From: Thomas Garnier <thgarnie@google.com>
To: Russell King - ARM Linux <linux@armlinux.org.uk>
Cc: Mark Rutland <mark.rutland@arm.com>,
Kernel Hardening <kernel-hardening@lists.openwall.com>,
Catalin Marinas <catalin.marinas@arm.com>,
Will Deacon <will.deacon@arm.com>,
LKML <linux-kernel@vger.kernel.org>,
David Howells <dhowells@redhat.com>,
Dave Hansen <dave.hansen@intel.com>,
Octavian Purdila <octavian.purdila@nxp.com>,
"H . Peter Anvin" <hpa@zytor.com>,
Miroslav Benes <mbenes@suse.cz>,
Chris Metcalf <cmetcalf@mellanox.com>,
Pratyush Anand <panand@redhat.com>,
Stephen Rothwell <sfr@canb.auug.org.au>,
Leonard Crestez <leonard.crestez@nxp.com>,
the arch/x86 maintainers <x86@kernel.org>,
Ingo Molnar <mingo@redhat.com>, Petr Mladek <pmladek@suse.com>,
Rik van Riel <riel@redhat.com>, Kees Cook <keescook@chromium.org>,
Arnd Bergmann <arnd@arndb.de>
Subject: Re: [kernel-hardening] Re: [PATCH v10 2/3] arm/syscalls: Check address limit on user-mode return
Date: Wed, 19 Jul 2017 11:50:39 -0700 [thread overview]
Message-ID: <CAJcbSZHMzrc8-VMQFXtacj7jYB9OqsqD92fXfQpL_DmomFK9Ow@mail.gmail.com> (raw)
In-Reply-To: <20170719183543.GT31807@n2100.armlinux.org.uk>
On Wed, Jul 19, 2017 at 11:35 AM, Russell King - ARM Linux
<linux@armlinux.org.uk> wrote:
> On Wed, Jul 19, 2017 at 10:20:35AM -0700, Thomas Garnier wrote:
>> On Wed, Jul 19, 2017 at 10:06 AM, Russell King - ARM Linux
>> <linux@armlinux.org.uk> wrote:
>> > On Wed, Jul 19, 2017 at 05:58:20PM +0300, Leonard Crestez wrote:
>> > Probably best to revert. I stopped looking at these patches during
>> > the discussion, as the discussion seemed to be mainly around other
>> > architectures, and I thought we had ARM settled.
>> >
>> > Looking at this patch now, there's several things I'm not happy with.
>> >
>> > The effect of adding a the new TIF flag for FSCHECK amongst the other
>> > flags is that we end up overflowing the 8-bit constant, and have to
>> > split the tests, meaning more instructions in the return path. Eg:
>> >
>> > - tst r1, #_TIF_SYSCALL_WORK | _TIF_WORK_MASK
>> > + tst r1, #_TIF_SYSCALL_WORK
>> > + bne fast_work_pending
>> > + tst r1, #_TIF_WORK_MASK
>> > bne fast_work_pending
>> >
>> > should be written:
>> >
>> > tst r1, #_TIF_SYSCALL_WORK
>> > tsteq r1, #_TIF_WORK_MASK
>> > bne fast_work_pending
>> >
>> > and:
>> >
>> > - tst r1, #_TIF_SYSCALL_WORK | _TIF_WORK_MASK
>> > + tst r1, #_TIF_SYSCALL_WORK
>> > + bne fast_work_pending
>> > + tst r1, #_TIF_WORK_MASK
>> >
>> > should be:
>> >
>> > tst r1, #_TIF_SYSCALL_WORK
>> > tsteq r1, #_TIF_WORK_MASK
>> >
>> > There's no need for extra branches.
>> >
>> > Now, the next issue is that I don't think this TIF-flag approach is
>> > good for ARM - alignment faults can happen any time due to misaligned
>> > packets in the networking code, and we really don't want to be doing
>> > this check in a place that we can loop.
>> >
>> > My original suggestion for ARM was to do the address limit check after
>> > all work had been processed, with interrupts disabled (so no
>> > possibility of this kind of loop happening.) However, that seems to
>> > have been replaced with this TIF approach, which is going to cause
>> > loops - I suspect if the probes code is enabled, this will suffer
>> > the same problem. Remember, the various probes stuff can walk
>> > userspace stacks, which means they'll be using set_fs().
>> >
>> > I don't see why we've ended up with this (imho) sub-standard TIF-flag
>> > approach, and I think it's going to be very problematical.
>> >
>> > Can we please go back to the approach I suggested back in March for
>> > ARM that doesn't suffer from this problem?
>>
>> During the extensive thread discussion, Linus asked to move away from
>> architecture specific changes to this work flag system. I am glad to
>> fix the assembly as you asked on a separate patch.
>
> Well, for the record, I don't think you've got to the bottom of the
> "infinite loop" potential of Linus' approach.
