RISC-V数据模型，-mabi=ilp32, ilp32f, ilp32d, lp64, lp64f, lp64d

谁践踏了优雅 2022-05-31 11:29 209阅读 0赞

RISC-V GCC通过`-mabi`选项指定数据模型和浮点参数传递规则。有效的选项值包括ilp32、ilp32f、ilp32d、lp64、lp64f 和 lp64d。前半部分指定数据模型，后半部分指定浮点参数传递规则。

mingdu.zheng **at** gmail *dot* com  
[http://blog.csdn.net/zoomdy/article/details/79353313][http_blog.csdn.net_zoomdy_article_details_79353313]

`i`指`int`，`l`指`long`，`p`指`pointer`即指针，`32/64`指前面给出的类型是32/64位的；`f`指`float`，指float型浮点数参数通过浮点数寄存器传递；`d`指`double`，指float型和double型浮点数参数通过浮点数寄存器传递。

数据模型：

<table> 
 <thead> 
  <tr> 
   <th>x</th> 
   <th>int字长</th> 
   <th>long字长</th> 
   <th>指针字长</th> 
  </tr> 
 </thead> 
 <tbody> 
  <tr> 
   <td>ilp32/ilp32f/ilp32d</td> 
   <td>32bits</td> 
   <td>32bits</td> 
   <td>32bits</td> 
  </tr> 
  <tr> 
   <td>lp64/lp64f/lp64d</td> 
   <td>32bits</td> 
   <td>64bits</td> 
   <td>64bits</td> 
  </tr> 
 </tbody> 
</table>

浮点参数传递规则：

<table> 
 <thead> 
  <tr> 
   <th>x</th> 
   <th>需要浮点扩展指令？</th> 
   <th>float参数</th> 
   <th>double参数</th> 
  </tr> 
 </thead> 
 <tbody> 
  <tr> 
   <td>ilp32/lp64</td> 
   <td>不需要</td> 
   <td>通过整数寄存器（a0-a1）传递</td> 
   <td>通过整数寄存器（a0-a3）传递</td> 
  </tr> 
  <tr> 
   <td>ilp32f/lp64f</td> 
   <td>需要F扩展</td> 
   <td>通过浮点寄存器（fa0-fa1）传递</td> 
   <td>通过整数寄存器（a0-a3）传递</td> 
  </tr> 
  <tr> 
   <td>ilp32d/lp64d</td> 
   <td>需要F扩展和D扩展</td> 
   <td>通过浮点寄存器（fa0-fa1）传递</td> 
   <td>通过浮点寄存器（fa0-fa1）传递</td> 
  </tr> 
 </tbody> 
</table>

浮点参数传递规则只跟`-mabi`选项有关，和`-march`选项没有直接关系，但是部分`-mabi`选项需要浮点寄存器，浮点寄存器是通过浮点扩展指令引入的，这就需要在`-march`选项中指定浮点扩展。

RISC-V GCC 使用手册对`-mabi`选项的说明：

> \-mabi=ABI-string  
> .  
> Specify integer and floating-point calling convention. ABI-string contains two parts: the size of integer types and the registers used for floating-point types. For example -march=rv64ifd -mabi=lp64d means that long and pointers are 64-bit (implicitly defining int to be 32-bit), and that floating-point values up to 64 bits wide are passed in F registers. Contrast this with -march=rv64ifd -mabi=lp64f, which still allows the compiler to generate code that uses the F and D extensions but only allows floating-point values up to 32 bits long to be passed in registers; or -march=rv64ifd -mabi=lp64, in which no floating-point arguments will be passed in registers.  
> .  
> The default for this argument is system dependent, users who want a specific calling convention should specify one explicitly. The valid calling conventions are: ilp32, ilp32f, ilp32d, lp64, lp64f, and lp64d. Some calling conventions are impossible to implement on some ISAs: for example, -march=rv32if -mabi=ilp32d is invalid because the ABI requires 64-bit values be passed in F registers, but F registers are only 32 bits wide.

[http_blog.csdn.net_zoomdy_article_details_79353313]: http://blog.csdn.net/zoomdy/article/details/79353313