 *input, u8 *refp, __be16 **rlco, __be16 *rlco_max,
 u32 fwht_encode_frame(struct fwht_raw_frame *frm,
 		      struct fwht_raw_frame *ref_frm,
 		      struct fwht_cframe *cf,
-		      bool is_intra, bool next_is_intra)
+		      bool is_intra, bool next_is_intra,
+		      unsigned int width, unsigned int height)
 {
-	unsigned int size = frm->height * frm->width;
+	unsigned int size = height * width;
 	__be16 *rlco = cf->rlc_data;
 	__be16 *rlco_max;
 	u32 encoding;
 
 	rlco_max = rlco + size / 2 - 256;
 	encoding = encode_plane(frm->luma, ref_frm->luma, &rlco, rlco_max, cf,
-				frm->height, frm->width,
+				height, width, frm->coded_width,
 				frm->luma_alpha_step, is_intra, next_is_intra);
 	if (encoding & FWHT_FRAME_UNENCODED)
 		encoding |= FWHT_LUMA_UNENCODED;
 	encoding &= ~FWHT_FRAME_UNENCODED;
 
 	if (frm->components_num >= 3) {
-		u32 chroma_h = frm->height / frm->height_div;
-		u32 chroma_w = frm->width / frm->width_div;
+		u32 chroma_h = height / frm->height_div;
+		u32 chroma_w = width / frm->width_div;
+		u32 chroma_coded_width = frm->coded_width / frm->width_div;
 		unsigned int chroma_size = chroma_h * chroma_w;
 
 		rlco_max = rlco + chroma_size / 2 - 256;
 		encoding |= encode_plane(frm->cb, ref_frm->cb, &rlco, rlco_max,
-					 cf, chroma_h, chroma_w,
+					 cf, chroma_h, chroma_w, chroma_coded_width,
 					 frm->chroma_step,
 					 is_intra, next_is_intra);
 		if (encoding & FWHT_FRAME_UNENCODED)

@@ -777,7 +783,7 @@ u32 fwht_encode_frame(struct fwht_raw_frame *frm,
 		encoding &= ~FWHT_FRAME_UNENCODED;
 		rlco_max = rlco + chroma_size / 2 - 256;
 		encoding |= encode_plane(frm->cr, ref_frm->cr, &rlco, rlco_max,
-					 cf, chroma_h, chroma_w,
+					 cf, chroma_h, chroma_w, chroma_coded_width,
 					 frm->chroma_step,
 					 is_intra, next_is_intra);
 		if (encoding & FWHT_FRAME_UNENCODED)

@@ -788,8 +794,8 @@ u32 fwht_encode_frame(struct fwht_raw_frame *frm,
 	if (frm->components_num == 4) {
 		rlco_max = rlco + size / 2 - 256;
 		encoding |= encode_plane(frm->alpha, ref_frm->alpha, &rlco,
-					 rlco_max, cf, frm->height, frm->width,
-					 frm->luma_alpha_step,
+					 rlco_max, cf, height, width,
+					 frm->coded_width, frm->luma_alpha_step,
 					 is_intra, next_is_intra);
 		if (encoding & FWHT_FRAME_UNENCODED)
 			encoding |= FWHT_ALPHA_UNENCODED;

@@ -801,7 +807,7 @@ u32 fwht_encode_frame(struct fwht_raw_frame *frm,
 }
 
 static void decode_plane(struct fwht_cframe *cf, const __be16 **rlco, u8 *ref,
-			 u32 height, u32 width, bool uncompressed)
+			 u32 height, u32 width, u32 coded_width, bool uncompressed)
 {
 	unsigned int copies = 0;
 	s16 copy[8 * 8];

@@ -813,6 +819,8 @@ static void decode_plane(struct fwht_cframe *cf, const __be16 **rlco, u8 *ref,
 		*rlco += width * height / 2;
 		return;
 	}
+	width = round_up(width, 8);
+	height = round_up(height, 8);
 
