-
Notifications
You must be signed in to change notification settings - Fork 120
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
feature(lvgl_port): RGB888 SIMD fill
- Loading branch information
Showing
12 changed files
with
1,613 additions
and
19 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
105 changes: 105 additions & 0 deletions
105
components/esp_lvgl_port/src/lvgl9/simd/lv_color_blend_to_rgb888_esp32.S
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,105 @@ | ||
| /* | ||
| * SPDX-FileCopyrightText: 2024 Espressif Systems (Shanghai) CO LTD | ||
| * | ||
| * SPDX-License-Identifier: Apache-2.0 | ||
| */ | ||
|
|
||
| // This is LVGL RGB888 simple fill for ESP32 processor | ||
|
|
||
| .section .text | ||
| .align 4 | ||
| .global lv_color_blend_to_rgb888_esp | ||
| .type lv_color_blend_to_rgb888_esp,@function | ||
| // The function implements the following C code: | ||
| // void lv_color_blend_to_rgb888(_lv_draw_sw_blend_fill_dsc_t * dsc); | ||
|
|
||
| // Input params | ||
| // | ||
| // dsc - a2 | ||
|
|
||
| // typedef struct { | ||
| // uint32_t opa; l32i 0 | ||
| // void * dst_buf; l32i 4 | ||
| // uint32_t dst_w; l32i 8 | ||
| // uint32_t dst_h; l32i 12 | ||
| // uint32_t dst_stride; l32i 16 | ||
| // const void * src_buf; l32i 20 | ||
| // uint32_t src_stride; l32i 24 | ||
| // const lv_opa_t * mask_buf; l32i 28 | ||
| // uint32_t mask_stride; l32i 32 | ||
| // } asm_dsc_t; | ||
|
|
||
| lv_color_blend_to_rgb888_esp: | ||
|
|
||
| entry a1, 32 | ||
|
|
||
| l32i.n a3, a2, 4 // a3 - dest_buff | ||
| l32i.n a4, a2, 8 // a4 - dest_w in uint24_t | ||
| l32i.n a5, a2, 12 // a5 - dest_h in uint16_t | ||
| l32i.n a6, a2, 16 // a6 - dest_stride in bytes | ||
| l32i.n a7, a2, 20 // a7 - src_buff (color) | ||
| l32i.n a8, a7, 0 // a8 - color as value | ||
|
|
||
| // a11 - dest_w_bytes = sizeof(uint24_t) * dest_w = 3 * a4 | ||
| slli a11, a4, 1 // a11 - dest_w_bytes = sizeof(uint16_t) * dest_w | ||
| add a11, a11, a4 // a11 - dest_w_bytes = a11 + a4 | ||
|
|
||
| // Prepare register combinations | ||
| // a13 - 0xBBRRGGBB a14 - 0xGGBBRRGG a15 - 0xRRGGBBRR | ||
| l8ui a13, a7, 0 // blue 000B | ||
| slli a13, a13, 24 // shift to B000 | ||
| or a13, a13, a8 // a13 BRGB | ||
|
|
||
| srli a14, a8, 8 // a14 00RG | ||
| slli a10, a8, 16 // a10 GB00 | ||
| or a14, a14, a10 // a14 GBRG | ||
|
|
||
| slli a15, a8, 8 // a15 RGB0 | ||
| l8ui a10, a7, 2 // a7 000R | ||
| or a15, a15, a10 // a15 RGBR | ||
|
|
||
| sub a6, a6, a11 // dest_stride = dest_stride - dest_w_bytes | ||
|
|
||
| // Prepare main loop length and dest_w_bytes | ||
| srli a9, a4, 2 // a9 = loop_len = dest_w / 4, calculate main loop_len for original dest_w | ||
| movi.n a8, 0x3 // a8 = 0x3, remainder mask | ||
| and a10, a4, a8 // a10 - remainder after division by 4 = a4 and 0x3 | ||
|
|
||
| .outer_loop: | ||
|
|
||
| // Run main loop which sets 12 bytes (4 rgb888) in one loop run | ||
| loopnez a9, ._main_loop | ||
| s32i.n a13, a3, 0 // save 32 bits from 32-bit color a13 to dest_buff a3, offset 0 | ||
| s32i.n a14, a3, 4 // save 32 bits from 32-bit color a14 to dest_buff a3, offset 4 | ||
| s32i.n a15, a3, 8 // save 32 bits from 32-bit color a15 to dest_buff a3, offset 8 | ||
| addi.n a3, a3, 12 // increment dest_buff pointer by 12 | ||
| ._main_loop: | ||
|
|
||
| bnei a10, 0x3, _less_than_3 // branch if less than 3 values left | ||
| s32i.n a13, a3, 0 // save 32 bits from a13 to dest_buff a3, offset 0 bytes | ||
| s32i.n a14, a3, 4 // save 32 bits from a14 to dest_buff a3, offset 4 bytes | ||
| s8i a15, a3, 8 // save 8 bits from a15 to dest_buff a3, offset 8 bytes | ||
| addi.n a3, a3, 9 // increment dest_buff pointer by 9 bytes | ||
| j _less_than_1 | ||
| _less_than_3: | ||
|
|
||
| bnei a10, 0x2, _less_than_2 // branch if less than 2 values left | ||
| s32i.n a13, a3, 0 // save 32 bits from a13 to dest_buff a3, offset 0 bytes | ||
| s16i a14, a3, 4 // save 16 bits from a14 to dest_buff a3, offset 4 bytes | ||
| addi.n a3, a3, 6 // increment dest_buff pointer by 6 bytes | ||
| j _less_than_1 | ||
| _less_than_2: | ||
|
|
||
| bnei a10, 0x1, _less_than_1 // branch if less than 1 value left | ||
| s16i a13, a3, 0 // save 16 bits from a13 to dest_buff a3, offset 0 bytes | ||
| s8i a15, a3, 2 // save 8 bits from a15 to dest_buff a3, offset 2 bytes | ||
| addi.n a3, a3, 3 // increment dest_buff pointer by 3 bytes | ||
| _less_than_1: | ||
|
|
||
| add a3, a3, a6 // dest_buff + dest_stride | ||
| addi.n a5, a5, -1 // decrease the outer loop | ||
| and a7, a8, a3 // a7 = dest_buff AND 0x3 (check if the address is 4-byte aligned) | ||
| bnez a5, .outer_loop | ||
|
|
||
| movi.n a2, 1 // return LV_RESULT_OK = 1 | ||
| retw.n // return |
Oops, something went wrong.