linux-kernel - Re: [PATCH 1/3] mm/mincore: make mincore() more conservative

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: <20190131094357.GQ18811@dhcp22.suse.cz>
Date:   Thu, 31 Jan 2019 10:43:57 +0100
From:   Michal Hocko <mhocko@...nel.org>
To:     Vlastimil Babka <vbabka@...e.cz>
Cc:     Andrew Morton <akpm@...ux-foundation.org>,
        Linus Torvalds <torvalds@...ux-foundation.org>,
        linux-kernel@...r.kernel.org, linux-mm@...ck.org,
        linux-api@...r.kernel.org, Peter Zijlstra <peterz@...radead.org>,
        Greg KH <gregkh@...uxfoundation.org>,
        Jann Horn <jannh@...gle.com>, Jiri Kosina <jkosina@...e.cz>,
        Dominique Martinet <asmadeus@...ewreck.org>,
        Andy Lutomirski <luto@...capital.net>,
        Dave Chinner <david@...morbit.com>,
        Kevin Easton <kevin@...rana.org>,
        Matthew Wilcox <willy@...radead.org>,
        Cyril Hrubis <chrubis@...e.cz>, Tejun Heo <tj@...nel.org>,
        "Kirill A . Shutemov" <kirill@...temov.name>,
        Daniel Gruss <daniel@...ss.cc>, Jiri Kosina <jikos@...nel.org>
Subject: Re: [PATCH 1/3] mm/mincore: make mincore() more conservative

On Wed 30-01-19 13:44:18, Vlastimil Babka wrote:
> From: Jiri Kosina <jkosina@...e.cz>
> 
> The semantics of what mincore() considers to be resident is not completely
> clear, but Linux has always (since 2.3.52, which is when mincore() was
> initially done) treated it as "page is available in page cache".
> 
> That's potentially a problem, as that [in]directly exposes meta-information
> about pagecache / memory mapping state even about memory not strictly belonging
> to the process executing the syscall, opening possibilities for sidechannel
> attacks.
> 
> Change the semantics of mincore() so that it only reveals pagecache information
> for non-anonymous mappings that belog to files that the calling process could
> (if it tried to) successfully open for writing.

I agree that this is a better way than the original 574823bfab82
("Change mincore() to count "mapped" pages rather than "cached" pages").
One thing is still not clear to me though. Is the new owner/writeable
check OK for the Netflix-like usecases? I mean does happycache have
appropriate access to the cache data? I have tried to re-read the
original thread but couldn't find any confirmation.

I nit below

> Originally-by: Linus Torvalds <torvalds@...ux-foundation.org>
> Originally-by: Dominique Martinet <asmadeus@...ewreck.org>
> Cc: Dominique Martinet <asmadeus@...ewreck.org>
> Cc: Andy Lutomirski <luto@...capital.net>
> Cc: Dave Chinner <david@...morbit.com>
> Cc: Kevin Easton <kevin@...rana.org>
> Cc: Matthew Wilcox <willy@...radead.org>
> Cc: Cyril Hrubis <chrubis@...e.cz>
> Cc: Tejun Heo <tj@...nel.org>
> Cc: Kirill A. Shutemov <kirill@...temov.name>
> Cc: Daniel Gruss <daniel@...ss.cc>
> Signed-off-by: Jiri Kosina <jkosina@...e.cz>
> Signed-off-by: Vlastimil Babka <vbabka@...e.cz>

other than that looks good to me.
Acked-by: Michal Hocko <mhocko@...e.com>

If this still doesn't help happycache kind of workloads then we should
add a capability check IMO but this looks like a decent foundation to
me.

> ---
>  mm/mincore.c | 15 ++++++++++++++-
>  1 file changed, 14 insertions(+), 1 deletion(-)
> 
> diff --git a/mm/mincore.c b/mm/mincore.c
> index 218099b5ed31..747a4907a3ac 100644
> --- a/mm/mincore.c
> +++ b/mm/mincore.c
> @@ -169,6 +169,14 @@ static int mincore_pte_range(pmd_t *pmd, unsigned long addr, unsigned long end,
>  	return 0;
>  }
>  
> +static inline bool can_do_mincore(struct vm_area_struct *vma)
> +{
> +	return vma_is_anonymous(vma) ||
> +		(vma->vm_file &&
> +			(inode_owner_or_capable(file_inode(vma->vm_file))
> +			 || inode_permission(file_inode(vma->vm_file), MAY_WRITE) == 0));
> +}

This is hard to read. Can we do
	if (vma_is_anonymous(vma))
		return true;
	if (!vma->vm_file)
		return false;
	return inode_owner_or_capable(file_inode(vma->vm_file)) ||
		inode_permission(file_inode(vma->vm_file), MAY_WRITE) == 0;

-- 
Michal Hocko
SUSE Labs