ground up写的真好

书中有几句话完全不该翻译，原文透出的那种确信和通透是应该保留下来，讲到文件操作这一章时，忍不住要做些笔记：
原文保留着，笔记写在旁边。

1.open操作,需要几个信息：文件名，操作方式，许可。

%eax 存系统调用的信号 ，为5.
%ebx 存文件第一个字符的地址。
%ecx 操作方式，read（0）/write（03101）。
%edx 许可 （0666）,不太了解，不过只是规则，需要时再查。

用法

程序：喂cpu！我给你发了一个中断（int $0x80），我要打开一个文件。
cpu：好的，中断收到，但我需要一些信息，你得一起给我，哦，你给了，我看看。
%eax中是5，所以你是要打开文件
%ebx中是一个4字节数据，我会把它当作文件的首地址
%ecx中是零，所以我直到你是要读这个文件
%edx中是0666，权限代码正确
我为你打开了这个文件，独有的识别码我放在了%eax中，请注意查收。

        movl $5,%eax
        movl 8(%ebp),%ebx
        movl $0,%ecx
        movl $0666,%edx
        int $0x80

程序：好的，收到，我会把它保存在%ebp位置的下方一个字节

        movl %eax,-4(%ebp)

程序：对了我刚才打开的是要读取的文件，我还要打开一个需要写入的文件，再接受一次中断吧！

        movl $5,%eax
        movl 12(%ebp),%ebx
        movl $03101,%ecx
        movl $0666,%edx
        int $0x80

cpu：%eax 5 打开 %ebx文件地址 %ecx 03101 写 %edx 0666 权限正确 搞定了！去%eax取这个文件识别码就好。

2.linux返回一串文件代号，存入%eax，这段信息将代表上面的文件。

3.下一步到了文件的具体操作。

read：

系统调用信号：3
%ebx ：文件代号
%ecx ： 缓冲区地址
%edx：缓冲区大小

write：

系统调用信号：4
%ebx ：文件代号
%ecx ： 缓冲区地址
%edx：缓冲区大小

4.用完的文件记得关掉

系统调用信号：6
%ebx：文件代号

1. Tell Linux the name of the file to open, and in what mode you want it opened
   (read, write, both read and write, create it if it doesn’t exist, etc.). This is
   handled with the open system call, which takes a filename, a number
   representing the mode, and a permission set as its parameters. %eax will hold
   the system call number, which is 5. The address of the first character of the
   filename should be stored in %ebx. The read/write intentions, represented as a
   number, should be stored in %ecx. For now, use 0 for files you want to read
   from, and 03101 for files you want to write to (you must include the leading
   zero)Finally, the permission set should be stored as a number in %edx. If
   you are unfamiliar with UNIX permissions, just use 0666 for the permissions
   (again, you must include the leading zero).
2. Linux will then return to you a file descriptor in %eax. Remember, this is a
   number that you use to refer to this file throughout your program.
3. Next you will operate on the file doing reads and/or writes, each time giving
   Linux the file descriptor you want to use. read is system call 3, and to call it
   you need to have the file descriptor in %ebx, the address of a buffer for storing
   the data that is read in %ecx, and the size of the buffer in %edx. Buffers will be
   explained in the Section called Buffers and .bss. read will return with either
   the number of characters read from the file, or an error code. Error codes can
   be distinguished because they are always negative numbers (more information
   on negative numbers can be found in Chapter 10). write is system call 4, and
   it requires the same parameters as the read system call, except that the buffer
   should already be filled with the data to write out. The write system call will
   give back the number of bytes written in %eax or an error code.
4. When you are through with your files, you can then tell Linux to close them.
   Afterwards, your file descriptor is no longer valid. This is done using close,
   system call 6. The only parameter to close is the file descriptor, which is
   placed in %ebx