New issue
Advanced search Search tips
Starred by 5 users

Issue metadata

Status: Fixed
Owner:
Closed: Jan 2015
Cc:
Area: AndroidFramework
Priority: Medium
Type: Defect



Sign in to add a comment

Android 4.3 Segfault in libskia.so

Project Member Reported by djsollen@google.com, Aug 6 2013

Issue description

There looks to be that possibly 2 separate bugs are tied to this android issue (https://code.google.com/p/android/issues/detail?id=58515).

1. My guess around the original bug filed is that we have some memory problem.  It could be a bug in the framework's management of Skia objects or it could be that they accidentally recycled a bitmap and are later attempting to use it.

2. The second problem (first seen in comment #2), could be that our 565 NEON code has a bug that is causing a crash.

 
I wish this was fixed.
Project Member

Comment 2 by scro...@google.com, Jan 29 2015

Cc: scro...@google.com
Is this still a problem?
Project Member

Comment 3 by djsollen@google.com, Jan 30 2015

Labels: -Area-AndroidNext Area-AndroidFramework
Status: Fixed
In our move to support aarch64 for ARM builds we did a significant refactor of the NEON procs.  Chrome has successfully been using the 565 and 8888 procs for multiple releases so I believe that part of the issue has been resolved.

As for the nullptr SIGSEV that was also lumped into the Android bug we have also refactored the android framework to avoid directly using the SkDevice and now force everything through a much more stable and reliable path.

Sign in to add a comment