Re: [PATCH] solib-svr4.c, several memory leaks

[Markus Deuling <deuling at de dot ibm dot com>]

msnyder@sonic.net schrieb:
> Jim, Kevin,
>
> Apparently we can't free the tmp name while bfd has the file open.
> Moving the xfree to later, after it has been closed, seems to fix
> the segfault.
>
> See attached patch.
> Michael
>
> > Folks,
> >
> > GDB started acting strange today, segfaulting when trying to open some
> > binaries, GDB itself included. Other binaries would work OK.
> >
> > Log:
> >
> > -------
> > ~/builds/gdb-head/> gdb-head-64bit/gdb/gdb gdb-head-64bit/gdb/gdb
> > GNU gdb 6.6.50.20070809-cvs
> > Copyright (C) 2007 Free Software Foundation, Inc.
> > GDB is free software, covered by the GNU General Public License, and you
> > are
> > welcome to change it and/or distribute copies of it under certain
> > conditions.
> > Type "show copying" to see the conditions.
> > There is absolutely no warranty for GDB.  Type "show warranty" for
> > details.
> > This GDB was configured as "powerpc64-unknown-linux-gnu"...
> > startUsing host libthread_db library "/lib64/ppc970/libthread_db.so.1".
> > (gdb) start
> > Breakpoint 1 at 0x10060b3c: file gdb-head/gdb/gdb.c, line 30.
> > Starting program: gdb-head-64bit/gdb/gdb
> > BFD: reopening : No such file or directory
> >
> > Segmentation fault
> > -------
> >
> >
> > I tracked down the root of the problem and this patch appears to be the
> > problem. Reverting it back fixes the problem.

Yes, it does. Please commit. These are my current testsuite results on x86 :-( :

# of expected passes		5706
# of unexpected failures	1233
<...>
# of unresolved testcases	4334


--
Markus Deuling
GNU Toolchain for Linux on Cell BE
deuling@de.ibm.com