linux-kernel - Re: [PATCH] mm: fix balloon_page

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: Windows password security audit tool. GUI, reports in PDF.

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

Message-Id: <20121127155201.ddfea7e1.akpm@linux-foundation.org>
Date:	Tue, 27 Nov 2012 15:52:01 -0800
From:	Andrew Morton <akpm@...ux-foundation.org>
To:	Rafael Aquini <aquini@...hat.com>
Cc:	linux-mm@...ck.org, linux-kernel@...r.kernel.org,
	Sasha Levin <levinsasha928@...il.com>,
	Mel Gorman <mel@....ul.ie>, Rik van Riel <riel@...hat.com>
Subject: Re: [PATCH] mm: fix balloon_page_movable() page->flags check

On Tue, 27 Nov 2012 21:31:10 -0200
Rafael Aquini <aquini@...hat.com> wrote:

> This patch fixes the following crash by fixing and enhancing the way 
> page->flags are tested to identify a ballooned page.
> 
> ---8<---
> BUG: unable to handle kernel NULL pointer dereference at 0000000000000194
> IP: [<ffffffff8122b354>] isolate_migratepages_range+0x344/0x7b0
> --->8---
> 
> The NULL pointer deref was taking place because balloon_page_movable()
> page->flags tests were incomplete and we ended up 
> inadvertently poking at private pages.
> 
> ....
>
>  /*
> + * __balloon_page_flags - helper to perform balloon @page ->flags tests.
> + *
> + * As balloon pages are got from Buddy, and we do not play with page->flags
> + * at driver level (exception made when we get the page lock for compaction),
> + * therefore we can safely identify a ballooned page by checking if the
> + * NR_PAGEFLAGS rightmost bits from the page->flags are all cleared.
> + * This approach also helps on skipping ballooned pages that are locked for
> + * compaction or release, thus mitigating their racy check at
> + * balloon_page_movable()
> + */
> +#define BALLOON_PAGE_FLAGS_MASK       ((1UL << NR_PAGEFLAGS) - 1)

hm, this seems a bit fragile.

What's actually going on here?  You're assuming that a page fresh from
buddy has all flags 0..NR_PAGEFLAGS cleared?

That may be true, I'm unsure.  Please take a look at
PAGE_FLAGS_CHECK_AT_FREE and PAGE_FLAGS_CHECK_AT_PREP and work out why
the heck these aren't the same thing!

Either way around, doing

	#define BALLOON_PAGE_FLAGS_MASK PAGE_FLAGS_CHECK_AT_PREP

seems rather more maintainable.

> +static inline bool __balloon_page_flags(struct page *page)
> +{
> +	return page->flags & BALLOON_PAGE_FLAGS_MASK ? false : true;

We have a neater way of doing the scalar-to-boolean conversion:

	return !(page->flags & BALLOON_PAGE_FLAGS_MASK);

> +}
> +
> +/*
>   * __is_movable_balloon_page - helper to perform @page mapping->flags tests
>   */
>  static inline bool __is_movable_balloon_page(struct page *page)
> @@ -135,8 +152,8 @@ static inline bool balloon_page_movable(struct page *page)
>  	 * Before dereferencing and testing mapping->flags, lets make sure
>  	 * this is not a page that uses ->mapping in a different way
>  	 */
> -	if (!PageSlab(page) && !PageSwapCache(page) && !PageAnon(page) &&
> -	    !page_mapped(page))
> +	if (__balloon_page_flags(page) && !page_mapped(page) &&
> +	    page_count(page) == 1)
>  		return __is_movable_balloon_page(page);
>  
>  	return false;
>
> ...
>
--
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/