github-code/lvgl
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

speed up true double buffering

Browse Source
This commit is contained in:
Gabor Kiss-Vamosi
2020-05-27 16:00:56 +02:00
parent 306405bcff
commit 70162562b2
1 changed files with 24 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files

28
src/lv_core/lv_refr.c
View File

@@ -17,6 +17,7 @@
#include "../lv_misc/lv_gc.h" #include "../lv_misc/lv_gc.h"
#include "../lv_draw/lv_draw.h" #include "../lv_draw/lv_draw.h"
#include "../lv_font/lv_font_fmt_txt.h" #include "../lv_font/lv_font_fmt_txt.h"
#include "../lv_gpu/lv_gpu_stm32_dma2d.h"
#if LV_USE_PERF_MONITOR #if LV_USE_PERF_MONITOR
#include "../lv_widgets/lv_label.h" #include "../lv_widgets/lv_label.h"
@@ -210,8 +211,14 @@ void _lv_disp_refr_task(lv_task_t * task)
/* With true double buffering the flushing should be only the address change of the /* With true double buffering the flushing should be only the address change of the
* current frame buffer. Wait until the address change is ready and copy the changed * current frame buffer. Wait until the address change is ready and copy the changed
* content to the other frame buffer (new active VDB) to keep the buffers synchronized*/ * content to the other frame buffer (new active VDB) to keep the buffers synchronized*/
while(vdb->flushing) while(vdb->flushing);
;
lv_color_t * copy_buf = NULL;
#if LV_USE_GPU_STM32_DMA2D
LV_UNUSED(copy_buf);
#else
copy_buf = _lv_mem_buf_get(disp_refr->driver.hor_res * sizeof(lv_color_t));
#endif
uint8_t * buf_act = (uint8_t *)vdb->buf_act; uint8_t * buf_act = (uint8_t *)vdb->buf_act;
uint8_t * buf_ina = (uint8_t *)vdb->buf_act == vdb->buf1 ? vdb->buf2 : vdb->buf1; uint8_t * buf_ina = (uint8_t *)vdb->buf_act == vdb->buf1 ? vdb->buf2 : vdb->buf1;
@@ -220,17 +227,30 @@ void _lv_disp_refr_task(lv_task_t * task)
uint16_t a; uint16_t a;
for(a = 0; a < disp_refr->inv_p; a++) { for(a = 0; a < disp_refr->inv_p; a++) {
if(disp_refr->inv_area_joined[a] == 0) { if(disp_refr->inv_area_joined[a] == 0) {
lv_coord_t y;
uint32_t start_offs = uint32_t start_offs =
(hres * disp_refr->inv_areas[a].y1 + disp_refr->inv_areas[a].x1) * sizeof(lv_color_t); (hres * disp_refr->inv_areas[a].y1 + disp_refr->inv_areas[a].x1) * sizeof(lv_color_t);
#if LV_USE_GPU_STM32_DMA2D
lv_gpu_stm32_dma2d_copy((lv_color_t *)buf_act + start_offs, disp_refr->driver.hor_res,
(lv_color_t *)buf_ina + start_offs, disp_refr->driver.hor_res,
lv_area_get_width(&disp_refr->inv_areas[a]),
lv_area_get_height(&disp_refr->inv_areas[a]));
#else
lv_coord_t y;
uint32_t line_length = lv_area_get_width(&disp_refr->inv_areas[a]) * sizeof(lv_color_t); uint32_t line_length = lv_area_get_width(&disp_refr->inv_areas[a]) * sizeof(lv_color_t);
for(y = disp_refr->inv_areas[a].y1; y <= disp_refr->inv_areas[a].y2; y++) { for(y = disp_refr->inv_areas[a].y1; y <= disp_refr->inv_areas[a].y2; y++) {
_lv_memcpy(buf_act + start_offs, buf_ina + start_offs, line_length); /* The frame buffer is probably in an external RAM where sequential access is much faster.
* So first copy a line into a buffer and write it back the ext. RAM */
_lv_memcpy(copy_buf, buf_ina + start_offs, line_length);
_lv_memcpy(buf_act + start_offs, copy_buf, line_length);
start_offs += hres * sizeof(lv_color_t); start_offs += hres * sizeof(lv_color_t);
} }
#endif
} }
} }
if(copy_buf) _lv_mem_buf_release(copy_buf);
} /*End of true double buffer handling*/ } /*End of true double buffer handling*/
/*Clean up*/ /*Clean up*/