$ra寄存器中存入的是pc的值（程序运行处的地址），调用函数时，在跳转前，必须保存当前地址（pc的值），以便后来返回。jal $ra 保存后跳转，jr $ra,返回到跳转前，通过$ra保存进入上层栈地址。
jal
直接跳转指令，并带有链接功能，指令的跳转地址在指令中，跳转发生时要把返回地址存放到R31寄存器（ra）中。
jr
使用寄存器的跳转指令，跳转地址在寄存器中。
所以根据ra定位core位置时，可以直接在ra位置前找jal便可以找到当前正在执行的函数。
转自http://user.it.uu.se/~justin/Teaching/NewDarkSlides/lecture5.pdf
• As in high level languages , when programming in assembly language you should split up your program into smaller functions, that you can reuse.
• One of the key ideas with functions is that you can call them
from any where and return back to where you called the function from.
• The MIPS processor has two instructions that enable you to call functions, jr and jal.
• Jump and link.
jal label
Copies the address of the next instruction into the register $ra
(register 31) and then jumps to the address label.
• jr $register jumps to the address in $register most common use
jr $ra
.data str: .asciiz "Hello mum!. " .text .globl main #necessary for the assembler main: jal message jal message li $v0,10 syscall #exit the program gracefully message: la $a0,str li $v0,4 syscall #Magic to printhings on the screen. jr $ra
• There are many way of passing values to functions, but there is a convention that most programs on the MIPS follow.
• $a0-$a3 (registers 4 to 7) arguments 1-4 of a function.
• $v0-$v1 (registers 2 and 3) results of a function.
li $a0,10 li $a1,21 li $a3,31 jal silly #Now the result of the function is is $v0. li $v0,10 syscall silly: add $t0,$a0,$a1 sub $v0,$a3,$t0 jr $ra