[Qemu-devel] What's the meaning and the functionality of the assembly instructions in the beginning of cpu_exec?

[曹莹]

I¡¯m working with Qemu source study, but now I have some difficulty in these assembly instructions at the beginning of cpu_exec function. My question as follows:

 

After entering function cpu_exec, there are some assembly codes:

 

register host_reg_t reg_AREG0 asm("ebp");

volatile host_reg_t saved_AREG0;

      

register host_reg_t reg_AREG1 asm("ebx");

volatile host_reg_t saved_AREG1;

      

register host_reg_t reg_AREG2 asm("esi");

volatile host_reg_t saved_AREG2;

      

register host_reg_t reg_AREG3 asm("edi");

volatile host_reg_t saved_AREG3;

 

and there is also:

__asm__ __volatile__ ("" : "=r" (reg_AREG0));

saved_AREG0 = reg_AREG0;

             

__asm__ __volatile__ ("" : "=r" (reg_AREG1));

saved_AREG1 = reg_AREG1;

             

__asm__ __volatile__ ("" : "=r" (reg_AREG2));

saved_AREG2 = reg_AREG2;

             

__asm__ __volatile__ ("" : "=r" (reg_AREG3));

saved_AREG3 = reg_AREG3;

 

Where are variables: reg_AREG-, saved_AREG- defined and how they are used in the later codes? I really understand these assembly codes¡¯ functionality. Do they use for Qemu virtual register mapping or translating optimization? Could anyone give me some advise, thank. Thank you very much for your help.

---

ÍøÒ×ΪÖÐСÆóÒµÃâ·ÑÌṩÆóÒµÓÊÏ䣨×ÔÖ÷ÓòÃû£©