Intercepting memory mapped files

Intercepting memory mapped files

[Kunal Baweja]

Hi,
I am currently working on implementing a simple encrypted file system using
a stackable file system WrapFS <http://wrapfs.filesystems.org/>. In order
to do so, as I read in the WrapFS mailing list
<http://www.fsl.cs.sunysb.edu/pipermail/wrapfs/2011-November/000058.html>
one needs to implement address space operations, which I have done and
posted on my github at this link
<https://github.com/bawejakunal/wrapfs-addresspace-operations/blob/master/mmap.c>
and it works fine. But now when I am trying to modify the data before being
written in wrapfs_lower() function I get a blank output in files written
out.
Please guide where am I going wrong in modifying the data to be written and
what could be the best way to decrypt it ?

Thanks

*Kunal Baweja*

Undergraduate B.E. (Hons.) Computer Science
Birla Institute Of Technology & Science, Pilani
K.K. Birla Goa Campus
+91 9049655665
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.kernelnewbies.org/pipermail/kernelnewbies/attachments/20150130/3ec8d9d9/attachment.html 

Intercepting memory mapped files

[Saket Sinha]

HI,

Kindly refer the following link that implements a minimal encryption over
wrapfs.
https://github.com/piekill/wrapfs

Go through the wrapfs_encrypt in wrapfs_readpage and wrapfs_decrypt in
wrapfs_writepage.

Regards,
Saket Sinha

On Fri, Jan 30, 2015 at 3:56 PM, Kunal Baweja <bawejakunal15@gmail.com>
wrote:

> Hi,
> I am currently working on implementing a simple encrypted file system
> using a stackable file system WrapFS <http://wrapfs.filesystems.org/>. In
> order to do so, as I read in the WrapFS mailing list
> <http://www.fsl.cs.sunysb.edu/pipermail/wrapfs/2011-November/000058.html>
> one needs to implement address space operations, which I have done and
> posted on my github at this link
> <https://github.com/bawejakunal/wrapfs-addresspace-operations/blob/master/mmap.c>
> and it works fine. But now when I am trying to modify the data before being
> written in wrapfs_lower() function I get a blank output in files written
> out.
> Please guide where am I going wrong in modifying the data to be written
> and what could be the best way to decrypt it ?
>
> Thanks
>
> *Kunal Baweja*
>
> Undergraduate B.E. (Hons.) Computer Science
> Birla Institute Of Technology & Science, Pilani
> K.K. Birla Goa Campus
> +91 9049655665
>
> _______________________________________________
> Kernelnewbies mailing list
> Kernelnewbies at kernelnewbies.org
> http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.kernelnewbies.org/pipermail/kernelnewbies/attachments/20150130/978f8ae8/attachment.html 

Intercepting memory mapped files

[Kunal Baweja]

Yeah I was trying something similar only but with a much simpler Caeser
cipher function of my own written as below:

void encrypt(char *data, char *encrypted, size_t size)
{
    unsigned int i;
    for(i=0;i<(unsigned int)size;i++)
        encrypted[i] = data[i] + 3;
    printk(KERN_INFO "%s",encrypted);
    return;
}

So in the given code wherever u see the calls to *wrapfs_encrypt*, I have
replaced with my simpler encryption function.

But I am not able to figure out why calling this function creates a
problem, the printk() statement inside the function shows me the correctly
encrypted contents but the file is still blank.

Thanks

*Kunal Baweja*

Undergraduate B.E. (Hons.) Computer Science
Birla Institute Of Technology & Science, Pilani
K.K. Birla Goa Campus
+91 9049655665

On Fri, Jan 30, 2015 at 4:03 PM, Saket Sinha <saket.sinha89@gmail.com>
wrote:

