ARMv7 ldr/str指令详解

因为ARM的算术运算不支持直接操作内存地址，所以要把内存里的数据先加载进寄存器。ldr指令就是干这事的，称为间接取址模式。

一共有3*3九种模式，先是直接偏移，先偏移，后偏移三大类，指的是如何对源操作数操作，是直接使用，还是在加载前对源操作数操作（比如地址加个数值），还是在加载后对操作数操作

每个大类里分三个小类，分别指源操作数是立即数，寄存器，还是标量寄存器（比如对寄存器里的数向左偏移两位，即乘4）

汇编指令和对应的C代码如下

 1 Immediate offset:
 2
 3   LDR    R0, [R1, #4]
 4   r0 = r1[1];
 5
 6     Remember, again, the offset is in bytes, so #4 would point to the second word (long int) in our array, thus in C the array indice [1] would provide the correct data.
 7
 8 Register offset:
 9
10   LDR    R0, [R1, R2]
11   r0 = r1[r2];
12
13 Scaled register offset:
14
15   LDR    R0, [R1, R2, LSL #4]
16   r0 = r1[(r2 << 4)];
17
18 Immediate pre-indexed:
19
20   LDR    R0, [R1, #4]!
21   r1 += 4; r0 = *r1;
22
23 Register pre-indexed:
24
25   LDR    R0, [R1, R2]!
26   r1 += r2; r0 = *r1;
27
28 Scaled register pre-indexed:
29
30   LDR    R0, [R1, R2, LSL #2]!
31   r1 += (r2 << #2); r0 = *r1;
32
33 Immediate post-indexed:
34
35   LDR    R0, [R1], #4
36   r0 = *r1; r1 += 4;
37
38 Register post-indexed:
39
40   LDR    R0, [R1], R2
41   r0 = *r1; r1 += r2;
42
43 Scaled register post-indexed:
44
45   LDR    R0, [R1, R2, LSL #2]!
46   r0 = *r1; r1 += (r2 << #2);

str指令是把寄存器里的数存到内存里，目标地址的计算方式与操作与ldr指令中源地址的计算方法与操作一样

Immediate offset:
  LDR    R0, [R1, #4]  r0 = r1[1];
    Remember, again, the offset is in bytes, so #4 would point to the second word (long int) in our array, thus in C the array indice [1] would provide the correct data. 
Register offset:
  LDR    R0, [R1, R2]  r0 = r1[r2];
Scaled register offset:
  LDR    R0, [R1, R2, LSL #4]  r0 = r1[(r2 << 4)];
Immediate pre-indexed:
  LDR    R0, [R1, #4]!  r1 += 4; r0 = *r1;
Register pre-indexed:
  LDR    R0, [R1, R2]!  r1 += r2; r0 = *r1;
Scaled register pre-indexed:
  LDR    R0, [R1, R2, LSL #2]!  r1 += (r2 << #2); r0 = *r1;
Immediate post-indexed:
  LDR    R0, [R1], #4]  r0 = *r1; r1 += 4;
Register post-indexed:
  LDR    R0, [R1], R2  r0 = *r1; r1 += r2;
Scaled register post-indexed:
  LDR    R0, [R1, R2, LSL #2]!  r0 = *r1; r1 += (r2 << #2);