Re: [PATCH 7/12] memcg add function to move account

[Balbir Singh <balbir@linux.vnet.ibm.com>]

KAMEZAWA Hiroyuki wrote:
> This patch provides a function to move account information of a page between
> mem_cgroups.
> 

What is the interface for moving the accounting? Is it an explicit call like
force_empty? The other concern I have is about merging two LRU lists, when we
move LRU pages from one mem_cgroup to another, where do we add them? To the head
or tail? I think we need to think about it and document it well.

The other thing is that once we have mult-hierarchy support (which we really
need), we need to move the accounting to the parent instead of root.

> This moving of page_cgroup is done under
>  - lru_lock of source/destination mem_cgroup is held.

I suppose you mean and instead of either for the lru_lock

>  - lock_page_cgroup() is held.
> 
> Then, a routine which touches pc->mem_cgroup without lock_page_cgroup() should
> confirm pc->mem_cgroup is still valid or not. Typlical code can be following.
> 
> (while page is not under lock_page())
> 	mem = pc->mem_cgroup;
> 	mz = page_cgroup_zoneinfo(pc)
> 	spin_lock_irqsave(&mz->lru_lock);
> 	if (pc->mem_cgroup == mem)
> 		...../* some list handling */
> 	spin_unlock_irq(&mz->lru_lock);
> 
> Or better way is
> 	lock_page_cgroup(pc);
> 	....
> 	unlock_page_cgroup(pc);
> 
> But you should confirm the nest of lock and avoid deadlock.
> (trylock is better if it's ok.)
> 
> If you find page_cgroup from mem_cgroup's LRU under mz->lru_lock,
> you don't have to worry about what pc->mem_cgroup points to.
> 
> Changelog: (v4) -> (v5)
>   - check for lock_page() is removed.
>   - rewrote description.
> 
> Changelog: (v2) -> (v4)
>   - added lock_page_cgroup().
>   - splitted out from new-force-empty patch.
>   - added how-to-use text.
>   - fixed race in __mem_cgroup_uncharge_common().
> 
> Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
> 
>  mm/memcontrol.c |   84 ++++++++++++++++++++++++++++++++++++++++++++++++++++++--
>  1 file changed, 81 insertions(+), 3 deletions(-)
> 
> Index: mmotm-2.6.27-rc7+/mm/memcontrol.c
> ===================================================================
> --- mmotm-2.6.27-rc7+.orig/mm/memcontrol.c
> +++ mmotm-2.6.27-rc7+/mm/memcontrol.c
> @@ -426,6 +426,7 @@ int task_in_mem_cgroup(struct task_struc
>  void mem_cgroup_move_lists(struct page *page, enum lru_list lru)
>  {
>  	struct page_cgroup *pc;
> +	struct mem_cgroup *mem;
>  	struct mem_cgroup_per_zone *mz;
>  	unsigned long flags;
> 
> @@ -444,9 +445,14 @@ void mem_cgroup_move_lists(struct page *
> 
>  	pc = page_get_page_cgroup(page);
>  	if (pc) {
> +		mem = pc->mem_cgroup;
>  		mz = page_cgroup_zoneinfo(pc);
>  		spin_lock_irqsave(&mz->lru_lock, flags);
> -		__mem_cgroup_move_lists(pc, lru);
> +		/*
> +		 * check against the race with move_account.
> +		 */
> +		if (likely(mem == pc->mem_cgroup))
> +			__mem_cgroup_move_lists(pc, lru);
>  		spin_unlock_irqrestore(&mz->lru_lock, flags);
>  	}
>  	unlock_page_cgroup(page);
> @@ -567,6 +573,70 @@ unsigned long mem_cgroup_isolate_pages(u
>  	return nr_taken;
>  }
> 
> +/**
> + * mem_cgroup_move_account - move account of the page
> + * @page ... the target page of being moved.
> + * @pc   ... page_cgroup of the page.
> + * @from ... mem_cgroup which the page is moved from.
> + * @to   ... mem_cgroup which the page is moved to.
> + *
> + * The caller must confirm following.
> + * 1. disable irq.
> + * 2. lru_lock of old mem_cgroup should be held.
> + * 3. pc is guaranteed to be valid and on mem_cgroup's LRU.
> + *
> + * Because we cannot call try_to_free_page() here, the caller must guarantee
> + * this moving of charge never fails. (if charge fails, this call fails.)
> + * Currently this is called only against root cgroup.
> + * which has no limitation of resource.
> + * Returns 0 at success, returns 1 at failure.
> + */
> +int mem_cgroup_move_account(struct page *page, struct page_cgroup *pc,
> +	struct mem_cgroup *from, struct mem_cgroup *to)
> +{
> +	struct mem_cgroup_per_zone *from_mz, *to_mz;
> +	int nid, zid;
> +	int ret = 1;
> +
> +	VM_BUG_ON(!irqs_disabled());
> +
> +	nid = page_to_nid(page);
> +	zid = page_zonenum(page);
> +	from_mz =  mem_cgroup_zoneinfo(from, nid, zid);
> +	to_mz =  mem_cgroup_zoneinfo(to, nid, zid);
> +
> +	if (res_counter_charge(&to->res, PAGE_SIZE)) {
> +		/* Now, we assume no_limit...no failure here. */
> +		return ret;
> +	}

