这是我自学 MIT6.S081 操作系统课程的 lab 代码笔记第四篇:Traps。此 lab 大致耗时:8小时。 课程地址:https://pdos.csail.mit.edu/6.S081/2020/schedule.html Lab 地址:https://pdos.csail.mit.edu/6.S081/2020/labs/traps.html 我的代码地址:https://github.com/Miigon/my-xv6-labs-2020/tree/traps Commits: https://github.com/Miigon/my-xv6-labs-2020/commits/traps 本文中代码注释是编写博客的时候加入的,原仓库中的代码可能缺乏注释或代码不完全相同。
Lab 4: Traps
This lab explores how system calls are implemented using traps. You will first do a warm-up exercises with stacks and then you will implement an example of user-level trap handling.
探索 trap 实现系统调用的方式。
注意本部分主要内容其实都在lecture里(lecture 5、lecture 6),实验不是非常复杂但是以理解概念为重,trap机制、trampoline作用、函数calling convention、调用栈、特权模式、riscv汇编,这些即使都不知道可能依然能完成 lab。但是不代表这些不重要,相反这些才是主要内容,否则 lab 就算跑起来也只是盲狙,没有真正达到学习效果。
RISC-V assembly (easy)
It will be important to understand a bit of RISC-V assembly, which you were exposed to in 6.004. There is a file user/call.c in your xv6 repo. make fs.img compiles it and also produces a readable assembly version of the program in user/call.asm.
Read the code in call.asm for the functions g, f, and main. The instruction manual for RISC-V is on the reference page. Here are some questions that you should answer (store the answers in a file answers-traps.txt):
阅读 call.asm,以及 RISC-V 指令集教程,回答问题。(学习 RISC-V 汇编)
代码语言:javascript复制Q: Which registers contain arguments to functions? For example, which register holds 13 in main's call to printf?
A: a0-a7; a2;
Q: Where is the call to function f in the assembly code for main? Where is the call to g? (Hint: the compiler may inline functions.)
A: There is none. g(x) is inlined within f(x) and f(x) is further inlined into main()
Q: At what address is the function printf located?
A: 0x0000000000000628, main calls it with pc-relative addressing.
Q: What value is in the register ra just after the jalr to printf in main?
A: 0x0000000000000038, next line of assembly right after the jalr
Q: Run the following code.
unsigned int i = 0x00646c72;
printf("H%x Wo%s", 57616, &i);
What is the output?
If the RISC-V were instead big-endian what would you set i to in order to yield the same output?
Would you need to change 57616 to a different value?
A: "He110 World"; 0x726c6400; no, 57616 is 110 in hex regardless of endianness.
Q: In the following code, what is going to be printed after 'y='? (note: the answer is not a specific value.) Why does this happen?
printf("x=%d y=%d", 3);
A: A random value depending on what codes there are right before the call.Because printf tried to read more arguments than supplied.
The second argument `3` is passed in a1, and the register for the third argument, a2, is not set to any specific value before the
call, and contains whatever there is before the call.
简单翻译:
代码语言:javascript复制Q: 哪些寄存器存储了函数调用的参数?举个例子,main 调用 printf 的时候,13 被存在了哪个寄存器中?
A: a0-a7; a2;
Q: main 中调用函数 f 对应的汇编代码在哪?对 g 的调用呢? (提示:编译器有可能会内链(inline)一些函数)
A: 没有这样的代码。 g(x) 被内链到 f(x) 中,然后 f(x) 又被进一步内链到 main() 中
Q: printf 函数所在的地址是?
A: 0x0000000000000628, main 中使用 pc 相对寻址来计算得到这个地址。
Q: 在 main 中 jalr 跳转到 printf 之后,ra 的值是什么?
A: 0x0000000000000038, jalr 指令的下一条汇编指令的地址。
Q: 运行下面的代码
unsigned int i = 0x00646c72;
printf("H%x Wo%s", 57616, &i);
输出是什么?
如果 RISC-V 是大端序的,要实现同样的效果,需要将 i 设置为什么?需要将 57616 修改为别的值吗?
A: "He110 World"; 0x726c6400; 不需要,57616 的十六进制是 110,无论端序(十六进制和内存中的表示不是同个概念)
Q: 在下面的代码中,'y=' 之后会答应什么? (note: 答案不是一个具体的值) 为什么?
printf("x=%d y=%d", 3);
A: 输出的是一个受调用前的代码影响的“随机”的值。因为 printf 尝试读的参数数量比提供的参数数量多。
第二个参数 `3` 通过 a1 传递,而第三个参数对应的寄存器 a2 在调用前不会被设置为任何具体的值,而是会
包含调用发生前的任何已经在里面的值。
Backtrace (moderate)
For debugging it is often useful to have a backtrace: a list of the function calls on the stack above the point at which the error occurred.
