Skip to contentSkip to conversation

system call logging in userspace

DISCUSSION Chad Hogan 2 participants 2 messages Apr 13, 2001 – Apr 16, 2001

lore
Share & tools
lore

Chad Hogan

author
#0 lore raw

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hello,

I'm not very experienced with dealing directly with the kernel, so I was hoping for a little advice...

I'd like to implement some sort of rudimentary (file)system-call logging. Specifically, I'd like information about write, open, creat, unlink, and maybe a few others to be pushed into userspace.  Mostly, I'd just like to know what files are being created, modified, and deleted as it happens.

It seems quite easy to me -- I was thinking of doing this with a module. I'll just grab the pointer from sys_call_table[__NR_open] and replace it with my own little wrapper that does nothing but call the original function, and then log the call in some manner.

================

asmlinkage int my_sys_open(const char *fname, int flags, int mode) {
         [preliminary stuff]

         returnval = real_sys_open(fname, flags, mode);

         [log information based on returnval, fname, whatever];

         return returnval;
}

int init_module()
{
         [other stuff]

         real_sys_open = sys_call_table[__NR_open]; sys_call_table[__NR_open] = my_sys_open; return 0;
}

init cleanup_module()
{
         sys_call_table[__NR_open] = real_sys_open; }

===========

The simplicity of the whole thing is what scares me a little bit.  Am I being horribly naive about something here?  It seems like an obviously useful module to have around, and yet I've never seen it and I couldn't find anyone who had done it already.  Is there a much better way to accomplish this than loading in a module?  Am I risking serious fs corruption?

It occurs to me that I may have some problems if something else changes the sys_call_table[__NR_open] and the two modules don't cooperate...

Thanks.
- --
Chad Hogan [email protected]
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.4 (FreeBSD)
Comment: For info see http://www.gnupg.org

iD8DBQE61kkHiSF5tViVwg0RAkMOAJ4rMTC/xvvknmiSf512Y5d06ezdpgCfZH+s rEQ6ltXalr2SVqFg7lhIFYc=
=iBPm
-----END PGP SIGNATURE-----

#1 lore raw

Man strace, or http://subterfugue.org

Quoted 70 lines

Hello,

I'm not very experienced with dealing directly with the kernel, so I was hoping for a little advice...

I'd like to implement some sort of rudimentary (file)system-call logging. Specifically, I'd like information about write, open, creat, unlink, and maybe a few others to be pushed into userspace.  Mostly, I'd just like to know what files are being created, modified, and deleted as it happens.

It seems quite easy to me -- I was thinking of doing this with a module. I'll just grab the pointer from sys_call_table[__NR_open] and replace it with my own little wrapper that does nothing but call the original function, and then log the call in some manner.

================

asmlinkage int my_sys_open(const char *fname, int flags, int mode) {
         [preliminary stuff]

         returnval = real_sys_open(fname, flags, mode);

         [log information based on returnval, fname, whatever];

         return returnval;
}

int init_module()
{
         [other stuff]

         real_sys_open = sys_call_table[__NR_open]; sys_call_table[__NR_open] = my_sys_open; return 0;
}

init cleanup_module()
{
         sys_call_table[__NR_open] = real_sys_open; }

===========

The simplicity of the whole thing is what scares me a little bit.  Am I being horribly naive about something here?  It seems like an obviously useful module to have around, and yet I've never seen it and I couldn't find anyone who had done it already.  Is there a much better way to accomplish this than loading in a module?  Am I risking serious fs corruption?

It occurs to me that I may have some problems if something else changes the sys_call_table[__NR_open] and the two modules don't cooperate...

Thanks.
- --
Chad Hogan [email protected]
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.4 (FreeBSD)
Comment: For info see http://www.gnupg.org

iD8DBQE61kkHiSF5tViVwg0RAkMOAJ4rMTC/xvvknmiSf512Y5d06ezdpgCfZH+s rEQ6ltXalr2SVqFg7lhIFYc=
=iBPm
-----END PGP SIGNATURE-----
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

Signature · 2 lines
Philips Velo 1: 1"x4"x8", 300gram, 60, 12MB, 40bogomips, linux, mutt,
details at http://atrey.karlin.mff.cuni.cz/~pavel/velo/index.html.