[libvirt] [PATCH] introduce VIR_CLOSE to be used rather than close()

[Eric Blake]

On 10/15/2010 11:32 AM, Eric Blake wrote:
>> +int virClose(int *fdptr);
>
> Needs an ATTRIBUTE_NONNULL(1). I'm also debating whether it needs
> ATTRIBUTE_RETURN_CHECK; normally, close() fails in very few situations,
> and in cleanup paths, you tend to already have another error more
> important so you can ignore the secondary close() failures. So I'm
> probably 80-20 against adding ATTRIBUTE_RETURN_CHECK.
>
>> +
>> +# define VIR_CLOSE(FD) \
>> + virClose(&(FD))

Thinking about it a bit more, what if we provide two variants?

int virClose(int *fdptr, bool preserve_errno)
{
     int saved_errno;
     int rc;

     if (*fdptr >= 0) {
         if (preserve_errno)
             saved_errno = errno;
         rc = close(*fdptr);
         *fdptr = -1;
         if (preserve_errno)
             errno = saved_errno;
     } else
         rc = 0;

     return rc;
}

int virClose(int *fdptr) ATTRIBUTE_RETURN_CHECK;

/* For use on normal paths; caller must check return value, and failure 
sets errno per close(). */
#define VIR_CLOSE(FD) virClose(&(FD),false);

/* For use on cleanup paths; errno is unaffected by close, and no return 
value to worry about. */
#define VIR_FORCE_CLOSE(FD) ignore_value(virClose(&(FD),true))

-- 
Eric Blake   eblake at redhat.com    +1-801-349-2682
Libvirt virtualization library http://libvirt.org