代码语言:javascript复制Implement a backtrace() function in kernel/printf.c. Insert a call to this function in sys_sleep, and then run bttest, which calls sys_sleep. Your output should be as follows:
backtrace:
0x0000000080002cda
0x0000000080002bb6
0x0000000080002898
代码语言:javascript复制After bttest exit qemu. In your terminal: the addresses may be slightly different but if you run addr2line -e kernel/kernel (or riscv64-unknown-elf-addr2line -e kernel/kernel) and cut-and-paste the above addresses as follows:
$ addr2line -e kernel/kernel
0x0000000080002de2
0x0000000080002f4a
0x0000000080002bfc
Ctrl-D
代码语言:javascript复制You should see something like this:
kernel/sysproc.c:74
kernel/syscall.c:224
kernel/trap.c:85
添加 backtrace 功能,打印出调用栈,用于调试。
在 defs.h 中添加声明
代码语言:javascript复制// defs.h
void printf(char*, ...);
void panic(char*) __attribute__((noreturn));
void printfinit(void);
void backtrace(void); // new
在 riscv.h 中添加获取当前 fp(frame pointer)寄存器的方法:
代码语言:javascript复制// riscv.h
static inline uint64
r_fp()
{
uint64 x;
asm volatile("mv %0, s0" : "=r" (x));
return x;
}
fp 指向当前栈帧的开始地址,sp 指向当前栈帧的结束地址。 (栈从高地址往低地址生长,所以 fp 虽然是帧开始地址,但是地址比 sp 高)
栈帧中从高到低第一个 8 字节 fp-8
是 return address,也就是当前调用层应该返回到的地址。
栈帧中从高到低第二个 8 字节 fp-16
是 previous address,指向上一层栈帧的 fp 开始地址。
剩下的为保存的寄存器、局部变量等。一个栈帧的大小不固定,但是至少 16 字节。
在 xv6 中,使用一个页来存储栈,如果 fp 已经到达栈页的上界,则说明已经到达栈底。
查看 call.asm,可以看到,一个函数的函数体最开始首先会扩充一个栈帧给该层调用使用,在函数执行完毕后再回收,例子:
代码语言:javascript复制int g(int x) {
0: 1141 addi sp,sp,-16 // 扩张调用栈,得到一个 16 字节的栈帧
2: e422 sd s0,8(sp) // 将返回地址存到栈帧的第一个 8 字节中
4: 0800 addi s0,sp,16
return x 3;
}
6: 250d addiw a0,a0,3
8: 6422 ld s0,8(sp) // 从栈帧读出返回地址
a: 0141 addi sp,sp,16 // 回收栈帧
c: 8082 ret // 返回
注意栈的生长方向是从高地址到低地址,所以扩张是 -16,而回收是 16。
更多关于寄存器、栈帧以及内存调用的细节,请查看 lecture 5,或 这个很有用的 slides。
实现 backtrace 函数:
代码语言:javascript复制// printf.c
void backtrace() {
uint64 fp = r_fp();
while(fp != PGROUNDUP(fp)) { // 如果已经到达栈底
uint64 ra = *(uint64*)(fp - 8); // return address
printf("%pn", ra);
fp = *(uint64*)(fp - 16); // previous fp
}
}
在 sys_sleep 的开头调用一次 backtrace()
代码语言:javascript复制// sysproc.c
uint64
sys_sleep(void)
{
int n;
uint ticks0;
backtrace(); // print stack backtrace.
if(argint(0, &n) < 0)
return -1;
// ......
return 0;
}
编译运行:
代码语言:javascript复制$ bttest
0x0000000080002dea
0x0000000080002cc4
0x00000000800028d0
Alarm (hard)
In this exercise you’ll add a feature to xv6 that periodically alerts a process as it uses CPU time. This might be useful for compute-bound processes that want to limit how much CPU time they chew up, or for processes that want to compute but also want to take some periodic action. More generally, you’ll be implementing a primitive form of user-level interrupt/fault handlers; you could use something similar to handle page faults in the application, for example. Your solution is correct if it passes alarmtest and usertests.
按照如下原型添加系统调用 sigalarm
和 sigreturn
(具体步骤不再赘述):
int sigalarm(int ticks, void (*handler)());
int sigreturn(void);
首先,在 proc 结构体的定义中,增加 alarm 相关字段:
- alarm_interval:时钟周期,0 为禁用
- alarm_handler:时钟回调处理函数
- alarm_ticks:下一次时钟响起前还剩下的 ticks 数
- alarm_trapframe:时钟中断时刻的 trapframe,用于中断处理完成后恢复原程序的正常执行
- alarm_goingoff:是否已经有一个时钟回调正在执行且还未返回(用于防止在 alarm_handler 中途闹钟到期再次调用 alarm_handler,导致 alarm_trapframe 被覆盖)
struct proc {
// ......
