Potential memory leak in tcpdump.c

[icy17]

Potential memory leak in tcpdump.c

[infrastation]

The first instance is currently at line 2355. Then there is a cleanup block with call to pcap_freecode() conditional on dflag and a number of calls to error(), none of which makes many enough cleanup moves before the exit. The second instance is currently at line 2759. It calls pcap_freecode() between the calls to pcap_compile(), but not after the last (or the only) call.

This way, there is indeed some amount of resources not explicitly freed before the exit. Given the number of various ways to allocate these, perhaps the best potential way to address this would be to have a separate resource deallocation function that cleanup() could explicitly use on a signal exit and all other exit paths could implicitly use via atexit().

Until anybody is willing to implement that, the resources will be freed automatically upon the termination of the process. If you are aware of more severe practical effects, please explain in detail.

[guyharris]

This way, there is indeed some amount of resources not explicitly freed before the exit.

The pcap_t is also not closed. In some situations closing a pcap_t performs some cleanup operations other than freeing memory, such as taking an interface out of monitor mode. (That's also a problem if the process doing the packet capture terminates abnormally.)