From: Kees Cook <keescook@chromium.org>
To: Ingo Molnar <mingo@kernel.org>
Cc: "Daniel Micay" <danielmicay@gmail.com>,
"Thomas Garnier" <thgarnie@google.com>,
"Martin Schwidefsky" <schwidefsky@de.ibm.com>,
"Heiko Carstens" <heiko.carstens@de.ibm.com>,
"Dave Hansen" <dave.hansen@intel.com>,
"Arnd Bergmann" <arnd@arndb.de>,
"Thomas Gleixner" <tglx@linutronix.de>,
"David Howells" <dhowells@redhat.com>,
"René Nyffenegger" <mail@renenyffenegger.ch>,
"Andrew Morton" <akpm@linux-foundation.org>,
"Paul E . McKenney" <paulmck@linux.vnet.ibm.com>,
"Eric W . Biederman" <ebiederm@xmission.com>,
"Oleg Nesterov" <oleg@redhat.com>,
"Pavel Tikhomirov" <ptikhomirov@virtuozzo.com>,
"Ingo Molnar" <mingo@redhat.com>,
"H . Peter Anvin" <hpa@zytor.com>,
"Andy Lutomirski" <luto@kernel.org>,
"Paolo Bonzini" <pbonzini@redhat.com>,
"Rik van Riel" <riel@redhat.>
Subject: Re: [kernel-hardening] Re: [PATCH v9 1/4] syscalls: Verify address limit before returning to user-mode
Date: Tue, 9 May 2017 09:30:02 -0700 [thread overview]
Message-ID: <CAGXu5jL-qvFxLkJZSosAovK4qL5eLPOD7orpei42x6mK_tBXhw@mail.gmail.com> (raw)
In-Reply-To: <20170509065619.wmqa6z6w3n6xpvrw@gmail.com>
On Mon, May 8, 2017 at 11:56 PM, Ingo Molnar <mingo@kernel.org> wrote:
>
> * Kees Cook <keescook@chromium.org> wrote:
>
>> > There's the option of using GCC plugins now that the infrastructure was
>> > upstreamed from grsecurity. It can be used as part of the regular build
>> > process and as long as the analysis is pretty simple it shouldn't hurt compile
>> > time much.
>>
>> Well, and that the situation may arise due to memory corruption, not from
>> poorly-matched set_fs() calls, which static analysis won't help solve. We need
>> to catch this bad kernel state because it is a very bad state to run in.
[attempting some thread-merging]
> Ok, so that's CVE-2010-4258, where an oops with KERNEL_DS set was used to escalate
> privileges, due to the kernel's oops handler not cleaning up the KERNEL_DS. The
> exploit used another bug, a crash in a network protocol handler, to execute the
> oops handler with KERNEL_DS set.
Right, I didn't mean to suggest that vulnerability would be fixed by
this solution. I was trying to show how there can be some pretty
complex interaction with exceptions/interrupts/etc that would make
pure static analysis still miss things.
> If memory corruption corrupted the task state into having addr_limit set to
> KERNEL_DS then there's already a fair chance that it's game over: it could also
> have set *uid to 0, or changed a sensitive PF_ flag, or a number of other
> things...
>
> Furthermore, think about it: there's literally an infinite amount of corrupted
> task states that could be a security problem and that could be checked after every
> system call. Do we want to check every one of them?
Right, but this "slippery slope" argument isn't the best way to reject
security changes. Let me take a step back and describe the threat, and
where we should likely spend time:
The primary threat with addr_limit getting changed is that a
narrowly-scoped attack (traditionally stack exhaustion or
adjacent-stack large-index writes) could be leveraged into opening the
entire kernel to writes (by allowing all syscalls with a
copy_to_user() call to suddenly be able to write to kernel memory).
So, really, the flaw is having addr_limit at all. Removing set_fs()
should, I think, allow this to become a const (or at least should get
us a lot closer).
The main path to corrupting addr_limit has been via stack corruption.
