[Bug sanitizer/108637] New: ASAN at -O2 misses a stack-use-after-scope

[Bug sanitizer/108637] New: ASAN at -O2 misses a stack-use-after-scope

[shaohua.li at inf dot ethz.ch]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108637

            Bug ID: 108637
           Summary: ASAN at -O2 misses a stack-use-after-scope
           Product: gcc
           Version: 13.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: sanitizer
          Assignee: unassigned at gcc dot gnu.org
          Reporter: shaohua.li at inf dot ethz.ch
                CC: dodji at gcc dot gnu.org, dvyukov at gcc dot gnu.org,
                    jakub at gcc dot gnu.org, kcc at gcc dot gnu.org, marxin at gcc dot gnu.org
  Target Milestone: ---

For the following code, ASAN at -O2 cannot detect the stack-use-after-scope.
This happens since GCC-11. GCC-10 and below work fine.

I understand that there is a type-incompatible warning. But I believe it should
not affect the ASAN anyway.

Compiler explorer: https://godbolt.org/z/WcbadP961

% cat a.c
char a, b, c;
short d;
int main() {
  {
    long e = a;
    for (; b <= 0; b++) {
      char *f;
      d = 2;
      for (; d; d--) {
        int g = 111;
        f = &g;
        if (e)
          c = 0;
      }
      *f += 1;
      a = *f;
    }
  }
}
%
% gcc-tk -fsanitize=address -O2 a.c && ./a.out
%
% gcc-tk -fsanitize=address -O1 a.c && ./a.out
=================================================================
==1==ERROR: AddressSanitizer: stack-use-after-scope on address 0x7fb285e00020
at pc 0x000000401272 bp 0x7ffda17739c0 sp 0x7ffda17739b8
READ of size 1 at 0x7fb285e00020 thread T0
    #0 0x401271 in main /a.c:15
    #1 0x7fb28883d082 in __libc_start_main
(/lib/x86_64-linux-gnu/libc.so.6+0x24082) (BuildId:
1878e6b475720c7c51969e69ab2d276fae6d1dee)
    #2 0x4010bd in _start (/output.s+0x4010bd) (BuildId:
cb6e8e9c15b7d8e5515d7e6b8591d2b77b5f3e21)

Address 0x7fb285e00020 is located in stack of thread T0 at offset 32 in frame
    #0 0x401185 in main /a.c:3

  This frame has 1 object(s):
    [32, 36) 'g' (line 10) <== Memory access at offset 32 is inside this
variable
HINT: this may be a false positive if your program uses some custom stack
unwind mechanism, swapcontext or vfork
      (longjmp and C++ exceptions *are* supported)
SUMMARY: AddressSanitizer: stack-use-after-scope /a.c:15 in main
Shadow bytes around the buggy address:
...
%

[Bug sanitizer/108637] ASAN at -O2 misses a stack-use-after-scope

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108637

Andrew Pinski <pinskia at gcc dot gnu.org> changed:

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

--- Comment #1 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
PRE removes the load/stores from/to *f .
Basically the compiler is able to remove the use-after-scope usage with -O2 and
above.

[Bug sanitizer/108637] ASAN at -O2 misses a stack-use-after-scope

[shaohua.li at inf dot ethz.ch]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108637

--- Comment #2 from Li Shaohua <shaohua.li at inf dot ethz.ch> ---
(In reply to Andrew Pinski from comment #1)
> PRE removes the load/stores from/to *f .
> Basically the compiler is able to remove the use-after-scope usage with -O2
> and above.

Well, this makes sense to me while in this case only -O2 not other opt levels
did not report this use-after-scope usage.

[Bug sanitizer/108637] ASAN at -O2 misses a stack-use-after-scope

[shaohua.li at inf dot ethz.ch]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108637

Li Shaohua <shaohua.li at inf dot ethz.ch> changed:

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

--- Comment #3 from Li Shaohua <shaohua.li at inf dot ethz.ch> ---
(In reply to Li Shaohua from comment #2)
> (In reply to Andrew Pinski from comment #1)
> > PRE removes the load/stores from/to *f .
> > Basically the compiler is able to remove the use-after-scope usage with -O2
> > and above.
> 
> Well, this makes sense to me while in this case only -O2 not other opt
> levels did not report this use-after-scope usage.

I looked into this again and found that for GCC-12, it was indeed the case. But
for the trunk, I indeed observed that PRE did not optimize away load/store.

Compiler explorer: https://godbolt.org/z/sb658P1cM

[Bug sanitizer/108637] ASAN at -O2 misses a stack-use-after-scope

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108637

Martin Liška <marxin at gcc dot gnu.org> changed:

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

--- Comment #4 from Martin Liška <marxin at gcc dot gnu.org> ---
It's reported since r13-3856-g837be6c7cfb49e16.