Occasionally, testing memcg's move_charge_at_immigrate on rc7 shows
a flurry of hundreds of warnings at kernel/res_counter.c:96, where
res_counter_uncharge_locked() does WARN_ON(counter->usage < val).
The first trace of each flurry implicates __mem_cgroup_cancel_charge()
of mc.precharge, and an audit of mc.precharge handling points to
mem_cgroup_move_charge_pte_range()'s THP handling in 12724850e806
"memcg: avoid THP split in task migration".
Checking !mc.precharge is good everywhere else, when a single page is
to be charged; but here the "mc.precharge -= HPAGE_PMD_NR" likely to
follow, is liable to result in underflow (a lot can change since the
precharge was estimated).
Simply check against HPAGE_PMD_NR: there's probably a better alternative,
trying precharge for more, splitting if unsuccessful; but this one-liner
is safer for now - no kernel/res_counter.c:96 warnings seen in 26 hours.
Signed-off-by: Hugh Dickins <[email protected]>
---
mm/memcontrol.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
--- 3.4-rc7/mm/memcontrol.c 2012-05-12 22:56:05.340002552 -0700
+++ linux/mm/memcontrol.c 2012-05-17 09:39:45.944034396 -0700
@@ -5481,7 +5481,7 @@ static int mem_cgroup_move_charge_pte_ra
* part of thp split is not executed yet.
*/
if (pmd_trans_huge_lock(pmd, vma) == 1) {
- if (!mc.precharge) {
+ if (mc.precharge < HPAGE_PMD_NR) {
spin_unlock(&vma->vm_mm->page_table_lock);
return 0;
}
(2012/05/19 3:28), Hugh Dickins wrote:
> Occasionally, testing memcg's move_charge_at_immigrate on rc7 shows
> a flurry of hundreds of warnings at kernel/res_counter.c:96, where
> res_counter_uncharge_locked() does WARN_ON(counter->usage < val).
>
> The first trace of each flurry implicates __mem_cgroup_cancel_charge()
> of mc.precharge, and an audit of mc.precharge handling points to
> mem_cgroup_move_charge_pte_range()'s THP handling in 12724850e806
> "memcg: avoid THP split in task migration".
>
> Checking !mc.precharge is good everywhere else, when a single page is
> to be charged; but here the "mc.precharge -= HPAGE_PMD_NR" likely to
> follow, is liable to result in underflow (a lot can change since the
> precharge was estimated).
>
> Simply check against HPAGE_PMD_NR: there's probably a better alternative,
> trying precharge for more, splitting if unsuccessful; but this one-liner
> is safer for now - no kernel/res_counter.c:96 warnings seen in 26 hours.
>
> Signed-off-by: Hugh Dickins <[email protected]>
Thank you.
Acked-by: KAMEZAWA Hiroyuki <[email protected]>