On architectures with CONFIG_THREAD_INFO_IN_TASK, this risk is greatly
reduced already, but it's not universally available yet. (And as long
as we're talking about stack attacks, CONFIG_VMAP_STACK makes
cross-stack overflows go away, and cross-stack indexing harder, but
that's not really about addr_limit since currently nothing with
VMAP_STACK doesn't already have THREAD_INFO_IN_TASK.)
So, left with a still exploitable target in memory that allows such an
expansion of attack method, I still think it's worth keeping this
patch series, but if we can drop set_fs() I could probably be
convinced the benefit of the series doesn't exceed the cost on
THREAD_INFO_IN_TASK-architectures (x86, arm64, s390). But that means
at least currently keeping it on arm, for example. If we can make
addr_limit const, well, we don't need the series at all.
-Kees
--
Kees Cook
Pixel Security
next prev parent reply other threads:[~2017-05-09 16:30 UTC|newest]
Thread overview: 89+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-04-28 15:32 [PATCH v9 1/4] syscalls: Verify address limit before returning to user-mode Thomas Garnier
[not found] ` <20170428153213.137279-1-thgarnie-hpIqsD4AKlfQT0dZR+AlfA@public.gmane.org>
2017-04-28 15:32 ` [PATCH v9 2/4] x86/syscalls: Optimize address limit check Thomas Garnier
2017-05-05 22:18 ` [PATCH v9 1/4] syscalls: Verify address limit before returning to user-mode Thomas Garnier
[not found] ` <CAJcbSZGQsRVg3QZ9QfLn2HBC+RP-7fUTab0bYDJ455d8y8GyNw-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
2017-05-08 7:33 ` Ingo Molnar
[not found] ` <20170508073352.caqe3fqf7nuxypgi-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
2017-05-08 7:52 ` Ingo Molnar
[not found] ` <20170508075209.7aluvpwildw325rf-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
2017-05-08 15:22 ` [kernel-hardening] " Daniel Micay
2017-05-08 15:26 ` Kees Cook
2017-05-08 19:51 ` Thomas Garnier
[not found] ` <CAGXu5jL61K0bRSEg9a_LswNyrt3K1J57REbWVcvAXT54zWwtMA-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
2017-05-09 6:56 ` Ingo Molnar
[not found] ` <20170509065619.wmqa6z6w3n6xpvrw-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
2017-05-09 11:10 ` Greg KH
[not found] ` <20170509111007.GA14702-U8xfFu+wG4EAvxtiuMwx3w@public.gmane.org>
2017-05-09 14:29 ` Thomas Garnier
[not found] ` <CAJcbSZFswDWZoK-1UK+xkRMJ4ttSYbtH2Y5WD5_aPR-8ru6t8A-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
2017-05-11 23:17 ` Thomas Garnier
[not found] ` <CAJcbSZEoRyewUtBHvqmNZL9FtT_q42Vmmd-EuC50x-ZRASiHHg-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
2017-05-11 23:44 ` Linus Torvalds
[not found] ` <CA+55aFwvQfs_X+paQF6Luc0Rq+W3J2fKuHRou7=ANcquDdXdDA-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
2017-05-12 5:28 ` Martin Schwidefsky
2017-05-12 5:34 ` Kees Cook
2017-05-12 5:54 ` Martin Schwidefsky
2017-05-12 19:01 ` Kees Cook
2017-05-12 19:08 ` Russell King - ARM Linux
