sharkd: support for MaxMindDB.

Call maxmind_db_lookup_process() [via host_name_lookup_process()] before processing request.
It's still buggy, webshark needs a refresh to show endpoint or conversation ipmap,
but it's better than nothing.

It seems to need refresh, cause maxmind_db_lookup_ipv[46] works only for cached entries.

Also cause mmdbresolve is pipe-based, every sharkd need to have it own mmdbresolve process.

Change-Id: I3d588d1fa9e9ba645cc3d0bff3d276b202159d21
Reviewed-on: https://code.wireshark.org/review/27335
Petri-Dish: Jakub Zawadzki <darkjames-ws@darkjames.pl>
Tested-by: Petri Dish Buildbot
Reviewed-by: Jakub Zawadzki <darkjames-ws@darkjames.pl>

1 files changed, 7 insertions, 0 deletions

diff --git a/sharkd.c b/sharkd.c
index 0070bcbc70..0b61074c8a 100644
--- a/sharkd.c
+++ b/sharkd.c
@@ -52,6 +52,7 @@
 #include "epan/register.h"
 #include <epan/epan_dissect.h>
 #include <epan/tap.h>
+#include <epan/uat-int.h>
 
 #include <codecs/codecs.h>
 
@@ -195,6 +196,12 @@ main(int argc, char *argv[])
   /* Build the column format array */
   build_column_format_array(&cfile.cinfo, prefs_p->num_cols, TRUE);
 
+#ifdef HAVE_MAXMINDDB
+  /* mmdbresolve is started from mmdb_resolve_start(), which is called from epan_load_settings via: read_prefs -> (...) uat_load_all -> maxmind_db_post_update_cb.
+   * Need to stop it, otherwise all sharkd will have same mmdbresolve process, including pipe descriptors to read and write. */
+  uat_clear(uat_get_table_by_name("MaxMind Database Paths"));
+#endif
+
   ret = sharkd_loop();
 clean_exit:
   col_cleanup(&cfile.cinfo);