linux-kernel - Re: [PATCH v12 4/7] mm: introduce compaction and migration for ballooned pages

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]

Date:	Tue, 20 Nov 2012 20:18:04 -0500
From:	Sasha Levin <levinsasha928@...il.com>
To:	Rafael Aquini <aquini@...hat.com>
CC:	linux-mm@...ck.org, linux-kernel@...r.kernel.org,
	virtualization@...ts.linux-foundation.org,
	Andrew Morton <akpm@...ux-foundation.org>,
	Mel Gorman <mel@....ul.ie>,
	"Michael S. Tsirkin" <mst@...hat.com>,
	Minchan Kim <minchan@...nel.org>,
	Rik van Riel <riel@...hat.com>,
	Rusty Russell <rusty@...tcorp.com.au>
Subject: Re: [PATCH v12 4/7] mm: introduce compaction and migration for ballooned
 pages

On 11/20/2012 09:14 AM, Rafael Aquini wrote:
> On Sun, Nov 18, 2012 at 09:59:47AM -0500, Sasha Levin wrote:
>> On Sat, Nov 17, 2012 at 4:54 PM, Rafael Aquini <aquini@...hat.com> wrote:
>>> On Sat, Nov 17, 2012 at 01:01:30PM -0500, Sasha Levin wrote:
>>>>
>>>> I'm getting the following while fuzzing using trinity inside a KVM tools guest,
>>>> on latest -next:
>>>>
>>>> [ 1642.783728] BUG: unable to handle kernel NULL pointer dereference at 0000000000000194
>>>> [ 1642.785083] IP: [<ffffffff8122b354>] isolate_migratepages_range+0x344/0x7b0
>>>>
>>>> My guess is that we see those because of a race during the check in
>>>> isolate_migratepages_range().
>>>>
>>>>
>>>> Thanks,
>>>> Sasha
>>>
>>> Sasha, could you share your .config and steps you did used with trinity? So I
>>> can attempt to reproduce this issue you reported.
>>
>> Basically try running trinity (with ./trinity -m --quiet --dangerous
>> -l off) inside a disposable guest as root.
>>
>> I manage to hit that every couple of hours.
>>
>> Config attached.
>>
> 
> Howdy Sasha,
> 
> After several hours since last Sunday running trinity tests on a traditional
> KVM-QEMU guest as well as running it on a lkvm guest (both running
> next-20121115) I couldn't hit a single time the crash you've reported,
> (un)fortunately.

Odd... I can see it happening here every couple of hours.

> Also, the .config you gave me, applied on top of next-20121115, haven't produced
> the same bin you've running and hitting the mentioned bug, apparently.
> 
> Here's the RIP for your crash:
> [ 1642.783728] BUG: unable to handle kernel NULL pointer dereference at
> 0000000000000194
> [ 1642.785083] IP: [<ffffffff8122b354>] isolate_migratepages_range+0x344/0x7b0
> 
> 
> And here's the symbol address for the next-20121115 with your .config I've been
> running tests on:
> [raquini@x61 linux]$ nm -n vmlinux | grep isolate_migratepages_range 
> ffffffff8122d890 T isolate_migratepages_range
> 
> Also, it seems quite clear I'm missing something from your tree, as applying the
> RIP displacement (0x344) to my local isolate_migratepages_range sym addr leads
> me to the _middle_ of a instruction opcode that does not dereference any
> pointers at all.

Yup, I carry another small fix to mpol (which is unrelated to this one).

> So, if you're consistently reproducing the same crash, consider to share with us
> a disassembled dump from the isolate_migratepages_range() you're running along
> with the crash stack-dump, please.

Sure!

The call chain is:

	isolate_migratepages_range
		balloon_page_movable
			__is_movable_balloon_page
				mapping_balloon

mapping_balloon() fails because it checks for mapping to be non-null (and it is -
it's usually a small value like 0x50), and then it dereferences that.

The relevant assembly is:

static inline int mapping_balloon(struct address_space *mapping)
{
        return mapping && test_bit(AS_BALLOON_MAP, &mapping->flags);
    17ab:       48 85 c0                test   %rax,%rax
    17ae:       0f 84 4c 02 00 00       je     1a00 <isolate_migratepages_range+0x590>
    17b4:       48 8b 80 40 01 00 00    mov    0x140(%rax),%rax
    17bb:       a9 00 00 00 20          test   $0x20000000,%eax
    17c0:       0f 84 3a 02 00 00       je     1a00 <isolate_migratepages_range+0x590>

It dies on 17b4.

Let me know if you need anything else from me, I can also add debug code into the
kernel if it would help you...


Thanks,
Sasha


--
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/