[not found] ` <CAGXu5jL9vUrn4kpjO+qa4cHmWBypeqP17OGbrMs=5Nz0YpQMZw-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
2017-05-12 19:08 ` Linus Torvalds
[not found] ` <CA+55aFzbiBqsYb7vwO=+L4Vp_GOgPu+DBOrq4fBnyzq5DbBehg-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
2017-05-12 19:30 ` Kees Cook
[not found] ` <CAGXu5j+xmyJ6RhtPw9rUgs7k3sZ1KKWffvyGGG--oPfu9W42ng-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
2017-05-12 20:21 ` Russell King - ARM Linux
2017-05-12 20:30 ` Peter Zijlstra
[not found] ` <20170512203044.GI4626-IIpfhp3q70z/8w/KjCw3T+5/BudmfyzbbVWyRVo5IupeoWH0uzbU5w@public.gmane.org>
2017-05-12 20:45 ` Russell King - ARM Linux
2017-05-12 21:00 ` Kees Cook
[not found] ` <CAGXu5jL6FPuShBpZfi6+XHqOk4gxocUJRYPHT5oR3HYh3xm+sA-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
2017-05-12 21:04 ` Kees Cook
2017-05-13 7:21 ` Christoph Hellwig
2017-05-12 21:06 ` Al Viro
[not found] ` <20170512210645.GS390-3bDd1+5oDREiFSDQTTA3OLVCufUGDwFn@public.gmane.org>
2017-05-12 21:16 ` [kernel-hardening] " Daniel Micay
2017-05-12 21:17 ` Kees Cook
[not found] ` <CAGXu5jJu=VTqp2tzkPB4RAVxdGC+_SSQwrUwdzWpu24AA-zEcg-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
2017-05-12 21:23 ` Daniel Micay
2017-05-12 21:41 ` Al Viro
2017-05-12 21:47 ` Rik van Riel
[not found] ` <1494625675.29205.21.camel-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
2017-05-12 22:57 ` Al Viro
[not found] ` <20170512214144.GT390-3bDd1+5oDREiFSDQTTA3OLVCufUGDwFn@public.gmane.org>
2017-05-12 21:50 ` Kees Cook
[not found] ` <CAGXu5j+EatK=DYONRkgovwLgytAnbG8jnAZaMSLckZFNVj3gig-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
2017-05-12 6:57 ` Ingo Molnar
2017-05-12 6:13 ` Andy Lutomirski
2017-05-12 6:58 ` Ingo Molnar
2017-05-12 17:05 ` Thomas Garnier
2017-05-09 16:30 ` Kees Cook [this message]
2017-05-08 12:46 ` [kernel-hardening] " Greg KH
[not found] ` <20170508124621.GA20705-U8xfFu+wG4EAvxtiuMwx3w@public.gmane.org>
2017-05-09 6:45 ` Ingo Molnar
[not found] ` <20170509064522.anusoikaalvlux3w-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
2017-05-09 8:56 ` Christoph Hellwig
2017-05-09 13:00 ` Andy Lutomirski
2017-05-09 13:02 ` [kernel-hardening] " Christoph Hellwig
2017-05-09 16:03 ` Christoph Hellwig
2017-05-09 16:50 ` Kees Cook
[not found] ` <CAGXu5jKHVMRMKDfn+=kkbm+JkWPhoEtDwKx=QXAYxg1p9bn7PQ-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
2017-05-09 22:52 ` Andy Lutomirski
[not found] ` <CALCETrV73=cDvaSLOMvb299yaGNJYME8LC-=P+N6p7R1NN97Yg-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
2017-05-09 23:31 ` Kees Cook
2017-05-10 1:59 ` Andy Lutomirski
2017-05-10 7:15 ` Christoph Hellwig
[not found] ` <CAGXu5jL6PeQmmdxh5h--fgrMK8DW_XZYpNfDOvvv_o9E3-Kxdw-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
2017-05-11 11:22 ` Borislav Petkov
2017-05-10 6:46 ` Christoph Hellwig
[not found] ` <20170509160322.GA15902-wEGCiKHe2LqWVfeAwA7xHQ@public.gmane.org>
