[llvm-commits] [PATCH] Try to figure out <size> argument in llvm.lifetime intrinsics

[Dmitri Gribenko]

On Sun, Nov 11, 2012 at 7:32 PM, Dmitry N. Mikushin <maemarcus at gmail.com> wrote:
> In case of stack reuse there should be another llvm.livetime start/end
> region, isn't it?

I was thinking about something like this:

$ cat a.c
#include <stdio.h>

void store_ptr(int *);
int *load_ptr(void);
int ret_int(void);
void use_ptr(int *);
extern int global;

int main()
{
  {
    int i[8];
    store_ptr(i);
  }
  {
    int j[8];
    j[0] = ret_int();
    use_ptr(j);
    int *k = load_ptr();
    *k = 42;
    use_ptr(j);
    printf("%d\n", global);
  }
}

$ cat b.c
static int *ptr;

void store_ptr(int *p) { ptr = p; }
int *load_ptr(void) { return ptr; }
int ret_int() { return 7; }
int global;
void use_ptr(int *p) { global = *p; }

Currently ASan does not catch an issue in this code, but with lifetime
markers emitted by Clang, it should.

Here 'i' and 'j' would occupy different addresses without stack reuse.
 In this case, 'i' would be still allocated while 'j' is alive, but
'i' would be poisoned.  With stack reuse, 'i' and 'j' will most likely
occupy the same block of memory.

Dmitri

-- 
main(i,j){for(i=2;;i++){for(j=2;j<i;j++){if(!(i%j)){j=0;break;}}if
(j){printf("%d\n",i);}}} /*Dmitri Gribenko <gribozavr at gmail.com>*/