drivers/amlogic/media/enhancement/amvecm/hdr/am_hdr10_tm.c - manifest_repos/kernel - Git at Google

 // SPDX-License-Identifier: (GPL-2.0+ OR MIT)
 /*
  * drivers/amlogic/media/enhancement/amvecm/hdr/gamut_convet.c
  *
  * Copyright (C) 2017 Amlogic, Inc. All rights reserved.
  *
  * This program is free software; you can redistribute it and/or modify
  * it under the terms of the GNU General Public License as published by
  * the Free Software Foundation; either version 2 of the License, or
  * (at your option) any later version.
  *
  * This program is distributed in the hope that it will be useful, but WITHOUT
  * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
  * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
  * more details.
  *
  */
 #include <linux/types.h>
 #include <linux/init.h>
 #include <linux/module.h>
 #include <linux/kernel.h>
 #include <linux/errno.h>
 #include <linux/debugfs.h>
 #include <linux/uaccess.h>
 #include <linux/amlogic/media/amvecm/amvecm.h>
 #include <linux/amlogic/media/amvecm/hdr10_tmo_alg.h>
 #include "am_hdr10_tm.h"
 #include "../set_hdr2_v0.h"
 #include "am_hdr10_tmo_fw.h"

 unsigned int hdr10_tm_dbg;
 module_param(hdr10_tm_dbg, uint, 0664);
 MODULE_PARM_DESC(hdr10_tm_dbg, "HDR10 tone mapping dbg\n");

 #define pr_hdr_tm(fmt, args...)\
 	do {\
 		if (hdr10_tm_dbg)\
 			pr_info(fmt, ## args);\
 	} while (0)

 unsigned int panell = 400;
 module_param(panell, uint, 0664);
 MODULE_PARM_DESC(panell, "display panel luminance\n");

 static unsigned int hdr10_tm_sel = 2; /*1 old algorithm, 2 hdr_tmo algorithm  default 2*/
 module_param(hdr10_tm_sel, uint, 0664);
 MODULE_PARM_DESC(hdr10_tm_sel, "hdr10_tm_sel\n");

 #define KNEE_POINT 2
 static unsigned int kp = KNEE_POINT;
 static unsigned int knee_point[KNEE_POINT] = {0, 0};
 module_param_array(knee_point, uint, &kp, 0664);
 MODULE_PARM_DESC(knee_point, "\n knee point xy\n");

 /*u32 ebz_p[MAX_BEIZER_ORDER] = {
  *	0, 1170, 1755, 2121, 2377, 2596, 2779, 2925, 3064, 3185,
  *	3320, 3514, 3799, 3957, 4095
  *};
  */

 /*time domain filter, avoid flicker*/
 /*scene change th: 2/3 scene diff*/
 u32 sc_th = 682;
 module_param(sc_th, uint, 0664);
 MODULE_PARM_DESC(sc_th, "scene change th\n");

 u32 hdr_tm_iir = 1;
 module_param(hdr_tm_iir, uint, 0664);
 MODULE_PARM_DESC(hdr_tm_iir, "HDR_TM_IIR\n");

 int is_hdr_tmo_support(void)
 {
 	if (get_cpu_type() >= MESON_CPU_MAJOR_ID_SM1)
 		return 1;
 	return 0;
 }

 int time_iir(u32 *maxl)
 {
 	int i;
 	static u32 hist_diff[3];
 	static u32 max_lum[3];
 	u32 diff;
 	u32 sum = 0;
 	int sumshft;
 	int norm14;
 	u32 bld_coef;
 	u32 dtl_lum = 0;
 	/*ret=1: scn change; ret=2: blend*/
 	int ret;

 	for (i = 0; i < 2; i++) {
 		hist_diff[i] = hist_diff[i + 1];
 		max_lum[i] = max_lum[i + 1];
 	}
 	hist_diff[2] = 0;
 	max_lum[2] = *maxl;

 	for (i = 0; i < 128; i++) {
 		sum += hdr_hist[15][i];
 		diff = (hdr_hist[15][i] > hdr_hist[14][i]) ?
 			(hdr_hist[15][i] - hdr_hist[14][i]) :
 			(hdr_hist[14][i] - hdr_hist[15][i]);
 		hist_diff[2] += diff;
 	}

 	if (sum == 0)
 		return 0;

 	sumshft =
 		(sum >= (1 << 24)) ? 8 :
 		(sum >= (1 << 22)) ? 6 :
 		(sum >= (1 << 20)) ? 4 :
 		(sum >= (1 << 18)) ? 2 :
 		(sum >= (1 << 16)) ? 0 :
 		(sum >= (1 << 14)) ? -2 :
 		(sum >= (1 << 12)) ? -4 :
 		(sum >= (1 << 10)) ? -6 :
 		(sum >= (1 << 8)) ? -8 :
 		(sum >= (1 << 6)) ? -10 :
 		(sum >= (1 << 4)) ? -12 : -16;

 	if (sumshft >= 0)
 		norm14 = (1 << 30) / (sum >> sumshft);
 	else if (sumshft > -16)
 		norm14 = (1 << (30 + sumshft)) / sum;
 	else
 		norm14 = 1 << 14;

 	if (sumshft >= 0) {
 		hist_diff[2] = ((hist_diff[2] >> sumshft) * norm14 +
 			(1 << 13)) >> 14;
 	} else {
 		hist_diff[2] = (((hist_diff[2] << (-sumshft)) * norm14) +
 			(1 << 13)) >> 14;
 	}

