Hi Christian, On Mon, May 29, 2023 at 07:18:46PM +0200, Christian Marangi wrote: > With the help of a Coverity Scan, it was pointed out that it's present a > memory leak in the corner case where find_proto is not NULL in the > function should_load_proto. find_proto return a struct xtables_match > pointer from xtables_find_match that is allocated but never freed. > > Correctly free the found proto in the corner case where find_proto > succeed. We have a '--valgrind' mode in shell testsuite and this has not been identified yet. So assuming your finding is correct, we lack a proper test case. Can you provide a reproducer (ideally as shell test-case)? Also I wonder if we can provide a Fixes: tag. The code in question is very old, so maybe hard to find. Thanks, Phil
