feat(dma2d): support RGB24 and BGR24 scramble conversion

Add async color convert coverage for direct RGB24/BGR24 byte-order swaps
and
thread both TX and RX CSC state through DMA2D transactions for future
extensibility.
This commit is contained in:
morris
2026-06-28 17:07:18 +08:00
parent 3a34711eac
commit 624a21f359
7 changed files with 345 additions and 80 deletions
@@ -48,10 +48,10 @@ TEST_CASE("async color convert basic callback", "[async_color_convert]")
uint16_t *src565 = heap_caps_aligned_calloc(64, pixel_num, sizeof(uint16_t),
MALLOC_CAP_INTERNAL | MALLOC_CAP_DMA | MALLOC_CAP_8BIT);
uint8_t *dst888 = heap_caps_aligned_calloc(64, pixel_num, 3,
MALLOC_CAP_INTERNAL | MALLOC_CAP_DMA | MALLOC_CAP_8BIT);
uint8_t *dst_bgr24 = heap_caps_aligned_calloc(64, pixel_num, 3,
MALLOC_CAP_INTERNAL | MALLOC_CAP_DMA | MALLOC_CAP_8BIT);
TEST_ASSERT_NOT_NULL(src565);
TEST_ASSERT_NOT_NULL(dst888);
TEST_ASSERT_NOT_NULL(dst_bgr24);
for (uint32_t i = 0; i < pixel_num; i++) {
src565[i] = (uint16_t)((i * 13) ^ 0x5AA5);
@@ -71,7 +71,7 @@ TEST_CASE("async color convert basic callback", "[async_color_convert]")
.src_height = height,
.src_x = 0,
.src_y = 0,
.dst_buffer = dst888,
.dst_buffer = dst_bgr24,
.dst_stride = width,
.dst_height = height,
.dst_x = 0,
@@ -96,10 +96,10 @@ TEST_CASE("async color convert basic callback", "[async_color_convert]")
TEST_ESP_OK(esp_async_color_convert_uninstall(conv_hdl));
free(src565);
free(dst888);
free(dst_bgr24);
}
TEST_CASE("async color convert roundtrip: RGB565<->RGB888", "[async_color_convert]")
TEST_CASE("async color convert roundtrip: RGB16<->BGR24", "[async_color_convert]")
{
const uint32_t width = 32;
const uint32_t height = 20;
@@ -107,12 +107,12 @@ TEST_CASE("async color convert roundtrip: RGB565<->RGB888", "[async_color_conver
uint16_t *src565 = heap_caps_aligned_calloc(64, pixel_num, sizeof(uint16_t),
MALLOC_CAP_INTERNAL | MALLOC_CAP_DMA | MALLOC_CAP_8BIT);
uint8_t *mid888 = heap_caps_aligned_calloc(64, pixel_num, 3,
MALLOC_CAP_INTERNAL | MALLOC_CAP_DMA | MALLOC_CAP_8BIT);
uint8_t *mid_bgr24 = heap_caps_aligned_calloc(64, pixel_num, 3,
MALLOC_CAP_INTERNAL | MALLOC_CAP_DMA | MALLOC_CAP_8BIT);
uint16_t *dst565 = heap_caps_aligned_calloc(64, pixel_num, sizeof(uint16_t),
MALLOC_CAP_INTERNAL | MALLOC_CAP_DMA | MALLOC_CAP_8BIT);
TEST_ASSERT_NOT_NULL(src565);
TEST_ASSERT_NOT_NULL(mid888);
TEST_ASSERT_NOT_NULL(mid_bgr24);
