Re: Make max_align_t respect _Float128 [version 2]

[Florian Weimer]

On 09/08/2016 02:26 PM, Bernd Schmidt wrote:
On 09/08/2016 01:21 AM, Paul Eggert wrote:
Sure, attached. On Fedora 24 x86-64 (GCC 6.1.1 20160621, valgrind
3.11.0), when I compile with "gcc -O2 flexouch.c" and run with "valgrind
./a.out", valgrind complains "Invalid read of size 2". This is because
GCC compiles "p->d[0] == 2 && p->d[1] == 3" into "cmpw $770, 8(%rax);
sete %al", which loads the uninitialized byte p->d[1] simultaneously
with the initialized byte p->d[0].

Interesting. That optimization doesn't really depend on d being a
flexible array, so you can also reproduce a (different) valgrind warning
with the following:

#include <stddef.h>
#include <stdlib.h>

struct s { int x; char d[2]; };

__attribute__((noinline,noclone)) void foo (struct s *p)
{
  p->d[0] = 1;
}

int
main (void)
{
  struct s *p = malloc (sizeof *p);
  foo (p);
  return p->d[0] == 2 && p->d[1] == 3;
}

Very interesting.  So the ASan failure reported for gnulib fts and this 
valgrind issue have separate causes (ASan does not care about undefined 
memory).

Thanks,
Florian