Kamezawa Hiroyuki [off-list ref] writes:

(2012/06/09 17:59), Aneesh Kumar K.V wrote:

quoted

From: "Aneesh Kumar K.V"<redacted>

This patch add support for cgroup removal. If we don't have parent
cgroup, the charges are moved to root cgroup.

Signed-off-by: Aneesh Kumar K.V<redacted>

I'm sorry if already pointed out....

quoted

---
  mm/hugetlb_cgroup.c |   81 +++++++++++++++++++++++++++++++++++++++++++++++++--
  1 file changed, 79 insertions(+), 2 deletions(-)

diff --git a/mm/hugetlb_cgroup.c b/mm/hugetlb_cgroup.c
index 48efd5a..9458fe3 100644
--- a/mm/hugetlb_cgroup.c
+++ b/mm/hugetlb_cgroup.c

@@ -99,10 +99,87 @@ static void hugetlb_cgroup_destroy(struct cgroup *cgroup)
  	kfree(h_cgroup);
  }

+
+static int hugetlb_cgroup_move_parent(int idx, struct cgroup *cgroup,
+				      struct page *page)
+{
+	int csize;
+	struct res_counter *counter;
+	struct res_counter *fail_res;
+	struct hugetlb_cgroup *page_hcg;
+	struct hugetlb_cgroup *h_cg   = hugetlb_cgroup_from_cgroup(cgroup);
+	struct hugetlb_cgroup *parent = parent_hugetlb_cgroup(cgroup);
+
+	if (!get_page_unless_zero(page))
+		goto out;

It seems this doesn't necessary...this is under hugetlb_lock().

already updated.

quoted

+
+	page_hcg = hugetlb_cgroup_from_page(page);
+	/*
+	 * We can have pages in active list without any cgroup
+	 * ie, hugepage with less than 3 pages. We can safely
+	 * ignore those pages.
+	 */
+	if (!page_hcg || page_hcg != h_cg)
+		goto err_out;
+
+	csize = PAGE_SIZE<<  compound_order(page);
+	if (!parent) {
+		parent = root_h_cgroup;
+		/* root has no limit */
+		res_counter_charge_nofail(&parent->hugepage[idx],
+					  csize,&fail_res);

                                              ^^^
space ?

I don't have code this way locally, may be a mail client error ?

quoted

+	}
+	counter =&h_cg->hugepage[idx];
+	res_counter_uncharge_until(counter, counter->parent, csize);
+
+	set_hugetlb_cgroup(page, parent);
+err_out:
+	put_page(page);
+out:
+	return 0;
+}
+
+/*
+ * Force the hugetlb cgroup to empty the hugetlb resources by moving them to
+ * the parent cgroup.
+ */
  static int hugetlb_cgroup_pre_destroy(struct cgroup *cgroup)
  {
-	/* We will add the cgroup removal support in later patches */
-	   return -EBUSY;
+	struct hstate *h;
+	struct page *page;
+	int ret = 0, idx = 0;
+
+	do {
+		if (cgroup_task_count(cgroup) ||
+		    !list_empty(&cgroup->children)) {
+			ret = -EBUSY;
+			goto out;
+		}

Is this check  going to moved to higher levels ? Do we still need
this. Or will that happen when pred_destroy becomes void ?

quoted

+		/*
+		 * If the task doing the cgroup_rmdir got a signal
+		 * we don't really need to loop till the hugetlb resource
+		 * usage become zero.
+		 */
+		if (signal_pending(current)) {
+			ret = -EINTR;
+			goto out;
+		}

I'll post a patch to remove this check from memcg because memcg's rmdir
always succeed now. So, could you remove this ?

Will drop this 

quoted

+		for_each_hstate(h) {
+			spin_lock(&hugetlb_lock);
+			list_for_each_entry(page,&h->hugepage_activelist, lru) {
+				ret = hugetlb_cgroup_move_parent(idx, cgroup, page);
+				if (ret) {

When 'ret' should be !0 ?
If hugetlb_cgroup_move_parent() always returns 0, the check will not be necessary.

I will make this void funciton.

-aneesh

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>