the "sys_mount" function allocates a block of memory whose size is 4096 bytes through buddy system(__get_free_page) or slab system(kmem_cache_alloc) for every parameters, then copy the parameters to the memory from user space.
the copy process of parameters is accomplished in "getname" and "copy_mount_options" functions.
now, i want to simplify the function according to the following method, but i don't know wether there is some potential risk.
1. don't to allocates memory in kernel space for each parameter.
2. sys_mount refer to the parameter through the memory pointer passed directly.
kernel version: 2.6.14
hard platform: ARM926ejs
the modified source of "sys_mount" function as following:
asmlinkage long sys_mount(char __user * dev_name, char __user * dir_name,
char __user * type, unsigned long flags,
void __user * data)
{
int retval;
unsigned long data_page;
unsigned long type_page;
unsigned long dev_page;
char *dir_page;
/* the first modification, comment the following code.
retval = copy_mount_options (type, &type_page); if (retval < 0)
return retval;
dir_page = getname(dir_name);
retval = PTR_ERR(dir_page);
if (IS_ERR(dir_page))
goto out1;
retval = copy_mount_options (dev_name, &dev_page); if (retval < 0)
goto out2;
retval = copy_mount_options (data, &data_page); if (retval < 0)
goto out3;
*/
lock_kernel();
/* the second modification, refer to the parameters by memory pointer inputed directly */ #if 0 retval = do_mount((char*)dev_page, dir_page, (char*)type_page,
flags, (void*)data_page);
#else
retval = do_mount((char*)dev_name, dir_name, (char*)type, flags, (void*)data); #endif unlock_kernel();
/*
free_page(data_page);
out3:
free_page(dev_page);
out2:
putname(dir_page);
out1:
free_page(type_page);
*/
return retval;
}
if there is no potential risk, the speed of "mount" command would be improved greatly.
Tell me plz if there is problem. thax.
--
To unsubscribe from this list: send an email with
"unsubscribe kernelnewbies" to ecartis@xxxxxxxxxxxx
Please read the FAQ at http://kernelnewbies.org/FAQ