TEST_ASSERT_NOT_NULL(dst565);
for (uint32_t i = 0; i < pixel_num; i++) {
@@ -127,13 +127,13 @@ TEST_CASE("async color convert roundtrip: RGB565<->RGB888", "[async_color_conver
async_color_convert_handle_t conv_hdl = NULL;
TEST_ESP_OK(esp_async_color_convert_install_dma2d(&config, &conv_hdl));
async_color_convert_request_t req_565_to_888 = {
async_color_convert_request_t req_565_to_bgr24 = {
.src_buffer = src565,
.src_stride = width,
.src_height = height,
.src_x = 0,
.src_y = 0,
.dst_buffer = mid888,
.dst_buffer = mid_bgr24,
.dst_stride = width,
.dst_height = height,
.dst_x = 0,
@@ -144,8 +144,8 @@ TEST_CASE("async color convert roundtrip: RGB565<->RGB888", "[async_color_conver
.dst_color_format = ESP_COLOR_FOURCC_BGR24,
};
async_color_convert_request_t req_888_to_565 = {
.src_buffer = mid888,
async_color_convert_request_t req_bgr24_to_565 = {
.src_buffer = mid_bgr24,
.src_stride = width,
.src_height = height,
.src_x = 0,
@@ -161,8 +161,8 @@ TEST_CASE("async color convert roundtrip: RGB565<->RGB888", "[async_color_conver
.dst_color_format = ESP_COLOR_FOURCC_RGB16,
};
TEST_ESP_OK(esp_color_convert_blocking(conv_hdl, &req_565_to_888, -1));
TEST_ESP_OK(esp_color_convert_blocking(conv_hdl, &req_888_to_565, -1));
TEST_ESP_OK(esp_color_convert_blocking(conv_hdl, &req_565_to_bgr24, -1));
TEST_ESP_OK(esp_color_convert_blocking(conv_hdl, &req_bgr24_to_565, -1));
// The final dst565 should be the same as the original src565 after round-trip conversion
TEST_ASSERT_EQUAL_MEMORY(src565, dst565, pixel_num * sizeof(uint16_t));
@@ -170,7 +170,7 @@ TEST_CASE("async color convert roundtrip: RGB565<->RGB888", "[async_color_conver
TEST_ESP_OK(esp_async_color_convert_uninstall(conv_hdl));
free(src565);
free(mid888);
free(mid_bgr24);
free(dst565);
}
@@ -305,7 +305,189 @@ static void uyvy_to_bgr24_reference_image(const uint8_t *src_uyvy, uint8_t *dst_
}
}
TEST_CASE("async color convert UYVY->RGB888 matches reference", "[async_color_convert]")
TEST_CASE("async color convert swaps RGB24 and BGR24 byte order", "[async_color_convert]")
{
const uint32_t width = 4;
const uint32_t height = 2;
const size_t pixel_count = width * height;
const size_t buf_size = pixel_count * 3;
static const uint8_t src_rgb24[] = {
0x10, 0x20, 0x30, 0x7F, 0x80, 0x81, 0xAA, 0x55, 0xFE, 0x01, 0xC0, 0x99,
0xDE, 0xAD, 0xBE, 0x00, 0x11, 0x22, 0x44, 0x88, 0xCC, 0xF0, 0x0D, 0x42,
};
static const uint8_t src_bgr24[] = {
