Skip to content

Commit 35a10dc

Browse files
pborcinpborcin
authored
Merge pull request #480 from pborcin/rgb888_accel
feature(lvgl_port): RGB888 SIMD fill
2 parents e4b0144 + a2694c6 commit 35a10dc

12 files changed

+1610
-20
lines changed

‎components/esp_lvgl_port/CMakeLists.txt‎

Lines changed: 1 addition & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -98,6 +98,7 @@ if((lvgl_ver VERSION_GREATER_EQUAL "9.1.0") AND (lvgl_ver VERSION_LESS "9.2.0"))
9898
# Force link .S files
9999
set_property(TARGET ${COMPONENT_LIB} APPEND PROPERTY INTERFACE_LINK_LIBRARIES "-u lv_color_blend_to_argb8888_esp")
100100
set_property(TARGET ${COMPONENT_LIB} APPEND PROPERTY INTERFACE_LINK_LIBRARIES "-u lv_color_blend_to_rgb565_esp")
101+
set_property(TARGET ${COMPONENT_LIB} APPEND PROPERTY INTERFACE_LINK_LIBRARIES "-u lv_color_blend_to_rgb888_esp")
101102
set_property(TARGET ${COMPONENT_LIB} APPEND PROPERTY INTERFACE_LINK_LIBRARIES "-u lv_rgb565_blend_normal_to_rgb565_esp")
102103
endif()
103104
endif()

‎components/esp_lvgl_port/include/esp_lvgl_port_lv_blend.h‎

Lines changed: 24 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -1,5 +1,5 @@
11
/*
2-
* SPDX-FileCopyrightText: 2024 Espressif Systems (Shanghai) CO LTD
2+
* SPDX-FileCopyrightText: 2024-2025 Espressif Systems (Shanghai) CO LTD
33
*
44
* SPDX-License-Identifier: Apache-2.0
55
*/
@@ -32,6 +32,11 @@ extern "C" {
3232
_lv_color_blend_to_rgb565_esp(dsc)
3333
#endif
3434

35+
#ifndef LV_DRAW_SW_COLOR_BLEND_TO_RGB888
36+
#define LV_DRAW_SW_COLOR_BLEND_TO_RGB888(dsc, dest_px_size) \
37+
_lv_color_blend_to_rgb888_esp(dsc, dest_px_size)
38+
#endif
39+
3540
#ifndef LV_DRAW_SW_RGB565_BLEND_NORMAL_TO_RGB565
3641
#define LV_DRAW_SW_RGB565_BLEND_NORMAL_TO_RGB565(dsc) \
3742
_lv_rgb565_blend_normal_to_rgb565_esp(dsc)
@@ -87,6 +92,24 @@ static inline lv_result_t _lv_color_blend_to_rgb565_esp(_lv_draw_sw_blend_fill_d
8792
return lv_color_blend_to_rgb565_esp(&asm_dsc);
8893
}
8994

95+
extern int lv_color_blend_to_rgb888_esp(asm_dsc_t *asm_dsc);
96+
97+
static inline lv_result_t _lv_color_blend_to_rgb888_esp(_lv_draw_sw_blend_fill_dsc_t *dsc, uint32_t dest_px_size)
98+
{
99+
if (dest_px_size != 3) {
100+
return LV_RESULT_INVALID;
101+
}
102+
asm_dsc_t asm_dsc = {
103+
.dst_buf = dsc->dest_buf,
104+
.dst_w = dsc->dest_w,
105+
.dst_h = dsc->dest_h,
106+
.dst_stride = dsc->dest_stride,
107+
.src_buf = &dsc->color,
108+
};
109+
110+
return lv_color_blend_to_rgb888_esp(&asm_dsc);
111+
}
112+
90113
extern int lv_rgb565_blend_normal_to_rgb565_esp(asm_dsc_t *asm_dsc);
91114

92115
static inline lv_result_t _lv_rgb565_blend_normal_to_rgb565_esp(_lv_draw_sw_blend_image_dsc_t *dsc)

‎components/esp_lvgl_port/src/lvgl9/simd/lv_color_blend_to_argb8888_esp32s3.S‎

Lines changed: 1 addition & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -32,8 +32,7 @@
3232

3333
lv_color_blend_to_argb8888_esp:
3434

35-
entry a1, 32
36-
ee.zero.q q0 // dummy TIE instruction, to enable the TIE
35+
entry a1, 32
3736

3837
l32i.n a3, a2, 4 // a3 - dest_buff
3938
l32i.n a4, a2, 8 // a4 - dest_w in uint32_t

‎components/esp_lvgl_port/src/lvgl9/simd/lv_color_blend_to_rgb565_esp32s3.S‎

Lines changed: 1 addition & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -31,8 +31,7 @@
3131

3232
lv_color_blend_to_rgb565_esp:
3333

34-
entry a1, 32
35-
ee.zero.q q0 // dummy TIE instruction, to enable the TIE
34+
entry a1, 32
3635

3736
l32i.n a3, a2, 4 // a3 - dest_buff
3837
l32i.n a4, a2, 8 // a4 - dest_w in uint16_t