 	/*normalize to 10bit*/
 	hist_diff[2] >>= 6;
 	if (hist_diff[2] > sc_th || (*maxl <= panell)) {
 		ret = 1;
 	} else {
 		bld_coef = hist_diff[2];
 		dtl_lum =
 			(max_lum[2] > max_lum[1]) ?
 			(max_lum[2] - max_lum[1]) :
 			(max_lum[1] - max_lum[2]);

 		if (dtl_lum < (1 << 3))
 			max_lum[2] =
 				(max_lum[2] > max_lum[1]) ?
 				(max_lum[1] + (dtl_lum >> 1)) :
 				(max_lum[1] - (dtl_lum >> 1));
 		else if (dtl_lum < (1 << 6))
 			max_lum[2] =
 				(max_lum[2] > max_lum[1]) ?
 				(max_lum[1] + (dtl_lum >> 3)) :
 				(max_lum[1] - (dtl_lum >> 3));
 		else
 			max_lum[2] =
 				(max_lum[2] > max_lum[1]) ?
 				(max_lum[1] + (dtl_lum >> 6)) :
 				(max_lum[1] - (dtl_lum >> 6));

 		*maxl = max_lum[2];
 		ret = 2;
 	}

 	pr_hdr_tm("ret = %d, dtl_lum = %d, max_lum[1] = %d, max_lum[2] = %d\n",
 		  ret, dtl_lum, max_lum[1], max_lum[2]);

 	return ret;
 }

 /*hdr10 tone mapping alg */
 int gen_oe_x_val(u64 *oe_x_val)
 {
 	unsigned int i = 0, j = 0;
 	int bin_num = 0;
 	int temp1 = 0, temp2 = 0;

 	oe_x_val[0] = 0;

 	for (i = 1; i < 8; i++) {
 		bin_num++;
 		temp1 = 1 << ((i - 1) + 4);
 		oe_x_val[i] = (u64)temp1;
 	}
 	for (j = 11; j < 20; j++) {
 		for (i = 0; i <  4; i++) {
 			bin_num++;
 			temp1 = 1 << (j - 2);
 			temp2 = 1 << j;
 			oe_x_val[bin_num] = (u64)temp1 * i + (u64)temp2;
 		}
 	}
 	for (j = 20; j < 31; j++) {
 		for (i = 0; i < 8; i++) {
 			bin_num++;
 			temp1 = 1 << (j - 3);
 			temp2 = 1 << j;
 			oe_x_val[bin_num] = (u64)temp1 * i + (u64)temp2;
 		}
 	}
 	for (i = 0; i < 16; i++) {
 		bin_num++;
 		oe_x_val[bin_num] = (1 << (31 - 4)) * i + (1 << 31);
 	}
 	oe_x_val[OE_X - 1] = MAX_32;

 	return 0;
 }

 void decasteliau_alg(u64 *ebzcur, u64 step, u64 range, u64 *p, u32 order)
 {
 	u64 point[MAX_BEIZER_ORDER + 1];
 	int i, j;

 	for (i = 0; i < order; i++)
 		point[i] = p[i];

 	for (i = 1; i < order; i++) {
 		for (j = 0; j < order - i; j++) {
 			point[j] = (range - step) * point[j] +
 				step * point[j + 1] + (range >> 1);
 			point[j] = div64_u64(point[j], range);
 		}
 	}
 	ebzcur[1] = point[0];
 }

 static u64 curvex[OE_X], curvey[OE_X];
 static u32 oo_gain[OE_X];

 int gen_knee_anchor(u32 maxl, u32 panell, u64 *sx, u64 *sy, u64 *anchor)
 {
 	int i;
 	int knee_bypass = 1229;
 	u64 anchor_linear[MAX_BEIZER_ORDER];
 	u64 norm, blendcoef;
 	//int tgtl = 10000;
 	struct basisootf_params basisootf_param;
 	u64 sx1, sy1;
 	u64 anchor0;

 	sx1 = *sx;
 	sy1 = *sy;

 	basisootf_params_init(&basisootf_param);

 	for (i = 0; i < MAX_BEIZER_ORDER - 1; i++)
 		anchor_linear[i] = (i + 1) *
 			(1 << MAX12_BIT) / MAX_BEIZER_ORDER;

 	pr_hdr_tm("knee anchor p1: %lld, %lld, %lld, %lld, %lld, %lld, %lld, %lld, %lld, %lld %lld\n",
 		  sx1, sy1,
 		  anchor_linear[0], anchor_linear[1],
 		  anchor_linear[2], anchor_linear[3],
 		  anchor_linear[4], anchor_linear[5],
 		  anchor_linear[6], anchor_linear[7],
 		  anchor_linear[8]);

