[<prev] [next>] [day] [month] [year] [list]
Message-ID: <10a6188e-ec3f-9dd9-8b14-206f1753697e@nwever.nl>
Date: Tue, 13 Dec 2016 23:59:53 +0100
From: Berend-Jan Wever <berendj@...ver.nl>
To: fulldisclosure@...lists.org, Bugtraq <bugtraq@...urityfocus.com>
Subject: [FD] MSIE 9 MSHTML CMarkup::ReloadInCompatView use-after-free
Since November I have been releasing details on all vulnerabilities I
found that I have not released before. This is the thirty-first entry
in the series. This information is available in more detail on my blog
at http://blog.skylined.nl/20161213001.html. There you can find a repro
that triggered this issue in addition to the information below.
If you find these releases useful, and would like to help me make time
to continue releasing this kind of information, you can make a donation
in bitcoin to 183yyxa9s1s1f7JBpPHPmzQ346y91Rx5DX.
Follow me on http://twitter.com/berendjanwever for daily browser bugs.
MSIE9 MSHTML CMarkup::ReloadInCompatView use-after-free
=======================================================
(The fix and CVE number for this issue are not known)
Synopsis
--------
A specially crafted web-page can trigger a use-after-free vulnerability
in Microsoft Internet Explorer 9. During a method call, the `this`
object can be freed and then continues to be used by the code that
implements the method. It appears that there is little to no time for an
attacker to attempt to control the contents of the freed memory before
the re-use, which would allow remote code execution.
Known affected software and attack vectors
------------------------------------------
* Microsoft Internet Explorer 9
An attacker would need to get a target user to open a specially
crafted web-page. Disabling JavaScript should prevent an attacker
from triggering the vulnerable code path.
Details
-------
By switching the a document's `designMode` property to `on` in a
deferred script, MSIE 9 can be made to reload a web page using
`CMarkup::ReloadInCompatView`. This method calls
`CDoc::CompatViewRefresh`, which indirectly calls
`CScriptCollection::~CScriptCollection`, which releases the `CMarkup`
object used as `this` in `CMarkup::ReloadInCompatView`.
The relevant stack for the freeing of this `CMarkup` object is:
```
76e8c484 kernel32!HeapFree+0x00000014
6780c4d8 MSHTML!CMarkup::`vector deleting destructor'+0x00000026
6776fb9b MSHTML!CScriptCollection::~CScriptCollection+0x00000152
67816a0d MSHTML!CScriptCollection::Release+0x00000053
6751f7e7 MSHTML!CWindow::SuperNavigateInternal+0x000004c4
675209f7 MSHTML!CWindow::SuperNavigate2WithBindFlags+0x00000032
679b05f8 MSHTML!CDoc::CompatViewRefresh+0x000000a0
679c00d4 MSHTML!CMarkup::ReloadInCompatView+0x0000021f
```
Immediately after returning to `CMarkup::ReloadInCompatView`, the code
will use the (now freed) `CMarkup` object. When page heap is enabled,
this lead to an immediate access violation.
Exploit
-------
I did not immediately find a way to control the freed memory before the
reuse following the `CDoc::CompatViewRefresh` call. I did not
immediately find other locations in the code where the same stale
pointer to the `CMarkup` object is used after it has been freed. It may
not be possible to exploit this use-after-free, as there does not appear
to be an easy window of opportunity to modify the freed memory before
its reuse.
However, when loading the repro in MSIE with page heap disabled, I do
see crashes from time to time, but in different locations in the code.
This indicates that one or more of the following should be true:
* There are ways to modify the freed `CMarkup` object before it is
reused.
* There are other locations where the freed `CMarkup` object is used
after it has been freed, and the freed `CMarkup` object can be
modified before this happens.
* There could be other stale pointers to freed memory that get reused,
and there are ways to modify the freed memory they point to before
that reuse.
As these other crash stacks do not include
`CMarkup::ReloadInCompatView`, it seems most likely that they are caused
by the second or third option, which could indicate that the bug is in
fact exploitable.
Time-line
---------
* 5 May 2014: This vulnerability was found through fuzzing.
* 14 May 2014: This vulnerability was submitted to ZDI.
* 3 July 2014: This vulnerability was rejected by ZDI.
* 9 July 2014: This vulnerability was submitted to EIP.
* July/August 2014: This vulnerability was rejected by EIP.
* 13 August 2014: This vulnerability was submitted to iDefense.
* Date unknown: This issue was withdrawn from iDefense.
* Date unknown: This vulnerability was address by Microsoft.
* 13 December 2016: Details of this vulnerability are released.
Cheers,
SkyLined
Download attachment "0x2557C5AA.asc" of type "application/pgp-keys" (2036 bytes)
Download attachment "signature.asc" of type "application/pgp-signature" (820 bytes)
_______________________________________________
Sent through the Full Disclosure mailing list
https://nmap.org/mailman/listinfo/fulldisclosure
Web Archives & RSS: http://seclists.org/fulldisclosure/
Powered by blists - more mailing lists