‎components/esp_lvgl_port/src/lvgl9/simd/lv_color_blend_to_rgb888_esp32.S‎

Lines changed: 105 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,105 @@
1+
/*
2+
* SPDX-FileCopyrightText: 2024 Espressif Systems (Shanghai) CO LTD
3+
*
4+
* SPDX-License-Identifier: Apache-2.0
5+
*/
6+
7+
// This is LVGL RGB888 simple fill for ESP32 processor
8+
9+
.section .text
10+
.align 4
11+
.global lv_color_blend_to_rgb888_esp
12+
.type lv_color_blend_to_rgb888_esp,@function
13+
// The function implements the following C code:
14+
// void lv_color_blend_to_rgb888(_lv_draw_sw_blend_fill_dsc_t * dsc);
15+
16+
// Input params
17+
//
18+
// dsc - a2
19+
20+
// typedef struct {
21+
// uint32_t opa; l32i 0
22+
// void * dst_buf; l32i 4
23+
// uint32_t dst_w; l32i 8
24+
// uint32_t dst_h; l32i 12
25+
// uint32_t dst_stride; l32i 16
26+
// const void * src_buf; l32i 20
27+
// uint32_t src_stride; l32i 24
28+
// const lv_opa_t * mask_buf; l32i 28
29+
// uint32_t mask_stride; l32i 32
30+
// } asm_dsc_t;
31+
32+
lv_color_blend_to_rgb888_esp:
33+
34+
entry a1, 32
35+
36+
l32i.n a3, a2, 4 // a3 - dest_buff
37+
l32i.n a4, a2, 8 // a4 - dest_w in uint24_t
38+
l32i.n a5, a2, 12 // a5 - dest_h in uint16_t
39+
l32i.n a6, a2, 16 // a6 - dest_stride in bytes
40+
l32i.n a7, a2, 20 // a7 - src_buff (color)
41+
l32i.n a8, a7, 0 // a8 - color as value
42+
43+
// a11 - dest_w_bytes = sizeof(uint24_t) * dest_w = 3 * a4
44+
slli a11, a4, 1 // a11 - dest_w_bytes = sizeof(uint16_t) * dest_w
45+
add a11, a11, a4 // a11 - dest_w_bytes = a11 + a4
46+
47+
// Prepare register combinations
48+
// a13 - 0xBBRRGGBB a14 - 0xGGBBRRGG a15 - 0xRRGGBBRR
49+
l8ui a13, a7, 0 // blue 000B
50+
slli a13, a13, 24 // shift to B000
51+
or a13, a13, a8 // a13 BRGB
52+
53+
srli a14, a8, 8 // a14 00RG
54+
slli a10, a8, 16 // a10 GB00
55+
or a14, a14, a10 // a14 GBRG
56+
57+
slli a15, a8, 8 // a15 RGB0
58+
l8ui a10, a7, 2 // a7 000R
59+
or a15, a15, a10 // a15 RGBR
60+
61+
sub a6, a6, a11 // dest_stride = dest_stride - dest_w_bytes
62+
63+
// Prepare main loop length and dest_w_bytes
64+
srli a9, a4, 2 // a9 = loop_len = dest_w / 4, calculate main loop_len for original dest_w
65+
movi.n a8, 0x3 // a8 = 0x3, remainder mask
66+
and a10, a4, a8 // a10 - remainder after division by 4 = a4 and 0x3
67+
68+
.outer_loop:
69+
70+
// Run main loop which sets 12 bytes (4 rgb888) in one loop run
71+
loopnez a9, ._main_loop
72+
s32i.n a13, a3, 0 // save 32 bits from 32-bit color a13 to dest_buff a3, offset 0
73+
s32i.n a14, a3, 4 // save 32 bits from 32-bit color a14 to dest_buff a3, offset 4
74+
s32i.n a15, a3, 8 // save 32 bits from 32-bit color a15 to dest_buff a3, offset 8
75+
addi.n a3, a3, 12 // increment dest_buff pointer by 12
76+
._main_loop:
77+
78+
bnei a10, 0x3, _less_than_3 // branch if less than 3 values left
79+
s32i.n a13, a3, 0 // save 32 bits from a13 to dest_buff a3, offset 0 bytes
80+
s32i.n a14, a3, 4 // save 32 bits from a14 to dest_buff a3, offset 4 bytes
81+
s8i a15, a3, 8 // save 8 bits from a15 to dest_buff a3, offset 8 bytes
82+
addi.n a3, a3, 9 // increment dest_buff pointer by 9 bytes
83+
j _less_than_1
84+
_less_than_3:
85+
86+
bnei a10, 0x2, _less_than_2 // branch if less than 2 values left
87+
s32i.n a13, a3, 0 // save 32 bits from a13 to dest_buff a3, offset 0 bytes
88+
s16i a14, a3, 4 // save 16 bits from a14 to dest_buff a3, offset 4 bytes
89+
addi.n a3, a3, 6 // increment dest_buff pointer by 6 bytes
90+
j _less_than_1
91+
_less_than_2:
92+
93+
bnei a10, 0x1, _less_than_1 // branch if less than 1 value left
94+
s16i a13, a3, 0 // save 16 bits from a13 to dest_buff a3, offset 0 bytes
95+
s8i a15, a3, 2 // save 8 bits from a15 to dest_buff a3, offset 2 bytes
96+
addi.n a3, a3, 3 // increment dest_buff pointer by 3 bytes
97+
_less_than_1:
98+
99+
add a3, a3, a6 // dest_buff + dest_stride
100+
addi.n a5, a5, -1 // decrease the outer loop
101+
and a7, a8, a3 // a7 = dest_buff AND 0x3 (check if the address is 4-byte aligned)
102+
bnez a5, .outer_loop
103+
104+
movi.n a2, 1 // return LV_RESULT_OK = 1
105+
retw.n // return

0 commit comments

Comments
 (0)
⚡