New issue
Advanced search Search tips

Issue 711903 link

Starred by 1 user

Issue metadata

Status: Duplicate
Merged: issue 712038
Owner:
Closed: Apr 2017
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: Linux
Pri: 2
Type: Bug



Sign in to add a comment

Direct-leak in sk_realloc_throw

Project Member Reported by ClusterFuzz, Apr 15 2017

Issue description

Components: Internals>Skia
Labels: M-60 Test-Predator-Wrong-CLs

Comment 2 by hcm@chromium.org, Apr 17 2017

Owner: fmalita@chromium.org
To Florin for a look...
Project Member

Comment 3 by ClusterFuzz, Apr 18 2017

ClusterFuzz has detected this issue as fixed in range 465194:465207.

Detailed report: https://clusterfuzz.com/testcase?key=6335096992038912

Fuzzer: attekett_surku_fuzzer
Job Type: linux_lsan_chrome_mp
Platform Id: linux

Crash Type: Direct-leak
Crash Address: 
Crash State:
  sk_realloc_throw
  reserve
  SkTextBlobBuilder::allocInternal
  
Sanitizer: address (ASAN)

Fixed: https://clusterfuzz.com/revisions?job=linux_lsan_chrome_mp&range=465194:465207

Reproducer Testcase: https://clusterfuzz.com/download/AMIfv97ulh3odRVjuq-I0_NGpauKI3Ew_lkxBQ2MRnxcS8vEhNoz_aP9XLoulfiRWgxsO3VAHNWacFmBtdKTjnNr8fRk1GxApRg2shOByxgnaJQ4osZc3Ji1GNBgXl7j6z2B0qlW7UUHS4pxmplXWMYwanVk7LToUCkJidYINY05IXJV0cZ3np8XCC_ER8uOVSwVTbmink4iUYyPeI8QLNEScLZC7e-9GS0dLKPB4G98EjvHXTOW1ZLqj93AcbZsljSOoxnyx6OdRpPQxKc0QRdEq_G6ML9pMC7Ms3jx3HC7bxwiE8WEZV0N6XstQHsJzTL5-87I4DMqwGtgII2TzVsCydj6f0Kw2UCxh92iaV1W5gA2otMUbj0?testcase_id=6335096992038912


Additional requirements: Requires Gestures

See https://dev.chromium.org/Home/chromium-security/bugs/reproducing-clusterfuzz-bugs for more information.

If you suspect that the result above is incorrect, try re-doing that job on the test case report page.
Cc: fmalita@chromium.org
Components: -Internals>Skia Blink>Paint
Owner: enne@chromium.org
Status: Assigned (was: Untriaged)
Can't repro at ToT.  Nothing related changed in Skia recently, and neither did the Blink text paint code in that regression range.

So I suspect the leak is at a higher level.

https://codereview.chromium.org/2768143002/ looks like a good candidate in that range.  It's also been reverted, so I triggered a progression task in CF and it found the leak fixed on revert.

Assigning to enne@, to take a look before relanding.

Comment 5 by enne@chromium.org, Apr 18 2017

Cc: danakj@chromium.org vmp...@chromium.org
This seems like a very similar bug to  issue 711964 .

I'd be really curious to know if there's some similar leak bugs in Skia or in SkPictureBuilder that were suddenly fixed and then unfixed.
Cc: hcm@google.com
Not that I know of, but I agree this leak is kind of odd (requires gesture "key,ctrl+Q" - is that quitting the browser?).
Project Member

Comment 7 by ClusterFuzz, Apr 18 2017

Labels: ClusterFuzz-Verified
Status: Verified (was: Assigned)
ClusterFuzz testcase 6335096992038912 is verified as fixed, so closing issue.

If this is incorrect, please add ClusterFuzz-Wrong label and re-open the issue.

Comment 8 by hcm@chromium.org, Apr 18 2017

@enne I think you're on our only other similar/in that space leak bug that's been filed recently- https://bugs.chromium.org/p/chromium/issues/detail?id=712038

Comment 9 by danakj@chromium.org, Apr 20 2017

Mergedinto: 712038
Status: Duplicate (was: Verified)
This is the same as 712038. The DrawTextBlobOp has text blob info with pointers, and it is lost due to alignment problems. I confirmed with the repro that 712038 fixes this.

Sign in to add a comment