windelbouwman / ppci

A compiler for ARM, X86, MSP430, xtensa and more implemented in pure Python
https://ppci.readthedocs.io/en/latest/
BSD 2-Clause "Simplified" License
337 stars 36 forks source link

Implementation of sbrk #113

Closed darleybarreto closed 4 years ago

darleybarreto commented 4 years ago

In order to implement dynamic allocated str (#91) and more, today I worked on bringing malloc support. I am almost there (I altered the necessary python code for that), the only thing remaining is sbrk function, following this implementation I would need to link the information about the end of the free memory on the host. What would be the best way of doing that?

windelbouwman commented 4 years ago

I'm not sure how to do that, but you could use a linker script to do so? From the memory map you could define a symbol and use that in an assembly level implementation of sbrk.

I'm not sure what you want to achieve, do you want to implement your own sbrk function?

darleybarreto commented 4 years ago

I want to implement malloc, so we could allocate arbitrary objects inside ppci (for unix based systems with brk similar syscalls). The "easy" way would be to implement brk syscall and sbrk wrapper, then malloc using sbrk.

windelbouwman commented 4 years ago

If you want to implement syscalls, I would checkout this sample: https://github.com/windelbouwman/ppci/blob/master/examples/linux64/hello-make/hello.c

darleybarreto commented 4 years ago

Oh yeah, I have seen this one, and I altered the implementation of do_inline_asm, so one can get the return of a syscall in a variable like this:

long syscall(long nr, long a, long b, long c)
{   
    long ret;

    asm(
        "mov rax, %0 \n"
        "mov rdi, %1 \n"
        "mov rsi, %2 \n"
        "mov rdx, %3 \n"
        "syscall \n"
        : "=r" (ret)
        : "r" (nr), "r" (a), "r" (b), "r" (c)
        : "rax", "rdi", "rsi", "rdx"
    );

    return ret;
}

In ret we would have the break point given by the system, so sbrk would work with this value to give the start memory address when called inside a malloc implementation.

darleybarreto commented 4 years ago

I worked out the syscalls and made a port of a simple malloc. So now we can do something like this

#include <stdio.h>
#include <stdlib.h>

int main(){
   int *int_vec = (int*)malloc(10*sizeof(int));
   for (int i =0; i < 10; i++){
      int_vec[i] = i*i + 2;
   }

   for (int i =0; i < 10; i++){
      printf("%d",int_vec[i]);
   }

   free(int_vec);

   exit(0);
}
windelbouwman commented 4 years ago

Cool! Nicely done!!

darleybarreto commented 4 years ago

I also took the liberty to change some file structures inside librt/libc, I guess you will check if it is ok when reviewing the PR. I've never submitted a PR before, should I do something w.r.t CI/CD?

windelbouwman commented 4 years ago

Sure, feel free to create a PR, lets see what happens :). No need to do any CI/CD. If you want, you can run tests locally with tox / pytest as described in the docs, but I guess this is done as well with CI/CD on github.

darleybarreto commented 4 years ago

Closed by #114