[Winpcap-users] Deadlock on pcap_open_live() ?

Gianluca Varenni gianluca.varenni at cacetech.com
Wed Mar 31 10:22:28 PDT 2010


I suspect that it's one of the bugs present in Microsoft NetMon, the COM 
component that we use under the hoods to sniff from dialup/VPN connections.

Are you able to create a small GTK based application showing the issue? I'd 
like to test it here.

Have a nice day
GV



--------------------------------------------------
From: "Dupuit Cyril" <cyrildupuit at hotmail.com>
Sent: Wednesday, March 31, 2010 9:58 AM
To: <winpcap-users at winpcap.org>
Subject: Re: [Winpcap-users] Deadlock on pcap_open_live() ?

> Hello,
>
> My application is used to sniff packets from ethernet like Wireshark.
>
> I offer a choice to the user which interface to use to capture packets.
> To do that, I use a specific thread used to compute some statistics to
> help him to do the best choice (with a GTK based dialog box).
>
> Sometimes, my application can send packets in a specific case (like
> broadcast or multicast).
>
> My application don't use directly COM objects but I am not shure that
> GTK uses or not COM interface.
>
> Hope that can help you to better understand my problem,
>
> Cyril
>
> Le 31/03/2010 17:18, Gianluca Varenni a écrit :
>> Is your application COM based?
>>
>> Have a nice day
>> GV
>>
>> --------------------------------------------------
>> From: "Dupuit Cyril" <cyrildupuit at hotmail.com>
>> Sent: Tuesday, March 30, 2010 2:28 PM
>> To: <winpcap-users at winpcap.org>
>> Subject: [Winpcap-users] Deadlock on pcap_open_live() ?
>>
>>
>>> Hello,
>>>
>>> I am developping an application that uses multiple threaded
>>> environnement on Windows XP.
>>>
>>> One of these threads opens a network interface to count the number of
>>> packets per second and compute statistics.
>>>
>>> This architecture works fine for a real network interface (ie realteck,
>>> 3 com, ...) but fails with the first interface (Adapter for generic
>>> dialup and VPN capture).
>>>
>>> I have made a little change on my code to use the same thread than my
>>> application.
>>> The result is good. My application works without deadlok.
>>>
>>> If I resume :
>>> - Multithreaded :
>>>     main thread                            Statistic thread
>>>            -> Do something...                      -> pcap_open_live()
>>> on the first interface (Adapter for generic dialup and VPN capture)
>>>
>>>  => Result is a deadlock in the pcap_open_live() function. The function
>>> is blocked.
>>>
>>> - Single threaded :
>>>    main thread
>>>         -> pcap_open_live() on the first interface.
>>>  => Result is an handle on the file openned (no deadlock detected).
>>>
>>> I have tried this solution only on WinPcap 4.1.1. I don't know if this
>>> problem appears in another version of WinPcap.
>>>
>>> Can you help me ?
>>>
>>> Thank you in adavance,
>>>
>>> Cyril
>>>
>>> _______________________________________________
>>> Winpcap-users mailing list
>>> Winpcap-users at winpcap.org
>>> https://www.winpcap.org/mailman/listinfo/winpcap-users
>>>
>> _______________________________________________
>> Winpcap-users mailing list
>> Winpcap-users at winpcap.org
>> https://www.winpcap.org/mailman/listinfo/winpcap-users
>>
>>
>>
>
> _______________________________________________
> Winpcap-users mailing list
> Winpcap-users at winpcap.org
> https://www.winpcap.org/mailman/listinfo/winpcap-users 



More information about the Winpcap-users mailing list