Re: [PATCH] iw_cm: Call provider listen_destroy function before changing state.

[Steve Wise <swise-7bPotxP6k4+P2YhJcF5u+vpXobYPEAuW@public.gmane.org>]

On 02/22/2012 02:02 PM, Steve Wise wrote:
> On 02/22/2012 01:50 PM, Hefty, Sean wrote:
>>> diff --git a/drivers/infiniband/core/iwcm.c b/drivers/infiniband/core/iwcm.c
>>> index 1a696f7..3b43bba 100644
>>> --- a/drivers/infiniband/core/iwcm.c
>>> +++ b/drivers/infiniband/core/iwcm.c
>>> @@ -340,11 +340,11 @@ static void destroy_cm_id(struct iw_cm_id *cm_id)
>>>       spin_lock_irqsave(&cm_id_priv->lock, flags);
>>>       switch (cm_id_priv->state) {
>>>       case IW_CM_STATE_LISTEN:
>>> -        cm_id_priv->state = IW_CM_STATE_DESTROYING;
>>>           spin_unlock_irqrestore(&cm_id_priv->lock, flags);
>> What does this spinlock protect against if all we do is read the state under it?
>>
>
> Probably nothing in this particular path.  It seems to serialize adding/removing references to the qp and manipulating 
> cm_id_priv->qp.  But I don't want to revisit the entire serialization of this module. :)
>
> But the point of the patch is to destroy the listening endpoint in the provider _before_ we stop accepting connect 
> requests.
>

Oh you were pointing out that I could rearrange the spin lock usage here since we no longer set the state in this one case?
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html