[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <4B0A5F8D.6090707@canonical.com>
Date: Mon, 23 Nov 2009 02:10:21 -0800
From: John Johansen <john.johansen@...onical.com>
To: Eric Paris <eparis@...hat.com>
CC: David Howells <dhowells@...hat.com>,
Andrew Morton <akpm@...ux-foundation.org>,
graff.yang@...il.com, linux-kernel@...r.kernel.org,
gyang@...ckfin.uclinux.org,
uclinux-dist-devel@...ckfin.uclinux.org,
Graff Yang <graf.yang@...log.com>,
linux-security-module@...r.kernel.org, john.johansen@...onical.com
Subject: Re: [PATCH] mm/nommu.c: Fix improperly call of security API in mmap
Eric Paris wrote:
> On Sat, 2009-11-21 at 00:16 +0000, David Howells wrote:
>> Eric Paris <eparis@...hat.com> wrote:
>>
>>> +/* sec_flags for security_file_mmap */
>>> +#define SECURITY_MMAP_ADDR_ONLY 0x01
>>> +#define SECURITY_MMAP_NOT_ADDR 0x02
>> Please add comments to these to indicate what they're intended to convey.
>> Would ADDR_ONLY be better as EXACT_ADDR?
>
> I think I should point out that this hook checks 2 things. Originally
> it was only used to check if a file should be allowed to be mmaped. It
> was later enhanced to check if the return address of mmap, if it is file
> backed or anonymous, is acceptable. These flags only influence the
> later.
>
> ADDR_ONLY means the security system should only check the address.
> NOT_ADDR means they security system should not check the address.
>
> You need ADDR_ONLY when the hook is called on map that is not file
> backed or where that has already been dealt with. You need NOT_ADDR
> only for nommu where the whole idea of mmap_min_addr is pointless.
>
> I'm not sure what comments would convey....
>
> /* security hook should only check the address */
> #define SECURITY_MMAP_ADDR_ONLY 0x01
> /* security hook should not check the address */
> #define SECURITY_MMAP_NOT_ADDR 0x02
>
> Does that add something?
>
> Still haven't heard where people scream they absolutely need this today,
> so I'm going to ask James to carry it in his for-next tree.
>
The comments convey a tad more but I don't think they are necessary, and
I concur, it would be good if it went into the for-next tree.
--
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/
Powered by blists - more mailing lists