linux-kernel - Re: [PATCH] drop_caches: Allow unmapping pages

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: <20190107141545.GX6310@bombadil.infradead.org>
Date:   Mon, 7 Jan 2019 06:15:45 -0800
From:   Matthew Wilcox <willy@...radead.org>
To:     Vincent Whitchurch <vincent.whitchurch@...s.com>
Cc:     akpm@...ux-foundation.org, viro@...iv.linux.org.uk,
        linux-fsdevel@...r.kernel.org, linux-mm@...ck.org,
        linux-kernel@...r.kernel.org, mcgrof@...nel.org,
        keescook@...omium.org, corbet@....net, linux-doc@...r.kernel.org,
        Vincent Whitchurch <rabinv@...s.com>
Subject: Re: [PATCH] drop_caches: Allow unmapping pages

On Mon, Jan 07, 2019 at 02:02:39PM +0100, Vincent Whitchurch wrote:
> +++ b/Documentation/sysctl/vm.txt
> @@ -222,6 +222,10 @@ To increase the number of objects freed by this operation, the user may run
>  number of dirty objects on the system and create more candidates to be
>  dropped.
>  
> +By default, pages which are currently mapped are not dropped from the
> +pagecache.  If you want to unmap and drop these pages too, echo 9 or 11 instead
> +of 1 or 3 respectively (set bit 4).

Typically we number bits from 0, so this would be bit 3, not 4.  I do see
elsewhere in this file somebody else got this wrong:

: with your system.  To disable them, echo 4 (bit 3) into drop_caches.

but that should also be fixed.

> +static int __invalidate_inode_page(struct page *page, bool unmap)
> +{
> +	struct address_space *mapping = page_mapping(page);
> +	if (!mapping)
> +		return 0;
> +	if (PageDirty(page) || PageWriteback(page))
> +		return 0;
> +	if (page_mapped(page)) {
> +		if (!unmap)
> +			return 0;
> +		if (!try_to_unmap(page, TTU_IGNORE_ACCESS))
> +			return 0;

You're going to get data corruption doing this.  try_to_unmap_one() does:

                /* Move the dirty bit to the page. Now the pte is gone. */
                if (pte_dirty(pteval))
                        set_page_dirty(page);

so PageDirty() can be false above, but made true by calling try_to_unmap().

I also think the way you've done this is expedient at the cost of
efficiency and layering violations.  I think you should first tear
down the mappings of userspace processes (which will reclaim a lot
of pages allocated to page tables), then you won't need to touch the
invalidate_inode_pages paths at all.