New issue
Advanced search Search tips
Note: Color blocks (like or ) mean that a user may not be available. Tooltip shows the reason.

Issue 733614 link

Starred by 1 user

Issue metadata

Status: Unconfirmed
Owner: ----
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: Linux
Pri: 2
Type: Bug



Sign in to add a comment

Malformed UI Latency message from renderer crashes browser

Reported by zhouat2...@gmail.com, Jun 15 2017

Issue description

UserAgent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10_12_5) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/59.0.3071.86 Safari/537.36

Steps to reproduce the problem:
1. hook the render process
2. fuzz the payload of ipc message
3. then send Message to browser 
4. crash the whole chrome

What is the expected behavior?

What went wrong?
Just one visit , crash the whole chrome .

Some one can construct malicious ipc message, which carry bad payload, then he put all thing  in his home page. Once other user visit his home page's url, the visitor's whole browser crashed!  

because the check in source file ,ref to here:

https://cs.chromium.org/chromium/src/ui/latency/latency_info.cc?l=226

Did this work before? N/A 

Chrome version: 59.0.3071.86  Channel: stable
OS Version: Ubuntu 14.04.5 LTS \n \l
Flash Version: 

In the poc video, i use `Chromium   51.0.2684.0 (Developer Build) (32-bit)`,  which is the  newest version  i can get from the url below . Which is not stripped, then i can  locate the send of ipc easily, ipc message is the second  argument. But it can affect all versions of chrome , and need some more  work to construct message by hand. 

https://commondatastorage.googleapis.com/chromium-browser-snapshots/index.html?prefix=Linux/
 
Cc: tdres...@chromium.org jbau...@chromium.org
Components: UI
Labels: Stability-Crash
Summary: Malformed UI Latency message from renderer crashes browser (was: Malicious payload of ipc message in render process, crash the whole chrome.)
Typically, denial of service issues are tracked as stability problems rather than security problems, although we should fix this up, per https://www.chromium.org/Home/chromium-security/education/security-tips-for-ipc#TOC-Safely-handle-known-bad-input

Were you able to send a malformed message that had an outcome other than a reliable exit of the process?

Comment 2 by est...@chromium.org, Jun 15 2017

Labels: Needs-Feedback

Comment 3 Deleted

Project Member

Comment 4 by sheriffbot@chromium.org, Jun 16 2017

Cc: est...@chromium.org
Labels: -Needs-Feedback
Thank you for providing more feedback. Adding requester "estark@chromium.org" to the cc list and removing "Needs-Feedback" label.

For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot

Comment 5 by est...@chromium.org, Jun 16 2017

Labels: Needs-Feedback
Thanks for the additional info. Is the crash from poc_zhouat.png the same as shown in the video, i.e. "Check failed" in latency_info.cc?
Project Member

Comment 6 by ClusterFuzz, Jun 16 2017

ClusterFuzz is analyzing your testcase. Developers can follow the progress at https://cluster-fuzz.appspot.com/testcase?key=6255979154112512

Comment 7 Deleted

Project Member

Comment 8 by sheriffbot@chromium.org, Jun 16 2017

Labels: -Needs-Feedback
Thank you for providing more feedback. Adding requester "estark@chromium.org" to the cc list and removing "Needs-Feedback" label.

For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot

Comment 9 by est...@chromium.org, Jun 16 2017

Labels: -Restrict-View-SecurityTeam -Via-Wizard-Security Security_Impact-None
Thanks. I'm removing security labels because the crash hits a CHECK, which is designed to crash the browser to prevent security problems. It would be better if it were not so easy to crash the browser, but we don't classify such problems as security bugs.

Comment 10 Deleted

Re comment #10: yes I agree Chrome should handle this case better which is why I'm leaving the bug open to be triaged as a functional bug. See https://dev.chromium.org/Home/chromium-security/security-faq#TOC-Are-denial-of-service-issues-considered-security-bugs- for our FAQ about DoS as security bugs.

Comment 12 Deleted

Labels: Hotlist-Input-Dev

Comment 14 Deleted

Labels: -Type-Bug-Security Type-Bug
Changing type to Bug since this bug doesn't seem to have security implications. 
Re #14, we don't give CVE numbers to non-security bugs.
Project Member

Comment 16 by ClusterFuzz, Jul 30

Labels: -Security_Impact-None Security_Impact-Head
Summary: <no crash state available> (was: Malformed UI Latency message from renderer crashes browser)
Testcase 6255979154112512 failed to reproduce the crash. Please inspect the program output at https://clusterfuzz.com/testcase?key=6255979154112512.
Summary: Malformed UI Latency message from renderer crashes browser (was: <no crash state available>)
Labels: Hotlist-DesktopUIChecked Hotlist-DesktopUIToolingRequired
Mass UI Triage.

Sign in to add a comment