 	if (maxl <= panell) {
 		sx1 = knee_bypass;
 		sy1 = knee_bypass;
 		for (i = 0; i < MAX_BEIZER_ORDER - 1; i++)
 			anchor[i] = anchor_linear[i];
 	} else {
 		norm = maxl;
 		blendcoef = panell;
 		sy1 = (blendcoef * knee_bypass +
 			(norm - blendcoef) * sy1 + (norm >> 1));
 		sy1 = div64_u64(sy1, norm);
 		sx1 = sy1 * panell;
 		sx1 = div64_u64(sx1, maxl);
 		for (i = 0; i < MAX_BEIZER_ORDER - 1; i++) {
 			anchor[i] =
 			(blendcoef * anchor_linear[i] +
 			(norm - blendcoef) * anchor[i] +
 			(norm >> 1));

 			anchor[i] = div64_u64(anchor[i], norm);
 		}

 		anchor0 = calcp1((u32)sx1, (u32)sy1,
 				 panell, maxl, &basisootf_param, NULL);

 		for (i = 1; i < MAX_BEIZER_ORDER - 1; i++)
 			anchor[i] = min(anchor[i], (i + 1) * anchor0);

 		if (anchor0 < anchor[1])
 			anchor[0] = anchor0;
 	}
 	*sx = sx1;
 	*sy = sy1;

 	pr_hdr_tm("knee anchor p2: %lld, %lld, %lld, %lld, %lld, %lld, %lld, %lld, %lld, %lld %lld\n",
 		  sx1, sy1,
 		  anchor[0], anchor[1],
 		  anchor[2], anchor[3],
 		  anchor[4], anchor[5],
 		  anchor[6], anchor[7],
 		  anchor[8]);

 	return 0;
 }

 int dynamic_hdr_sdr_ootf(u32 maxl, u32 panell, u64 sx, u64 sy, u64 *anchor)
 {
 	int i;
 	u32 tgtl = 10000;
 	static u64 oe_x[OE_X] = {0};
 	int order = MAX_BEIZER_ORDER + 1;
 	u64 step;
 	u64 kx, ky;
 	u64 ebzcur[2];
 	u64 range_ebz_x, range_ebz_y;
 	u64 finalanchor[MAX_BEIZER_ORDER + 1];
 	u64 temp;
 	u64 sx_primy;
 	u64 kx_primy;

 	gen_oe_x_val(oe_x);

 	gen_knee_anchor(maxl, panell, &sx, &sy, anchor);

 	pr_hdr_tm("d1: %d, %d, %lld, %lld, %lld, %lld, %lld, %lld, %lld, %lld, %lld, %lld %lld\n",
 		  maxl, panell,
 		  sx, sy,
 		  anchor[0], anchor[1],
 		  anchor[2], anchor[3],
 		  anchor[4], anchor[5],
 		  anchor[6], anchor[7],
 		  anchor[8]);

 	for (i = 0; i < order - 2; i++)
 		finalanchor[i + 1] = anchor[i] << (MAX32_BIT - MAX12_BIT);

 	finalanchor[0] = 0;
 	finalanchor[10] = MAX_32;

 	if (tgtl <= maxl)
 		maxl = tgtl - 1;

 	kx = sx << (MAX32_BIT - MAX12_BIT);
 	ky = sy << (MAX32_BIT - MAX12_BIT);

 	kx = kx * maxl;
 	kx = div64_u64(kx, tgtl);

 	if (maxl <= panell) {
 		sx_primy = (u64)(1 << MAX12_BIT) * panell +  (tgtl >> 1);
 		sx_primy = div64_u64(sx_primy, tgtl);
 	} else {
 		sx_primy = (u64)(1 << MAX12_BIT) * maxl +  (tgtl >> 1);
 		sx_primy = div64_u64(sx_primy, tgtl);
 	}
 	kx_primy = sx_primy << (MAX32_BIT - MAX12_BIT);

 	range_ebz_x = kx_primy - kx;
 	range_ebz_y = MAX_32 - ky;

