[Bug tree-optimization/100780] New: __builtin___sprintf_chk not optimized when it could be

[msebor at gcc dot gnu.org via Gcc-bugs]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=100780

            Bug ID: 100780
           Summary: __builtin___sprintf_chk not optimized when it could be
           Product: gcc
           Version: 11.1.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: tree-optimization
          Assignee: unassigned at gcc dot gnu.org
          Reporter: msebor at gcc dot gnu.org
  Target Milestone: ---

Calls to __builtin___sprintf_chk that can be proven not to overflow the
destination can be folded to calls to plain sprintf (or even memcpy).  Only the
simplest are, but the more interesting ones aren't, even though the sprintf
pass has all the smarts to do it.

$ cat a.c && gcc -O2 -S -Wall -fdump-tree-optimized=/dev/stdout a.c 
extern char a[32];

void f (void)
{
  __builtin___sprintf_chk (a, 0, sizeof a, "%s", "1234");  // optimized
}

void g (void)
{
  __builtin___sprintf_chk (a, 0, sizeof a, "%i", 1234);    // not optimized
}

;; Function f (f, funcdef_no=0, decl_uid=1944, cgraph_uid=1, symbol_order=0)

void f ()
{
  <bb 2> [local count: 1073741824]:
  __builtin_memcpy (&a, "1234", 5); [tail call]
  return;

}



;; Function g (g, funcdef_no=1, decl_uid=1947, cgraph_uid=2, symbol_order=1)

void g ()
{
  <bb 2> [local count: 1073741824]:
  __builtin___sprintf_chk (&a, 0, 32, "%i", 1234); [tail call]
  return;

}