> HI,
>
> Kindly refer the following link that implements a minimal encryption over
> wrapfs.
> https://github.com/piekill/wrapfs
>
> Go through the wrapfs_encrypt in wrapfs_readpage and wrapfs_decrypt in
> wrapfs_writepage.
>
> Regards,
> Saket Sinha
>
> On Fri, Jan 30, 2015 at 3:56 PM, Kunal Baweja <bawejakunal15@gmail.com>
> wrote:
>
>> Hi,
>> I am currently working on implementing a simple encrypted file system
>> using a stackable file system WrapFS <http://wrapfs.filesystems.org/>.
>> In order to do so, as I read in the WrapFS mailing list
>> <http://www.fsl.cs.sunysb.edu/pipermail/wrapfs/2011-November/000058.html>
>> one needs to implement address space operations, which I have done and
>> posted on my github at this link
>> <https://github.com/bawejakunal/wrapfs-addresspace-operations/blob/master/mmap.c>
>> and it works fine. But now when I am trying to modify the data before being
>> written in wrapfs_lower() function I get a blank output in files written
>> out.
>> Please guide where am I going wrong in modifying the data to be written
>> and what could be the best way to decrypt it ?
>>
>> Thanks
>>
>> *Kunal Baweja*
>>
>> Undergraduate B.E. (Hons.) Computer Science
>> Birla Institute Of Technology & Science, Pilani
>> K.K. Birla Goa Campus
>> +91 9049655665
>>
>> _______________________________________________
>> Kernelnewbies mailing list
>> Kernelnewbies at kernelnewbies.org
>> http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies
>>
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.kernelnewbies.org/pipermail/kernelnewbies/attachments/20150130/1048de10/attachment-0001.html 

Intercepting memory mapped files

[devendra.aaru]

On Fri, Jan 30, 2015 at 6:48 AM, Kunal Baweja <bawejakunal15@gmail.com>
wrote:

> Yeah I was trying something similar only but with a much simpler Caeser
> cipher function of my own written as below:
>
> void encrypt(char *data, char *encrypted, size_t size)
> {
>     unsigned int i;
>     for(i=0;i<(unsigned int)size;i++)
>         encrypted[i] = data[i] + 3;
>

This is not caesar's cipher. If your buffer contains alphabets then only
you can apply caesar cipher.

It should be something like

          encrypted[i] = (data[i] - 3) % 26;



>     printk(KERN_INFO "%s",encrypted);
>     return;
> }
>
> So in the given code wherever u see the calls to *wrapfs_encrypt*, I have
> replaced with my simpler encryption function.
>
> But I am not able to figure out why calling this function creates a
> problem, the printk() statement inside the function shows me the correctly
> encrypted contents but the file is still blank.
>
> Thanks
>
> *Kunal Baweja*
>
> Undergraduate B.E. (Hons.) Computer Science
> Birla Institute Of Technology & Science, Pilani
> K.K. Birla Goa Campus
> +91 9049655665
>
> On Fri, Jan 30, 2015 at 4:03 PM, Saket Sinha <saket.sinha89@gmail.com>
> wrote:
>
>> HI,
>>
>> Kindly refer the following link that implements a minimal encryption over
>> wrapfs.
>> https://github.com/piekill/wrapfs
>>
>> Go through the wrapfs_encrypt in wrapfs_readpage and wrapfs_decrypt in
>> wrapfs_writepage.
>>
>> Regards,
>> Saket Sinha
>>
>> On Fri, Jan 30, 2015 at 3:56 PM, Kunal Baweja <bawejakunal15@gmail.com>
>> wrote:
>>
>>> Hi,
>>> I am currently working on implementing a simple encrypted file system
>>> using a stackable file system WrapFS <http://wrapfs.filesystems.org/>.
>>> In order to do so, as I read in the WrapFS mailing list
>>> <http://www.fsl.cs.sunysb.edu/pipermail/wrapfs/2011-November/000058.html>
>>> one needs to implement address space operations, which I have done and
>>> posted on my github at this link
>>> <https://github.com/bawejakunal/wrapfs-addresspace-operations/blob/master/mmap.c>
>>> and it works fine. But now when I am trying to modify the data before being
>>> written in wrapfs_lower() function I get a blank output in files written
>>> out.
>>> Please guide where am I going wrong in modifying the data to be written
>>> and what could be the best way to decrypt it ?
>>>
>>> Thanks
>>>
>>> *Kunal Baweja*
>>>
>>> Undergraduate B.E. (Hons.) Computer Science
>>> Birla Institute Of Technology & Science, Pilani
>>> K.K. Birla Goa Campus
>>> +91 9049655665
>>>
>>> _______________________________________________
>>> Kernelnewbies mailing list
>>> Kernelnewbies at kernelnewbies.org
>>> http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies
>>>
>>>
>>
>
> _______________________________________________
> Kernelnewbies mailing list
> Kernelnewbies at kernelnewbies.org
> http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.kernelnewbies.org/pipermail/kernelnewbies/attachments/20150130/f63f93e7/attachment.html 

