hi ,Hello,
yes, it is.
I started to read docs and examples about LKM, and i
have some doubts, if someone could help me. :)
1 - In the nowsday is necessary use get_user(),
put_user(), copy_from_user(),copy_to_user() to
transform from kernel to user memory and vice-versa ?
Or it's automatically actually ? I'm asking it because
i can manipulate normal data received directilly as
argument from syscall without use this functions (as
described in text that i'm reading). :)
2 - I saw that i must use get_fs() and set_fs() to
access DS (data segment) from kernel or User space,
and in article it's describe that it can be used to
"execute" libc functions over kernel. So i tryed it:
int uid;
mm_segment_t old_fs;
old_fs = get_fs();
set_fs(get_ds());
uid = getuid();
printk("\nUID %d",uid);
set_fs(old_fs);
But when i try insmod it:
libc-set_fs.o: unresolved symbol getuid
yes, u would get that error because, you should be using sys_getuid() rather than getuid() , since the kernel doesnt know what getuid() is since, it is a userspace function, and after insmod or rather during insmod your module is being linked with the kernel ,libc would internally call sys_getuid() which is transparent to the user.
Ok, i know that i can get my uid via current struct,
but i want to know how to i execute libc functions
over kernel. How to do it ? Can someone give me a
example ? :)
int uid;
mm_segment_t old_fs;
old_fs = get_fs();
set_fs(get_ds());
uid = sys_getuid(); // kernel knows about sys_getuid() only !
printk("\nUID %d",uid);
set_fs(old_fs);
cheers,
Amith