 	/*
 	 * When decoding each macroblock the rlco pointer will be increased

@@ -822,13 +830,13 @@ static void decode_plane(struct fwht_cframe *cf, const __be16 **rlco, u8 *ref,
 	 */
 	for (j = 0; j < height / 8; j++) {
 		for (i = 0; i < width / 8; i++) {
-			u8 *refp = ref + j * 8 * width + i * 8;
+			u8 *refp = ref + j * 8 * coded_width + i * 8;
 
 			if (copies) {
 				memcpy(cf->de_fwht, copy, sizeof(copy));
 				if (stat & PFRAME_BIT)
-					add_deltas(cf->de_fwht, refp, width);
-				fill_decoder_block(refp, cf->de_fwht, width);
+					add_deltas(cf->de_fwht, refp, coded_width);
+				fill_decoder_block(refp, cf->de_fwht, coded_width);
 				copies--;
 				continue;
 			}

@@ -847,35 +855,39 @@ static void decode_plane(struct fwht_cframe *cf, const __be16 **rlco, u8 *ref,
 			if (copies)
 				memcpy(copy, cf->de_fwht, sizeof(copy));
 			if (stat & PFRAME_BIT)
-				add_deltas(cf->de_fwht, refp, width);
-			fill_decoder_block(refp, cf->de_fwht, width);
+				add_deltas(cf->de_fwht, refp, coded_width);
+			fill_decoder_block(refp, cf->de_fwht, coded_width);
 		}
 	}
 }
 
 void fwht_decode_frame(struct fwht_cframe *cf, struct fwht_raw_frame *ref,
-		       u32 hdr_flags, unsigned int components_num)
+		       u32 hdr_flags, unsigned int components_num,
+		       unsigned int width, unsigned int height, unsigned int coded_width)
 {
 	const __be16 *rlco = cf->rlc_data;
 
-	decode_plane(cf, &rlco, ref->luma, cf->height, cf->width,
+	decode_plane(cf, &rlco, ref->luma, height, width, coded_width,
 		     hdr_flags & FWHT_FL_LUMA_IS_UNCOMPRESSED);
 
 	if (components_num >= 3) {
-		u32 h = cf->height;
-		u32 w = cf->width;
+		u32 h = height;
+		u32 w = width;
+		u32 c = coded_width;
 
 		if (!(hdr_flags & FWHT_FL_CHROMA_FULL_HEIGHT))
 			h /= 2;
-		if (!(hdr_flags & FWHT_FL_CHROMA_FULL_WIDTH))
+		if (!(hdr_flags & FWHT_FL_CHROMA_FULL_WIDTH)) {
 			w /= 2;
-		decode_plane(cf, &rlco, ref->cb, h, w,
+			c /= 2;
+		}
+		decode_plane(cf, &rlco, ref->cb, h, w, c,
 			     hdr_flags & FWHT_FL_CB_IS_UNCOMPRESSED);
-		decode_plane(cf, &rlco, ref->cr, h, w,
+		decode_plane(cf, &rlco, ref->cr, h, w, c,
 			     hdr_flags & FWHT_FL_CR_IS_UNCOMPRESSED);
 	}
 
 	if (components_num == 4)
-		decode_plane(cf, &rlco, ref->alpha, cf->height, cf->width,
+		decode_plane(cf, &rlco, ref->alpha, height, width, coded_width,
 			     hdr_flags & FWHT_FL_ALPHA_IS_UNCOMPRESSED);
 }

index 90ff8962fca7..8163014c368c 100644
@@ -81,6 +81,12 @@
 #define FWHT_FL_COMPONENTS_NUM_MSK	GENMASK(17, 16)
 #define FWHT_FL_COMPONENTS_NUM_OFFSET	16
 
+/* A macro to calculate the needed padding in order to make sure
+ * both luma and chroma components resolutions are rounded up to
+ * closest multiple of 8
+ */
+#define vic_round_dim(dim, div) (round_up((dim) / (div), 8) * (div))
+
 struct fwht_cframe_hdr {
 	u32 magic1;
 	u32 magic2;

