[Bug middle-end/70871] New: questionable optimisation in fold-const.c

[jan.sm...@alcatel-lucent.com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70871

            Bug ID: 70871
           Summary: questionable optimisation in fold-const.c
           Product: gcc
           Version: 5.3.1
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: middle-end
          Assignee: unassigned at gcc dot gnu.org
          Reporter: jan.sm...@alcatel-lucent.com
  Target Milestone: ---

typedef struct X
{   
    void *a; 
    long int b;
    int c;
    int d;   
    unsigned int e[3];
    unsigned short f;
    unsigned char g;
    unsigned char h;
    unsigned char i;
    unsigned char j;
    unsigned char k;
    unsigned char l;
    unsigned char m;
    unsigned char n;
};



int somefunc(void)
{
    struct X *pX = (0L);

    extern void call(void*);
    call(&pX);

    if(!pX->n  && !pX->l)
        return 0;

    return 1;
}

for -O1 and higher

 if(!pX->n  && !pX->l)

results in (tree-original)

 if ((BIT_FIELD_REF <*pX, 64, 320> & 280379743272960) == 0)


which eventually results in 
 ASAN_CHECK (6, _1, 8, 8) 
when using -fsantize=address

Wheras ASAN really should be reporting a 2x 1 byte load and not 1x 8.

Thanks