Re: [PATCH v2] Make xmalloc and xrealloc thread-safe

[Nicolas Pitre <nico@fluxnic.net>]

On Wed, 7 Apr 2010, Fredrik Kuivinen wrote:

> As I mentioned in another mail in this thread, our mutex
> implementation on WIN32 already is recursive. It is implemented on top
> of the CRITICAL_SECTION type, which is recursive. See
> http://msdn.microsoft.com/en-us/library/ms682530%28VS.85%29.aspx

Ahhhh.  Goodie.

> We only need something like the following (on top of Nico's previous
> patch). Warning: It hasn't even been compile tested on WIN32.
> 
[...]
> diff --git a/thread-utils.c b/thread-utils.c
> index 4f9c829..3c8d817 100644
> --- a/thread-utils.c
> +++ b/thread-utils.c
> @@ -1,4 +1,5 @@
>  #include "cache.h"
> +#include <pthread.h>

This will fail compilation on Windows surely?

>  #if defined(hpux) || defined(__hpux) || defined(_hpux)
>  #  include <sys/pstat.h>
> @@ -43,3 +44,24 @@ int online_cpus(void)
>  
>  	return 1;
>  }
> +
> +int init_recursive_mutex(pthread_mutex_t *m)
> +{
> +#ifdef _WIN32
> +	/* The mutexes in the WIN32 pthreads emulation layer are
> +	 * recursive, so we don't have to do anything extra here. */
> +	return pthread_mutex_init(m, NULL);
> +#else
> +	pthread_mutexattr_t a;
> +	int ret;
> +	if (pthread_mutexattr_init(&a))
> +		die("pthread_mutexattr_init failed: %s", strerror(errno));
> +
> +	if (pthread_mutexattr_settype(&a, PTHREAD_MUTEX_RECURSIVE))
> +		die("pthread_mutexattr_settype failed: %s", strerror(errno));
> +
> +	ret = pthread_mutex_init(m, &a);
> +	pthread_mutexattr_destroy(&a);
> +	return ret;

Are you sure the pthread_mutexattr_t object can be destroyed even if the 
mutex is still in use?  Is the attribute object "attached" to the mutex 
or merely used as a template?


Nicolas