Please BUG_ON() if the charging fails, we can be sure we catch assumptions that
are broken.

> +	if (!try_lock_page_cgroup(page)) {
> +		res_counter_uncharge(&to->res, PAGE_SIZE);
> +		return ret;
> +	}
> +
> +	if (page_get_page_cgroup(page) != pc) {
> +		res_counter_uncharge(&to->res, PAGE_SIZE);
> +		goto out;
> +	}
> +
> +	if (spin_trylock(&to_mz->lru_lock)) {

The spin_trylock is to avoid deadlocks, right?

> +		__mem_cgroup_remove_list(from_mz, pc);
> +		css_put(&from->css);
> +		res_counter_uncharge(&from->res, PAGE_SIZE);
> +		pc->mem_cgroup = to;
> +		css_get(&to->css);
> +		__mem_cgroup_add_list(to_mz, pc);
> +		ret = 0;
> +		spin_unlock(&to_mz->lru_lock);
> +	} else {
> +		res_counter_uncharge(&to->res, PAGE_SIZE);
> +	}
> +out:
> +	unlock_page_cgroup(page);
> +
> +	return ret;
> +}
> +
>  /*
>   * Charge the memory controller for page usage.
>   * Return
> @@ -754,16 +824,24 @@ __mem_cgroup_uncharge_common(struct page
>  	if ((ctype == MEM_CGROUP_CHARGE_TYPE_MAPPED)
>  	    && ((PageCgroupCache(pc) || page_mapped(page))))
>  		goto unlock;
> -
> +retry:
> +	mem = pc->mem_cgroup;
>  	mz = page_cgroup_zoneinfo(pc);
>  	spin_lock_irqsave(&mz->lru_lock, flags);
> +	if (ctype == MEM_CGROUP_CHARGE_TYPE_MAPPED &&
> +	    unlikely(mem != pc->mem_cgroup)) {
> +		/* MAPPED account can be done without lock_page().
> +		   Check race with mem_cgroup_move_account() */

Coding style above is broken. Can this race really occur? Why do we get mem
before acquiring the mz->lru_lock? We don't seem to be using it.

> +		spin_unlock_irqrestore(&mz->lru_lock, flags);
> +		goto retry;
> +	}
>  	__mem_cgroup_remove_list(mz, pc);
>  	spin_unlock_irqrestore(&mz->lru_lock, flags);
> 
>  	page_assign_page_cgroup(page, NULL);
>  	unlock_page_cgroup(page);
> 
> -	mem = pc->mem_cgroup;
> +
>  	res_counter_uncharge(&mem->res, PAGE_SIZE);
>  	css_put(&mem->css);
> 
> 


-- 
	Balbir