[DragonFlyBSD - Bug #2376] (Closed) determine real cause of ldd crash

John Marino via Redmine bugtracker-admin at leaf.dragonflybsd.org
Tue Aug 14 07:49:47 PDT 2012


Issue #2376 has been updated by John Marino.

Status changed from New to Closed

RTLD was at fault after all.  The full object structure wasn't guaranteed to be fully allocated when it started digesting ELF notes.

FreeBSD hit the bug too (svn 239019)
Fixed here: http://gitweb.dragonflybsd.org/dragonfly.git/commitdiff/4db7b2663090f11386e806c421560a4d0923a7fb
----------------------------------------
Bug #2376: determine real cause of ldd crash
http://bugs.dragonflybsd.org/issues/2376

Author: John Marino
Status: Closed
Priority: Normal
Assignee: 
Category: 
Target version: 


Problem: ldd core dumps in specific cases within RTLD

To reproduce:
> cd /usr/pkgsrc/audio/arts
> bmake install
> ldd /usr/pkg/bin/artswrapper

It dumps within rtld reading digest notes.  That's because the obj pointer has a null value even though it was just created (signal 11).

Trying to step through with gdb will result in a "signal 5", stopping at ldd.c line 205, right after a fork.

I need some help determining if the problem is ldd.c or something in the kernel.  I don't think rtld is really at fault here, although I'm not confident about the TLS/resident part.






-- 
You have received this notification because you have either subscribed to it, or are involved in it.
To change your notification preferences, please click here: http://bugs.dragonflybsd.org/my/account





More information about the Bugs mailing list