Fix tst-obprintf - and mcheck in general

[Andreas Jaeger]

On 04/26/2012 03:15 PM, Andreas Jaeger wrote:
> On Thursday, April 26, 2012 15:01:16 Richard Guenther wrote:
>> On Thu, 26 Apr 2012, Andreas Schwab wrote:
>>> Andreas Jaeger<aj@suse.com>  writes:
>>>> One option is to compile with -fno-builtin-malloc, the alternative
>>>> would be to use some otherways to initialize the malloc
>>>> implementation.
>>>
>>> Does it work to use the equivalent of math_opt_barrier from
>>> <math_private.h>?
>>
>> Yes, that it should work, too.
>
> Yes, the following works as well.
>
> What's the best place to put this barrier in? Put it into include/malloc.h?

Which is the better way? A barrier like below or setting 
-fno-builtin-malloc?

Andreas

> Andreas
>
> diff --git a/malloc/mcheck.c b/malloc/mcheck.c
> index 9213740..7491c49 100644
> --- a/malloc/mcheck.c
> +++ b/malloc/mcheck.c
> @@ -370,6 +370,12 @@ mabort (enum mcheck_status status)
>   #endif
>   }
>
> +#ifndef malloc_opt_barrier
> +#define malloc_opt_barrier(x) \
> +({ __typeof (x) __x = x; __asm ("" : "+m" (__x)); __x; })
> +
> +#endif
> +
>   int
>   mcheck (func)
>        void (*func) (enum mcheck_status);
> @@ -381,6 +387,7 @@ mcheck (func)
>       {
>         /* We call malloc() once here to ensure it is initialized.  */
>         void *p = malloc (0);
> +      p = malloc_opt_barrier (p);
>         free (p);
>
>         old_free_hook = __free_hook;
>


-- 
  Andreas Jaeger aj@{suse.com,opensuse.org} Twitter/Identica: jaegerandi
   SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg, Germany
    GF: Jeff Hawn,Jennifer Guild,Felix Imendörffer,HRB16746 (AG Nürnberg)
     GPG fingerprint = 93A3 365E CE47 B889 DF7F  FED1 389A 563C C272 A126