Hi David,
Yes, I have studied something about security. The problem is that I
cannot express correctly in english (it’s not my native language) what I
want to say.
The filter driver and the two-sector-code were working with a hardcoded
key, because I wanted to obviate the authentication phase.
Now, the filter driver works. My next goal is to remove the harcoded
key from the code (from filter and boot) and to add the authentication.
When the user logs in, his password will be converted to the key. The
code that hooks int 13h will use that key to process the accesses to the
disk.
But, when xp loads my driver (this driver has now no hardcoded key) I
need to send the key (from the authentication) in the int 13h code to
the loading driver. How can I do this?
Thanks,
Rafa.
P.S.: Thank you to everyone that helped me with the corruption problem.
Now I see that was a foolish question but I can assure you, David, that
I reviewed the database list and I didn’t find something useful for my
problem. Possibly I was looking for in the wrong way. I’m sorry.
Oh, Peter, I didn’t think about those cases, hibernation and similar.
Thank you, I will reactivate that code. It was disabled when I was
investigating the filesystem corruption. When the filter was corrected,
I left the code disabled because I thought that was better keeping it so.
David J. Craig wrote:
Have you ever done any study about security? A ‘password’ is useless. If
the code knows the encryption key, then your protection is non-existent.
You must find a way to pass the encryption key from the user input to the
code so disk access can only be obtained if the user provides the key. Once
upon a time in MS-DOS, I wrote a disk drive encryptor. The user password
was used to generate the encryption key. The key was not present and could
not be derived from studying the code. This did mean the ‘user password’
was really a ‘system password’ that had to be shared by all users of that
computer. Today I would consider using SmartCards to store the encryption
key so it would only boot when the card was present and its pin number
properly entered. This is a two factor access method which means a user
must ‘have’ something and ‘know’ something before access is granted.
“Rafa” wrote in message news:xxxxx@ntdev…
>
>>Hi all,
>>
>>thank you for the previous solution. Now the filter driver works perfectly
>>and there are no corruptions in the filesystem.
>>
>>I’m planning to introduce an authentication phase before the xp boots. Now
>>is harcoded in the two-sector-code and the filter driver.
>>
>>With the adequate password, my int 13h will decrypt the sectors read but
>>how can I pass the password to the filter driver? I cannot write it to the
>>disk. Using a special service when calling the int 13h seems valid but I
>>suppose that when my driver is loaded, it cannot call this interrupt.
>>
>>So, anyone knows a good method to pass the password to the filter driver?
>>
>>Thanks in advance,
>>
>>Rafa.
>>
>
>
>
>
> —
> Questions? First check the Kernel Driver FAQ at http://www.osronline.com/article.cfm?id=256
>
> You are currently subscribed to ntdev as: xxxxx@dilmun.ls.fi.upm.es
> To unsubscribe send a blank email to xxxxx@lists.osr.com
>