@@ -95,7 +101,6 @@ struct fwht_cframe_hdr {
 };
 
 struct fwht_cframe {
-	unsigned int width, height;
 	u16 i_frame_qp;
 	u16 p_frame_qp;
 	__be16 *rlc_data;

@@ -106,12 +111,12 @@ struct fwht_cframe {
 };
 
 struct fwht_raw_frame {
-	unsigned int width, height;
 	unsigned int width_div;
 	unsigned int height_div;
 	unsigned int luma_alpha_step;
 	unsigned int chroma_step;
 	unsigned int components_num;
+	unsigned int coded_width;
 	u8 *luma, *cb, *cr, *alpha;
 };
 

@@ -125,8 +130,10 @@ struct fwht_raw_frame {
 u32 fwht_encode_frame(struct fwht_raw_frame *frm,
 		      struct fwht_raw_frame *ref_frm,
 		      struct fwht_cframe *cf,
-		      bool is_intra, bool next_is_intra);
+		      bool is_intra, bool next_is_intra,
+		      unsigned int width, unsigned int height);
 void fwht_decode_frame(struct fwht_cframe *cf, struct fwht_raw_frame *ref,
-		       u32 hdr_flags, unsigned int components_num);
+		       u32 hdr_flags, unsigned int components_num,
+		       unsigned int width, unsigned int height, unsigned int coded_width);
 
 #endif

index 8cb0212df67f..652296bdd250 100644
@@ -56,7 +56,7 @@ const struct v4l2_fwht_pixfmt_info *v4l2_fwht_get_pixfmt(u32 idx)
 
 int v4l2_fwht_encode(struct v4l2_fwht_state *state, u8 *p_in, u8 *p_out)
 {
-	unsigned int size = state->width * state->height;
+	unsigned int size;
 	const struct v4l2_fwht_pixfmt_info *info = state->info;
 	struct fwht_cframe_hdr *p_hdr;
 	struct fwht_cframe cf;

@@ -66,8 +66,9 @@ int v4l2_fwht_encode(struct v4l2_fwht_state *state, u8 *p_in, u8 *p_out)
 
 	if (!info)
 		return -EINVAL;
-	rf.width = state->width;
-	rf.height = state->height;
+
+	size = state->coded_width * state->coded_height;
+	rf.coded_width = state->coded_width;
 	rf.luma = p_in;
 	rf.width_div = info->width_div;
 	rf.height_div = info->height_div;

@@ -163,15 +164,14 @@ int v4l2_fwht_encode(struct v4l2_fwht_state *state, u8 *p_in, u8 *p_out)
 		return -EINVAL;
 	}
 
-	cf.width = state->width;
-	cf.height = state->height;
 	cf.i_frame_qp = state->i_frame_qp;
 	cf.p_frame_qp = state->p_frame_qp;
 	cf.rlc_data = (__be16 *)(p_out + sizeof(*p_hdr));
 
 	encoding = fwht_encode_frame(&rf, &state->ref_frame, &cf,
 				     !state->gop_cnt,
-				     state->gop_cnt == state->gop_size - 1);
+				     state->gop_cnt == state->gop_size - 1,
+				     state->visible_width, state->visible_height);
 	if (!(encoding & FWHT_FRAME_PCODED))
 		state->gop_cnt = 0;
 	if (++state->gop_cnt >= state->gop_size)

@@ -181,8 +181,8 @@ int v4l2_fwht_encode(struct v4l2_fwht_state *state, u8 *p_in, u8 *p_out)
 	p_hdr->magic1 = FWHT_MAGIC1;
 	p_hdr->magic2 = FWHT_MAGIC2;
 	p_hdr->version = htonl(FWHT_VERSION);
-	p_hdr->width = htonl(cf.width);
-	p_hdr->height = htonl(cf.height);
+	p_hdr->width = htonl(state->visible_width);
+	p_hdr->height = htonl(state->visible_height);
 	flags |= (info->components_num - 1) << FWHT_FL_COMPONENTS_NUM_OFFSET;
 	if (encoding & FWHT_LUMA_UNENCODED)
 		flags |= FWHT_FL_LUMA_IS_UNCOMPRESSED;