int alarm_interval; // Alarm interval (0 for disabled)
void(*alarm_handler)(); // Alarm handler
int alarm_ticks; // How many ticks left before next alarm goes off
struct trapframe *alarm_trapframe; // A copy of trapframe right before running alarm_handler
int alarm_goingoff; // Is an alarm currently going off and hasn't not yet returned? (prevent re-entrance of alarm_handler)
};
sigalarm 与 sigreturn 具体实现:
代码语言:javascript复制// sysproc.c
uint64 sys_sigalarm(void) {
int n;
uint64 fn;
if(argint(0, &n) < 0)
return -1;
if(argaddr(1, &fn) < 0)
return -1;
return sigalarm(n, (void(*)())(fn));
}
uint64 sys_sigreturn(void) {
return sigreturn();
}
代码语言:javascript复制// trap.c
int sigalarm(int ticks, void(*handler)()) {
// 设置 myproc 中的相关属性
struct proc *p = myproc();
p->alarm_interval = ticks;
p->alarm_handler = handler;
p->alarm_ticks = ticks;
return 0;
}
int sigreturn() {
// 将 trapframe 恢复到时钟中断之前的状态,恢复原本正在执行的程序流
struct proc *p = myproc();
*p->trapframe = *p->alarm_trapframe;
p->alarm_goingoff = 0;
return 0;
}
在 proc.c 中添加初始化与释放代码:
代码语言:javascript复制// proc.c
static struct proc*
allocproc(void)
{
// ......
found:
p->pid = allocpid();
// Allocate a trapframe page.
if((p->trapframe = (struct trapframe *)kalloc()) == 0){
release(&p->lock);
return 0;
}
// Allocate a trapframe page for alarm_trapframe.
if((p->alarm_trapframe = (struct trapframe *)kalloc()) == 0){
release(&p->lock);
return 0;
}
p->alarm_interval = 0;
p->alarm_handler = 0;
p->alarm_ticks = 0;
p->alarm_goingoff = 0;
// ......
return p;
}
static void
freeproc(struct proc *p)
{
// ......
if(p->alarm_trapframe)
kfree((void*)p->alarm_trapframe);
p->alarm_trapframe = 0;
// ......
p->alarm_interval = 0;
p->alarm_handler = 0;
p->alarm_ticks = 0;
p->alarm_goingoff = 0;
p->state = UNUSED;
}
在 usertrap() 函数中,实现时钟机制具体代码:
代码语言:javascript复制void
usertrap(void)
{
int which_dev = 0;
// ......
if(p->killed)
exit(-1);
// give up the CPU if this is a timer interrupt.
// if(which_dev == 2) {
// yield();
// }
// give up the CPU if this is a timer interrupt.
if(which_dev == 2) {
if(p->alarm_interval != 0) { // 如果设定了时钟事件
if(--p->alarm_ticks <= 0) { // 时钟倒计时 -1 tick,如果已经到达或超过设定的 tick 数
if(!p->alarm_goingoff) { // 确保没有时钟正在运行
p->alarm_ticks = p->alarm_interval;
// jump to execute alarm_handler
*p->alarm_trapframe = *p->trapframe; // backup trapframe
p->trapframe->epc = (uint64)p->alarm_handler;
p->alarm_goingoff = 1;
}
// 如果一个时钟到期的时候已经有一个时钟处理函数正在运行,则会推迟到原处理函数运行完成后的下一个 tick 才触发这次时钟
}
}
yield();
}
usertrapret();
}
这样,在每次时钟中断的时候,如果进程有已经设置的时钟(alarm_interval != 0
),则进行 alarm_ticks 倒数。当 alarm_ticks 倒数到小于等于 0 的时候,如果没有正在处理的时钟,则尝试触发时钟,将原本的程序流保存起来(*alarm_trapframe = *trapframe
),然后通过修改 pc 寄存器的值,将程序流转跳到 alarm_handler 中,alarm_handler 执行完毕后再恢复原本的执行流(*trapframe = *alarm_trapframe
)。这样从原本程序执行流的视角,就是不可感知的中断了。
编译运行:
代码语言:javascript复制$ alarmtest
test0 start
.............alarm!
test0 passed
test1 start
..alarm!
..alarm!
..alarm!
..alarm!
..alarm!
..alarm!
..alarm!
.alarm!
...alarm!
..alarm!
test1 passed
test2 start
..............alarm!
test2 passed
Optional challenge exercises
Print the names of the functions and line numbers in backtrace() instead of numerical addresses (hard).
xv6 默认的编译模式会在生成的可执行文件中,含有调试信息,其中包含了所有符号的名称以及其对应的地址。理论上 backtrace 可以做类似 addr2line 的操作,通过解析可执行文件本身附带的调试信息获得地址对应的源码文件以及行号。这里跳过该 challenge.