mm/hugetlb: optimize the surplus state transfer code in move_hugetlb_state()

We should not transfer the per-node surplus state when we do not cross the
node in order to save some cpu cycles

Link: https://lkml.kernel.org/r/[email protected]
Signed-off-by: Miaohe Lin <[email protected]>
Reviewed-by: Mike Kravetz <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
This commit is contained in:
Miaohe Lin authored and Linus Torvalds committed 2021-05-05 11:27:20 -07:00
1 parent 04adbc3f7b
commit 5af1ab1d24
1 file changed
+6
+6
View File
@@ -5682,6 +5682,12 @@ void move_hugetlb_state(struct page *oldpage, struct page *newpage, int reason)
SetHPageTemporary(oldpage);
ClearHPageTemporary(newpage);
/*
* There is no need to transfer the per-node surplus state
* when we do not cross the node.
*/
if (new_nid == old_nid)
return;
spin_lock(&hugetlb_lock);
if (h->surplus_huge_pages_node[old_nid]) {
h->surplus_huge_pages_node[old_nid]--;