 	if (maxl <= panell) {
 		for (i = 0; i < OE_X; i++) {
 			if (oe_x[i] <= kx_primy) {
 				temp = tgtl << TM_GAIN_BIT;
 				oo_gain[i] = div64_u64(temp, panell);
 			} else {
 				curvey[i] = oe_x[148] << TM_GAIN_BIT;
 				curvex[i] = oe_x[i];
 				oo_gain[i] = div64_u64(curvey[i], curvex[i]);
 			}
 			if (oo_gain[i] < (1 << TM_GAIN_BIT))
 				oo_gain[i] = 1 << TM_GAIN_BIT;
 			if (oo_gain[i] > (1 << MAX12_BIT) - 1)
 				oo_gain[i] = (1 << MAX12_BIT) - 1;
 		}
 	} else {
 		for (i = 0; i < OE_X; i++) {
 			if (oe_x[i] <= kx) {
 				temp = tgtl << TM_GAIN_BIT;
 				oo_gain[i] = div64_u64(temp, panell);
 			} else if (oe_x[i] <=  kx_primy) {
 				step = oe_x[i] - kx;
 				decasteliau_alg(&ebzcur[0], step,
 						range_ebz_x,
 						finalanchor, order);
 				curvey[i] = ky +
 					((range_ebz_y * ebzcur[1] + MAX_32 / 2)
 					>> MAX32_BIT);
 				temp = curvey[i] << TM_GAIN_BIT;
 				curvex[i] = kx + step;
 				oo_gain[i] = div64_u64(temp, curvex[i]);
 			}  else {
 				curvey[i] = oe_x[148] << TM_GAIN_BIT;
 				curvex[i] = oe_x[i];
 				oo_gain[i] = div64_u64(curvey[i], curvex[i]);
 			}
 			if (oo_gain[i] < (1 << TM_GAIN_BIT))
 				oo_gain[i] = 1 << TM_GAIN_BIT;
 			if (oo_gain[i] > (1 << MAX12_BIT) - 1)
 				oo_gain[i] = (1 << MAX12_BIT) - 1;
 		}
 	}

 	for (i = 0; i < OE_X; i++) {
 		if ((i + 1) % 10 == 0)
 			pr_hdr_tm("%d, %d, %d, %d, %d, %d, %d, %d, %d, %d\n",
 				  oo_gain[i - 9], oo_gain[i - 8],
 				  oo_gain[i - 7], oo_gain[i - 6],
 				  oo_gain[i - 5], oo_gain[i - 4],
 				  oo_gain[i - 3], oo_gain[i - 2],
 				  oo_gain[i - 1], oo_gain[i]);
 		if (i == 148)
 			pr_hdr_tm("%d, %d, %d, %d, %d, %d, %d, %d, %d\n",
 				  oo_gain[i - 8],
 				  oo_gain[i - 7], oo_gain[i - 6],
 				  oo_gain[i - 5], oo_gain[i - 4],
 				  oo_gain[i - 3], oo_gain[i - 2],
 				  oo_gain[i - 1], oo_gain[i]);
 	}
 	pr_hdr_tm("\n");

 	return 0;
 }

 int hdr10_tm_dynamic_proc(struct vframe_master_display_colour_s *p)
 {
 	int i;
 	u32 maxl;
 	u32 primry_maxl;
 	u32 panel_luma;
 	u64 sx, sy;
 	u64 anchor[MAX_BEIZER_ORDER - 1];
 	u32 P_init[MAX_BEIZER_ORDER - 1] = {
 		181, 406, 607, 796, 834, 863, 890, 917, 938
 	};
 	int scn_chang_flag = 1;
 	struct aml_tmo_reg_sw *pre_tmo_reg;

 	if (p->luminance[0] > 10000)
 		p->luminance[0] /= 10000;
 	/*no luminance*/
 	if (p->luminance[0] == 0)
 		p->luminance[0] = 1200;

 	primry_maxl = p->luminance[0];

 	/*use 95% maxl because of high percert flicker*/
 	maxl = (percentile[8] > primry_maxl) ? primry_maxl : percentile[8];
 	pr_hdr_tm("maxl = %d, percentile[8] = %d, primry_maxl =%d\n",
 		  maxl, percentile[8], primry_maxl);

 	if (hdr_tm_iir)
 		scn_chang_flag = time_iir(&maxl);

 	if (scn_chang_flag == 0)
 		pr_hdr_tm("invalid hist\n");
 	else if (scn_chang_flag == 1)
 		pr_hdr_tm("scn change\n");
 	else if (scn_chang_flag == 2)
 		pr_hdr_tm("maxl blend: maxl = %d\n", maxl);

 	panel_luma = panell;

 	sx = knee_point[0];
 	sy = knee_point[1];

 	/*invalid case*/
 	if (maxl == 0) {
 		maxl = 1200;
 		pr_hdr_tm("invalid maxl\n");
 	}

 	for (i = 0; i < MAX_BEIZER_ORDER - 1; i++)
 		anchor[i] = P_init[i] << 2;

 	pre_tmo_reg = tmo_fw_param_get();

 	if (!is_hdr_tmo_support() || !pre_tmo_reg->pre_hdr10_tmo_alg) {
 		/*used old hdr alg*/
 		pr_hdr_tm("used old hdr alg - ");
 		if (!pre_tmo_reg->pre_hdr10_tmo_alg)
 			pr_hdr_tm("hdr alg ko insmod failed.\n");
 		else
 			pr_hdr_tm("the chip is not support.\n");
 		hdr10_tm_sel = 1;
 	}

