Het is niet mijn bedoeling iets af te kraken (misschien begon ik het topic wel een beetje op die manier). Het is duidelijk dat meneer Steve Gibson het niet voor het geld doet...ik vind het gewoon ff een interessante discussie, aangezien zijn site schijnbaar veel gebruikt wordt en dus als betrouwbaar wordt beschouwd(?).
Als je vervolgens de quotes uit de newsgroups leest: [quote]
Thread: Announcing the "SocketToMe Utility Toolkit"
From: AK1400 ak1400@hotmai1.com
Date: 2001-08-11 00:32
Groups: grc.news.feedback
SocketToMe does what it says it should do. I can create a raw socket using admin privileges and then when I try again using user privs it claims to still be able to create the raw socket.
But... ...running one of my own raw socket applications on Win2K Pro (SP2) it will of course work fine when run with admin privileges
- I can send out TCP SYN packets and sniff the network and all the other stuff you can do with Win2K raw sockets. However, after having already created and used raw sockets when run as admin, running the same program with bog standard user privileges (shift-right click, Run As...) my program will CREATE the raw socket without complaining but will fail on the bind() command that is required to use raw sockets. You MUST bind() the socket to an interface on the local host or else you won't be able to send anything.
So the question is, does SocketToMe actually try to bind() the raw socket?
If it is not doing this then the raw socket it creates is useless and cannot be used for anything.
-Robin
[/qoute]
en daarop is vervolgens Steve's antwoord:
Re: Announcing the "SocketToMe Utility Toolkit"
From: Steve Gibson support@grc.com
Date: 2001-08-11 01:07
Groups: grc.news.feedback
> So the question is, does SocketToMe actually try to bind()
> the raw socket? If it is not doing this then the raw socket
> it creates is useless and cannot be used for anything. N
Now it's MY turn to say "Whoops!" <> You're right Robin ...
I have just been creating a raw socket, then using "setsockopt" to attempt to get IP_HDRINCL accepted. But that's as far as I went! Damn.
I'll add the code and duplicate your results right away! Thanks!!!!
en een tijdje later:
Re: Announcing the "SocketToMe Utility Toolkit"
From: Steve Gibson support@grc.com
Date:2001-08-11 01:42
Groups: grc.news.feedback
Robin ... In thinking about this some more, it's not clear to me what it even means to "bind" a raw socket ... when you're going to be providing your own IP and TCP headers.
Certainly for receiving incoming data, binding the local socket would get the stack to perform some inbound filtering for you ... but I think that an unbound raw socket can simply use "sendto" to emit the payload it has assembled. (But I do need to make sure that can be done.) Anyway ... thanks for the heads up ... I clearly need to verify the ability to generate potentially malicious traffic without privileges.
Dan begin ik me toch wel ernstig af te vragen of die man wel zo'n expert is...hier nog een voorbeeld
http://radsoft.net/news/roundups/grc/20010821,00.html
Ik betwijfel het...
Maar idd: als een bep. site je niet bevalt, moet je er niet naar toe gaan...dat is absoluut waar. Maar als een site, die over het algemeen wordt gezien (?) als 1 van DE sites om de veiligheid van je internetverbinding mee te testen valse informatie zou bevatten, zou dat toch geen goeie zaak zijn, lijkt me...wel dan?