Re: [PATCH] Fix i386 memmove issue [BZ #22644]

[Andreas Schwab <schwab@suse.de>]

On Feb 19 2018, Andrew Senkevich <andrew.n.senkevich@gmail.com> wrote:

> diff --git a/string/test-memmove.c b/string/test-memmove.c
> index edc7a4c..8dc152b
> --- a/string/test-memmove.c
> +++ b/string/test-memmove.c
> @@ -245,6 +245,49 @@ do_random_tests (void)
>      }
>  }
>
> +#if __SIZEOF_POINTER__ == 4
> +static void
> +do_test2 (void)
> +{
> +  uint32_t i;
> +  uint32_t num = 0x20000000;
> +  uint32_t * large_buf = mmap (0, sizeof(uint32_t) * num, PROT_READ |
> PROT_WRITE,
> +        MAP_PRIVATE | MAP_ANON, -1, 0);
> +  if (large_buf == MAP_FAILED)
> +    error (EXIT_FAILURE, errno, "large mmap failed");
> +
> +  if (!((uint32_t)(large_buf) < (0x80000000 - 128) && (0x80000000 +
> 128) < (uint32_t)(&large_buf[num])))
> +    {
> +      error (0, 0,"allocated large memory doesn't cross 0x80000000 boundary");
> +      ret = 1;
> +      return;

Please properly fold long lines, and remove the redundant parens.  Also,
there is no guarantee that the address range is unallocated.

Andreas.

-- 
Andreas Schwab, SUSE Labs, schwab@suse.de
GPG Key fingerprint = 0196 BAD8 1CE9 1970 F4BE  1748 E4D4 88E3 0EEA B9D7
"And now for something completely different."