Trust No Program
This topic is locked: you cannot edit posts or make replies.
nick s


Joined: 20 Dec 2008
Posts: 329
Reply with quote
tzuk wrote:
If it's a 19144 bytes long file with CRC = 9dc00fdc then I have it.

That matches my copy.

_________________
Nick
View user's profileSend private message
tzuk


Joined: 22 Jun 2004
Posts: 15003
Reply with quote
Thanks for the confirmation. I'll check it soon.

_________________
tzuk
View user's profileSend private message
tzuk


Joined: 22 Jun 2004
Posts: 15003
Reply with quote
I looked into it. What normally happens with HIPS-type drivers is that they inject into the kernel somewhat similar to this:

Application --> Open Process API --> ...
... --> HIPS Driver --> Operating System --> Nt Open Process --> ...
... --> Sandboxie

But XueTr injects like this:

Application --> Open Process API --> ...
... --> Operating System --> Nt Open Process --> XeuTr --> ...
... --> Sandboxie

This difference is confusing to Sandboxie, because it looks like some driver (XueTr in this case) is issuing the Open Process request. Sandboxie always permits a request that it thinks came from a driver.

I'm not sure that I can resolve this conflict without explicitly identifying the XueTr driver by name, which is an option, but not the best one.
View user's profileSend private message
tzuk


Joined: 22 Jun 2004
Posts: 15003
Reply with quote
Update. I redesigned the way Sandboxie figures out if the call came from the application or a driver that happens to be operating in the context of that application. It resolves this particular conflict with XueTr. The new way should actually be more robust than the old way. I will release beta version 3.41.02 in a day or two, and you guys would be able to confirm it at that time.
View user's profileSend private message
nick s


Joined: 20 Dec 2008
Posts: 329
Reply with quote
tzuk wrote:
Update. I redesigned the way Sandboxie figures out if the call came from the application or a driver that happens to be operating in the context of that application. It resolves this particular conflict with XueTr. The new way should actually be more robust than the old way. I will release beta version 3.41.02 in a day or two, and you guys would be able to confirm it at that time.

Excellent. Thanks for the update.
View user's profileSend private message
raid


Joined: 23 Aug 2008
Posts: 58
Location: TN, USA
Reply with quote
tzuk wrote:
Update. I redesigned the way Sandboxie figures out if the call came from the application or a driver that happens to be operating in the context of that application. It resolves this particular conflict with XueTr. The new way should actually be more robust than the old way. I will release beta version 3.41.02 in a day or two, and you guys would be able to confirm it at that time.


Sweet.

Also, is it possible for sandboxie to support "simulated" injection of one of it's files? I have a few malware samples that like to take over processes and then drop the files, and sandboxie isn't okay with this because it wants to hijack a kernel sandboxie driver; Only, I don't want it to be able to do it for real. Smile

_________________
Everything is so different, yet I am the same...
View user's profileSend private message
aigle


Joined: 16 Mar 2006
Posts: 21
Reply with quote
raid wrote:

Also, is it possible for sandboxie to support "simulated" injection of one of it's files? I have a few malware samples that like to take over processes and then drop the files, and sandboxie isn't okay with this because it wants to hijack a kernel sandboxie driver; Only, I don't want it to be able to do it for real. Smile
Are you saying that these samples are able to bypass SBIE? Surprised
View user's profileSend private message
ssj100


Joined: 23 Apr 2009
Posts: 843
Reply with quote
aigle wrote:
raid wrote:

Also, is it possible for sandboxie to support "simulated" injection of one of it's files? I have a few malware samples that like to take over processes and then drop the files, and sandboxie isn't okay with this because it wants to hijack a kernel sandboxie driver; Only, I don't want it to be able to do it for real. Smile
Are you saying that these samples are able to bypass SBIE? Surprised


Not at all. He's simply saying that if he tries to run those malware files sandboxed, nothing will happen, because Sandboxie will block it. However, I think he wants Sandboxie to be able to let the malware run sandboxed to observe its behaviour, but to keep it contained of course.
View user's profileSend private message
Guest


Reply with quote
Is this only releated to this POC used in the prescence of XueTr ? or could it be exploited by some rare types of malware w/o the aforementioned program being on the system?
tzuk


Joined: 22 Jun 2004
Posts: 15003
Reply with quote
Anonymous wrote:
Is this only releated to this POC used in the prescence of XueTr ? or could it be exploited by some rare types of malware w/o the aforementioned program being on the system?


This problem can only occur when XueTr is running outside the sandbox.

raid wrote:
Also, is it possible for sandboxie to support "simulated" injection of one of it's files?


That depends on the injection I suppose. For example, if the malware tries to inject Explorer.exe you might simulate it by running Explorer.exe in the sandbox.
View user's profileSend private message
tzuk


Joined: 22 Jun 2004
Posts: 15003
Reply with quote
Should be fixed in version 3.41.02.
View user's profileSend private message
nick s


Joined: 20 Dec 2008
Posts: 329
Reply with quote
tzuk wrote:
Should be fixed in version 3.41.02.

Confirmed. Thanks for the fix.
View user's profileSend private message
nick s


Joined: 20 Dec 2008
Posts: 329
Reply with quote
...and thanks as well to a256886572008 for reporting the issue.
View user's profileSend private message
raid


Joined: 23 Aug 2008
Posts: 58
Location: TN, USA
Reply with quote
aigle wrote:
raid wrote:

Also, is it possible for sandboxie to support "simulated" injection of one of it's files? I have a few malware samples that like to take over processes and then drop the files, and sandboxie isn't okay with this because it wants to hijack a kernel sandboxie driver; Only, I don't want it to be able to do it for real. Smile
Are you saying that these samples are able to bypass SBIE? Surprised


By no means. They die on the spot, actually.
View user's profileSend private message
raid


Joined: 23 Aug 2008
Posts: 58
Location: TN, USA
Reply with quote
ssj100 wrote:
aigle wrote:
raid wrote:

Also, is it possible for sandboxie to support "simulated" injection of one of it's files? I have a few malware samples that like to take over processes and then drop the files, and sandboxie isn't okay with this because it wants to hijack a kernel sandboxie driver; Only, I don't want it to be able to do it for real. Smile
Are you saying that these samples are able to bypass SBIE? Surprised


Not at all. He's simply saying that if he tries to run those malware files sandboxed, nothing will happen, because Sandboxie will block it. However, I think he wants Sandboxie to be able to let the malware run sandboxed to observe its behaviour, but to keep it contained of course.


Thats entirely correct. Sorry I didn't clarify myself better.
View user's profileSend private message
A virus pass sandboxie
You cannot post new topics in this forum
You can reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
All times are GMT  
Page 2 of 3  

Use the RSS feed to watch this topic for replies
  
  
 This topic is locked: you cannot edit posts or make replies.  

Sandboxie is Copyright © 2004-2012 by Sandboxie Holdings LLC.  All rights reserved.
Sandboxie.com | Contact Author
This site has been viewed 208,373,169 times since June 2004