// For DEPTH_TO_SPACE layer: block_size
DNNModel *ff_dnn_load_model_native(const char *model_filename, DNNFunctionType func_type, const char *options, AVFilterContext *filter_ctx)
{
+#define DNN_NATIVE_MAGIC "FFMPEGDNNNATIVE"
DNNModel *model = NULL;
- char header_expected[] = "FFMPEGDNNNATIVE";
- char *buf;
- size_t size;
+ // sizeof - 1 to skip the terminating '\0' which is not written in the file
+ char buf[sizeof(DNN_NATIVE_MAGIC) - 1];
int version, header_size, major_version_expected = 1;
NativeModel *native_model = NULL;
AVIOContext *model_file_context;
/**
* check file header with string and version
*/
- size = sizeof(header_expected);
- buf = av_malloc(size);
- if (!buf) {
+ if (avio_read(model_file_context, buf, sizeof(buf)) != sizeof(buf) ||
+ memcmp(buf, DNN_NATIVE_MAGIC, sizeof(buf)))
goto fail;
- }
-
- // size - 1 to skip the ending '\0' which is not saved in file
- avio_get_str(model_file_context, size - 1, buf, size);
- dnn_size = size - 1;
- if (strncmp(buf, header_expected, size) != 0) {
- av_freep(&buf);
- goto fail;
- }
- av_freep(&buf);
+ dnn_size = sizeof(buf);
version = (int32_t)avio_rl32(model_file_context);
dnn_size += 4;