>
> Eg, perf will likely trigger this same issue. Eg, perf record -a -g
> will attempt to record the callchain both in kernel space and userspace
> each time a perf interrupt happens. If the perf interrupt frequency is
> sufficiently high that we have multiple interrupts during the execution
> of do_work_pending() and its called functions, then that will turn this
> into an infinite loop yet again.
Do you think it applies to the patch I just sent? The other approach
is to check at the entrance, ignore _TIF_FSCHECK on the loop and clear
it on exit.
>
> --
> RMK's Patch system: http://www.armlinux.org.uk/developer/patches/
> FTTC broadband for 0.8mile line: currently at 9.6Mbps down 400kbps up
> according to speedtest.net.
--
Thomas
next prev parent reply other threads:[~2017-07-19 18:50 UTC|newest]
Thread overview: 23+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-06-15 1:12 [PATCH v10 1/3] x86/syscalls: Check address limit on user-mode return Thomas Garnier
2017-06-15 1:12 ` [PATCH v10 2/3] arm/syscalls: " Thomas Garnier
[not found] ` <20170615011203.144108-2-thgarnie-hpIqsD4AKlfQT0dZR+AlfA@public.gmane.org>
2017-06-20 20:18 ` Kees Cook
[not found] ` <CAGXu5jLR7io8u-M8tqbYW22C+sb2a2wSYLRBqJ_dguT4x+1tsQ-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
2017-06-20 20:31 ` Thomas Garnier
2017-06-21 9:08 ` Will Deacon
2017-07-18 14:36 ` Leonard Crestez
2017-07-18 16:04 ` Thomas Garnier
[not found] ` <CAJcbSZEr8HPBwH1oVaHqPzAY4MS_=yqMoqPhcauuKu3cikB3uQ-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
2017-07-18 17:18 ` Leonard Crestez
2017-07-18 19:04 ` Thomas Garnier
[not found] ` <CAJcbSZFr9KJTfGfiZo2fThoDkAE-D1OFf2YtELq4P6jX8syesQ-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
2017-07-19 14:58 ` Leonard Crestez
[not found] ` <1500476300.22834.13.camel-3arQi8VN3Tc@public.gmane.org>
2017-07-19 16:51 ` Thomas Garnier
2017-07-19 17:06 ` Russell King - ARM Linux
2017-07-19 17:20 ` [kernel-hardening] " Thomas Garnier
[not found] ` <CAJcbSZHi6454skNpG8ecMnq90LdUfcxy2RYZD+7og1C1PeypvQ-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
2017-07-19 18:35 ` Russell King - ARM Linux
2017-07-19 18:50 ` Thomas Garnier [this message]
2017-06-15 1:12 ` [PATCH v10 3/3] arm64/syscalls: " Thomas Garnier
[not found] ` <20170615011203.144108-3-thgarnie-hpIqsD4AKlfQT0dZR+AlfA@public.gmane.org>
2017-06-21 8:16 ` Catalin Marinas
2017-06-21 13:57 ` Thomas Garnier
[not found] ` <20170615011203.144108-1-thgarnie-hpIqsD4AKlfQT0dZR+AlfA@public.gmane.org>
2017-06-20 20:24 ` [PATCH v10 1/3] x86/syscalls: " Kees Cook
2017-06-28 17:52 ` Kees Cook
[not found] ` <CAGXu5jKrJv0y70e5JiafKGcGzWoJPZM_HruZ=Y0rM1m0J4tZAA-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
2017-07-06 20:38 ` Thomas Garnier
[not found] ` <CAJcbSZE6Og4gwhFwhy_-Jaq6GovwN3y1B6O89JmkpXHtVfDLBA-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
2017-07-06 20:48 ` Thomas Gleixner
2017-07-06 20:52 ` 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=CAJcbSZHMzrc8-VMQFXtacj7jYB9OqsqD92fXfQpL_DmomFK9Ow@mail.gmail.com \
--to=thgarnie@google.com \
--cc=arnd@arndb.de \
--cc=catalin.marinas@arm.com \
--cc=cmetcalf@mellanox.com \
--cc=dave.hansen@intel.com \
--cc=dhowells@redhat.com \
--cc=hpa@zytor.com \
--cc=keescook@chromium.org \
--cc=kernel-hardening@lists.openwall.com \
--cc=leonard.crestez@nxp.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux@armlinux.org.uk \
--cc=mark.rutland@arm.com \
--cc=mbenes@suse.cz \
--cc=mingo@redhat.com \
--cc=octavian.purdila@nxp.com \
--cc=panand@redhat.com \
--cc=pmladek@suse.com \
--cc=riel@redhat.com \
--cc=sfr@canb.auug.org.au \
--cc=will.deacon@arm.com \
--cc=x86@kernel.org \
/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).