/*
* Copyright 2016 Advanced Micro Devices , Inc .
*
* Permission is hereby granted , free of charge , to any person obtaining a
* copy of this software and associated documentation files ( the " Software " ) ,
* to deal in the Software without restriction , including without limitation
* the rights to use , copy , modify , merge , publish , distribute , sublicense ,
* and / or sell copies of the Software , and to permit persons to whom the
* Software is furnished to do so , subject to the following conditions :
*
* The above copyright notice and this permission notice shall be included in
* all copies or substantial portions of the Software .
*
* THE SOFTWARE IS PROVIDED " AS IS " , WITHOUT WARRANTY OF ANY KIND , EXPRESS OR
* IMPLIED , INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY ,
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT . IN NO EVENT SHALL
* THE COPYRIGHT HOLDER ( S ) OR AUTHOR ( S ) BE LIABLE FOR ANY CLAIM , DAMAGES OR
* OTHER LIABILITY , WHETHER IN AN ACTION OF CONTRACT , TORT OR OTHERWISE ,
* ARISING FROM , OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
* OTHER DEALINGS IN THE SOFTWARE .
*
* Authors : AMD
*
*/
#include "dm_services.h"
#include "core_types.h"
#include "reg_helper.h"
#include "dcn401/dcn401_dpp.h"
#include "basics/conversion.h"
#include "dcn10/dcn10_cm_common.h"
#define NUM_PHASES 64
#define HORZ_MAX_TAPS 8
#define VERT_MAX_TAPS 8
#define BLACK_OFFSET_RGB_Y 0 x0
#define BLACK_OFFSET_CBCR 0 x8000
#define REG(reg)\
dpp->tf_regs->reg
#define CTX \
dpp->base.ctx
#undef FN
#define FN(reg_name, field_name) \
dpp->tf_shift->field_name, dpp->tf_mask->field_name
#define NUM_ELEMENTS(a) (sizeof (a) / sizeof ((a)[0 ]))
enum dcn401_coef_filter_type_sel {
SCL_COEF_LUMA_VERT_FILTER = 0 ,
SCL_COEF_LUMA_HORZ_FILTER = 1 ,
SCL_COEF_CHROMA_VERT_FILTER = 2 ,
SCL_COEF_CHROMA_HORZ_FILTER = 3 ,
SCL_COEF_SC_VERT_FILTER = 4 ,
SCL_COEF_SC_HORZ_FILTER = 5
};
enum dscl_autocal_mode {
AUTOCAL_MODE_OFF = 0 ,
/* Autocal calculate the scaling ratio and initial phase and the
* DSCL_MODE_SEL must be set to 1
*/
AUTOCAL_MODE_AUTOSCALE = 1 ,
/* Autocal perform auto centering without replication and the
* DSCL_MODE_SEL must be set to 0
*/
AUTOCAL_MODE_AUTOCENTER = 2 ,
/* Autocal perform auto centering and auto replication and the
* DSCL_MODE_SEL must be set to 0
*/
AUTOCAL_MODE_AUTOREPLICATE = 3
};
enum dscl_mode_sel {
DSCL_MODE_SCALING_444_BYPASS = 0 ,
DSCL_MODE_SCALING_444_RGB_ENABLE = 1 ,
DSCL_MODE_SCALING_444_YCBCR_ENABLE = 2 ,
DSCL_MODE_SCALING_YCBCR_ENABLE = 3 ,
DSCL_MODE_LUMA_SCALING_BYPASS = 4 ,
DSCL_MODE_CHROMA_SCALING_BYPASS = 5 ,
DSCL_MODE_DSCL_BYPASS = 6
};
void dpp401_set_cursor_attributes(
struct dpp *dpp_base,
struct dc_cursor_attributes *cursor_attributes)
{
struct dcn401_dpp *dpp = TO_DCN401_DPP(dpp_base);
enum dc_cursor_color_format color_format = cursor_attributes->color_format;
int cur_rom_en = 0 ;
if (color_format == CURSOR_MODE_COLOR_PRE_MULTIPLIED_ALPHA ||
color_format == CURSOR_MODE_COLOR_UN_PRE_MULTIPLIED_ALPHA) {
if (cursor_attributes->attribute_flags.bits.ENABLE_CURSOR_DEGAMMA) {
cur_rom_en = 1 ;
}
}
REG_UPDATE_3(CURSOR0_CONTROL,
CUR0_MODE, color_format,
CUR0_EXPANSION_MODE, 0 ,
CUR0_ROM_EN, cur_rom_en);
if (color_format == CURSOR_MODE_MONO) {
/* todo: clarify what to program these to */
REG_UPDATE(CURSOR0_COLOR0,
CUR0_COLOR0, 0 x00000000);
REG_UPDATE(CURSOR0_COLOR1,
CUR0_COLOR1, 0 xFFFFFFFF);
}
dpp_base->att.cur0_ctl.bits.expansion_mode = 0 ;
