Notes |
|
(0001279)
|
user4
|
2009-06-15 11:54
|
|
Issue seems to be related to the usage of Net::Pcap and the loop function.
See in particular http://search.cpan.org/~saper/Net-Pcap-0.16/Pcap.pm#LIMITATIONS [^] which mentions:
"At present, only one callback function and user data scalar can be current at any time as they are both stored in global variables."
This could maybe explain why the db_connect function is called with the wrong arguments when pfmon is running several arp threads on different interfaces.
One possible solution might be to create a pfarplistener daemon (similar to pfdhcplistener) and remove this functionality from pfmon (would be a nice cleanup anyway) and pfarplistener would have to be called with one (1) interface ... this way we are getting around the several Net::Pcap threads |
|
|
(0001286)
|
treker
|
2009-06-19 23:51
|
|
|
|
|
Hi Treker,
I'm triaging for 1.8.8 and I'm interested in pushing your fix in there. Unfortunately, I know little of the arp mode and I've never done a setup in arp mode.
I'm kind of hoping you can give me the go / no-go for this patch as it is right now. What I don't understand is why using Net::PcapUtils instead of Net::Pcap would work-around the problem since Net::PcapUtils is identified as a wrapper around Net::Pcap?
Also why are you using PcapUtils' next insead of loop like it was done before? Maybe that's the part that fixed the issue and if it is, what's the consequence? Could it be that you don't get the error but you don't get a packet either..?
Anyway, the point is, if you tested it and you are sure that it works, then I'll merge it so let me know.
Thanks, |
|
|
|
Waiting for feedback on issue. Re-targeting for 1.9.1 for now. |
|
|
|
Reminder sent to: treker Hi Treker,
Can you check note 0001515 I left in this bug. I am asking more information on your patch.
Thanks! |
|
|
|
Starting to think that I'm experiencing this bug in another location in the code. See 0001065. |
|
|
|
Fixed a long time ago by getting rid of the ARP mode ;) |
|