On Wed, Sep 07, 2016 at 08:37:23PM +0530, Parav Pandit wrote:

Hi Leon,

quoted

quoted

Signed-off-by: Parav Pandit <redacted>
+static struct rdmacg_resource_pool *
+get_cg_rpool_locked(struct rdma_cgroup *cg, struct rdmacg_device *device)
+{
+     struct rdmacg_resource_pool *rpool;
+
+     rpool = find_cg_rpool_locked(cg, device);
+     if (rpool)
+             return rpool;
+
+     rpool = kzalloc(sizeof(*rpool), GFP_KERNEL);
+     if (!rpool)
+             return ERR_PTR(-ENOMEM);
+
+     rpool->device = device;
+     set_all_resource_max_limit(rpool);
+
+     INIT_LIST_HEAD(&rpool->cg_node);
+     INIT_LIST_HEAD(&rpool->dev_node);
+     list_add_tail(&rpool->cg_node, &cg->rpools);
+     list_add_tail(&rpool->dev_node, &device->rpools);
+     return rpool;
+}

<...>

quoted

+     for (p = cg; p; p = parent_rdmacg(p)) {
+             rpool = get_cg_rpool_locked(p, device);
+             if (IS_ERR_OR_NULL(rpool)) {

get_cg_rpool_locked always returns !NULL (error, or pointer)

Can this change go as incremental change after this patch, since this
patch is close to completion?
Or I need to revise v13?

Sure, it is cleanup. It is not worth of respinning.

quoted

quoted

+                     ret = PTR_ERR(rpool);
+                     goto err;

I didn't review the whole series yet.

Did you get a chance to review the series?

We need to decide on fundamental question before reviewing it, which is
"how to fit rdmacg to new ABI model".

Thanks