oss-sec mailing list archives
Re: graphicsmagick: memory allocation failure in MagickRealloc (memory.c)
From: Bob Friesenhahn <bfriesen () simple dallas tx us>
Date: Thu, 1 Dec 2016 21:20:13 -0600 (CST)
On Thu, 1 Dec 2016, Agostino Sarubbo wrote:
If suitable for a CVE please assign one. Thanks. Description: Graphicsmagick is an Image Processing System. This is an old memory failure, discovered time ago. The maintainer, Mr. Bob Friesenhahn was able to reproduce the issue; I’m quoting his feedback about: The problem is that the embedded JPEG data claims to have dimensions 59395×56833 and this is only learned after we are in the JPEG reader. But for some reasons (maybe not easy to fix) it is still not fixed.
We did make an unreleased fix (Mercurial changeset 14953:38d0f281e8c8, and earlier changeset 14831:28c0bb8bf89a), but perhaps not the way you like. The fix which was made was to require that the embedded JPEG data has the same dimensions as the containing JNG file. The existing resource limit mechanism would then allow the user to constrain the size of the JNG image. The default constraints in a 64-bit build are larger than what the JPEG format supports.
It does not seem correct to change the default limits of the software in order to make fuzzing easier.
Bob -- Bob Friesenhahn bfriesen () simple dallas tx us, http://www.simplesystems.org/users/bfriesen/ GraphicsMagick Maintainer, http://www.GraphicsMagick.org/
Current thread:
- graphicsmagick: memory allocation failure in MagickRealloc (memory.c) Agostino Sarubbo (Dec 01)
- Re: graphicsmagick: memory allocation failure in MagickRealloc (memory.c) Bob Friesenhahn (Dec 01)
- Re: graphicsmagick: memory allocation failure in MagickRealloc (memory.c) cve-assign (Dec 04)