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

Issue 682169 link

Starred by 1 user

Issue metadata

Status: WontFix
Owner:
Closed: Jan 2017
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: Linux , Android , Windows , Chrome , Mac
Pri: 1
Type: Bug-Security



Sign in to add a comment

Crash in SkPixelInfo::CopyPixels

Project Member Reported by ClusterFuzz, Jan 18 2017

Issue description

Project Member

Comment 1 by sheriffbot@chromium.org, Jan 18 2017

Labels: M-56
Project Member

Comment 2 by sheriffbot@chromium.org, Jan 18 2017

Labels: ReleaseBlock-Stable
Project Member

Comment 3 by sheriffbot@chromium.org, Jan 18 2017

Labels: Pri-1

Comment 4 by palmer@chromium.org, Jan 18 2017

Components: Internals>Skia
Labels: OS-Android OS-Chrome OS-Mac OS-Windows
Owner: reed@chromium.org
Status: Assigned (was: Untriaged)
A Skia bug is likely to be cross-platform; setting flags.

Comment 5 by reed@chromium.org, Jan 19 2017

Cc: reed@google.com

Comment 6 by reed@chromium.org, Jan 19 2017

Cc: -reed@google.com reed@chromium.org
Owner: reed@google.com

Comment 7 by reed@google.com, Jan 19 2017

How do I repro this locally? I don't see noel-image-flip mentioned on this help page https://chromium.googlesource.com/chromium/src/testing/libfuzzer/+/HEAD/reproducing.md

Comment 8 by aarya@google.com, Jan 19 2017

You just run the testcase with chrome (built with chromeos=1), see command line in report. this is not a libfuzzer testcase.

Comment 9 by reed@google.com, Jan 19 2017

This test is asking chrome to allocate (at least) 2 buffers, each 1-gig (webp image claims to be 1048876 x 250 at 4-bytes-per-pixel).

Given that on each crash, the address is nicely on a page-boundary, I expect this is a case of malloc over committing.

Comment 10 by reed@google.com, Jan 19 2017

Cc: -reed@chromium.org herb@chromium.org mtklein@chromium.org
Cc: awhalley@chromium.org
+awhalley@ for consult.

We're going to ship to stable / stable candidates in < 1 week, and this is a medium severity issue.  Can we remove the RB-Stable tag please to avoid blocking the release?  If we get a patch in time, or desperately want to pull this into a refresh, we can discuss that, but I'm worried more about risk from introducing a last minute change than a medium severity issue detected internally (though please let me know if you disagree).

Comment 12 by herb@chromium.org, Jan 20 2017

Cc: -herb@chromium.org herb@google.com

Comment 13 by herb@google.com, Jan 20 2017

I don't seem to be able to access https://cluster-fuzz.appspot.com/testcase?key=4742495444664320 with either herb@google.com or herb@chromium.org.

Comment 14 by aarya@google.com, Jan 20 2017

Cc: -herb@google.com reed@chromium.org
Owner: herb@google.com
Cc: mbarbe...@chromium.org
Labels: -M-56 M-57
I'm OK to push this to 57.

mbarbella@ would you mind checking herb@'s access to the clusterfuzz issue?

Comment 16 by aarya@google.com, Jan 20 2017

I already marked herb@google.com as owner, now he should be able to access.
I suspect this is another OOM issue (similar to  bug 633475  and  bug 598724 ). ASan always outputs SEGV for these even though the real signal is SIGBUS.

Comment 18 by herb@google.com, Jan 23 2017

Status: WontFix (was: Assigned)
As per #9, I think this is an OOM issue. Closing. Reopen if this has a frequency > 1.
Labels: -ReleaseBlock-Stable
Project Member

Comment 20 by sheriffbot@chromium.org, May 2 2017

Labels: -Restrict-View-SecurityTeam allpublic
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

Sign in to add a comment