[PATCH 1/3] posix: Prevent pthread_setspecific from returning EAGAIN.

[Gedare Bloom]

Looks good.

On Fri, Mar 21, 2014 at 5:26 AM, Christian Mauderer
<christian.mauderer at embedded-brains.de> wrote:
> From: Christian Mauderer <Christian.Mauderer at embedded-brains.de>
>
> The man-page for pthread_setspecific does not define the EAGAIN return value.
> Further without this patch it was not possible to set keys that have been
> already set a new value.
> ---
>  cpukit/posix/src/keysetspecific.c | 59 +++++++++++++++++++++++----------------
>  1 file changed, 35 insertions(+), 24 deletions(-)
>
> diff --git a/cpukit/posix/src/keysetspecific.c b/cpukit/posix/src/keysetspecific.c
> index c6e2d30..9afd35c 100644
> --- a/cpukit/posix/src/keysetspecific.c
> +++ b/cpukit/posix/src/keysetspecific.c
> @@ -37,37 +37,48 @@ int pthread_setspecific(
>    POSIX_Keys_Control          *the_key;
>    Objects_Locations            location;
>    POSIX_Keys_Key_value_pair   *value_pair_ptr;
> +  RBTree_Node                 *p;
> +  POSIX_Keys_Key_value_pair    search_node;
>
>    the_key = _POSIX_Keys_Get( key, &location );
>    switch ( location ) {
>
>      case OBJECTS_LOCAL:
> -      value_pair_ptr = _POSIX_Keys_Key_value_pair_allocate();
> -
> -      if ( !value_pair_ptr ) {
> -        _Objects_Put( &the_key->Object );
> -
> -        return ENOMEM;
> -      }
> -
> -      value_pair_ptr->key = key;
> -      value_pair_ptr->thread_id = _Thread_Executing->Object.id;
> -      value_pair_ptr->value = value;
> -      if ( _RBTree_Insert( &_POSIX_Keys_Key_value_lookup_tree,
> -                           &(value_pair_ptr->Key_value_lookup_node) ) ) {
> -        _Freechain_Put( (Freechain_Control *)&_POSIX_Keys_Keypool,
> -                        (void *) value_pair_ptr );
> -        _Objects_Put( &the_key->Object );
> -
> -        return EAGAIN;
> +      search_node.key = key;
> +      search_node.thread_id = _Thread_Executing->Object.id;
> +      p = _RBTree_Find( &_POSIX_Keys_Key_value_lookup_tree,
> +                                    &search_node.Key_value_lookup_node );
> +
> +      if ( p ) {
> +        value_pair_ptr = _RBTree_Container_of( p,
> +                                          POSIX_Keys_Key_value_pair,
> +                                          Key_value_lookup_node );
> +
> +        value_pair_ptr->value = value;
> +      } else {
> +        value_pair_ptr = _POSIX_Keys_Key_value_pair_allocate();
> +
> +        if ( !value_pair_ptr ) {
> +          _Objects_Put( &the_key->Object );
> +
> +          return ENOMEM;
> +        }
> +
> +        value_pair_ptr->key = key;
> +        value_pair_ptr->thread_id = _Thread_Executing->Object.id;
> +        value_pair_ptr->value = value;
> +        /* The insert can only go wrong if the same node is already in a unique
> +         * tree. This has been already checked with the _RBTree_Find() */
> +        (void) _RBTree_Insert( &_POSIX_Keys_Key_value_lookup_tree,
> +                             &(value_pair_ptr->Key_value_lookup_node) );
> +
> +        /** append rb_node to the thread API extension's chain */
> +        _Chain_Append_unprotected(
> +          &_Thread_Executing->Key_Chain,
> +          &value_pair_ptr->Key_values_per_thread_node
> +        );
>        }
>
> -      /** append rb_node to the thread API extension's chain */
> -      _Chain_Append_unprotected(
> -        &_Thread_Executing->Key_Chain,
> -        &value_pair_ptr->Key_values_per_thread_node
> -      );
> -
>        _Objects_Put( &the_key->Object );
>
>        return 0;
> --
> 1.8.4.5
>
> _______________________________________________
> rtems-devel mailing list
> rtems-devel at rtems.org
> http://www.rtems.org/mailman/listinfo/rtems-devel