如何在camera 驱动中配置HDR的曝光模块

（仅作记录）

1. 平台：rk3588

2. sensor：imx415

3. 配置所需文件：

IMX415_AppNote_DOL_E_Rev1.0

IMX415-AAQR-C_TechnicalDatasheet_E_Rev0.4

HDR功能，其曝光增益和线性模式不一样，有各种的计算公式以及逻辑条件限制，具体看datasheet里面的restriction。

以HDR2为例，首先看下datasheet里的 restriction 如下图所示；

 

 上面那两张图就是配置hdrae所需要的，此外可以先理解一下RHS1 SHR0 SHR1，看下图

 可以如下简单理解：

FSC = vts 

SHR0= FSC - l_exp_time ；

RHS1= s_exp_time + SHR1 ；

接下来看下代码，根据图中的限制条件，一一列出；其实也挺简单的。

	/* Restrictions
	 *   FSC = 2 * VMAX and FSC should be 4n;
	 *   exp_l = FSC - SHR0 + Toffset;
	 *   exp_l should be even value;
	 *
	 *   SHR0 = FSC - exp_l + Toffset;
	 *   SHR0 <= (FSC -8);
	 *   SHR0 >= RHS1 + 9;
	 *   SHR0 should be 2n;
	 *
	 *   exp_s = RHS1 - SHR1 + Toffset;
	 *   exp_s should be even value;
	 *
	 *   RHS1 < BRL * 2;
	 *   RHS1 <= SHR0 - 9;
	 *   RHS1 >= SHR1 + 8;
	 *   SHR1 >= 9;
	 *   RHS1(n+1) >= RHS1(n) + BRL * 2 -FSC + 2;
	 *
	 *   SHR1 should be 2n+1 and RHS1 should be 4n+1;
	 */

	/* The HDR mode vts is double by default to workaround T-line */

	fsc = imx415->cur_vts;
	shr0 = fsc - l_exp_time;

	if (imx415->cur_mode->height == 2192) {
		rhs1_max = min(RHS1_MAX_X2(BRL_ALL), ((shr0 - 9u) / 4 * 4 + 1));
		rhs1_min = max(SHR1_MIN_X2 + 8u, rhs1_old + 2 * BRL_ALL - fsc + 2);
	} else {
		rhs1_max = min(RHS1_MAX_X2(BRL_BINNING), ((shr0 - 9u) / 4 * 4 + 1));
		rhs1_min = max(SHR1_MIN_X2 + 8u, rhs1_old + 2 * BRL_BINNING - fsc + 2);
	}
	rhs1_min = (rhs1_min + 3) / 4 * 4 + 1;
	rhs1 = (SHR1_MIN_X2 + s_exp_time + 3) / 4 * 4 + 1;/* shall be 4n + 1 */
	dev_dbg(&client->dev,
		"line(%d) rhs1 %d, rhs1 min %d rhs1 max %d\n",
		__LINE__, rhs1, rhs1_min, rhs1_max);
	if (rhs1_max < rhs1_min) {
		dev_err(&client->dev,
			"The total exposure limit makes rhs1 max is %d,but old rhs1 limit makes rhs1 min is %d\n",rhs1_max, rhs1_min);
		return -EINVAL;
	}
	rhs1 = clamp(rhs1, rhs1_min, rhs1_max);
	dev_dbg(&client->dev,
		"line(%d) rhs1 %d, short time %d rhs1_old %d, rhs1_new %d\n",
		__LINE__, rhs1, s_exp_time, rhs1_old, rhs1);

	rhs1_old = rhs1;

	if (rhs1 - s_exp_time <= SHR1_MIN_X2) {
		shr1 = SHR1_MIN_X2;
		s_exp_time = rhs1 - shr1;
	} else {
		shr1 = rhs1 - s_exp_time;
	}

	if (shr0 < rhs1 + 9)
		shr0 = rhs1 + 9;
	else if (shr0 > fsc - 8)
		shr0 = fsc - 8;

注释：
rhs1_min = (rhs1_min + 3) / 4 * 4 + 1;
rhs1 = (SHR1_MIN_X2 + s_exp_time + 3) / 4 * 4 + 1;/* shall be 4n + 1 */
像rhs1_min + 3主要是4对齐，建议往大的对齐，例如 （11+3）/4*4+1=12+1; 11/4*4+1=8+1; 
所以11当然是往大的12对齐而不是8.
 

提一下线性模式的曝光配置，不同sensor，写法不一样的；有的sensor是直接通过shr0来设置曝光的（shr寄存器）；有的直接像曝光寄存器写入曝光值。如下分别用imx335和imx378为例

case V4L2_CID_EXPOSURE:
		/* 4 least significant bits of expsoure are fractional part */
		ret = imx577_write_reg(imx577->client,
					IMX577_REG_EXPOSURE,
					IMX577_REG_VALUE_16BIT,
					ctrl->val);
		break;

case V4L2_CID_EXPOSURE:
		if (imx335->cur_mode->hdr_mode != NO_HDR)
			return ret;
		shr0 = imx335->cur_vts - ctrl->val;
		ret = imx335_write_reg(imx335->client, IMX335_LF_EXPO_REG_L,
				       IMX335_REG_VALUE_08BIT,
				       IMX335_FETCH_EXP_L(shr0));
		ret |= imx335_write_reg(imx335->client, IMX335_LF_EXPO_REG_M,
				       IMX335_REG_VALUE_08BIT,
				       IMX335_FETCH_EXP_M(shr0));
		ret |= imx335_write_reg(imx335->client, IMX335_LF_EXPO_REG_H,
				       IMX335_REG_VALUE_08BIT,
				       IMX335_FETCH_EXP_H(shr0));
		dev_dbg(&client->dev, "set exposure(shr0) %d = cur_vts(%d) - val(%d)\n",
			shr0, imx335->cur_vts, ctrl->val);
		break;

以上需要自己去体会哈！