0x30, 0x20, 0x10, 0x81, 0x80, 0x7F, 0xFE, 0x55, 0xAA, 0x99, 0xC0, 0x01,
0xBE, 0xAD, 0xDE, 0x22, 0x11, 0x00, 0xCC, 0x88, 0x44, 0x42, 0x0D, 0xF0,
};
TEST_ASSERT_EQUAL(sizeof(src_rgb24), buf_size);
TEST_ASSERT_EQUAL(sizeof(src_bgr24), buf_size);
uint8_t *rgb24 = heap_caps_aligned_calloc(64, 1, buf_size,
MALLOC_CAP_INTERNAL | MALLOC_CAP_DMA | MALLOC_CAP_8BIT);
uint8_t *bgr24 = heap_caps_aligned_calloc(64, 1, buf_size,
MALLOC_CAP_INTERNAL | MALLOC_CAP_DMA | MALLOC_CAP_8BIT);
uint8_t *dst_bgr24 = heap_caps_aligned_calloc(64, 1, buf_size,
MALLOC_CAP_INTERNAL | MALLOC_CAP_DMA | MALLOC_CAP_8BIT);
uint8_t *dst_rgb24 = heap_caps_aligned_calloc(64, 1, buf_size,
MALLOC_CAP_INTERNAL | MALLOC_CAP_DMA | MALLOC_CAP_8BIT);
TEST_ASSERT_NOT_NULL(rgb24);
TEST_ASSERT_NOT_NULL(bgr24);
TEST_ASSERT_NOT_NULL(dst_bgr24);
TEST_ASSERT_NOT_NULL(dst_rgb24);
memcpy(rgb24, src_rgb24, buf_size);
memcpy(bgr24, src_bgr24, buf_size);
memset(dst_bgr24, 0xA5, buf_size);
memset(dst_rgb24, 0x5A, buf_size);
async_color_convert_config_t config = {
.backlog = 1,
.intr_priority = 0,
.dma_burst_size = 16,
};
async_color_convert_handle_t conv_hdl = NULL;
TEST_ESP_OK(esp_async_color_convert_install_dma2d(&config, &conv_hdl));
async_color_convert_request_t req_rgb_to_bgr = {
.src_buffer = rgb24,
.src_stride = width,
.src_height = height,
.src_x = 0,
.src_y = 0,
.dst_buffer = dst_bgr24,
.dst_stride = width,
.dst_height = height,
.dst_x = 0,
.dst_y = 0,
.copy_width = width,
.copy_height = height,
.src_color_format = ESP_COLOR_FOURCC_RGB24,
.dst_color_format = ESP_COLOR_FOURCC_BGR24,
};
async_color_convert_request_t req_bgr_to_rgb = {
.src_buffer = bgr24,
.src_stride = width,
.src_height = height,
.src_x = 0,
.src_y = 0,
.dst_buffer = dst_rgb24,
.dst_stride = width,
.dst_height = height,
.dst_x = 0,
.dst_y = 0,
.copy_width = width,
.copy_height = height,
.src_color_format = ESP_COLOR_FOURCC_BGR24,
.dst_color_format = ESP_COLOR_FOURCC_RGB24,
};
TEST_ESP_OK(esp_color_convert_blocking(conv_hdl, &req_rgb_to_bgr, -1));
TEST_ASSERT_EQUAL_MEMORY(src_bgr24, dst_bgr24, buf_size);
TEST_ESP_OK(esp_color_convert_blocking(conv_hdl, &req_bgr_to_rgb, -1));
TEST_ASSERT_EQUAL_MEMORY(src_rgb24, dst_rgb24, buf_size);
TEST_ESP_OK(esp_async_color_convert_uninstall(conv_hdl));
free(rgb24);
free(bgr24);
free(dst_bgr24);
free(dst_rgb24);
}
// Verifies the scramble route and BGR24/RGB24->UYVY conversion compose correctly.