 	if (hdr10_tm_sel == 1) {
 		dynamic_hdr_sdr_ootf(maxl, panel_luma, sx, sy, anchor);
 		memcpy(oo_y_lut_hdr_sdr, oo_gain, sizeof(u32) * OE_X);
 	} else if (hdr10_tm_sel == 2) {
 		hdr10_tmo_gen(oo_gain);
 		memcpy(oo_y_lut_hdr_sdr, oo_gain, sizeof(u32) * OE_X);
 	} else {
 		memcpy(oo_y_lut_hdr_sdr, oo_y_lut_hdr_sdr_def,
 			sizeof(u32) * OE_X);
 	}

 	if (hdr10_tm_dbg > 0)
 		hdr10_tm_dbg--;

 	return 0;
 }
	// SPDX-License-Identifier: (GPL-2.0+ OR MIT)
	/*
	* drivers/amlogic/media/enhancement/amvecm/hdr/gamut_convet.c
	*
	* Copyright (C) 2017 Amlogic, Inc. All rights reserved.
	*
	* This program is free software; you can redistribute it and/or modify
	* it under the terms of the GNU General Public License as published by
	* the Free Software Foundation; either version 2 of the License, or
	* (at your option) any later version.
	*
	* This program is distributed in the hope that it will be useful, but WITHOUT
	* ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
	* FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
	* more details.
	*
	*/
	#include <linux/types.h>
	#include <linux/init.h>
	#include <linux/module.h>
	#include <linux/kernel.h>
	#include <linux/errno.h>
	#include <linux/debugfs.h>
	#include <linux/uaccess.h>
	#include <linux/amlogic/media/amvecm/amvecm.h>
	#include <linux/amlogic/media/amvecm/hdr10_tmo_alg.h>
	#include "am_hdr10_tm.h"
	#include "../set_hdr2_v0.h"
	#include "am_hdr10_tmo_fw.h"

	unsigned int hdr10_tm_dbg;
	module_param(hdr10_tm_dbg, uint, 0664);
	MODULE_PARM_DESC(hdr10_tm_dbg, "HDR10 tone mapping dbg\n");

	#define pr_hdr_tm(fmt, args...)\
	do {\
	if (hdr10_tm_dbg)\
	pr_info(fmt, ## args);\
	} while (0)

	unsigned int panell = 400;
	module_param(panell, uint, 0664);
	MODULE_PARM_DESC(panell, "display panel luminance\n");

	static unsigned int hdr10_tm_sel = 2; /1 old algorithm, 2 hdr_tmo algorithm default 2/
	module_param(hdr10_tm_sel, uint, 0664);
	MODULE_PARM_DESC(hdr10_tm_sel, "hdr10_tm_sel\n");

	#define KNEE_POINT 2
	static unsigned int kp = KNEE_POINT;
	static unsigned int knee_point[KNEE_POINT] = {0, 0};
	module_param_array(knee_point, uint, &kp, 0664);
	MODULE_PARM_DESC(knee_point, "\n knee point xy\n");

	/*u32 ebz_p[MAX_BEIZER_ORDER] = {
	* 0, 1170, 1755, 2121, 2377, 2596, 2779, 2925, 3064, 3185,
	* 3320, 3514, 3799, 3957, 4095
	*};
	*/

	/time domain filter, avoid flicker/
	/scene change th: 2/3 scene diff/
	u32 sc_th = 682;
	module_param(sc_th, uint, 0664);
	MODULE_PARM_DESC(sc_th, "scene change th\n");

	u32 hdr_tm_iir = 1;
	module_param(hdr_tm_iir, uint, 0664);
	MODULE_PARM_DESC(hdr_tm_iir, "HDR_TM_IIR\n");

	int is_hdr_tmo_support(void)
	{
	if (get_cpu_type() >= MESON_CPU_MAJOR_ID_SM1)
	return 1;
	return 0;
	}

	int time_iir(u32 *maxl)
	{
	int i;
	static u32 hist_diff[3];
	static u32 max_lum[3];
	u32 diff;
	u32 sum = 0;
	int sumshft;
	int norm14;
	u32 bld_coef;
	u32 dtl_lum = 0;
	/ret=1: scn change; ret=2: blend/
	int ret;

	for (i = 0; i < 2; i++) {
	hist_diff[i] = hist_diff[i + 1];
	max_lum[i] = max_lum[i + 1];
	}
	hist_diff[2] = 0;
	max_lum[2] = *maxl;

	for (i = 0; i < 128; i++) {
	sum += hdr_hist[15][i];
	diff = (hdr_hist[15][i] > hdr_hist[14][i]) ?
	(hdr_hist[15][i] - hdr_hist[14][i]) :
	(hdr_hist[14][i] - hdr_hist[15][i]);
	hist_diff[2] += diff;
	}

	if (sum == 0)
	return 0;

	sumshft =
	(sum >= (1 << 24)) ? 8 :
	(sum >= (1 << 22)) ? 6 :
	(sum >= (1 << 20)) ? 4 :
	(sum >= (1 << 18)) ? 2 :
	(sum >= (1 << 16)) ? 0 :
	(sum >= (1 << 14)) ? -2 :
	(sum >= (1 << 12)) ? -4 :
	(sum >= (1 << 10)) ? -6 :
	(sum >= (1 << 8)) ? -8 :
	(sum >= (1 << 6)) ? -10 :
	(sum >= (1 << 4)) ? -12 : -16;

