[openib-general] Re: [PATCH 1/2] iWARP Connection Manager.

Sean Hefty mshefty at ichips.intel.com
Thu Jun 1 14:09:12 PDT 2006


Steve Wise wrote:
>>>+int iw_cm_disconnect(struct iw_cm_id *cm_id, int abrupt)
>>>+{
>>>+	struct iwcm_id_private *cm_id_priv;
>>>+	unsigned long flags;
>>>+	int ret = 0;
>>>+
>>>+	cm_id_priv = container_of(cm_id, struct iwcm_id_private, id);
>>>+	/* Wait if we're currently in a connect or accept downcall */
>>>+	wait_event(cm_id_priv->connect_wait, 
>>>+		   !test_bit(IWCM_F_CONNECT_WAIT, &cm_id_priv->flags));
>>
>>Am I understanding this check correctly?  You're checking to see if the user has 
>>called iw_cm_disconnect() at the same time that they called iw_cm_connect() or 
>>iw_cm_accept().  Are connect / accept blocking, or are you just waiting for an 
>>event?
> 
> 
> The CM must wait for the low level provider to finish a connect() or
> accept() operation before telling the low level provider to disconnect
> via modifying the iwarp QP.  Regardless of whether they block, this
> disconnect can happen concurrently with the connect/accept so we need to
> hold the disconnect until the connect/accept completes.
> 
> 
>>>+EXPORT_SYMBOL(iw_cm_disconnect);
>>>+static void destroy_cm_id(struct iw_cm_id *cm_id)
>>>+{
>>>+	struct iwcm_id_private *cm_id_priv;
>>>+	unsigned long flags;
>>>+	int ret;
>>>+
>>>+	cm_id_priv = container_of(cm_id, struct iwcm_id_private, id);
>>>+	/* Wait if we're currently in a connect or accept downcall. A
>>>+	 * listening endpoint should never block here. */
>>>+	wait_event(cm_id_priv->connect_wait, 
>>>+		   !test_bit(IWCM_F_CONNECT_WAIT, &cm_id_priv->flags));
>>
>>Same question/comment as above.
>>
> 
> 
> Same answer.  

There's a difference between trying to handle the user calling 
disconnect/destroy at the same time a call to accept/connect is active, versus 
the user calling disconnect/destroy after accept/connect have returned.  In the 
latter case, I think you're fine.  In the first case, this is allowing a user to 
call destroy at the same time that they're calling accept/connect. 
Additionally, there's no guarantee that the F_CONNECT_WAIT flag has been set by 
accept/connect by the time disconnect/destroy tests it.

- Sean



More information about the general mailing list