linux-kernel - Re: [RFC][BUGFIX][PATCH 2/2] memcg: fix file mapped underflow at migration (v3)

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:	Fri, 23 Apr 2010 17:08:46 +0900
From:	Daisuke Nishimura <nishimura@....nes.nec.co.jp>
To:	KAMEZAWA Hiroyuki <kamezawa.hiroyu@...fujitsu.com>
Cc:	LKML <linux-kernel@...r.kernel.org>, linux-mm <linux-mm@...ck.org>,
	Mel Gorman <mel@....ul.ie>, Rik van Riel <riel@...hat.com>,
	Minchan Kim <minchan.kim@...il.com>,
	Balbir Singh <balbir@...ux.vnet.ibm.com>,
	KOSAKI Motohiro <kosaki.motohiro@...fujitsu.com>,
	Christoph Lameter <cl@...ux-foundation.org>,
	Andrea Arcangeli <aarcange@...hat.com>,
	Andrew Morton <akpm@...ux-foundation.org>,
	Daisuke Nishimura <nishimura@....nes.nec.co.jp>
Subject: Re: [RFC][BUGFIX][PATCH 2/2] memcg: fix file mapped underflow at
 migration (v3)

I'm sorry for my late reply.

On Tue, 20 Apr 2010 18:19:25 +0900, KAMEZAWA Hiroyuki <kamezawa.hiroyu@...fujitsu.com> wrote:
> On Tue, 20 Apr 2010 13:20:50 +0900
> Daisuke Nishimura <nishimura@....nes.nec.co.jp> wrote:
> 
> > > It will have no meanings for migrating
> > > file caches, but it may have some meanings for easy debugging. 
> > > I think "mark it always but it's used only for anonymous page" is reasonable
> > > (if it causes no bug.)
> > > 
> > Anyway, I don't have any strong objection.
> > It's all right for me as long as it is well documented or commented.
> > 
> Okay, before posting as v4, here is draft version.
> 
Thank you for adding good comments about what it does and why we need it.
I like the direction that we set MIGRATION flags only on the old page.
And this patch looks good to me, except that checkpatch warns some problems
about indent :)

I have one question.

>  /* remove redundant charge if migration failed*/
>  void mem_cgroup_end_migration(struct mem_cgroup *mem,
> -		struct page *oldpage, struct page *newpage)
> +	struct page *oldpage, struct page *newpage)
>  {
> -	struct page *target, *unused;
> +	struct page *used, *unused;
>  	struct page_cgroup *pc;
> -	enum charge_type ctype;
>  
>  	if (!mem)
>  		return;
> +	/* blocks rmdir() */
>  	cgroup_exclude_rmdir(&mem->css);
>  	/* at migration success, oldpage->mapping is NULL. */
>  	if (oldpage->mapping) {
> -		target = oldpage;
> -		unused = NULL;
> +		used = oldpage;
> +		unused = newpage;
>  	} else {
> -		target = newpage;
> +		used = newpage;
>  		unused = oldpage;
>  	}
> -
> -	if (PageAnon(target))
> -		ctype = MEM_CGROUP_CHARGE_TYPE_MAPPED;
> -	else if (page_is_file_cache(target))
> -		ctype = MEM_CGROUP_CHARGE_TYPE_CACHE;
> -	else
> -		ctype = MEM_CGROUP_CHARGE_TYPE_SHMEM;
> -
> -	/* unused page is not on radix-tree now. */
> -	if (unused)
> -		__mem_cgroup_uncharge_common(unused, ctype);
> -
> -	pc = lookup_page_cgroup(target);
>  	/*
> -	 * __mem_cgroup_commit_charge() check PCG_USED bit of page_cgroup.
> -	 * So, double-counting is effectively avoided.
> +	 * We disallowed uncharge of pages under migration because mapcount
> +	 * of the page goes down to zero, temporarly.
> +	 * Clear the flag and check the page should be charged.
>  	 */
> -	__mem_cgroup_commit_charge(mem, pc, ctype);
> -
> +	pc = lookup_page_cgroup(unused);
> +	/* This flag itself is not racy, so, check it before lock */
> +	if (PageCgroupMigration(pc)) {
> +		lock_page_cgroup(pc);
> +		ClearPageCgroupMigration(pc);
> +		unlock_page_cgroup(pc);
> +	}
The reason why "This flag itself is not racy" is that we update the flag only
while the page is isolated ?
Then, we doesn't need page_cgroup lock, do we ? PCG_USED bit will avoid
double-uncharge.

Thanks,
Daisuke Nishimura.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/