	if (sumshft >= 0)
	norm14 = (1 << 30) / (sum >> sumshft);
	else if (sumshft > -16)
	norm14 = (1 << (30 + sumshft)) / sum;
	else
	norm14 = 1 << 14;

	if (sumshft >= 0) {
	hist_diff[2] = ((hist_diff[2] >> sumshft) * norm14 +
	(1 << 13)) >> 14;
	} else {
	hist_diff[2] = (((hist_diff[2] << (-sumshft)) * norm14) +
	(1 << 13)) >> 14;
	}

	/normalize to 10bit/
	hist_diff[2] >>= 6;
	if (hist_diff[2] > sc_th \|\| (*maxl <= panell)) {
	ret = 1;
	} else {
	bld_coef = hist_diff[2];
	dtl_lum =
	(max_lum[2] > max_lum[1]) ?
	(max_lum[2] - max_lum[1]) :
	(max_lum[1] - max_lum[2]);

	if (dtl_lum < (1 << 3))
	max_lum[2] =
	(max_lum[2] > max_lum[1]) ?
	(max_lum[1] + (dtl_lum >> 1)) :
	(max_lum[1] - (dtl_lum >> 1));
	else if (dtl_lum < (1 << 6))
	max_lum[2] =
	(max_lum[2] > max_lum[1]) ?
	(max_lum[1] + (dtl_lum >> 3)) :
	(max_lum[1] - (dtl_lum >> 3));
	else
	max_lum[2] =
	(max_lum[2] > max_lum[1]) ?
	(max_lum[1] + (dtl_lum >> 6)) :
	(max_lum[1] - (dtl_lum >> 6));

	*maxl = max_lum[2];
	ret = 2;
	}

	pr_hdr_tm("ret = %d, dtl_lum = %d, max_lum[1] = %d, max_lum[2] = %d\n",
	ret, dtl_lum, max_lum[1], max_lum[2]);

	return ret;
	}

	/hdr10 tone mapping alg /
	int gen_oe_x_val(u64 *oe_x_val)
	{
	unsigned int i = 0, j = 0;
	int bin_num = 0;
	int temp1 = 0, temp2 = 0;

	oe_x_val[0] = 0;

	for (i = 1; i < 8; i++) {
	bin_num++;
	temp1 = 1 << ((i - 1) + 4);
	oe_x_val[i] = (u64)temp1;
	}
	for (j = 11; j < 20; j++) {
	for (i = 0; i < 4; i++) {
	bin_num++;
	temp1 = 1 << (j - 2);
	temp2 = 1 << j;
	oe_x_val[bin_num] = (u64)temp1 * i + (u64)temp2;
	}
	}
	for (j = 20; j < 31; j++) {
	for (i = 0; i < 8; i++) {
	bin_num++;
	temp1 = 1 << (j - 3);
	temp2 = 1 << j;
	oe_x_val[bin_num] = (u64)temp1 * i + (u64)temp2;
	}
	}
	for (i = 0; i < 16; i++) {
	bin_num++;
	oe_x_val[bin_num] = (1 << (31 - 4)) * i + (1 << 31);
	}
	oe_x_val[OE_X - 1] = MAX_32;

	return 0;
	}

	void decasteliau_alg(u64 ebzcur, u64 step, u64 range, u64 p, u32 order)
	{
	u64 point[MAX_BEIZER_ORDER + 1];
	int i, j;

	for (i = 0; i < order; i++)
	point[i] = p[i];

	for (i = 1; i < order; i++) {
	for (j = 0; j < order - i; j++) {
	point[j] = (range - step) * point[j] +
	step * point[j + 1] + (range >> 1);
	point[j] = div64_u64(point[j], range);
	}
	}
	ebzcur[1] = point[0];
	}

	static u64 curvex[OE_X], curvey[OE_X];
	static u32 oo_gain[OE_X];

	int gen_knee_anchor(u32 maxl, u32 panell, u64 sx, u64 sy, u64 *anchor)
	{
	int i;
	int knee_bypass = 1229;
	u64 anchor_linear[MAX_BEIZER_ORDER];
	u64 norm, blendcoef;
	//int tgtl = 10000;
	struct basisootf_params basisootf_param;
	u64 sx1, sy1;
	u64 anchor0;

	sx1 = *sx;
	sy1 = *sy;

	basisootf_params_init(&basisootf_param);

	for (i = 0; i < MAX_BEIZER_ORDER - 1; i++)
	anchor_linear[i] = (i + 1) *
	(1 << MAX12_BIT) / MAX_BEIZER_ORDER;

