From 2e679ba8647aff7b0114de2940a03e4186ccc3dc Mon Sep 17 00:00:00 2001
From: Luca Deri <deri@ntop.org>
Date: Wed, 28 May 2025 22:10:10 +0200
Subject: Added ndpi_load_protocols_dir() API call for loading IP-based
 protocol detection Added --protocols-list-dir <dir> to ndpiReader for loading
 IP_based protocol

---
 src/lib/ndpi_main.c | 161 +++++++++++++++++++++++++++++++++++++++++++++++++++-
 1 file changed, 160 insertions(+), 1 deletion(-)

(limited to 'src/lib/ndpi_main.c')

diff --git a/src/lib/ndpi_main.c b/src/lib/ndpi_main.c
index 1adae4a3e..1370909eb 100644
--- a/src/lib/ndpi_main.c
+++ b/src/lib/ndpi_main.c
@@ -3450,7 +3450,9 @@ static const char *categories[NDPI_PROTOCOL_NUM_CATEGORIES] = {
   "Hw/Sw",
   "Dating",
   "Travel",
-  "Food"
+  "Food",
+  "Bots",
+  "Scanners"
 };
 
 #if !defined(NDPI_CFFI_PREPROCESSING) && defined(__linux__)
@@ -5247,6 +5249,99 @@ int load_category_file_fd(struct ndpi_detection_module_struct *ndpi_str,
 
 /* ******************************************************************** */
 
+int load_protocol_id_file_fd(struct ndpi_detection_module_struct *ndpi_str,
+			     FILE *fd, u_int16_t protocol_id) {
+  char buffer[256], *line;
+  u_int num_loaded = 0;
+  unsigned int failed_lines = 0;
+  unsigned int lines_read = 0;
+
+  (void)lines_read;
+
+  if(!ndpi_str || !fd || !ndpi_str->protocols)
+    return(0);
+
+  while(1) {
+    int len;
+
+    line = fgets(buffer, sizeof(buffer), fd);
+
+    if(line == NULL)
+      break;
+
+    lines_read++;
+    len = strlen(line);
+
+    if(len <= 1 || len == sizeof(buffer) - 1) {
+      NDPI_LOG_ERR(ndpi_str, "[NDPI] Failed to read file line #%u, line too short/long\n",
+                   lines_read);
+      failed_lines++;
+      continue;
+    } else if (line[0] == '#')
+      continue;
+
+    int i = 0;
+
+    for (i = 0; i < len; ++i) {
+      if (line[i] == '\r' || line[i] == '\n') {
+        line[i] = '\0';
+        break;
+      }
+    }
+
+    if(strchr(line, ':') != NULL) {
+      if(ndpi_add_host_ip_subprotocol(ndpi_str, line, protocol_id, 1 /* IPv6 */) == 0)
+	num_loaded++;
+      else
+	failed_lines++;
+    } else if(strchr(line, '.') != NULL) {
+      /* IPv4 */
+      if(ndpi_add_host_ip_subprotocol(ndpi_str, line, protocol_id, 0 /* IPv4 */) == 0)
+	num_loaded++;
+      else
+	failed_lines++;
+    } else {
+      /* No clue */
+      failed_lines++;
+      continue;
+    }
+  }
+
+  if(failed_lines)
+    return(-1 * failed_lines);
+  
+  return(num_loaded);
+}
+
+/* ******************************************************************** */
+
+/*
+  Loads a file (separated by <cr>) of IP addresses associated with the
+  specified protocol
+*/
+int ndpi_load_protocol_id_file(struct ndpi_detection_module_struct *ndpi_str,
+			       char *path, u_int16_t protocol_id) {
+  int rc;
+  FILE *fd;
+
+  if(!ndpi_str || !path)
+    return(-1);
+
+  fd = fopen(path, "r");
+  if(fd == NULL) {
+    NDPI_LOG_ERR(ndpi_str, "Unable to open file %s [%s]\n", path, strerror(errno));
+    return -1;
+  }
+
+  rc = load_protocol_id_file_fd(ndpi_str, fd, protocol_id);
+
+  fclose(fd);
+
+  return rc;
+}
+
+/* ******************************************************************** */
+
 /*
   Load files (whose name is <categoryid>_<label>.<extension>) stored
   in a directory and bind each domain to the specified category.
@@ -5311,6 +5406,70 @@ int ndpi_load_categories_dir(struct ndpi_detection_module_struct *ndpi_str,
 
 /* ******************************************************************** */
 
+/*
+  Load files (whose name is <protocolid>_<label>.<extension>) stored
+  in a directory and bind each domain to the specified protocol.
+
+  It can be used to load all files store in the lists/ directory
+
+  It returns the number of loaded files or -1 in case of failure
+*/
+int ndpi_load_protocols_dir(struct ndpi_detection_module_struct *ndpi_str,
+			    char *dir_path) {
+!  DIR *dirp;
+  struct dirent *dp;
+  int failed_files = 0;
+  int num_loaded = 0;
+
+  if(!ndpi_str || !dir_path)
+    return(0);
+
+  dirp = opendir(dir_path);
+  if (dirp == NULL)
+    return(0);
+
+  while((dp = readdir(dirp)) != NULL) {
+    char *underscore, *extn;
+
+    if(dp->d_name[0] == '.') continue;
+    extn = strrchr(dp->d_name, '.');
+
+    if((extn == NULL) || strcmp(extn, ".list"))
+      continue;
+
+    /* Check if the format is <proto it>_<string>.<extension> */
+    if((underscore = strchr(dp->d_name, '_')) != NULL) {
+      int proto_id;
+      const char *errstrp;
+
+      underscore[0] = '\0';
+      proto_id = ndpi_strtonum(dp->d_name, 1, NDPI_LAST_IMPLEMENTED_PROTOCOL - 1, &errstrp, 10);
+      if(errstrp == NULL) {
+	/* Valid file */
+	char path[512];
+
+	underscore[0] = '_';
+	snprintf(path, sizeof(path), "%s/%s", dir_path, dp->d_name);
+
+	if (ndpi_load_protocol_id_file(ndpi_str, path, proto_id) < 0) {
+	  NDPI_LOG_ERR(ndpi_str, "Failed to load '%s'\n", path);
+	  failed_files++;
+	}else
+	  num_loaded++;
+      }
+    }
+  }
+
+  (void)closedir(dirp);
+
+  if(failed_files)
+    return(-1 * failed_files);
+
+  return(num_loaded);
+}
+
+/* ******************************************************************** */
+
 static int ndpi_load_risky_domain(struct ndpi_detection_module_struct *ndpi_str,
 				  char* domain_name) {
   if(ndpi_str->risky_domain_automa.ac_automa == NULL) {
-- 
cgit v1.2.3