https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101494
Martin Sebor <msebor at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Last reconfirmed| |2021-07-21
Status|UNCONFIRMED |NEW
CC| |msebor at gcc dot gnu.org
Ever confirmed|0 |1
--- Comment #2 from Martin Sebor <msebor at gcc dot gnu.org> ---
Confirmed. I think this exposes two underlying bugs: one that the
initialization isn't detected and another that the second argument to attribute
access isn't respected. A slightly enhanced test case:
$ cat b.c && gcc -O2 -S -Wall b.c
__attribute__ ((access (read_only, 1, 2))) void f (const void*, int);
void g (void)
{
char *p = __builtin_alloca (1);
*p = 0;
f (p, 0); // bogus -Wmaybe-uninitialized
}
void h (void)
{
char *p = __builtin_malloc (1);
f (p, 0); // bogus -Wmaybe-uninitialized
}
b.c: In function ‘g’:
b.c:7:3: warning: ‘p’ is used uninitialized [-Wuninitialized]
7 | f (p, 0); // bogus -Wmaybe-uninitialized
| ^~~~~~~~
b.c:1:49: note: in a call to ‘f’ declared with attribute ‘access (read_only, 1,
2)’ here
1 | __attribute__ ((access (read_only, 1, 2))) void f (const void*, int);
| ^
b.c: In function ‘h’:
b.c:13:3: warning: ‘p’ is used uninitialized [-Wuninitialized]
13 | f (p, 0); // bogus -Wmaybe-uninitialized
| ^~~~~~~~
b.c:1:49: note: in a call to ‘f’ declared with attribute ‘access (read_only, 1,
2)’ here
1 | __attribute__ ((access (read_only, 1, 2))) void f (const void*, int);
| ^