Intercepting memory mapped files

[Kunal Baweja]

Hi devendra, it's not necessary to wrap around the characters in caeser
cipher by taking a modulus like you pointed out, there are many variants of
it. Anyways my actual problem is that the encrypted data is not being
written into the file whereas if I don't make a call to the encrypt
function then the contents are written perfectly to the files.

On 5:25pm, Fri 30-Jan-2015 devendra.aaru <devendra.aaru@gmail.com> wrote:

> On Fri, Jan 30, 2015 at 6:48 AM, Kunal Baweja <bawejakunal15@gmail.com>
> wrote:
>
>> Yeah I was trying something similar only but with a much simpler Caeser
>> cipher function of my own written as below:
>>
>> void encrypt(char *data, char *encrypted, size_t size)
>> {
>>     unsigned int i;
>>     for(i=0;i<(unsigned int)size;i++)
>>         encrypted[i] = data[i] + 3;
>>
>
> This is not caesar's cipher. If your buffer contains alphabets then only
> you can apply caesar cipher.
>
> It should be something like
>
>           encrypted[i] = (data[i] - 3) % 26;
>
>
>
>>     printk(KERN_INFO "%s",encrypted);
>>     return;
>> }
>>
>> So in the given code wherever u see the calls to *wrapfs_encrypt*, I
>> have replaced with my simpler encryption function.
>>
>> But I am not able to figure out why calling this function creates a
>> problem, the printk() statement inside the function shows me the correctly
>> encrypted contents but the file is still blank.
>>
>> Thanks
>>
>> *Kunal Baweja*
>>
>> Undergraduate B.E. (Hons.) Computer Science
>> Birla Institute Of Technology & Science, Pilani
>> K.K. Birla Goa Campus
>> +91 9049655665
>>
>> On Fri, Jan 30, 2015 at 4:03 PM, Saket Sinha <saket.sinha89@gmail.com>
>> wrote:
>>
>>> HI,
>>>
>>> Kindly refer the following link that implements a minimal encryption
>>> over wrapfs.
>>> https://github.com/piekill/wrapfs
>>>
>>> Go through the wrapfs_encrypt in wrapfs_readpage and wrapfs_decrypt in
>>> wrapfs_writepage.
>>>
>>> Regards,
>>> Saket Sinha
>>>
>>> On Fri, Jan 30, 2015 at 3:56 PM, Kunal Baweja <bawejakunal15@gmail.com>
>>> wrote:
>>>
>>>> Hi,
>>>> I am currently working on implementing a simple encrypted file system
>>>> using a stackable file system WrapFS <http://wrapfs.filesystems.org/>.
>>>> In order to do so, as I read in the WrapFS mailing list
>>>> <http://www.fsl.cs.sunysb.edu/pipermail/wrapfs/2011-November/000058.html>
>>>> one needs to implement address space operations, which I have done and
>>>> posted on my github at this link
>>>> <https://github.com/bawejakunal/wrapfs-addresspace-operations/blob/master/mmap.c>
>>>> and it works fine. But now when I am trying to modify the data before being
>>>> written in wrapfs_lower() function I get a blank output in files written
>>>> out.
>>>> Please guide where am I going wrong in modifying the data to be written
>>>> and what could be the best way to decrypt it ?
>>>>
>>>> Thanks
>>>>
>>>> *Kunal Baweja*
>>>>
>>>> Undergraduate B.E. (Hons.) Computer Science
>>>> Birla Institute Of Technology & Science, Pilani
>>>> K.K. Birla Goa Campus
>>>> +91 9049655665
>>>>
>>>> _______________________________________________
>>>> Kernelnewbies mailing list
>>>> Kernelnewbies at kernelnewbies.org
>>>> http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies
>>>>
>>>>
>>>
>>
>> _______________________________________________
>> Kernelnewbies mailing list
>> Kernelnewbies at kernelnewbies.org
>> http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies
>>
>>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.kernelnewbies.org/pipermail/kernelnewbies/attachments/20150130/ce97bc9a/attachment.html 