@@ -202,15 +202,13 @@ int v4l2_fwht_encode(struct v4l2_fwht_state *state, u8 *p_in, u8 *p_out)
 	p_hdr->ycbcr_enc = htonl(state->ycbcr_enc);
 	p_hdr->quantization = htonl(state->quantization);
 	p_hdr->size = htonl(cf.size);
-	state->ref_frame.width = cf.width;
-	state->ref_frame.height = cf.height;
 	return cf.size + sizeof(*p_hdr);
 }
 
 int v4l2_fwht_decode(struct v4l2_fwht_state *state, u8 *p_in, u8 *p_out)
 {
-	unsigned int size = state->width * state->height;
-	unsigned int chroma_size = size;
+	unsigned int size;
+	unsigned int chroma_size;
 	unsigned int i;
 	u32 flags;
 	struct fwht_cframe_hdr *p_hdr;

@@ -218,13 +216,15 @@ int v4l2_fwht_decode(struct v4l2_fwht_state *state, u8 *p_in, u8 *p_out)
 	u8 *p;
 	unsigned int components_num = 3;
 	unsigned int version;
+	const struct v4l2_fwht_pixfmt_info *info;
 
 	if (!state->info)
 		return -EINVAL;
 
+	info = state->info;
+	size = state->coded_width * state->coded_height;
+	chroma_size = size;
 	p_hdr = (struct fwht_cframe_hdr *)p_in;
-	cf.width = ntohl(p_hdr->width);
-	cf.height = ntohl(p_hdr->height);
 
 	version = ntohl(p_hdr->version);
 	if (!version || version > FWHT_VERSION) {

@@ -234,12 +234,11 @@ int v4l2_fwht_decode(struct v4l2_fwht_state *state, u8 *p_in, u8 *p_out)
 	}
 
 	if (p_hdr->magic1 != FWHT_MAGIC1 ||
-	    p_hdr->magic2 != FWHT_MAGIC2 ||
-	    (cf.width & 7) || (cf.height & 7))
+	    p_hdr->magic2 != FWHT_MAGIC2)
 		return -EINVAL;
 
 	/* TODO: support resolution changes */
-	if (cf.width != state->width || cf.height != state->height)
+	if (ntohl(p_hdr->width) != state->visible_width || ntohl(p_hdr->height) != state->visible_height)
 		return -EINVAL;
 
 	flags = ntohl(p_hdr->flags);

@@ -260,7 +259,8 @@ int v4l2_fwht_decode(struct v4l2_fwht_state *state, u8 *p_in, u8 *p_out)
 	if (!(flags & FWHT_FL_CHROMA_FULL_HEIGHT))
 		chroma_size /= 2;
 
-	fwht_decode_frame(&cf, &state->ref_frame, flags, components_num);
+	fwht_decode_frame(&cf, &state->ref_frame, flags, components_num,
+			  state->visible_width, state->visible_height, state->coded_width);
 
 	/*
 	 * TODO - handle the case where the compressed stream encodes a

index ed53e28d4f9c..2a09ad13ddd6 100644
@@ -23,8 +23,10 @@ struct v4l2_fwht_pixfmt_info {
 
 struct v4l2_fwht_state {
 	const struct v4l2_fwht_pixfmt_info *info;
-	unsigned int width;
-	unsigned int height;
+	unsigned int visible_width;
+	unsigned int visible_height;
+	unsigned int coded_width;
+	unsigned int coded_height;
 	unsigned int gop_size;
 	unsigned int gop_cnt;
 	u16 i_frame_qp;

250 lines not on this page · raw message