	pr_hdr_tm("knee anchor p1: %lld, %lld, %lld, %lld, %lld, %lld, %lld, %lld, %lld, %lld %lld\n",
	sx1, sy1,
	anchor_linear[0], anchor_linear[1],
	anchor_linear[2], anchor_linear[3],
	anchor_linear[4], anchor_linear[5],
	anchor_linear[6], anchor_linear[7],
	anchor_linear[8]);

	if (maxl <= panell) {
	sx1 = knee_bypass;
	sy1 = knee_bypass;
	for (i = 0; i < MAX_BEIZER_ORDER - 1; i++)
	anchor[i] = anchor_linear[i];
	} else {
	norm = maxl;
	blendcoef = panell;
	sy1 = (blendcoef * knee_bypass +
	(norm - blendcoef) * sy1 + (norm >> 1));
	sy1 = div64_u64(sy1, norm);
	sx1 = sy1 * panell;
	sx1 = div64_u64(sx1, maxl);
	for (i = 0; i < MAX_BEIZER_ORDER - 1; i++) {
	anchor[i] =
	(blendcoef * anchor_linear[i] +
	(norm - blendcoef) * anchor[i] +
	(norm >> 1));

	anchor[i] = div64_u64(anchor[i], norm);
	}

	anchor0 = calcp1((u32)sx1, (u32)sy1,
	panell, maxl, &basisootf_param, NULL);

	for (i = 1; i < MAX_BEIZER_ORDER - 1; i++)
	anchor[i] = min(anchor[i], (i + 1) * anchor0);

	if (anchor0 < anchor[1])
	anchor[0] = anchor0;
	}
	*sx = sx1;
	*sy = sy1;

	pr_hdr_tm("knee anchor p2: %lld, %lld, %lld, %lld, %lld, %lld, %lld, %lld, %lld, %lld %lld\n",
	sx1, sy1,
	anchor[0], anchor[1],
	anchor[2], anchor[3],
	anchor[4], anchor[5],
	anchor[6], anchor[7],
	anchor[8]);

	return 0;
	}

	int dynamic_hdr_sdr_ootf(u32 maxl, u32 panell, u64 sx, u64 sy, u64 *anchor)
	{
	int i;
	u32 tgtl = 10000;
	static u64 oe_x[OE_X] = {0};
	int order = MAX_BEIZER_ORDER + 1;
	u64 step;
	u64 kx, ky;
	u64 ebzcur[2];
	u64 range_ebz_x, range_ebz_y;
	u64 finalanchor[MAX_BEIZER_ORDER + 1];
	u64 temp;
	u64 sx_primy;
	u64 kx_primy;

	gen_oe_x_val(oe_x);

	gen_knee_anchor(maxl, panell, &sx, &sy, anchor);

	pr_hdr_tm("d1: %d, %d, %lld, %lld, %lld, %lld, %lld, %lld, %lld, %lld, %lld, %lld %lld\n",
	maxl, panell,
	sx, sy,
	anchor[0], anchor[1],
	anchor[2], anchor[3],
	anchor[4], anchor[5],
	anchor[6], anchor[7],
	anchor[8]);

	for (i = 0; i < order - 2; i++)
	finalanchor[i + 1] = anchor[i] << (MAX32_BIT - MAX12_BIT);

	finalanchor[0] = 0;
	finalanchor[10] = MAX_32;

	if (tgtl <= maxl)
	maxl = tgtl - 1;

	kx = sx << (MAX32_BIT - MAX12_BIT);
	ky = sy << (MAX32_BIT - MAX12_BIT);

	kx = kx * maxl;
	kx = div64_u64(kx, tgtl);

	if (maxl <= panell) {
	sx_primy = (u64)(1 << MAX12_BIT) * panell + (tgtl >> 1);
	sx_primy = div64_u64(sx_primy, tgtl);
	} else {
	sx_primy = (u64)(1 << MAX12_BIT) * maxl + (tgtl >> 1);
	sx_primy = div64_u64(sx_primy, tgtl);
	}
	kx_primy = sx_primy << (MAX32_BIT - MAX12_BIT);

	range_ebz_x = kx_primy - kx;
	range_ebz_y = MAX_32 - ky;

	if (maxl <= panell) {
	for (i = 0; i < OE_X; i++) {
	if (oe_x[i] <= kx_primy) {
	temp = tgtl << TM_GAIN_BIT;
	oo_gain[i] = div64_u64(temp, panell);
	} else {
	curvey[i] = oe_x[148] << TM_GAIN_BIT;
	curvex[i] = oe_x[i];
	oo_gain[i] = div64_u64(curvey[i], curvex[i]);
	}
	if (oo_gain[i] < (1 << TM_GAIN_BIT))
	oo_gain[i] = 1 << TM_GAIN_BIT;
	if (oo_gain[i] > (1 << MAX12_BIT) - 1)
	oo_gain[i] = (1 << MAX12_BIT) - 1;
	}
	} else {
	for (i = 0; i < OE_X; i++) {
	if (oe_x[i] <= kx) {
	temp = tgtl << TM_GAIN_BIT;
	oo_gain[i] = div64_u64(temp, panell);
	} else if (oe_x[i] <= kx_primy) {
	step = oe_x[i] - kx;
	decasteliau_alg(&ebzcur[0], step,
	range_ebz_x,
	finalanchor, order);
	curvey[i] = ky +
	((range_ebz_y * ebzcur[1] + MAX_32 / 2)
	>> MAX32_BIT);
	temp = curvey[i] << TM_GAIN_BIT;
	curvex[i] = kx + step;
	oo_gain[i] = div64_u64(temp, curvex[i]);
	} else {
	curvey[i] = oe_x[148] << TM_GAIN_BIT;
	curvex[i] = oe_x[i];
	oo_gain[i] = div64_u64(curvey[i], curvex[i]);
	}
	if (oo_gain[i] < (1 << TM_GAIN_BIT))
	oo_gain[i] = 1 << TM_GAIN_BIT;
	if (oo_gain[i] > (1 << MAX12_BIT) - 1)
	oo_gain[i] = (1 << MAX12_BIT) - 1;
	}
	}