Intercepting memory mapped files

[Valdis.Kletnieks at vt.edu]

On Fri, 30 Jan 2015 17:18:21 +0530, Kunal Baweja said:

> void encrypt(char *data, char *encrypted, size_t size)
> {
>     unsigned int i;
>     for(i=0;i<(unsigned int)size;i++)
>         encrypted[i] = data[i] + 3;
>     printk(KERN_INFO "%s",encrypted);
>     return;
> }
>
> So in the given code wherever u see the calls to *wrapfs_encrypt*, I have
> replaced with my simpler encryption function.
>
> But I am not able to figure out why calling this function creates a
> problem, the printk() statement inside the function shows me the correctly
> encrypted contents but the file is still blank.

Stupid question, but... once you call encrypt(data, encrypted, size),
do you then write out the 'encrypted' buffer, or do you write out 'data'?

(Bonus points for avoiding a memory leak here :)
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 848 bytes
Desc: not available
Url : http://lists.kernelnewbies.org/pipermail/kernelnewbies/attachments/20150130/12ef82ab/attachment.bin 

Intercepting memory mapped files

[Kunal Baweja]

Obviously I am writing the encrypted data :-/ I will post my code on github
in an hour so that everyone gets to know what exactly I am doing :-)

On 6:44pm, Fri 30-Jan-2015 null <Valdis.Kletnieks@vt.edu> wrote:

> On Fri, 30 Jan 2015 17:18:21 +0530, Kunal Baweja said:
>
> > void encrypt(char *data, char *encrypted, size_t size)
> > {
> >     unsigned int i;
> >     for(i=0;i<(unsigned int)size;i++)
> >         encrypted[i] = data[i] + 3;
> >     printk(KERN_INFO "%s",encrypted);
> >     return;
> > }
> >
> > So in the given code wherever u see the calls to *wrapfs_encrypt*, I have
> > replaced with my simpler encryption function.
> >
> > But I am not able to figure out why calling this function creates a
> > problem, the printk() statement inside the function shows me the
> correctly
> > encrypted contents but the file is still blank.
>
> Stupid question, but... once you call encrypt(data, encrypted, size),
> do you then write out the 'encrypted' buffer, or do you write out 'data'?
>
> (Bonus points for avoiding a memory leak here :)
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.kernelnewbies.org/pipermail/kernelnewbies/attachments/20150130/c5d5950a/attachment.html 

Intercepting memory mapped files

[Valdis.Kletnieks at vt.edu]

On Fri, 30 Jan 2015 06:55:53 -0500, "devendra.aaru" said:

> This is not caesar's cipher. If your buffer contains alphabets then only
> you can apply caesar cipher.
>
> It should be something like
>
>           encrypted[i] = (data[i] - 3) % 26;

Consider the sequence of 4 hex bytes 0x17314B65.  Your code will output
that as 0x14141414.  What happens when you try to decrypt that?

Issue two:  Your code fails to re-add an  'a' or 'A' as appropriate, so
instead of rotating through ASCII hex codes 41-5A and 61-7A,it smashes
then down to 0-19.

If you're going that route, you need to go the *whole* way down that route:

temp = data[i];
if (temp >= 'A' && temp <= 'Z') encrypted[i] = (temp -3) % 26 + 'A';
else if (temp >= 'a' && temp <= 'z') encrypted[i] = (temp -3) % 26 + 'a';
else encrypted[i] = temp;

Moral of the story: Designing secure usable crypto is a *lot* harder
than it looks.