2017-05-10 2:11 ` Al Viro
2017-05-10 2:45 ` Al Viro
2017-05-10 3:12 ` Al Viro
[not found] ` <20170510031254.GC390-3bDd1+5oDREiFSDQTTA3OLVCufUGDwFn@public.gmane.org>
2017-05-10 3:21 ` Al Viro
[not found] ` <20170510032137.GD390-3bDd1+5oDREiFSDQTTA3OLVCufUGDwFn@public.gmane.org>
2017-05-10 3:39 ` Al Viro
2017-05-10 6:54 ` Christoph Hellwig
2017-05-10 6:53 ` Christoph Hellwig
2017-05-10 7:27 ` Al Viro
2017-05-10 7:35 ` Christoph Hellwig
2017-05-10 6:49 ` Christoph Hellwig
2017-05-10 7:28 ` Arnd Bergmann
2017-05-10 7:35 ` Christoph Hellwig
2017-05-09 16:05 ` Brian Gerst
[not found] ` <CALCETrUh8NO2scaqEM48K70Fo2+V3=Cpyk4JurCDiCYp4nm_+g-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
2017-05-10 7:37 ` [kernel-hardening] " Arnd Bergmann
2017-05-10 8:08 ` Al Viro
2017-05-10 8:14 ` Christoph Hellwig
2017-05-11 0:18 ` Andy Lutomirski
2017-05-12 7:00 ` Ingo Molnar
2017-05-12 7:15 ` Al Viro
2017-05-12 7:35 ` Christoph Hellwig
2017-05-12 8:07 ` Christoph Hellwig
2017-05-12 8:23 ` Greg KH
[not found] ` <20170512071549.GP390-3bDd1+5oDREiFSDQTTA3OLVCufUGDwFn@public.gmane.org>
2017-05-12 7:43 ` [kernel-hardening] " Arnd Bergmann
2017-05-12 8:11 ` Christoph Hellwig
2017-05-12 8:16 ` Al Viro
2017-05-12 8:11 ` Al Viro
[not found] ` <20170512081154.GQ390-3bDd1+5oDREiFSDQTTA3OLVCufUGDwFn@public.gmane.org>
2017-05-12 8:20 ` Arnd Bergmann
2017-05-12 23:20 ` Andy Lutomirski
2017-05-08 13:09 ` Kees Cook
2017-05-08 14:02 ` Ingo Molnar
[not found] ` <20170508140230.23kxf2kfeazeo4zr-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
2017-05-08 14:06 ` Jann Horn
[not found] ` <CAG48ez0Hz=CimkPwuq903tgJkGj8gXUtiQJJb-P2zUes6bd6Hw-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
2017-05-08 20:48 ` Al Viro
[not found] ` <20170508204858.GT29622-3bDd1+5oDREiFSDQTTA3OLVCufUGDwFn@public.gmane.org>
2017-05-12 23:15 ` Andy Lutomirski
2017-05-08 15:24 ` Kees Cook
[not found] ` <CAGXu5jJ4iY7QZ9wRu5dmm7RHtLh_V6TQh4huWwLCYPKOr63aiA-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
2017-05-09 6:34 ` Ingo Molnar
2017-04-28 15:32 ` [PATCH v9 3/4] arm/syscalls: Optimize address limit check Thomas Garnier
2017-04-28 15:32 ` [PATCH v9 4/4] arm64/syscalls: " Thomas Garnier
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=CAGXu5jL-qvFxLkJZSosAovK4qL5eLPOD7orpei42x6mK_tBXhw@mail.gmail.com \
--to=keescook@chromium.org \
--cc=akpm@linux-foundation.org \
--cc=arnd@arndb.de \
--cc=danielmicay@gmail.com \
--cc=dave.hansen@intel.com \
--cc=dhowells@redhat.com \
--cc=ebiederm@xmission.com \
--cc=heiko.carstens@de.ibm.com \
--cc=hpa@zytor.com \
--cc=luto@kernel.org \
--cc=mail@renenyffenegger.ch \
--cc=mingo@kernel.org \
--cc=mingo@redhat.com \
--cc=oleg@redhat.com \
--cc=paulmck@linux.vnet.ibm.com \
--cc=pbonzini@redhat.com \
--cc=ptikhomirov@virtuozzo.com \
--cc=riel@redhat. \
--cc=schwidefsky@de.ibm.com \
--cc=tglx@linutronix.de \
--cc=thgarnie@google.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).