lvgl V7中lv_canvas_clean函数实现

在LVGL V7中，`lv_canvas_clean()`函数的实现如下所示：

void lv_canvas_clean(lv_obj_t * canvas)
{
    LV_ASSERT_OBJ(canvas, LV_OBJX_NAME);

    lv_canvas_ext_t * ext = lv_obj_get_ext_attr(canvas);
    if(ext->dsc.header.always_zero == 0) return;

    lv_color_t color = lv_color_hex(ext->dsc.header.color);
    lv_disp_t * disp = lv_obj_get_disp(canvas);
    lv_disp_buf_t * draw_buf = lv_disp_get_draw_buf(disp);

    uint32_t w = lv_obj_get_width(canvas);
    uint32_t h = lv_obj_get_height(canvas);
    uint32_t x = lv_obj_get_x(canvas);
    uint32_t y = lv_obj_get_y(canvas);

    if(ext->clip_area.width != 0 && ext->clip_area.height != 0) {
        lv_area_t rel_area;
        lv_obj_get_coords(canvas, &rel_area);
        lv_area_intersect(&rel_area, &ext->clip_area, &rel_area);

        if(lv_area_get_width(&rel_area) == 0 || lv_area_get_height(&rel_area) == 0) return;

        lv_area_t abs_area;
        lv_obj_get_abs_coords(canvas, &abs_area);

        lv_disp_drv_t * drv = lv_disp_get_drv(disp);

        uint32_t y1 = abs_area.y1 + rel_area.y1;
        uint32_t x1 = abs_area.x1 + rel_area.x1;

        uint32_t len = lv_area_get_width(&rel_area);
        uint32_t px_offset = (y1 * draw_buf->w + x1);

        for(uint32_t i = 0; i < lv_area_get_height(&rel_area); i++) {
            lv_disp_flush_dsc_t flush_dsc;
            lv_disp_flush_dsc_init(&flush_dsc);
            flush_dsc.area.x1 = x1;
            flush_dsc.area.x2 = x1 + len - 1;
            flush_dsc.area.y1 = y1 + i;
            flush_dsc.area.y2 = y1 + i;

            lv_disp_flush_drv_t flush_drv;
            lv_disp_flush_drv_init(&flush_drv);
            flush_drv.draw_buf   = draw_buf;
            flush_drv.flush_cb   = drv->flush_cb;
            flush_drv.user_data  = drv->user_data;
            flush_drv.wait_cb    = drv->wait_cb;
            flush_drv.monitor_cb = drv->monitor_cb;
            flush_drv.hor_res    = drv->hor_res;
            flush_drv.ver_res    = drv->ver_res;

            /* Fill the line */
            lv_memset_00(&draw_buf->buf_act[px_offset], len * sizeof(lv_color_t));
            lv_disp_flush(&flush_drv, &flush_dsc);
        }
    } else {
        lv_disp_fill(&color, NULL, w * h, x, y);
    }
}

该函数主要有以下几个步骤：

1. 获取canvas控件对象的扩展属性`ext`。
2. 判断`ext->dsc.header.always_zero`是否为0。如果为0，表示该canvas是不透明的，已经被完全覆盖，无需再次清空。如果不为0，则需要进行清空操作。
3. 获取canvas的宽度`w`、高度`h`、x坐标`x`和y坐标`y`。
4. 如果canvas有剪切区域，则获取剪切区域的相对位置和绝对位置，并且遍历剪切区域的每一行，将该行的像素点设置为背景色。
5. 如果canvas没有剪切区域，则直接使用`lv_disp_fill()`函数进行填充。

总的来说，`lv_canvas_clean()`函数的实现比较简单，主要是对canvas的背景色进行填充，以达到清空canvas的效果。