Also, I'm going to put in a plug for Bruce Schneier's 'Applied Cryptography,
Second Edition' - a must-read if you're planning to do actual crypto work.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 848 bytes
Desc: not available
Url : http://lists.kernelnewbies.org/pipermail/kernelnewbies/attachments/20150130/9e9cb1a9/attachment.bin 

Intercepting memory mapped files

[Valdis.Kletnieks at vt.edu]

On Fri, 30 Jan 2015 13:17:10 +0000, Kunal Baweja said:

> Obviously I am writing the encrypted data :-/

Just checking.  You'd be *amazed* at how many times in the last 35 years
of code hacking, I've asked questions like that and there's a 10 second
pause before the person goes into Emily Litella mode... "Nevermind...".

(For those who don't get it - https://screen.yahoo.com/weekend-emily-litella-violins-tv-000000080.html
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 848 bytes
Desc: not available
Url : http://lists.kernelnewbies.org/pipermail/kernelnewbies/attachments/20150130/d67de01b/attachment-0001.bin 

Intercepting memory mapped files

[Kunal Baweja]

Ok for the sake of correctness and robustness of caeser cipher I have made
the suggested changes to it and the code is posted here:
https://github.com/bawejakunal/wrapfs-caeser/blob/master/mmap.c Now coming
back to the actual problem can someone please help me as to why isn't the
encrypted data being written to the files. The printk() statements do show
me the encrypted data though.

*Kunal Baweja*

Undergraduate B.E. (Hons.) Computer Science
Birla Institute Of Technology & Science, Pilani
K.K. Birla Goa Campus
+91 9049655665

On Fri, Jan 30, 2015 at 7:05 PM, <Valdis.Kletnieks@vt.edu> wrote:

> On Fri, 30 Jan 2015 13:17:10 +0000, Kunal Baweja said:
>
> > Obviously I am writing the encrypted data :-/
>
> Just checking.  You'd be *amazed* at how many times in the last 35 years
> of code hacking, I've asked questions like that and there's a 10 second
> pause before the person goes into Emily Litella mode... "Nevermind...".
>
> (For those who don't get it -
> https://screen.yahoo.com/weekend-emily-litella-violins-tv-000000080.html
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.kernelnewbies.org/pipermail/kernelnewbies/attachments/20150130/c3abe1bf/attachment.html 

Intercepting memory mapped files

[Kunal Baweja]

Ok somehow after introducing the robustness in caeser cipher it is working
correctly. Thanks. But the questions still remains why wasn't the data
being written to the file previously ?

*Kunal Baweja*

Undergraduate B.E. (Hons.) Computer Science
Birla Institute Of Technology & Science, Pilani
K.K. Birla Goa Campus
+91 9049655665

On Fri, Jan 30, 2015 at 7:31 PM, Kunal Baweja <bawejakunal15@gmail.com>
wrote:

> Ok for the sake of correctness and robustness of caeser cipher I have made
> the suggested changes to it and the code is posted here:
> https://github.com/bawejakunal/wrapfs-caeser/blob/master/mmap.c Now
> coming back to the actual problem can someone please help me as to why
> isn't the encrypted data being written to the files. The printk()
> statements do show me the encrypted data though.
>
> *Kunal Baweja*
>
> Undergraduate B.E. (Hons.) Computer Science
> Birla Institute Of Technology & Science, Pilani
> K.K. Birla Goa Campus
> +91 9049655665
>
> On Fri, Jan 30, 2015 at 7:05 PM, <Valdis.Kletnieks@vt.edu> wrote:
>
>> On Fri, 30 Jan 2015 13:17:10 +0000, Kunal Baweja said:
>>
>> > Obviously I am writing the encrypted data :-/
>>
>> Just checking.  You'd be *amazed* at how many times in the last 35 years
>> of code hacking, I've asked questions like that and there's a 10 second
>> pause before the person goes into Emily Litella mode... "Nevermind...".
>>
>> (For those who don't get it -
>> https://screen.yahoo.com/weekend-emily-litella-violins-tv-000000080.html
>>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.kernelnewbies.org/pipermail/kernelnewbies/attachments/20150130/861778f2/attachment.html