On Tue, 13 Dec 2011 15:55:25 +0800
Bob Liu <lliubbo@xxxxxxxxx> wrote:
> As the document memcg_test.txt said:
> In do_swap_page(), following events occur when pte is unchanged.
> (1) the page (SwapCache) is looked up.
> (2) lock_page()
> (3) try_charge_swapin()
> (4) reuse_swap_page() (may call delete_swap_cache())
> (5) commit_charge_swapin()
> (6) swap_free().
>
> And below situation:
> (C) The page has been charged before (2) and reuse_swap_page() doesn't
> call delete_from_swap_cache().
>
> In this case, __mem_cgroup_commit_charge_swapin() may uncharge memsw twice.
> See below two uncharge place:
>
> __mem_cgroup_commit_charge_swapin {
> => __mem_cgroup_commit_charge_lrucare
> => __mem_cgroup_commit_charge() <== PageCgroupUsed
> => __mem_cgroup_cancel_charge()
> <== 1.uncharge memsw here
>
> if (do_swap_account && PageSwapCache(page)) {
> if (swap_memcg) {
> if (!mem_cgroup_is_root(swap_memcg))
> res_counter_uncharge(&swap_memcg->memsw,
> PAGE_SIZE);
> <== 2.uncharged memsw again here
>
> mem_cgroup_swap_statistics(swap_memcg, false);
> mem_cgroup_put(swap_memcg);
> }
> }
> }
How this happens ?
1. all swap-cache handling is serialized by lock_page().
2. If the page_cgroup is marked as PCG_USED, record in swap_cgroup must be cleared.
and swap_memcg never be found.
There is no real bug. If you want to add VM_BUG_ON() as
commit_charge() decreases memsw count but swap_cgroup is found.
ok, please write a patch.
Nack to this patch.
Thanks,
-Kame
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@xxxxxxxxx. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Fight unfair telecom internet charges in Canada: sign http://stopthemeter.ca/
Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>