最近在和同学们一起学习Linux驱动开发方面的知识,在学习过程中经常会遇到一些关于内核的知识,牵扯到一些数据结构和函数, 在遇到时我们没有绕过去,而是直接去看源代码,在此期间收获了不少,发现内核写的确实很好很强大,要一下搞定它,那是痴心妄想,所以我从一些小的地方开始,在遇到内核源码时就去读它。今天读到一个有用的函数:copy_to_user(),其在内核定义如下:
835 /** 836 * copy_to_user: - Copy a block of data into user space. 837 * @to: Destination address, in user space. 838 * @from: Source address, in kernel space. 839 * @n: Number of bytes to copy. 840 * 841 * Context: User context only. This function may sleep. 842 * 843 * Copy data from kernel space to user space. 844 * 845 * Returns number of bytes that could not be copied. 846 * On success, this will be zero. 847 */ 848 unsigned long 849 copy_to_user(void __user *to, const void *from, unsigned long n) 850 { 851 if (access_ok(VERIFY_WRITE, to, n)) 852 n = __copy_to_user(to, from, n); 853 return n; 854 } |
相信稍微懂英语的人都能读懂前面的注释,其功能是将内核空间的内容复制到用户空间,所复制的内容是从from来,到to去,复制n个字节。
其中又牵扯到两个函数:access_ok()和__copy_to_user(),好我们继续往下深入,先来看看第一个函数access_ok()的源码:
85/** 86 * access_ok: - Checks if a user space pointer is valid 87 * @type: Type of access: %VERIFY_READ or %VERIFY_WRITE. Note that 88 * %VERIFY_WRITE is a superset of %VERIFY_READ - if it is safe 89 * to write to a block, it is always safe to read from it. 90 * @addr: User space pointer to start of block to check 91 * @size: Size of block to check 92 * 93 * Context: User context only. This function may sleep. 94 * 95 * Checks if a pointer to a block of memory in user space is valid. 96 * 97 * Returns true (nonzero) if the memory block may be valid, false (zero) 98 * if it is definitely invalid. 99 * 100 * Note that, depending on architecture, this function probably just 101 * checks that the pointer is in the user space range - after calling 102 * this function, memory access functions may still return -EFAULT. 103 */ 104#ifdef CONFIG_MMU 105#define access_ok(type,addr,size) (likely(__range_ok(addr,size) == 0)) 106#else 107static inline int access_ok(int type, const void *addr, unsigned long size) 108{ 109 extern unsigned long memory_start, memory_end; 110 unsigned long val = (unsigned long)addr; 111 112 return ((val >= memory_start) && ((val + size) < memory_end)); 113} 114#endif /* CONFIG_MMU */ |
其功能是检查用户空间是否合法,它的第一个参数:type,有两种类型:VERIFY_READ 和VERIFY_WRITE,前者为可读,后者可写,注意:如果标志为可写(VERIFY_WRITE)时,必然可读!因为可写是可读的超集(%VERIFY_WRITE is a superset of %VERIFY_READ)。
检查过程如下:addr为起始地址,size为所要复制的大小,那么从addr到addr+size则是所要检查的空间,如果它的范围在memory_start和memory_end之间的话,则返回真。至于memory_start详细信息,我没有读。
到此为止,如果检查合法,那么OK,我们来实现真正的复制功能:__copy_to_user(),其源码定义如下:
80static inline __kernel_size_t __copy_to_user(void __user *to, const void *from, 81 __kernel_size_t n) 82{ 83 return __copy_user((void __force *)to, from, n); 84} |
哈哈,又遇到一个函数:__copy_user(),这个函数才真正在做底层的复制工作。想继续深入的:看这里
()。
我的分析到此为止。
阅读(2046) | 评论(2) | 转发(0) |