TEST_CASE("async color convert RGB24 and BGR24 inputs produce identical UYVY output", "[async_color_convert]")
{
const uint32_t width = 4;
const uint32_t height = 2;
const size_t pixel_count = width * height;
const size_t rgb_size = pixel_count * 3;
const size_t uyvy_size = pixel_count * 2;
static const uint8_t src_rgb24[] = {
0x10, 0x20, 0x30, 0x7F, 0x80, 0x81, 0xAA, 0x55, 0xFE, 0x01, 0xC0, 0x99,
0xDE, 0xAD, 0xBE, 0x00, 0x11, 0x22, 0x44, 0x88, 0xCC, 0xF0, 0x0D, 0x42,
};
static const uint8_t src_bgr24[] = {
0x30, 0x20, 0x10, 0x81, 0x80, 0x7F, 0xFE, 0x55, 0xAA, 0x99, 0xC0, 0x01,
0xBE, 0xAD, 0xDE, 0x22, 0x11, 0x00, 0xCC, 0x88, 0x44, 0x42, 0x0D, 0xF0,
};
const color_conv_std_rgb_yuv_t conv_std = COLOR_CONV_STD_RGB_YUV_BT601;
TEST_ASSERT_EQUAL(sizeof(src_rgb24), rgb_size);
TEST_ASSERT_EQUAL(sizeof(src_bgr24), rgb_size);
uint8_t *rgb24 = heap_caps_aligned_calloc(64, 1, rgb_size,
MALLOC_CAP_INTERNAL | MALLOC_CAP_DMA | MALLOC_CAP_8BIT);
uint8_t *bgr24 = heap_caps_aligned_calloc(64, 1, rgb_size,
MALLOC_CAP_INTERNAL | MALLOC_CAP_DMA | MALLOC_CAP_8BIT);
uint8_t *dst_from_rgb24 = heap_caps_aligned_calloc(64, 1, uyvy_size,
MALLOC_CAP_INTERNAL | MALLOC_CAP_DMA | MALLOC_CAP_8BIT);
uint8_t *dst_from_bgr24 = heap_caps_aligned_calloc(64, 1, uyvy_size,
MALLOC_CAP_INTERNAL | MALLOC_CAP_DMA | MALLOC_CAP_8BIT);
TEST_ASSERT_NOT_NULL(rgb24);
TEST_ASSERT_NOT_NULL(bgr24);
TEST_ASSERT_NOT_NULL(dst_from_rgb24);
TEST_ASSERT_NOT_NULL(dst_from_bgr24);
memcpy(rgb24, src_rgb24, rgb_size);
memcpy(bgr24, src_bgr24, rgb_size);
async_color_convert_config_t config = {
.backlog = 2,
.intr_priority = 0,
.dma_burst_size = 16,
};
async_color_convert_handle_t conv_hdl = NULL;
TEST_ESP_OK(esp_async_color_convert_install_dma2d(&config, &conv_hdl));
memset(dst_from_rgb24, 0xA5, uyvy_size);
memset(dst_from_bgr24, 0x5A, uyvy_size);
async_color_convert_request_t req_rgb24_to_uyvy = {
.src_buffer = rgb24,
.src_stride = width,
.src_height = height,
.src_x = 0,
.src_y = 0,
.dst_buffer = dst_from_rgb24,
.dst_stride = width,
.dst_height = height,
.dst_x = 0,
.dst_y = 0,
.copy_width = width,
.copy_height = height,
.src_color_format = ESP_COLOR_FOURCC_RGB24,
.dst_color_format = ESP_COLOR_FOURCC_UYVY,
.color_conv_std = conv_std,
};
async_color_convert_request_t req_bgr24_to_uyvy = {
.src_buffer = bgr24,
.src_stride = width,
.src_height = height,
.src_x = 0,
.src_y = 0,
.dst_buffer = dst_from_bgr24,
.dst_stride = width,
.dst_height = height,
.dst_x = 0,
.dst_y = 0,
.copy_width = width,
.copy_height = height,
.src_color_format = ESP_COLOR_FOURCC_BGR24,
.dst_color_format = ESP_COLOR_FOURCC_UYVY,
.color_conv_std = conv_std,
};
TEST_ESP_OK(esp_color_convert_blocking(conv_hdl, &req_rgb24_to_uyvy, -1));
TEST_ESP_OK(esp_color_convert_blocking(conv_hdl, &req_bgr24_to_uyvy, -1));
TEST_ASSERT_EQUAL_MEMORY(dst_from_bgr24, dst_from_rgb24, uyvy_size);
TEST_ESP_OK(esp_async_color_convert_uninstall(conv_hdl));
free(rgb24);
free(bgr24);
free(dst_from_rgb24);
free(dst_from_bgr24);
}
TEST_CASE("async color convert UYVY->BGR24 matches reference", "[async_color_convert]")
{
const uint32_t src_stride = 32;
const uint32_t dst_stride = 64;