dpp_base->att.cur0_ctl.bits.cur0_rom_en = cur_rom_en;
dpp_base->att.cur0_ctl.bits.mode = color_format;
}
void dpp401_set_cursor_position(
struct dpp *dpp_base,
const struct dc_cursor_position *pos,
const struct dc_cursor_mi_param *param,
uint32_t width,
uint32_t height)
{
struct dcn401_dpp *dpp = TO_DCN401_DPP(dpp_base);
uint32_t cur_en = pos->enable ? 1 : 0 ;
if (dpp_base->pos.cur0_ctl.bits.cur0_enable != cur_en) {
REG_UPDATE(CURSOR0_CONTROL, CUR0_ENABLE, cur_en);
dpp_base->pos.cur0_ctl.bits.cur0_enable = cur_en;
}
}
void dpp401_set_optional_cursor_attributes(
struct dpp *dpp_base,
struct dpp_cursor_attributes *attr)
{
struct dcn401_dpp *dpp = TO_DCN401_DPP(dpp_base);
if (attr) {
REG_UPDATE(CURSOR0_FP_SCALE_BIAS_G_Y, CUR0_FP_BIAS_G_Y, attr->bias);
REG_UPDATE(CURSOR0_FP_SCALE_BIAS_G_Y, CUR0_FP_SCALE_G_Y, attr->scale);
REG_UPDATE(CURSOR0_FP_SCALE_BIAS_RB_CRCB, CUR0_FP_BIAS_RB_CRCB, attr->bias);
REG_UPDATE(CURSOR0_FP_SCALE_BIAS_RB_CRCB, CUR0_FP_SCALE_RB_CRCB, attr->scale);
}
}
/* Program Cursor matrix block in DPP CM */
static void dpp401_program_cursor_csc(
struct dpp *dpp_base,
enum dc_color_space color_space,
const struct dpp_input_csc_matrix *tbl_entry)
{
struct dcn401_dpp *dpp = TO_DCN401_DPP(dpp_base);
uint32_t mode_select = 0 ;
struct color_matrices_reg cur_matrix_regs;
unsigned int i;
const uint16_t *regval = NULL;
int arr_size = sizeof (dpp_input_csc_matrix) / sizeof (struct dpp_input_csc_matrix);
if (color_space < COLOR_SPACE_YCBCR601) {
REG_SET(CUR0_MATRIX_MODE, 0 , CUR0_MATRIX_MODE, CUR_MATRIX_BYPASS);
return ;
}
/* If adjustments not provided use hardcoded table for color space conversion */
if (tbl_entry == NULL) {
for (i = 0 ; i < arr_size; i++)
if (dpp_input_csc_matrix[i].color_space == color_space) {
regval = dpp_input_csc_matrix[i].regval;
break ;
}
if (regval == NULL) {
BREAK_TO_DEBUGGER();
REG_SET(CUR0_MATRIX_MODE, 0 , CUR0_MATRIX_MODE, CUR_MATRIX_BYPASS);
return ;
}
} else {
regval = tbl_entry->regval;
}
REG_GET(CUR0_MATRIX_MODE, CUR0_MATRIX_MODE_CURRENT, &mode_select);
//If current set in use not set A, then use set A, otherwise use set B
if (mode_select != CUR_MATRIX_SET_A)
mode_select = CUR_MATRIX_SET_A;
else
mode_select = CUR_MATRIX_SET_B;
cur_matrix_regs.shifts.csc_c11 = dpp->tf_shift->CUR0_MATRIX_C11_A;
cur_matrix_regs.masks.csc_c11 = dpp->tf_mask->CUR0_MATRIX_C11_A;
cur_matrix_regs.shifts.csc_c12 = dpp->tf_shift->CUR0_MATRIX_C12_A;
cur_matrix_regs.masks.csc_c12 = dpp->tf_mask->CUR0_MATRIX_C12_A;
if (mode_select == CUR_MATRIX_SET_A) {
cur_matrix_regs.csc_c11_c12 = REG(CUR0_MATRIX_C11_C12_A);
cur_matrix_regs.csc_c33_c34 = REG(CUR0_MATRIX_C33_C34_A);
} else {
cur_matrix_regs.csc_c11_c12 = REG(CUR0_MATRIX_C11_C12_B);
cur_matrix_regs.csc_c33_c34 = REG(CUR0_MATRIX_C33_C34_B);
}
cm_helper_program_color_matrices(
dpp->base.ctx,
regval,
&cur_matrix_regs);
//select coefficient set to use
REG_SET(CUR0_MATRIX_MODE, 0 , CUR0_MATRIX_MODE, mode_select);
}
/* Program Cursor matrix block in DPP CM */
void dpp401_set_cursor_matrix(
struct dpp *dpp_base,
enum dc_color_space color_space,
struct dc_csc_transform cursor_csc_color_matrix)
{
//Since we don't have cursor matrix information, force bypass mode by passing in unknown color space
dpp401_program_cursor_csc(dpp_base, COLOR_SPACE_UNKNOWN, NULL);
}
Messung V0.5 in Prozent C=97 H=97 G=96
¤ Dauer der Verarbeitung: 0.7 Sekunden
(vorverarbeitet am 2026-10-12)
¤
*© Formatika GbR, Deutschland