Re: Memory leak in libpcap (top of tree) and/or kernel

[Guy Harris <guy () alum mit edu>]

On Jul 10, 2009, at 5:09 PM, Gianluca Varenni wrote:

> I think the routine usb_cleanup_linux_mmap(pcap_t* handle) should be  
> modified as follows:

        ...

> I'm not too happy about querying again the fd for the RING_SIZE (it  
> should be stored in the pcap_t structure), but I tried it on my VM  
> and it seems to work ok. BTW, it's interesting there is a note about  
> the missing munmap in the code...
>
> What do you think?

I've checked in and pushed a change to do the unmap; I got rid of the  
comment, as the answer to the question is apparently "yes", and also  
took the opportunity to add to the pcap_md structure a pointer and  
size_t, the former of which points to the memory-mapped buffer and the  
latter of which is the size of the buffer (i.e., it now stores the  
size of the mmapped region in the pcap_t structure), and used them  
both for the regular Linux mmapped capture and the Linux USB mmapped  
capture.  (The FreeBSD mmapped capture already had pointers and  
lengths for the two regions.)
-
This is the tcpdump-workers list.
Visit https://cod.sandelman.ca/ to unsubscribe.