[Bug analyzer/104860] New: RFE: -Wanalyzer-possible-null-argument and -Wanalyzer-null-argument should respect __attribute__((access, ...))

[Bug analyzer/104860] New: RFE: -Wanalyzer-possible-null-argument and -Wanalyzer-null-argument should respect __attribute__((access, ...))

[dmalcolm at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=104860

            Bug ID: 104860
           Summary: RFE: -Wanalyzer-possible-null-argument and
                    -Wanalyzer-null-argument should respect
                    __attribute__((access, ...))
           Product: gcc
           Version: 12.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: analyzer
          Assignee: dmalcolm at gcc dot gnu.org
          Reporter: dmalcolm at gcc dot gnu.org
  Target Milestone: ---

Similar to PR analyzer/104793, but relating to NULL/possibly NULL pointers (and
affects reads as well as writes):

typedef __SIZE_TYPE__ size_t;

int getrandom (void *__buffer, size_t __length,
               unsigned int __flags)
  __attribute__ ((access (__write_only__, 1, 2)));

#define GRND_RANDOM 0x02

void test (int flag)
{
  char *buf = __builtin_malloc (1024);

  if (getrandom(buf, 16, GRND_RANDOM))
    __builtin_printf("%s\n", buf);

  __builtin_free (buf);
}


The call to malloc could fail, but we don't yet complain about the
possibly-NULL param to getrandom, that's marked with __attribute__ ((access,
...))

[Bug analyzer/104860] RFE: -Wanalyzer-possible-null-argument and -Wanalyzer-null-argument should respect __attribute__((access, ...))

[dmalcolm at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=104860

--- Comment #1 from David Malcolm <dmalcolm at gcc dot gnu.org> ---
Questions posted to GCC list about this: "__attribute__ ((access, ...)) vs
__attribute__ ((nonnull))"
  https://gcc.gnu.org/pipermail/gcc/2022-March/238389.html

[Bug analyzer/104860] RFE: -Wanalyzer-possible-null-argument and -Wanalyzer-null-argument should respect __attribute__((access, ...))

[dmalcolm at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=104860

David Malcolm <dmalcolm at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         Resolution|---                         |INVALID
             Status|UNCONFIRMED                 |RESOLVED

--- Comment #2 from David Malcolm <dmalcolm at gcc dot gnu.org> ---
The consensus in that discussion is that "access" and "nonnull" are separate
concepts, and thus the user needs to provide both if they mean both.

Closing this out as INVALID.