Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Tue, 28 Jun 2011 17:12:13 -0700
From: Linus Torvalds <>
To: Andrew Morton <>
Cc: Vasiliy Kulikov <>,,
Subject: Re: [Security] CVE request: kernel: taskstats/procfs io infoleak
 (was: taskstats authorized_keys presence infoleak PoC)

On Tue, Jun 28, 2011 at 3:53 PM, Andrew Morton
<> wrote:
> a) I haven't thought very hard about it, but isn't it the case that
>   fuzzifying the byte counts in this manner will still permit the
>   length of these things to be determined, albeit with a larger data
>   set?

Well, if we're talking things like passwords read from /dev/tty, there
really _isn't_ a larger data set to be had. Which is why I suspect
that it would be fine to just mask the low bits and give 1kB
resolution in general.

It's not like I could imagine an app like "iotop" would ever care
about individual bytes, so there's no reason to expose things at that

> b) Where does the problem lie?  Is it with the kernel, which exposes
>   accurate accounting?  Or is it with userspace, which accidentally
>   exposes sensitive information by failing to account for the kernel's
>   exposure of accurate accounting information?

Well, if you do a read of a password from a tty, there really isn't
much you can do about the tty IO count showing up. If the rest of the
IO is packetized some way, you can probably figure out the parts that
are individual bytes.

> c) Should this information be world-readable?  Perhaps we should add
>   more rational privileges here.  Back-compatibility issues.

I already applied the /proc part. That seemed like a nobrainer. The
taskstat part look slike it might break iotop to tighten the security,
so there I'm thinking the granularity approach would be a sufficient

> If rounding the counts to a 1k granularity will indeed defeat the
> attack (I'm unsure) then I'd suggest that a fix would be to perform
> that fuzzification if the receiving process doesn't have suitable
> permissions.  So if the user is reading his own stats or is root, he
> still gets byte-resolution results.  This keeps the stats as useful as
> we can make them and reduces the back-compatibility damage.



Powered by blists - more mailing lists

Please check out the Open Source Software Security Wiki, which is counterpart to this mailing list.

Confused about mailing lists and their use? Read about mailing lists on Wikipedia and check out these guidelines on proper formatting of your messages.