	for (i = 0; i < OE_X; i++) {
	if ((i + 1) % 10 == 0)
	pr_hdr_tm("%d, %d, %d, %d, %d, %d, %d, %d, %d, %d\n",
	oo_gain[i - 9], oo_gain[i - 8],
	oo_gain[i - 7], oo_gain[i - 6],
	oo_gain[i - 5], oo_gain[i - 4],
	oo_gain[i - 3], oo_gain[i - 2],
	oo_gain[i - 1], oo_gain[i]);
	if (i == 148)
	pr_hdr_tm("%d, %d, %d, %d, %d, %d, %d, %d, %d\n",
	oo_gain[i - 8],
	oo_gain[i - 7], oo_gain[i - 6],
	oo_gain[i - 5], oo_gain[i - 4],
	oo_gain[i - 3], oo_gain[i - 2],
	oo_gain[i - 1], oo_gain[i]);
	}
	pr_hdr_tm("\n");

	return 0;
	}

	int hdr10_tm_dynamic_proc(struct vframe_master_display_colour_s *p)
	{
	int i;
	u32 maxl;
	u32 primry_maxl;
	u32 panel_luma;
	u64 sx, sy;
	u64 anchor[MAX_BEIZER_ORDER - 1];
	u32 P_init[MAX_BEIZER_ORDER - 1] = {
	181, 406, 607, 796, 834, 863, 890, 917, 938
	};
	int scn_chang_flag = 1;
	struct aml_tmo_reg_sw *pre_tmo_reg;

	if (p->luminance[0] > 10000)
	p->luminance[0] /= 10000;
	/no luminance/
	if (p->luminance[0] == 0)
	p->luminance[0] = 1200;

	primry_maxl = p->luminance[0];

	/use 95% maxl because of high percert flicker/
	maxl = (percentile[8] > primry_maxl) ? primry_maxl : percentile[8];
	pr_hdr_tm("maxl = %d, percentile[8] = %d, primry_maxl =%d\n",
	maxl, percentile[8], primry_maxl);

	if (hdr_tm_iir)
	scn_chang_flag = time_iir(&maxl);

	if (scn_chang_flag == 0)
	pr_hdr_tm("invalid hist\n");
	else if (scn_chang_flag == 1)
	pr_hdr_tm("scn change\n");
	else if (scn_chang_flag == 2)
	pr_hdr_tm("maxl blend: maxl = %d\n", maxl);

	panel_luma = panell;

	sx = knee_point[0];
	sy = knee_point[1];

	/invalid case/
	if (maxl == 0) {
	maxl = 1200;
	pr_hdr_tm("invalid maxl\n");
	}

	for (i = 0; i < MAX_BEIZER_ORDER - 1; i++)
	anchor[i] = P_init[i] << 2;

	pre_tmo_reg = tmo_fw_param_get();

	if (!is_hdr_tmo_support() \|\| !pre_tmo_reg->pre_hdr10_tmo_alg) {
	/used old hdr alg/
	pr_hdr_tm("used old hdr alg - ");
	if (!pre_tmo_reg->pre_hdr10_tmo_alg)
	pr_hdr_tm("hdr alg ko insmod failed.\n");
	else
	pr_hdr_tm("the chip is not support.\n");
	hdr10_tm_sel = 1;
	}

	if (hdr10_tm_sel == 1) {
	dynamic_hdr_sdr_ootf(maxl, panel_luma, sx, sy, anchor);
	memcpy(oo_y_lut_hdr_sdr, oo_gain, sizeof(u32) * OE_X);
	} else if (hdr10_tm_sel == 2) {
	hdr10_tmo_gen(oo_gain);
	memcpy(oo_y_lut_hdr_sdr, oo_gain, sizeof(u32) * OE_X);
	} else {
	memcpy(oo_y_lut_hdr_sdr, oo_y_lut_hdr_sdr_def,
	sizeof(u32) * OE_X);
	}

	if (hdr10_tm_dbg > 0)
	hdr10_tm_dbg--;

	return 0;
	}